Percutaneous pinning for treating distal radial fractures in adults

Summary of findings for the main comparison. Summary of findings: percutaneous pinning versus cast or brace immobilisation only

Percutaneous pinning compared with cast or brace immobilisation only for treating distal radius fractures in adults
Patient or population: adults with dorsally displaced distal radius fractures Settings: emergency department, hospital Intervention: reduction and percutaneous pinning, usually supplemented by cast immobilisation Comparison: reduction and cast (or brace) immobilisation alone
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	cast only	percutaneous pinning
Short‐term patient‐reported hand/wrist/upper limb function. DASH scores (30 to 150; worst disability) At 1.5 months	Median 97 (study result)	Median 85 (study result)	Not available	60 (1 study)	⊕⊝⊝⊝ very low^a	There are very limited data for this outcome. The study reported the difference was not statistically significant.
Medium‐term patient‐reported hand/wrist/upper limb function. DASH scores (30 to 150; worst disability) At 6 months	Median 84	Median 67	Not available	60 (1 study)	⊕⊝⊝⊝ very low^a	There are very limited data for this outcome. The study reported the difference was statistically significant. However, it may not be clinically important.^b Another study (43 participants) reported similar functional results in the two groups, probably at 12 months.
Long‐term patient‐reported hand/wrist/upper limb function. Over 12 months follow‐up	See comment	See comment	‐	‐	‐	This was not reported in any of the 11 studies (917 participants) for this comparison.
Number of people incurring one or more complications requiring substantive treatment	See comment	See comment	‐	799 (9 studies)	⊕⊝⊝⊝ very low^c	Data are available for the most common individual complications in this category: treated redisplacement, pin tract infection, CRPS type 1 and persistent finger stiffness.^d
Overall number of people incurring one or more complications	See comment	See comment	‐	799 (9 studies)	⊕⊝⊝⊝ very low^c	As well as those complications listed above, data were available for other surgery‐related complications (e.g. K‐wire migration) and other complications occurring in both groups, particularly those related to the median nerve (e.g. carpal tunnel syndrome).^e
Grip strength At 12 months	See comment	See comment	‐	238 (4 studies)	⊕⊝⊝⊝ very low^f	None of the four studies provided data for pooling. Results were reported as similar in the two groups in three trials and favouring pinning in the fourth trial.
Medium‐term quality of life measured via the Short Form 36 ‐ Physical score (0: worst to 100: best health) At 4 months	Mean SF‐36 in the plaster cast group was 38.2	Mean SF‐36 in the intervention group was 4.00 higher (1.59 lower to 9.59 higher)		54 (1 study)	⊕⊝⊝⊝ very low^g	A similar lack of between‐group difference was found for mental scores (MD 0.62, 95% CI ‐5.32 to 6.56).
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval;MD: Mean difference; RR: Risk Ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
^a Very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision. ^b Based on a scoring scheme of 0 to 100 (worst outcome), a minimal clinically important difference (MCID) of 15 is recommended in DASH/QuickDASH; this translates to 18 for a scale of 30 to 150; which is greater than the 16 difference in median scores at six months. ^c All the evidence for individual complications was considered very low quality, downgraded two levels for very serious risk of bias and one or two levels for serious or very serious imprecision. ^d Redisplacement resulting in secondary treatment only occurred in the nonsurgical treatment group: 38/316 (12%); range 3.3% to 75% (reported in 6 studies). Pin tract infection requiring antibiotics and, often, early wire removal only occurred in the pinning group: 22/285 (7.7%); range 0% to 15% (reported in 7 studies). There was one case of persistent deep infection. Incidence of RSD (nowadays, complex regional pain syndrome type 1), reported in 4 studies, was similar in the two groups: 13/123 (10.6%) versus 17/125 (13.6%); this was dominated by the results for short‐term RSD in one study: 12/48 versus 15/50. Persistent finger stiffness was less common after pinning versus cast alone after cast removal or at 12 weeks: 29/147 versus 53/149; RR 0.52, 95% CI 0.36 to 0.76; 2 studies; the treatment implications of this were not stated. ^e Reported events for other outcomes were generally few and reported in single trials only. These were mainly surgery‐related. The exception was complications relating to the median nerve, such as carpal tunnel syndrome, which occurred in both groups but were uncommon overall, amounting to 14 cases in 383 participants (3.7%) reported in five studies. ^f The quality of the evidence was very low given the qualitative and incomplete nature of the reported results and the very serious risk of bias. ^g Very low quality, downgraded one level for serious risk of bias and two levels for very serious imprecision.

Background

Description of the condition

Fractures of the distal radius, often referred to as "wrist fractures", are common in both children and adults. They are usually defined as occurring within three centimetres of the radiocarpal joint, where the distal (lower) end of the radius meets with two (the lunate and the scaphoid) of the eight bones forming the carpus (wrist). The majority are closed injuries, the overlying skin remaining intact.

In this review, we consider the treatment of distal radial fracture in adults only, in whom they are one of the most common fractures, predominantly in white and older populations in high‐income countries (MacIntyre 2016; Sahlin 1990; Singer 1998; Van Staa 2001). In women, the incidence of these fractures increases with age, starting at around 40 years of age. Before this age, the incidence is higher in men (Mosenthal 2019; Singer 1998). In contrast, between 60 to 94 years of age, females predominate. A multicentre study in the United Kingdom of patients aged 35 years and above with Colles' fracture (see below) reported an annual incidence of 9/10,000 in men and 37/10,000 in women (O'Neill 2001). This is just one of the 22 epidemiological studies describing the incidence in five global regions presented in MacIntyre 2016. Although MacIntyre 2016 warns against direct comparisons, they point out that the incidence of distal radius fracture is higher in Scandinavia, New Zealand and North America, than in regions of Asia and Africa. MacIntyre 2016 considered that incidence rates were increasing, whereas a study of the Swedish National Patient Registry found a decline in annual incidence in women from 77/10,000 in 2005 to 63/10,000 in 2013 and 18/10,00 to 15/10,000 in men; both populations were aged over 50 (Mellstrand Navarro 2019).

Young adults more often sustain this injury as a result of high‐energy trauma, such as a traffic accident or sports. In older adults, especially females, the fracture usually results from low‐energy or moderate trauma, such as falling from standing height. This reflects the greater fragility of the bone, resulting from postmenopausal osteoporosis or disuse. It has been estimated that, at 50 years of age, a white woman in the USA or Northern Europe has a 15% lifetime risk of a distal radius fracture whereas a man has a lifetime risk of just over two per cent (Cummings 1985). More recent estimates (Van Staa 2001) of the lifetime risks of radius or ulna fracture at 50 years of age remain consistent with these figures: 16.6% for women versus 2.9% for men.

Distal radial fractures are usually treated on an outpatient basis and nonsurgically but often require hospital admission, including when undergoing surgery. The number of people receiving surgery for these fractures has been increasing over the years in many countries: for example, Mellstrand Navarro 2019 reported the proportion of patients undergoing surgery for these fractures in Sweden had increased by 6.7% for women and 4.2% for men between 2005 and 2013.

Classification

Surgeons have classified fractures by anatomical configuration or fracture pattern, to help in their management. Simple classifications were based on clinical appearance and often named after those who described them. The term "Colles' fracture" (Figure 1) is still used for a fracture of the distal radius in which there is an obvious and typical clinical deformity, commonly referred to as a dinner fork deformity, due to the shape of the resultant forearm (Colles 1814). The introduction of X‐rays and other imaging methods made it clear that the characteristic deformity (dorsal angulation, dorsoradial displacement, radial shortening) may be associated with a range of different patterns of fracture which may be important in determining the outcome of treatment, and therefore the way in which treatment is conducted. For example, the fracture through the distal radius in a Colles' fracture may be extra‐articular (leaving the joint surface of the radius intact) or intra‐articular (the joint surface is disrupted, sometimes in a complex manner). Numerous classifications have been devised to define and group different fracture patterns. One of the most commonly used is that of Frykman, which distinguishes between extra‐articular and intra‐articular fractures of the radiocarpal and distal radio‐ulnar joints, and the presence or absence of an associated distal ulnar fracture (Frykman 1967). Another commonly used system is the AO (Arbeitsgemeinschaft fur Osteosynthesefragen) system (Muller 1991; AO 2018), which divides the fractures into three major groups: group A (extra‐articular), group B (simple/partial intra‐articular), and group C (complex/complete intra‐articular). These three groups are then subdivided yielding 27 different fracture types. Other classification systems have attempted to link fracture type more directly with fracture management. For instance, Cooney 1993 proposed a 'Universal Classification' based on fracture displacement, articular involvement, reducibility (whether the fracture can be reduced; that is whether the bone fragments can be put back in place) and stability (whether, once reduced, the fragments will remain so).

Figure 1

$Example of a dorsally displaced "Colles'" distal radius fracture$

Example of a dorsally displaced "Colles'" distal radius fracture

Description of the intervention

For the purposes of this review, the nonsurgical treatment of distal radial fractures in adults involves reduction of the fracture, when displaced, and stabilisation in a plaster cast or other external brace. However, for fractures considered to be unstable (i.e. likely to lose position once reduced), the results of such treatment are not deemed consistently satisfactory. This has resulted in the development of other strategies involving surgery (insertion of metalwork into the bone) aiming to provide an improved position and more reliable stabilisation of the fracture.

One such surgical treatment strategy is percutaneous pinning, which involves the percutaneous (through the skin) insertion of pins or wires, which may or may not be threaded. This is considered less invasive, quicker or often less technically demanding than open surgery, where the fractured bone is exposed to direct view. In percutaneous pinning, the reduction of the fracture is closed (see Handoll 2003b); although pins ‐ such as Kirschner wires ‐ may be used to manipulate the fracture fragments. In a comprehensive account of percutaneous pinning of fractures of the distal radius, Rayhack 1993 refers to a "myriad of options, decisions and questions that must be addressed". These decisions include: the reduction technique; the method and extent of skin incision; the use and type of radiographic control; the configuration of the pins; the number, size and type of pins; whether the pin ends are left exposed out of the skin or not; and the type and duration of immobilisation after pinning. Later decisions involve the timing and methods for pin removal. A variety of different pinning techniques or methods have been described in the literature (Fernandez 1999; Rayhack 1993). In many of these, pins are placed in the bone and used to fix the distal radial fragment(s). In contrast, in Kapandji's double intrafocal pinning (Kapandji 1988), and subsequent developments of this technique, pins or wires are used to manipulate and then support or 'buttress' the distal radial fragment(s).

How the intervention might work

We consider three key comparisons below.

As indicated above, the attraction of surgery, compared with nonsurgical treatment involving cast immobilisation, is that the reduction of the fracture could be more accurate and the pins help hold the fracture in place while it heals. Surgery, however, is more invasive and comes with the risk of additional complications, notably pin track infection (see below). It is also a more costly primary intervention.

Various techniques of wiring are likely to have different complication profiles but also may result in differentially more stable or accurate fracture constructs. However, while still under debate, the implications of moderate deviation in anatomical (radiographic) parameters from average normal values on long‐term patient‐rated function remain uncertain (Downing 2008; Johnson 2019b).

Immobilisation after pinning can help rest the wrist and relieve pain in the first few days after intervention; beyond this, prolonged immobilisation for a number of weeks after surgery is aimed at supplementing the stability of the fracture and wire construct. A longer duration of immobilisation carries the additional risk of stiffness and could hamper pin‐site management.

Complications

Complications from this injury are diverse and frequent (Altissimi 1986; Atkins 1989; Cooney 1980).

Some are associated with the injury itself. As well as concomitant injuries to soft tissues, fracture displacement can further compromise blood vessels, tendons or nerves, with median nerve dysfunction being the most common early complication (Belsole 1993). Complex regional pain syndrome type 1, often termed reflex sympathetic dystrophy (RSD), but also referred to as algodystrophy, Sudeck's atrophy or shoulder‐hand syndrome (Fernandez 1996), is a major complication requiring many months of physiotherapy in serious cases to alleviate symptoms (pain and tenderness, impairment of joint mobility, swelling, dystrophy, vasomotor instability). The aetiology is often unclear.

Malunion is a complication that is poorly defined, but reflects the loss of anatomical alignment of the injured bones, either as a result of the injury or after treatment. The question of whether anatomical restoration of the distal radius is necessary to achieve good long term function and how much of a deformity can be accepted has not been answered. There is some consensus that intra‐articular step‐offs should generally be corrected, as they predispose to the development of radiologically identified degenerative changes (Downing 2008). However, an evidenced‐based radiographic threshold for intervention in extra‐articular malunion has not been defined; the clinical decision‐making is also dependent on patient factors such as mental capacity, functional demands, comorbidities and age (Johnson 2019a). Post‐traumatic arthritis can occur several months or years after injury (Knirk 1986; Taleisnik 1984), with varying impact on wrist and hand function. Furthermore, studies have failed to investigate thresholds linked to patient‐reported satisfaction with appearance.

Complications can also directly result from treatment interventions. Examples of such include residual finger stiffness, which may be due to faulty application of plaster casts (Gartland 1951) and pin track infection, or nerve or tendon injury, as a result of percutaneous pinning. Loss of fracture position can occur with any treatment modality and may or may not require further intervention.

Why it is important to do this review

Percutaneous pinning is one of the key methods for surgical treatment of distal radial fractures. It is attractive because it is less invasive and mostly relatively simple and quicker to perform compared with other fixation methods involving an open surgical approach to the bone, such as plating. Since the previous version of this review (Handoll 2007), studies have illustrated a rapid increase in the use of plate fixation and often a decline in percutaneous pinning (e.g. Mellstrand Navarro 2019 reporting trends in Sweden between 2005 and 2013; and Mosenthal 2019 reporting trends between in USA between 2007 and 2014). Huetteman 2019, reporting trends in USA between 2010 and 2015, found that the decline in pinning was across all age groups in adults. In the UK, this trend has been countered to some extent by the findings of DRAFFT 2014, a large multicentre pragmatic randomised controlled trial, involving 461 adults, set in the United Kingdom National Health Service. DRAFFT 2014 concluded that there was no difference in patient‐reported functional outcome in participants with dorsally displaced distal radius fractures which could be reduced closed, when treated with Kirschner wires or volar locking plates. Costa 2016 reported that for patients undergoing surgery in England, the proportion of patients having K‐wire fixation rose to 42% with a concurrent fall in the proportion having fixation with a plate to 48% following publication of the trial. Though this review does not compare percutaneous pinning with other modalities of surgical fixation such as plating, the possibility of an increased relevance of percutaneous pinning to current clinical practice worldwide for these fractures, necessitated an update of the available evidence on this group of interventions.

The previous version of this review, published in 2007, concluded that "Though there is some evidence to support its use, the precise role and methods of percutaneous pinning are not established". As well as need for a search update and to incorporate the new evidence, we also considered it important to update the review methods, including systematically appraising the quality of the evidence using GRADE (Guyatt 2008; Schünemann 2019).

Objectives

We aimed to compare the relative effects of the following for adults with these injuries.

Any method of percutaneous pinning versus nonsurgical treatment involving plaster cast or brace use alone.
Any method of percutaneous pinning versus any other method of percutaneous pinning.
Any technique or type of material or device used for percutaneous pinning versus any other technique or type of material or device at surgery or postoperatively.
Any type or duration of postoperative immobilisation versus any other type or duration of immobilisation including none.
Any method or timing of pin or wire removal versus any other method or timing of pin or wire removal including no removal.

We considered these effects primarily in terms of patient‐rated functional outcome and the incidence of people with complications which resulted in secondary treatment and overall.

We planned to study the outcomes in different age groups and for different types of fracture patterns; in particular, whether extra‐articular or intra‐articular.

Methods

Criteria for considering studies for this review

Types of studies

Any randomised or quasi‐randomised (method of allocating participants to a treatment which is not strictly random e.g. by date of birth, hospital record number, alternation) controlled clinical trials of percutaneous pinning for treating distal radial fractures in adults were considered.

Types of participants

Patients of either sex who had completed skeletal growth, with a fracture of the distal radius. Our primary focus was on dorsally displaced fractures. Percutaneous pinning may be considered as primary treatment or may take place after the failure of initial nonsurgical management, generally within two to three weeks. Trials with a mixed population of adults and children were included provided the proportion of children was clearly small (< 5%); otherwise, they would have been excluded unless separate data for adults could have been obtained.

Types of interventions

This included the following comparisons.

Surgical interventions involving percutaneous pinning by itself versus nonsurgical (sometimes referred to as "conservative") interventions such as plaster cast immobilisation.
Different methods of percutaneous pinning in the treatment of fractures of the distal radius. Thus, comparisons evaluating:

1. different types of devices used for pinning;
2. use of different surgical techniques associated with percutaneous pinning, including type and extent of skin incision and measures to assist pin removal;
3. type and duration of immobilisation after percutaneous pinning;
4. type and timing of pin or wire removal.

Trials were excluded that compared percutaneous pinning with other methods of surgical fixation, such as external fixation, trials evaluating the use of supplementary percutaneous pinning in addition to another method of surgical fixation, and surgical versus nonsurgical treatment trials where the type of surgery was chosen by the surgeon, where percutaneous pinning was one of different surgical fixation methods used. We also excluded trials on pin site maintenance or other measures to prevent wound infection (already covered in Lethaby 2013).

Types of outcome measures

We aimed to divide all outcomes into short‐term (within three months), medium‐term (greater than three months and up to 12 months) and long‐term (greater than 12 months).

Primary outcomes

Function
1. Patient‐reported outcome measure (PROM) scores of hand/wrist/upper limb function (short‐, medium‐, long‐term), e.g. Patient Rated Wrist Evaluation (PRWE) Questionnaire (MacDermid 1998; MacDermid 2000), Disability of the Arm, Shoulder and Hand (DASH) (Hudak 1996) and QuickDASH Outcome Measures (Beaton 2005), Patient Evaluation Measure (PEM) (Macey 1995), Michigan Hand Questionnaire (MHQ) (Chung 1998) and Short MHQ (Waljee 2011).
Complications
1. Numbers of participants who incurred any complication that would typically receive substantive treatment (e.g. secondary surgical intervention, antibiotics for pin‐track infection; physiotherapy for Complex Regional Pain Syndrome (CRPS)). Finger stiffness during cast use, for which specific additional treatment, such as an unscheduled cast change or physiotherapy, was not undertaken, would generally not fall into this category.
2. Overall number of participants with complications.
3. Numbers of participants with individual complications

As found in the previous version of this review, the reporting of complications was incomplete and trials reported on the numbers of participants with individual complications, such as infection, rather than providing the total numbers of participants with complications requiring treatment. We have thus presented all data on complications as well as producing, where not reported, overall totals for participants with complications, provided unit of analyses and selective outcome reporting issues could be avoided. Our assumptions and interpretation, where reported data were incomplete, are documented in the review.

With regards to Complex Regional Pain Syndrome, CRPS Type 1 is now generally considered the equivalent of the older 'reflex sympathetic dystrophy' (RSD) and other terms such as 'algodystrophy', 'Sudeck’s atrophy' and 'shoulder hand syndrome'. In clinical practice, the term CRPS is often used inconsistently and, as the trials did not provide definitions, we have taken the approach to maintain the terminology used in source documents when reporting individual trial reports.

Secondary outcomes

Return to function as expressed by:
1. Return to (previous) occupation;
2. Return to patient key activities that are important to them.
Clinical:
1. Grip strength;
2. Range of motion;
3. Hand or wrist function performance scores, e.g. Jebsen‐Taylor (Jebsen 1969).
Quality of life:
1. Measures of health‐related quality of life (short‐, medium‐, long‐term) e.g. EQ‐5D (EuroQol Group 1990; Rundgren 2018), SF‐36 (Ware 1993), SF‐12 (Kosinski 2007).
Pain (short‐, medium‐, long‐term):
1. Visual analogue scales (VAS) or other stand‐alone pain scores;
2. Reports of life or function‐affecting pain measured as the pain component of a composite score, e.g. Gartland and Werley score (Gartland 1951).
Patient‐reported satisfaction:
1. Overall satisfaction;
2. Satisfaction with cosmetic appearance (often reported in terms of dissatisfaction).
Physician‐reported and/or composite outcome scores:
1. Total scores of physician‐reported and/or composite outcomes scores (short‐, medium‐, long‐term) e.g. Gartland and Werley score (Gartland 1951). In preference, for this outcome, we planned to report on 'poor and fair' categories.
Anatomical:
1. Radiographic parameters: e.g. dorsal angulation, radial length, radial inclination, ulnar variance and for intra‐articular fractures: the presence or absence of intra‐articular step or gap of the articular surface (Fernandez 1996; Kreder 1996a), based on plain radiographs. Definitions of four of the most commonly reported radiological parameters are presented in Table 1.
Economic outcomes:
1. Healthcare costs;
2. Resource use.

Table 1. Definition of radiological parameters

Parameter	Definition (Radiographic view)	Approximate normative values
Dorsal angulation (dorsal or volar or palmar tilt)	Angle between a) the line which connects the most distal points of the dorsal and volar cortical rims of the radius and b) the line drawn perpendicular to the longitudinal axis of the radius Lateral view	11–12 degrees, where positive values indicate a palmar angulation and negative values indicate angulation dorsally beyond neutral
Radial length	Distance between a) a line drawn at the tip of the radial styloid process, perpendicular to the longitudinal axis of the radius and b) a second perpendicular line at the level of the distal articular surface of the ulnar head Posteroanterior (PA) view	11‐12 mm
Radial angle or radial inclination	Angle between a) the line drawn from the tip of the radial styloid process to the ulnar corner of the articular surface of the distal end of the radius and b) the line drawn perpendicular to the longitudinal axis of the radius. Posteroanterior (PA) view.	22‐23 degrees
Ulnar variance	Vertical distance between a) a line drawn parallel to the proximal surface of the lunate facet of the distal radius and b) a line parallel to the articular surface of the ulnar head. Posteroanterior (PA) view obtained with the wrist in neutral forearm rotation, the elbow flexed 90° and the shoulder abducted 90°	Varies in different populations, with a mean usually neutral to slightly negative (e.g. to ‐1 mm)

Summary of findings table

The main findings to be included in the 'Summary of findings' tables were as follows.

Function: PROM scores ‐ short‐, medium‐, long‐term. Preference would have been given to PRWE (MacDermid 2000) if multiple PROM scores reported.
Complications: overall numbers of participants with complications that would typically receive substantive treatment; overall numbers of participants with one or more complications.
Clinical: grip strength in the medium term, ideally 12 months.
Quality of life: in the medium term, ideally 12 months.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (11 June 2019), the Cochrane Central Register of Controlled Trials (CENTRAL) (CRS Web 11 June 2019, Issue 6), MEDLINE (including Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily, Ovid MEDLINE and Versions(R)) (1946 to 12 June 2019), and Embase (1974 to 12 June 2019). For this update, the search results were limited from 2006 onwards. Details of the search strategies used for the previous version of the review are given in Handoll 2007. We did not apply any language restrictions.

In MEDLINE (Ovid Online), a subject‐specific strategy was combined with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials (sensitivity‐maximising version) (Lefebvre 2011). Search strategies for CENTRAL, MEDLINE, and Embase are reported in Appendix 1.

We also searched ClinicalTrials.gov (12 June 2019) and the WHO International Clinical Trials Registry Platform (WHO ICTRP) (12 June 2019) for ongoing and recently completed trials (Appendix 1).

Searching other resources

We checked reference lists of articles.

We searched abstracts of the following conferences:

American Academy of Orthopaedic Surgeons (AAOS) annual meeting (2014, 2015, 2016 (posters only), 2017, 2018*, 2019),
American Orthopaedic Trauma Association (AOTA) annual meetings abstracts (1996 to 2018),
Bone and Joint Journal (BJJ) Orthopaedic Proceedings (April 2018),
British Society for Surgery of the Hand (BSSH) (2012 to April 2019),
British Trauma Society (BTS) annual scientific meeting (2014, 2015, 2016, 2018),
Société Internationale de Chirurgie Orthopédique et de Traumatologie (SICOT) meetings in 2014 (SICOT 2014), 2015 (SICOT 2015), 2016 (SICOT 2016), 2017 (SICOT 2017) and 2018 (SICOT 2018),
Federation of the European Societies for Surgery of the Hand (FESSH) XXI meeting June 2016 (published in Journal of Hand Surgery. European Volume 2016 41 Suppl 1: FESSH 2016); Federation of the European Societies for Surgery of the Hand (FESSH) XXII and EFSHT XII combined meeting June 2017 (published in Journal of Hand Surgery. European Volume 2017 42E Suppl 1: FESSH 2017); Federation of the European Societies for Surgery of the Hand (FESSH) XXIII meeting June 2018 (published in Journal of Hand Surgery. European Volume 2018 43 Issue 2 suppl: FESSH 2018).

* The 2018 AAOS proceedings were searched by the editorial base and included in the CRS.

Data collection and analysis

None of the authors were investigators of any of the trials eligible for this review. Should this have occurred, measures would have been taken to ensure independent assessment of eligibility and, if included, processing.

Selection of studies

For this update, pairs of review authors (AK and HH; AK and AS) independently screened search results and assessed potentially eligible studies for inclusion. The initial decisions of trial eligibility were based on citations and, where available, abstracts and indexing terms. We obtained full articles and, where necessary to ascertain trial methods and status, individual review authors, mainly AK, sent requests for information to trial investigators. Study inclusion was by consensus. Titles of journals, names of authors or supporting institutions were not masked at any stage.

Data extraction and management

For newly included trials, pairs of review authors (AK and HH; AK and AS) extracted trial details and data using a data extraction form. All disagreements were resolved by discussion. Where possible, we contacted trialists of trials that were not reported in full journal publications for additional information and data.

Results were collected for the final follow‐up time point for which these were available. We also recorded where clinically important differences had been reported at intermediate follow‐up assessments, but were no longer apparent at final follow‐up.

All three authors entered the data into Review Manager 5 (RevMan 2014).

Assessment of risk of bias in included studies

The same pairs of review authors as for data extraction independently assessed risk of bias for newly included trials. Two authors (AK and HH) assessed risk of bias for trials that had been included in the previous version of the review. All differences were resolved by discussion. We used the tool outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b). This tool incorporates assessment of randomisation (sequence generation and allocation concealment), blinding (of participants and treatment providers, and outcome assessment), completeness of outcome data, selection of outcomes reported and other sources of bias. We considered 'subjective' outcomes (e.g. patient‐rated functional outcome scores, pain) and 'objective' outcomes (e.g. functional impairment, complications) separately in our assessment of blinding (performance bias). In our assessment of blinding (detection bias), we further split objective outcomes into those of functional impairment (e.g. grip strength) and the rest (e.g. complications). We considered short‐term (up to three months follow‐up) and longer‐term (three months or longer follow‐up) outcomes in our assessment of completeness of outcome data. We assessed two additional sources of bias: bias resulting from major imbalances in key baseline characteristics (e.g. age, gender, type of fracture, type of definitive treatment); and performance bias, particularly 'differential expertise' bias resulting from lack of comparability in clinician's experience with the interventions under test.

Additionally, we assessed four other aspects of trial design and reporting that would help us judge the applicability of the trial findings. The four aspects were: definition of the study population; description of the interventions; definition of primary outcome measures; and length of follow‐up.

Measures of treatment effect

Where available, quantitative data, both dichotomous and continuous, that were reported in individual trial reports for outcomes listed in the inclusion criteria were presented in the analyses. Risk ratios and 95% confidence intervals (CIs) were calculated for dichotomous outcomes and mean differences and 95% CIs were calculated for continuous outcomes.

Unit of analysis issues

We remained aware of potential unit of analysis issues arising from inclusion of participants with bilateral fractures, and presentation of outcomes, such as total complications, by the number of events rather than participants with these outcomes. We did not identify any trial reporting the inclusion of bilateral fractures, which, if few in number, would be unlikely to make much difference to study findings. We avoided the second stated unit of analysis issue, mainly by reporting on incidences of individual complications.

Dealing with missing data

Where appropriate for binary outcomes, we performed intention‐to‐treat analyses to include all people randomised to the intervention groups. In our protocol for this review, we indicated that we would, where appropriate, investigate the effect of missing data (dropouts and exclusions) by conducting best and worst case scenarios. So far, we have found no included data set where this would have been appropriate. We were alert to the potential mislabelling or nonidentification of standard errors and standard deviations. We calculated missing standard deviations from exact P values where provided. We did not assume values in order to present standard deviations in the analyses.

Assessment of heterogeneity

The decision to pool the results of individual studies depended on an assessment of clinical and methodological heterogeneity. If we considered studies sufficiently homogeneous for data pooling, we assessed statistical heterogeneity by visual inspection of the forest plots, and by using the Chi² test with a significance level of P value less than 0.1, and the I² statistic. We based our interpretation of the I² statistic results on those suggested by Higgins 2011a (Section 9.5.2):

0% to 40%: might not be important;
30% to 60%: may represent moderate heterogeneity;
50% to 90%; may represent substantial heterogeneity;
75% to 100%: considerable (very substantial) heterogeneity

Assessment of reporting biases

There were insufficient data (a minimum of 10 trials is recommended) to assess publication bias, for example, by preparing a funnel plot.

Data synthesis

Where appropriate, results of comparable groups of trials were pooled. As stipulated in the protocol, we used the fixed‐effect model and 95% confidence intervals.

Subgroup analysis and investigation of heterogeneity

There were no data available to carry out our prespecified subgroup analyses by age and gender and type of fracture (primarily extra‐articular versus intra‐articular fractures). While we planned to present separate subgroups according to whether there was a fundamental difference in pinning method (such as transfixation versus Kapandji's intrafocal method), the availability and nature of the outcome data did not support this. To investigate whether the results of subgroups were significantly different, we planned to inspect the overlap of CIs and perform the test for subgroup differences available in Review Manager 5 (RevMan 2014).

Sensitivity analysis

There were insufficient data available to carry out prespecified sensitivity analyses examining various aspects of trial and review methodology, including the effects of missing data, study quality (specifically allocation concealment, outcome assessor blinding and reportage of surgical experience), and inclusion of trials only reported in abstracts.

'Summary of findings' tables and quality assessment of the evidence

We produced 'Summary of findings' tables for one comparison for which a more substantive body of evidence had accrued. We used the GRADE approach to assess the quality of evidence related to each of the key outcomes listed in the Types of outcome measures for each comparison (see the Cochrane Handbook for Systematic Reviews of Interventions Section 12.2, Schunemann 2011).

Results

Description of studies

Results of the search

The search was updated from January 2006 to June 2019. A total of 1604 records resulted from searches of the following databases: Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (24), the Cochrane Central Register of Controlled Trials (435), MEDLINE (632), Embase (265), ClinicalTrials.gov (86) and the WHO ICTRP (162). After deduplication of the main databases, performed by the BJMT Information Specialist, we screened 1297 records from these. For the search results from conference proceedings, we did not usually record the overall number of potentially eligible studies for those conferences proceedings where we used more general search terms or we had searched the whole conference proceedings, such as for the BSSH, where search facilities were unavailable. For the purposes of compiling the PRISMA diagram, we noted that we identified 647 potentially eligible studies from other sources, primarily conference proceedings, and articles relating to the six studies awaiting assessment in the previous version of this review (Handoll 2007).

The search update (to June 2019 for the main databases) identified a total of 44 articles for potential inclusion, for which full reports were obtained, where possible. We linked any references pertaining to the same study under a single study ID. Upon further analysis, 13 studies were included (Delgado 2009; Gravier 2006; Hargreaves 2004; Mardani 2011; Murphy 2008; Saddiki 2012; Shannon 2003; Snow 2007; Soleiman pour 2011; Venkatesh 2016; Waheed 2004, Wong 2010; Zyluk 2007), seven were excluded (Chung 2019; Gunay 2015; IRCT2013120814271N2; IRCT2016061828510N1; ISRCTN37842313; NCT02353338; Tomaszuik 2017), and five are ongoing studies (DRAFFT 2; IRCT20160508027797N4; IRCT2016112727797N1;NCT03311633; RBR‐8mq8bs). A further three studies (Kanakeshwar 2017; Mirhamidi 2013; Russe 2000) await classification. Five of the six studies awaiting classification in Handoll 2007 are now included: Gravier 2006 (formerly Gravier 2005), Hargreaves 2004, Shannon 2003, Snow 2007 (formerly Snow 2006), and Waheed 2004; the sixth study remains in this category (Russe 2000).

Overall, there are now 26 included studies, 12 excluded studies, three studies awaiting classification and five ongoing trials. A flow diagram summarising the study selection process is shown in Figure 2. The results from the previous search (up to 2006) are given in Appendix 2.

Figure 2

Study flow diagram for updated review

Included studies

Twenty of the 26 included studies were fully reported in medical journals. Trial reports of the other six trials (Delgado 2009; Korner 1999; Murphy 2008; Shannon 2003; Verhulst 1990; Waheed 2004) were only available as conference abstracts.

In this update and the first version of the review, we sent requests for further information, including publication status, on 12 trials and received responses for seven trials (Azzopardi 2005; Gravier 2006; Gupta 1999; Hargreaves 2004; Snow 2007; Strohm 2004; Zyluk 2007). However, additional unpublished information was obtained for only two trials (Gupta 1999; Zyluk 2007), as detailed in the Notes section for each trial in Characteristics of included studies.

Details of the methods, participants, interventions and outcomes of individual trials are provided in Characteristics of included studies.

Setting

The publication dates of the main reports of these trials span 16 years, Verhulst 1990 being the earliest. Aside from Lenoble 1995, which had two centres, the studies were all single centre studies, mainly in teaching hospitals. The trials each took place in one of 11 countries (Belgium (3), China (1), France (5), Germany (2), India (2), Iran (2), Ireland (3), Morocco (1), Poland (1), Spain (2), UK (4)). Translations were obtained for five trials, were from French in three trials (Fikry 1998; Gravier 2006; Milliez 1992), from Persian for Soleiman pour 2011 and from Polish for Zyluk 2007.

Participants

Sex and age

The 26 included trials involved a total of 1946 participants, most of whom were female. Three studies, reported in conference abstracts only, provided no information on sex or age (Korner 1999; Murphy 2008; Verhulst 1990); although Verhulst 1990 referred to an "elderly" population. Two further studies, both published in full articles, did not report on sex (Venkatesh 2016) or age (Soleiman pour 2011). It is noteworthy that the data for Stoffelen 1998 were best guesses because of the disparities in the baseline characteristics data provided in the three papers for this trial. The majority of participants were female in 16 trials, with the percentages ranging from 67% to 89%. Of the six trials featuring more male participants, the percentages of female participants were 21% in Delgado 2009, which focused on heavy labour workers in Spain, 25% in Fikry 1998, and between 42% to 46% in the other four trials (Hargreaves 2004; Mardani 2011; Soleiman pour 2011; Stoffelen 1998). The trial populations in the populations of three of these trials were markedly younger, with the mean age being 34 years in Fikry 1998, 35 years in Hargreaves 2004 and 40 years in Delgado 2009. Otherwise, where reported, the mean ages ranged from 48 years in Venkatesh 2016 (no sex data provided) to 71.5 years in Azzopardi 2005. The youngest participant was 7 years in Hargreaves 2004 and oldest was 92 years in Strohm 2004. It is clear that the vast majority of participants in the included trials were skeletally mature; with confirmation based on trial inclusion criteria or baseline characteristics information available for 22 trials. The four exceptions are Hargreaves 2004, where an unknown proportion were children; Korner 1999 and Murphy 2008, where the absence of any information on age means inclusion of children cannot be ruled out; and, to a very limited extent, Strohm 2004, where the youngest participant was 15 years of age. Three trials (Azzopardi 2005; Rodriguez‐Merchan 1997; Wong 2010) further restricted the trial population to more mature adults: above 60, 45 and 65 years respectively. An upper age limit of 65 years was applied in Rodriguez‐Merchan 1997, and 80 years in Gravier 2006 and Stoffelen 1998.

Fractures

A variety of descriptions was used to define the types of fracture in the included trials. There was explicit reference to inclusion of dorsally displaced or Colles' fractures (or exclusion of anterior displacement) in 16 trials (Allain 1999; Azzopardi 2005; Fikry 1998; Gupta 1999 Milliez 1992; Rodriguez‐Merchan 1997; Saddiki 2012; Shankar 1992; Soleiman pour 2011; Stoffelen 1998; Strohm 2004; Verhulst 1990; Wong 2010; Zyluk 2007). There was no explicit inclusion of volar displaced fractures such as Barton's (Smith 1988). Trials often used descriptive terms such as displaced, unstable and comminuted, and referred to classification systems such as Frykman (Frykman 1967) or the AO system (Muller 1991). While Hargreaves 2004 and Strohm 2004 included a very few people with open fractures, it is likely that all the fractures in the other trials were closed, and certain in those 11 trials (Allain 1999; Azzopardi 2005; Casteleyn 1992; Lenoble 1995; Mardani 2011; Milliez 1992; Saddiki 2012; Soleiman pour 2011; Venkatesh 2016; Wong 2010; Zyluk 2007) which stipulated the exclusion of open fractures. Six trials described 'unstable' fractures, of which four trials described radiological criteria for defining an unstable fracture, thus defining the extent of displacement required for trial entry (Azzopardi 2005; Rodriguez‐Merchan 1997; Snow 2007; Wong 2010); whereas the other two did not (Delgado 2009; Shannon 2003). Mardani 2011 limited inclusion to stable fractures. Only Snow 2007 explicitly and solely included fractures that had redisplaced after a primary reduction. Of the 18 trials applying or reporting fracture type according to an established (not own) classification system, three trials grouped fractures according to more than one classification system. The Frykman system was used by 10 trials, the AO system by nine trials and Castaing system and Fernandez classification used by one trial each. Ten trials (Allain 1999; Casteleyn 1992; Korner 1999; Lenoble 1995; Milliez 1992; Saddiki 2012; Soleiman pour 2011; Snow 2007; Strohm 2004; Zyluk 2007) explicitly included both extra‐articular and intra‐articular fractures, although the extent of intra‐articular involvement was often limited; such as the exclusion of radiocarpal joint articular fractures in Casteleyn 1992, of fractures with more than two articular fractures in Allain 1999 and Lenoble 1995; and of comminuted fractures in Soleiman pour 2011. Seven trials only included or involved extra‐articular fractures (Azzopardi 2005; Gravier 2006; Gupta 1999; Mardani 2011; Stoffelen 1998; Venkatesh 2016; Wong 2010), whereas two trials (Rodriguez‐Merchan 1997; Shankar 1992) included intra‐articular fractures only. The involvement of the articular surface of the distal radius fracture was not clear in seven trials: Delgado 2009 (there was mention of intra‐articular fracture, however); Fikry 1998, where dislocated radiocarpal joint fractures were excluded; Hargreaves 2004; Murphy 2008; Shannon 2003; Verhulst 1990, which only referred to Colles' type fractures; or Waheed 2004.

Explicit exclusion of other injuries and conditions

Ten trials explicitly excluded people with multiple trauma, including other fractures (Allain 1999; Casteleyn 1992; Fikry 1998; Lenoble 1995; Mardani 2011; Milliez 1992; Snow 2007; Stoffelen 1998; Wong 2010; Zyluk 2007). Seven trials, including five of the first group, explicitly excluded people with previous wrist, elbow or upper limb fracture or malunion (Allain 1999; Azzopardi 2005; Lenoble 1995; Mardani 2011; Milliez 1992; Saddiki 2012; Snow 2007). Nine trials, only two of which did not appear in one or both of other categories, excluded for other conditions or reasons: dementia or psychiatric illness or problems (Azzopardi 2005; Mardani 2011; Wong 2010); severe brain injury (Stoffelen 1998); vascular or nerve complications (Milliez 1992; Soleiman pour 2011; Venkatesh 2016); unable or unwilling to participate in rehabilitation (Fikry 1998; Snow 2007: unable to physically or mentally participate in follow‐up). In all, 12 trials including all six trials only reported in abstracts did not report on other exclusion criteria.

Comparisons

The 26 included trials have been grouped according to the comparisons addressed by each trial. While most of the trials evaluated aspects of definitive treatment, three trials (Allain 1999; Milliez 1992; Soleiman pour 2011) tested the duration of immobilisation after surgery and Snow 2007 compared cast immobilisation of wrist in dorsiflexion versus palmar flexion after intrafocal pinning in people whose fractures had redisplaced.

Any method of percutaneous pinning versus nonsurgical treatment involving plaster or brace use

Eleven trials involving 917 participants compared closed reduction and internal fixation involving percutaneous pinning with closed reduction and plaster cast immobilisation (Azzopardi 2005; Delgado 2009; Gupta 1999; Mardani 2011; Rodriguez‐Merchan 1997; Shankar 1992; Stoffelen 1998; Venkatesh 2016; Verhulst 1990; Wong 2010; Zyluk 2007). A concise summary of the participants, fracture type, timing and details of the interventions for the 11 trials is given in Table 2. There was explicit reference to inclusion of dorsally displaced or Colles' fractures (or exclusion of anterior displacement) in eight trials (Azzopardi 2005; Gupta 1999 Rodriguez‐Merchan 1997; Shankar 1992; Stoffelen 1998; Verhulst 1990; Wong 2010; Zyluk 2007). In Mardani 2011 and Venkatesh 2016 the inclusion of dorsally displaced fractures could confidently be implied as dorsal by the context of the study report. One trial, reported solely in abstract form (Delgado 2009), did not report on fracture direction or displacement. There was no explicit inclusion of volar displaced fractures, such as Smith's or Barton's (Smith 1988) in any trial.

See: Summary of findings for the main comparison Summary of findings: percutaneous pinning versus cast or brace immobilisation only

Table 2. Percutaneous pinning versus nonsurgical treatment: participants & interventions

Trial	Participants	Fracture type and classification	Timing of surgery and reduction method	Fixation	Nonsurgical treatment
Azzopardi 2005	57; 89% female (of 54); mean age 71.5 years	Unstable (dorsal angulation 20+ degrees) extra‐articular fractures. AO type A3^a; Frykman I and II	Timing not stated. Closed reduction	Crossed pins, one from the radial styloid and the other from the dorso‐ulnar side of the distal fragment into the distal radial shaft. Cast immobilisation for 5 weeks	Closed reduction under general anaesthesia and fluoroscopy with three‐point fixation obtained in a "well‐moulded" short‐arm cast for 5 weeks
Delgado 2009	43; 21% female; mean age 40 years	Unstable fractures of distal radius (intra‐articular fracture mentioned in abstract report), no further details	Timing not stated. Closed reduction	No details on wiring type or duration of immobilisation	Closed reduction. No details on cast type or duration of immobilisation
Gupta 1999	50; 74% female; mean age 56 years	Colles'. All extra‐articular. Frykman I and II	Timing not stated. Closed reduction	Crossed pins, one from the radial styloid and the other from the dorso‐ulnar side of the distal fragment into the distal radial shaft. Cast immobilisation for 6 weeks	Closed reduction under traction and plaster cast for 6 weeks (change in position and new plaster at 3 weeks)
Mardani 2011	198; 44% female; mean age 50.8 years	Displaced distal radius fracture with congruous joint with less than 2 mm joint gap [described as Fernandez classification type 1]. Dorsal displacement can be inferred by content of study report.	Timing not stated. Closed reduction	Percutaneous pinning with smooth 1.5 mm or 2 mm pin and immobilised with short arm cast. Cast immobilisation for 6 or 8 weeks	Closed reduction under general anaesthesia and long arm cast applied by the same orthopaedist. Cast immobilisation for 6 or 8 weeks
Rodriguez‐Merchan 1997	40; 73% female; mean age 57 years	Comminuted unstable fractures. All intra‐articular. Displaced (10+ degrees dorsal angulation/3+ mm radial shortening). Frykman III to VIII	Probably next day after presentation at hospital. Closed reduction	Crossed pins, two pins from the radial side and one from the ulnar side of the distal fragment into the distal shaft. Cast immobilisation for 7 weeks.	Closed reduction under local anaesthesia and plaster cast for 7 weeks
Shankar 1992	45; 88% female; age range 17‐88 years	Comminuted Colles', Frykman IV to VIII intra‐articular fractures	Timing not stated. Closed reduction	Two pins inserted obliquely from the radial side across the inferior radio‐ulnar joint and into the ulnar medial cortex. Pins incorporated into plaster. Cast immobilisation for 5 to 6 weeks	Closed reduction under general anaesthesia and image intensifier control plaster with cast for 5 to 6 weeks
Stoffelen 1998	98; % female unknown (discrepancies between trial reports but fewer females); mean age 58 years	Colles'. Extra‐articular fractures with dorsal displacement. Frykman I and II	Timing not stated. Probably closed reduction	Triple intrafocal Kapandji pinning: 3 pins, inserted at the fracture site and driven into the radial shaft, act as buttresses to the distal fragment. Cast immobilisation for 1 week	Closed reduction. Above‐elbow plaster cast for 3 weeks; below‐elbow for 3 weeks
Venkatesh 2016	70; % female unknown; mean age 47.75 years	Extra‐articular fractures of distal radius (AO types 23‐A2, 23‐A3^a). Dorsal displacement can be inferred by content of study report and treatment in Colles' cast	Not older than 2 weeks. Closed reduction	5 mm K‐wire through the radial styloid process piercing the far medial cortex of proximal fragment. An additional K‐wire was passed through the ulnar side of the radius engaging the opposite cortex when deemed necessary by the operating surgeon. Below elbow cast in neutral for 6 weeks	Closed reduction under general anaesthesia. Below elbow cast maintaining the wrist in palmar flexion and ulnar deviation (Colles' cast) for 6 weeks
Verhulst 1990	130; % female unknown; "elderly"	Colles'	Timing not stated. Closed reduction	Single Kirschner wire (pin) inserted through radial styloid, across fracture and into the distal radial shaft. Cast immobilisation for 4 weeks (mean)	Closed reduction and plaster cast for 5 weeks (mean)
Wong 2010	62; 82% (of 60) female; mean age 70.5 years	Unstable (dorsal angulation > 20 degrees and radial shortening > 5 mm) extra‐articular fracture of the distal radius	Fracture less than 2 weeks old. Closed reduction	Bier's block; 3 percutaneous K‐wires ("tripod" construct described in report); removable palmar splint for resting purpose; immediate gentle mobilisation under instructions of a physiotherapist was allowed; wires were removed "after fracture consolidation"	Closed reduction under haematoma block, Jones's methods used for reduction, below elbow plaster. No details on duration of immobilisation
Zyluk 2007	124; 82% (of 60) female; mean age 61 years	Displaced distal radius fracture. AO A2, A3, B1, B2, C1, C2^a and Frykman I‐VII	Timing not stated. Closed reduction	Regional block; percutaneous fixation with 2 to 4 k‐wires introduced via the radial styloid so that they pierce the opposite cortex. In some cases, one K‐wire was introduced "from the base of the arm". Wires were buried and a plaster immobilising "only the wrist" was applied. Wires removed under local anaesthesia in clinic at 6 weeks	Closed reduction under local anaesthesia and immobilisation in a "Vienna" type above elbow plaster/backslab, shortened to below elbow at 2 weeks. Mobilised at 5 weeks

^a AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)

Where known, the majority of participants were female in Azzopardi 2005; Gupta 1999; Rodriguez‐Merchan 1997; Shankar 1992; Wong 2010 and Zyluk 2007; and male in Delgado 2009 and Mardani 2011. Where stated, mean age in the trials ranged from 40 years in Delgado 2009 and 71.5 years in Azzopardi 2005. Six trials only included or involved extra‐articular fractures (Azzopardi 2005; Gupta 1999; Mardani 2011; Stoffelen 1998; Venkatesh 2016; Wong 2010), whereas Rodriguez‐Merchan 1997; Shankar 1992) included intra‐articular fractures only. The involvement of the articular surface of the distal radius fracture was not clear in Delgado 2009 and Verhulst 1990. Zyluk 2007 included both extra‐articular and intra‐articular fractures.

Each trial employed a different pinning technique or duration of immobilisation, or both. Various numbers of pins were used to transfix the distal radius fragment to the radial shaft in eight trials: Mardani 2011 and Verhulst 1990 used one pin; Venkatesh 2016 used one or two pins; Azzopardi 2005 and Gupta 1999 used two pins; Rodriguez‐Merchan 1997 and Wong 2010 used three pins; and Zyluk 2007 used two to four4 pins. In Shankar 1992, two pins were placed obliquely through the distal radial fragments across the inferior radio‐ulnar joint and fixed in the ulna. Stoffelen 1998 used Kapandji's triple intrafocal technique, where three pins introduced at the fracture site and driven into the radial shaft supported the fracture fragment. Stoffelen 1998, Wong 2010 and Zyluk 2007 also differed from the other trials in that there was a difference in the duration of plaster immobilisation between the two intervention groups. Wong 2010 opted for immediate mobilisation of the pinning group, a removable palmar splint being used for resting.

Four trials (Azzopardi 2005, Mardani 2011, Shankar 1992 and Venkatesh 2016) specified that closed reduction casting was performed under general anaesthetic in the casting arm, with the rest of the trials specifying either local anaesthesia (Rodriguez‐Merchan 1997; Zyluk 2007), hematoma block (Wong 2010) or giving no details of anaesthetic/analgesic technique (Delgado 2009; Gupta 1999, Stoffelen 1998 and Verhulst 1990). Two trials specified that closed reduction and casting was performed under fluoroscopic control (Azzopardi 2005; Shankar 1992).

Delgado 2009, which was reported only in an abstract, provided no details on wiring technique nor cast type and no duration of immobilisation for either group.

Any method of percutaneous pinning versus any other method of percutaneous pinning

Six trials made one of five comparisons of different methods of percutaneous wiring. Kapandji intrafocal pinning featured in five trials (Fikry 1998; Gravier 2006; Lenoble 1995; Saddiki 2012; Strohm 2004) but not in Shannon 2003.

Kapandji intrafocal pinning (two or three wires) versus trans‐styloid fixation (two wires)

One trial compared Kapandji fixation with two or three Kirschner wires followed by immediate mobilisation versus trans‐styloid fixation with two Kirschner wires and plaster cast immobilisation for about 45 days in 120 adults with a closed extra‐articular or simple (two part only) intra‐articular, dorsally displaced distal radius fracture with posteromedial fragment (Lenoble 1995). Wire removal was around 45 days in both groups. Baseline characteristics were provided only for the 96 participants assessed at final follow‐up; of these, most were female (65/96, 68%) with a mean age of 57 years.

Modified Kapandji (dorsal Kapandji wires and trans‐styloid fixation) pinning versus Kapandji intrafocalpinning

A single‐centre study conducted in France compared a modified Kapandji technique that included insertion of a lateral 2/100 diameter transfocal styloid pin versus intrafocal Kapandji pinning (Gravier 2006). Further details of the interventions are provided in Characteristics of included studies. A posterior back slab was then applied for three weeks followed by passive rehabilitation. Pins were removed on day 45 postoperatively. The full report of the trial reported the randomisation of 85 study participants, who were mostly female (57/85, 67%) with a mean age 55 years, had dorsally displaced extra‐articular fractures of the distal radius, with or without an associated ulnar styloid fracture. An earlier abstract reported a population of 95 participants, with an average age of 49 years.

Kapandji intrafocalpinning (three wires) versus Py's isoelastic pinning (two wires)

Two trials undertaken in France compared Kapandji intrafocal pinning with three Kirschner wires versus with Py's isoelastic pinning, where two wires were inserted across the fracture and along the medullary canal, in 207 people with dorsally displaced distal radius fractures (Fikry 1998; Saddiki 2012). In Fikry 1998, both groups had immobilisation in a forearm cast for four weeks with pin removal at eight weeks. In Saddiki 2012, participants were in plaster for three weeks with wires removed at six weeks. Saddiki 2012 included 97 participants with either intra‐ or extra‐articular dorsally displaced distal radius fracture requiring surgical treatment. Though some involvement of the articular surface of the distal radius was possible in Fikry 1998, dislocated radiocarpal joint and "comminuted" fractures were excluded.The demographics differed between the two trials. Of the 88 participants followed‐up in Fikry 1998, their mean age was 34 years and 25% were female. The mean age of participants in Saddiki 2012 was 63 years and 86% were female.

Modified Kapandji intrafocal pinning (three wires) versus Willenegger pinning (two wires)

In Strohm 2004, one participant group was treated with a modified Kapandji method, involving intrafocal pinning with two Kirschner wires through the fracture gap with a third wire inserted via the radial styloid, followed by six weeks of immobilisation in a volar splint. Physiotherapy, with the volar splint removed, was started at three weeks. The other participant group received Willenegger pinning, where two wires introduced via the styloid process were inserted across the fracture, followed by immobilisation in a forearm cast for six weeks. The 100, mainly female (85%), participants with an average age of 65 years had Colles' type fractures, either extra‐ and intra‐articular (AO types A2, A3 or C1).

"Spring‐loaded intramedullary" pinning (not Py) versus "traditional transcortical" pinning (three wires used for both)

Shannon 2003, a single‐centre trial from Ireland, reported their study in a conference abstract only and we have been unable to obtain further information. They compared the insertion of three k‐wires using "a novel spring‐loaded intramedullary technique", where the k‐wires were engaged in the opposite cortex and driven down the medullary canal, versus insertion in "the traditional transcortical fashion". All fractures were reduced, and three 1.6 mm K‐wires were used. No details on the immobilisation or rehabilitation regimen were provided. The 46 participants were mainly female (37/46, 80%) with a mean age of 58 years. Fractures were described as "unstable" fractures of the distal radius; the AO classification was used but the population was not described.

Any technique or type of material or device used for percutaneous pinning versus any other technique or type of material or device at surgery or postoperatively

Biodegradable pins or wires versus metal pins or wires

Two trials undertaken in European teaching hospitals (Casteleyn 1992; Korner 1999) evaluated the use of biodegradable pinning in 70 people with extra‐articular and intra‐articular fractures. None of the 30 mainly female (77%) participants (mean age 61 years, range 22 to 85) of Casteleyn 1992 had radiocarpal joint fractures, with only Frykmann type I, II, V and VI fractures included, balanced across randomisation groups. No details of participant age or sex were available for Korner 1999, though it is stated that the intervention group included AO type A2, A3 and B1 fractures, with no detail of fracture characteristics for the comparison group. Different pinning techniques were used in the two trials: Kapandji's intrafocal pinning with two wires was used in Casteleyn 1992, while two pins or Kirschner wires were inserted via the radial styloid and across the fracture (modified Willenegger method) in Korner 1999. There was no cast immobilisation in Casteleyn 1992; there was no mention of postsurgical care in Korner 1999. Korner 1999 was available only in abstract form.

Pinning with buried wires versus exposed percutaneous wires

Three single‐centre trials, conducted in hospitals in Ireland and the UK, evaluated the technique of burying wires versus leaving them exposed in 168 people with isolated distal radius fractures requiring Kirchner wire fixation (Hargreaves 2004; Murphy 2008; Waheed 2004). Only Hargreaves 2004 was reported in a full‐text article. Participants in Hargreaves 2004 included both children and adults (mean age 35 years) of which 45% were female. All participants in Waheed 2004 were adults (mean age 57 years) of which 73% were female. There was minimal information on fracture type; notably two participants in Hargreaves 2004 had open fractures. Murphy 2008 provided no details on the population and neither Murphy 2008 nor Waheed 2004 described the surgery or interventions. Usually fixation was by two crossed K‐wires in Hargreaves 2004, which provided a detailed account of the interventions and other care, including the administration of a single dose of antibiotics at induction of anaesthesia. Wire removal was probably at around six weeks for all three trials.

Any type or duration of postoperative immobilisation versus any other type or duration of immobilisation including none

Cast immobilisation of wrist in dorsiflexion versus palmar flexion after intrafocal pinning

Snow 2007 compared cast immobilisation with the wrist in 30 degrees dorsiflexion versus 30 degrees palmar flexion after closed manipulation and percutaneous intrafocal Kirschner wire fixation in 63 participants whose fracture had redisplaced within 14 days. Fractures were extra‐articular or intra‐articular and participants were mainly female (82.5%) with a mean age of 60 years. After removal of the K‐wires at the outpatient clinic at three weeks, participants were then placed back into their selected casts for a further two weeks. The postoperative care of both groups was identical.

Duration of postoperative immobilisation after percutaneous pinning

The duration of immobilisation after percutaneous pinning was tested by three single‐centre trials, two of which were quasi‐randomised, in 170 people with extra‐articular or intra‐articular distal radius fractures (Allain 1999; Milliez 1992; Soleiman pour 2011). Allain 1999 and Milliez 1992 compared one week versus six weeks cast immobilisation in 120 people and Soleiman pour 2011 compared one week versus four weeks immobilisation in 50 people. The participants of Allain 1999 and Milliez 1992 were mainly female (74% overall) with a mean age of 55 years. Soleiman pour 2011 included more males (54% of the population) but did not report on age. Trans‐styloid fixation with two Kirschner wires was used in Allain 1999 and Kapandji intrafocal pinning with three Kirschner wires was used in Milliez 1992; in both these trials, the early mobilisation group was advised against heavy work and lifting heavy loads but did not appear to receive any other specific instructions to those given after cast removal in both groups. Although there was mention of Kapanji in the text, Soleiman pour 2011 did not specify the surgical technique used nor provide information on the postsurgical regimen.

Any method or timing of pin or wire removal versus any other method or timing of pin or wire removal including no removal

There were no included trials for this topic.

Outcomes

The outcomes collected by individual trials are listed in the 'Outcomes' section of Characteristics of included studies, where these are grouped under '(1) Functional', which includes function assessed via PROMs, return to former activities, grip strength and range of motion, and composite measures such as the Gartland and Werley score; '(2) Clinical', which includes pain and complications; '(3) Anatomical'; and, exceptionally, '(4) Other'. In the following, we focused on function measured via PROM scores, complications, grip strength and quality of life as these were the outcomes chosen for presentation in 'Summary of findings' tables.

Length of final follow‐up ranged from around six weeks in five trials (Gravier 2006; Hargreaves 2004; Murphy 2008; Shannon 2003; Waheed 2004) to two years (Lenoble 1995; Verhulst 1990), but was of variable duration in Fikry 1998, Korner 1999, Strohm 2004 and Zyluk 2007.

The three trials measuring a PROM for function all recorded DASH scores (Delgado 2009; Saddiki 2012; Zyluk 2007). All trials reported to some extent on complications except Gravier 2006 and Venkatesh 2016. For most trials, the number of participants incurring a complication in each group was not reported and could not be deduced with confidence from the reported individual complications because of concerns over unit of analysis issues and/or incomplete data collection and reporting. Grip strength was reported in 13 trials, and would have been measured in several others as it formed part of composite scores. Just two trials reported on heath‐related quality of life: SF‐36 was reported by Azzopardi 2005, and the WHO quality of life by Wong 2010.

Excluded studies

Twelve studies were excluded for reasons stated in Characteristics of excluded studies. Six studies were found not to be randomised trials (Biedermann 2001; Delattre 1994; Galli 2002; Gunay 2015; Seifert 1998; Tomaszuik 2017); two trials were abandoned (Harper 2000; ISRCTN37842313); three trials did not test an appropriate comparison (Chung 2019; IRCT2016061828510N1; NCT02353338) and we judged that further clarification or a trial report of IRCT2013120814271N2, which was sparsely and unclearly reported in a trial registration document, was unlikely to materialise.

Ongoing studies

Details of the five ongoing studies are given below.

DRAFFT 2: pragmatic multicentre trial comparing percutaneous pinning plus cast immobilisation versus moulded cast immobilisation in 506 participants (probably actual recruitment). Recruitment was completed in March 2019. Trial completion, which included a 12‐month follow‐up, is scheduled for September 2020.
IRCT20160508027797N4: trial registered in March 2019 comparing 3‐pin percutaneous pinning with cast immobilisation in 60 participants (stated target). With a follow‐up of 42 weeks, it is likely this trial is still ongoing.
IRCT2016112727797N1: this completed trial compared three crossed pinning and four‐pin radioulnar transfixation methods in 50 participants. This small trial testing a new comparison in this review was published in mid August 2019, after the cut‐off time for inclusion (Hosseinzadeh 2019).
NCT03311633: trial registered in October 2017 comparing three versus six weeks of percutaneous pinning in 60 participants (stated target). All will receive six weeks cast immobilisation. The expected date of trial completed is February 2020.
RBR‐8mq8bs: trial registered in August 2017 compared pinning with buried wires versus exposed percutaneous wires in 200 participants (stated target). Correspondence on 30 August 2019 from the lead author indicated that the trial report was poised for submission.

Studies awaiting assessment

Details of the three trials pending assessment are given below.

Kanakeshwar 2017: this published abstract reported a comparison of three versus two k‐wires in 145 participants. We have received no response to our request for clarification of trial methods, including interventions, and results.
Mirhamidi 2013: a published journal article reported this trial of Kapandji pinning versus extrafocal pinning in 45 participants. We have received no response to our request for clarification of trial methods, including treatment allocation and further data.
Russe 2000: this multicentre trial incompletely reported in four abstracts, each with different numbers of participants, compared bioabsorbable versus metal wires. Although we received information in 2006 that it was a randomised trial with 115 participants, we received no further response to requests for further information and data.

Risk of bias in included studies

The 'Risk of bias' judgements on 12 items for the individual trials are summarised in Figure 3 and described in the 'Risk of bias' tables in the Characteristics of included studies. Frequently assessments resulted in an 'Unclear' (?) verdict; this often reflected a lack of information upon which to judge the item (Figure 4). However, we usually took a lack of information on blinding to imply that there was no blinding.

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 4

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

All trials were at high risk of bias, invariably performance bias that for most trials reflected the impracticality of blinding care providers or participants to the treatment allocation, and generally detection bias, although a few trials succeeded in blinding of some outcome assessment. As detailed below, most trials were rated at high risk of selective reporting bias.

Allocation

Five trials judged at high risk of selection bias, both relating to random sequence generation and allocation concealment used quasi‐randomised methods based on admission sequence or alternation (Gravier 2006; Gupta 1999; Milliez 1992; Stoffelen 1998), or participant chart numbers (Allain 1999). The three trials judged at low risk of random sequence generation bias, described tossing a coin (Azzopardi 2005), computer generation (Venkatesh 2016) and drawing of lots by the participant (Zyluk 2007). The other 18 trials did not describe their method of sequence generation and were rated at unclear bias for this item.

Only Wong 2010, which described the use of sequentially numbered opaque sealed envelopes, was judged at low risk of selection bias relating to allocation concealment. Of the other 20 trials rated at unclear risk of bias for this item, six mentioned the use of envelopes but did not provide sufficient details to confirm secure allocation concealment. This included Rodriguez‐Merchan 1997, which while describing the use of blinded consecutively numbered envelopes, did not describe whether these were sealed. The other five trials referred to sealed envelopes (Casteleyn 1992; Hargreaves 2004), closed envelopes (Shannon 2003; Snow 2007) or gave no details (Zyluk 2007). Fourteen trials gave no details of their method for safeguarding allocation concealment; although blinded randomisation was claimed by Strohm 2004.

Blinding

Blinding of participants and personnel was generally not practical and not claimed for these trials. All 19 trials reporting subjective outcomes were judged at high risk of performance bias for these outcomes. Aside from Gravier 2006, which was judged at unclear risk for radiological assessment only, the other 25 trials were judged at high risk of performance bias for objective outcomes.

Total blinding of outcome assessment is impractical for trials testing surgical interventions but it is possible for some outcomes and more so at longer term follow‐up. Of the 19 trials reporting on subjective outcomes, only Wong 2010 was considered at low risk of detection bias as there appeared to be effective blinding in place. We rated 13 trials at high risk of detection bias for these outcomes and the other five trials at unclear risk, generally as their outcome scores seemed less susceptible to bias (Allain 1999; Fikry 1998; Lenoble 1995; Rodriguez‐Merchan 1997; Saddiki 2012).

Of the 22 trials reporting on measures of functional impairment, three were rated at low risk of detection bias as there was blinded assessment of these outcomes (Allain 1999; Snow 2007; Wong 2010). We rated 17 trials at unclear risk of bias as we considered these outcomes were less susceptible to bias despite the lack of blinding. However, two trials were judged at high risk of bias (Venkatesh 2016; Zyluk 2007).

The risk of detection bias for complications was considered high for 21 trials and unclear for three trials, where the susceptibility for bias appeared less for different reasons. Notably, Gravier 2006 and Venkatesh 2016 did not report on complications.

Incomplete outcome data

Overall, participant flow was poorly reported; where reported, loss to follow‐up, sometimes including exclusions because of treatment failure, ranged from none to excesses of 20% in Fikry 1998 and Lenoble 1995; 32% in Saddiki 2012; and 52% in Zyluk 2007. The risk of attrition bias at short‐ and long‐term follow‐up was judged the same for those 18 trials for which short‐ and long‐term follow‐up data were available. Risk of attrition bias was judged at low risk in five trials (Casteleyn 1992; Gupta 1999; Hargreaves 2004; Soleiman pour 2011; Wong 2010); at high risk in 11 trials (Delgado 2009; Fikry 1998; Gravier 2006; Korner 1999; Lenoble 1995; Mardani 2011; Saddiki 2012; Shannon 2003; Strohm 2004; Verhulst 1990; Zyluk 2007) and at unclear risk in the remaining 10 trials. Aside from high loss to follow‐up, incomplete data and data inconsistencies or discrepancies were common reasons for rating trials at high risk of attrition bias.

Selective reporting

We judged 20 trials at high risk of selective reporting bias and six at unclear risk (Allain 1999; Azzopardi 2005; Casteleyn 1992; Gupta 1999; Wong 2010; Zyluk 2007). We found no published protocol for any trial. Trial registrations, both prospective, were found for just two trials (Azzopardi 2005; Snow 2007). As well as lack of information on the intended outcomes prior to starting the trial, missing outcomes, incomplete and inadequate reporting of results, extra outcomes not listed in methods, disparities between methods and results such as length of follow‐up or variable follow‐up were assessed as indicative of a high risk of selective reporting bias. No trial reported reasons for changes to protocol.

Other potential sources of bias

We assessed this in terms of bias resulting from major imbalances in key baseline characteristics and performance bias, particularly resulting from lack or potential lack of comparability in clinician's experience with the interventions under test.

We judged 19 trials at unclear risk of bias relating to important imbalances in baseline characteristics because of the absence of data on these, incompletely reported characteristics data for key characteristics or because of small imbalances that potentially could have affected the results. We judged that Stoffelen 1998 was at high risk for this item because of the major between‐group differences reported for sex and high‐energy injuries. We judged six trials at low risk of bias for this item (Gupta 1999; Hargreaves 2004; Mardani 2011; Milliez 1992; Rodriguez‐Merchan 1997; Wong 2010).

Comparability of care programmes, comprising interventions and related aspects other than the trial interventions (such as type of anaesthesia, timing of the interventions, comparability of the experience of the health professionals applying the interventions (operator bias), and rehabilitation) was hard to assess. We judged that five trials were at low risk of bias for this item (Gravier 2006; Mardani 2011; Snow 2007; Venkatesh 2016; Wong 2010) and the remaining 21 trials at unclear risk, predominantly because of incomplete information. We noted that distinctive differences between the two groups, where an intervention other than the trial intervention was used exclusively in one group, can change the actual comparison under test but was not considered a risk of bias. For instance, in Lenoble 1995, the Kapandji pinning group had immediate mobilisation but not the trans‐styloid fixation group; in Strohm 2004, physiotherapy was provided after three weeks of immobilisation to the modified Kapandji pinning group only.

Effects of interventions

Any method of percutaneous pinning versus nonsurgical treatment involving plaster or brace use

Closed reduction and percutaneous pinning was compared with nonsurgical treatment usually comprising closed reduction and plaster immobilisation alone in 11 trials involving 917 participants (Azzopardi 2005; Delgado 2009; Gupta 1999; Mardani 2011; Rodriguez‐Merchan 1997; Shankar 1992; Stoffelen 1998; Venkatesh 2016; Verhulst 1990; Wong 2010; Zyluk 2007). Table 2 summarises the participants, fracture type, timing and details of the interventions for the 11 trials. Details on length of participant follow‐up and losses to follow‐up for all trials are summarised in Table 3. Nine trials (776 participants) applied across‐fracture pinning whereas Stoffelen 1998 used Kapandji's triple intrafocal technique with early mobilisation in 98 participants. Delgado 2009 provided no details on wiring technique. The anaesthetic/analgesic technique in the nonsurgical arm also varied, with four trials employing general anaesthesia. Results for outcomes were often incomplete for many of the trials and opportunities to pool data were limited. There were no data available to carry out our prespecified subgroup analyses by age and gender and type of fracture.

Table 3. Percutaneous pinning versus nonsurgical treatment: length and completeness of follow‐up

Study ID^a	Length of follow‐up in months	Completeness of follow‐up
Mardani 2011	3	No participant flow diagram. Some loss, as the authors described mean postoperative visits of 4.4 for group 1 and 3.6 for group 2 out of a possible 5.
Gupta 1999	6	No losses (reported)
Shankar 1992	6	No losses (inferred)
Venkatesh 2016	6	10 participants lost to follow‐up (5 from each group) were excluded from the analysis.
Zyluk 2007	6	64 (31 from the pinning and 33 from the nonsurgical group)
Azzopardi 2005	12	9 (3 deaths and 6 lost to follow‐up). No information on which group they were from
Delgado 2009	12	Not known
Rodriguez‐Merchan 1997	12	No losses (inferred)
Stoffelen 1998	12	No losses (inferred)
Verhulst 1990	24	Not known
Wong 2010	19.5 (13 to 24)	2 participants were excluded from the analysis; these were two deaths from heart disease "which was not related to postoperative complications"; hence they appeared to be from the surgery group.

^a Ordered by length of follow‐up

Primary outcomes

Patient‐reported function

DASH, a PROM score for upper‐limb function, was used in the two trials, Delgado 2009 and Zyluk 2007. Delgado 2009 reported similar functional results in the two groups, probably at 12 months. Zyluk 2007 reported median DASH scores (scale used 30 to 150; higher scores = worse outcome) and ranges at 1.5, 3 and 6 months for 60 participants. Zyluk 2007 indicated that the difference in scores, which favoured the pinning group, was not statistically significant at 1.5 months (median 85 versus 97) but was at six months (48 versus 64); very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision (see Analysis 1.1). In addition to the uncertainty surrounding these findings, it is not certain whether the difference between the two groups in DASH scores was clinically important. Based on a scoring scheme of 0 to 100 (worst outcome), a minimal clinically important difference (MCID) of 15 is recommended in DASH/QuickDASH; this translates to 18 for a scale of 30 to 150, which is greater than the 16 difference in median scores at six months. Zyluk 2007 also collapsed the DASH scores (range 30 to 150) into three categories (normal: 30 to 40, satisfactory: 41 to 60 and reduced function: > 60). At six months, there were fewer participants in the pinning group with reduced function: 8/30 versus 21/30, RR 0.38, 95% CI 0.20 to 0.72; very low‐quality evidence downgraded two levels for very serious risk of bias, one level for serious imprecision and one level for serious indirectness given this improper categorisation of the DASH scores (Analysis 1.2).

Complications

With the exception of Delgado 2009, which claimed to have recorded complications but did not report them, and Venkatesh 2016, which did not mention any complications in their report, the other included trials reported to some extent on complications. The available data are presented in Analysis 1.3. Null events have also been entered when reported. Only data on individual complications were reported. All evidence for individual complications was considered very low quality, downgraded two levels for very serious risk of bias and one or two levels for serious or very serious imprecision.

We were not able to report on the overall numbers of participants who incurred one of more complications that would typically received substantive treatment. The main complications that would have contributed to this are redisplacement requiring secondary treatment, pin track infection that would have been treated with antibiotics and usually early wire removal, and CRPS (referred to as reflex sympathetic dystrophy in the majority of older studies). We didn't include finger stiffness during cast use. However, finger stiffness after cast removal at six weeks in Stoffelen 1998 and persistent stiffness at 12 weeks in Mardani 2011 would likely to have required additional intervention in the form of physiotherapy and is included. Hence, complications that would typically have received substantive treatment are reported as follows.

Redisplacement resulting in secondary treatment (consisting of reduction or reduction and K‐wire fixation) only occurred in the nonsurgical treatment group: 0/275 versus 38/316 (12%); RR 0.09, 95% CI 0.03 to 0.27; 6 studies. The percentages ranged from 3.3% (3/85) in Verhulst 1990 to 75% (15/20) in Rodriguez‐Merchan 1997. This last trial included intra‐articular fractures only.
Self‐evidently, pin tract infection requiring antibiotics and, where reported, wire removal only occurred in the pinning group: 22/285 (7.7%) versus 0/286; RR 8.33, 95% CI 2.54 to 27.28; 7 studies). The percentages ranged from 0% in Gupta 1999 to 15% (15/99) in Mardani 2011. There was one case of persistent deep infection in Zyluk 2007,
The incidence of RSD was similar in the two groups: 13/123 versus 17/125, RR 2.08, 95% CI 0.40 to 10.85; 4 studies. There was a high incidence (12/48 versus 15/50) of short‐term RSD in Stoffelen 1998; this persisted in six participants overall (4/48 versus 2/50). There was one case of shoulder hand syndrome in Shankar 1992.
More participants in the cast only treatment group had persistent finger stiffness at six weeks in Stoffelen 1998 and at three months in Mardani 2011: 29/147 versus 53/149; RR 0.52, 95% CI 0.36 to 0.76; 2 studies.

Other pinning ‐ or typically surgery‐related complications were: K‐wire migration, which occurred in 5/70 (7%) participants (data from Gupta 1999 and Verhulst 1990); one case of K‐wire removal due to metal sensitivity in Shankar 1992; three cases of stab wound infection in Verhulst 1990; and six cases of superficial radial nerve injury, one of which persisted, in the pinning group of Stoffelen 1998. Stoffelen 1998 suggested these nerve injuries may have resulted from injuries caused by the radially‐sited pin. There was no report of tendon rupture or injury in either treatment group; which was confirmed in two studies (Azzopardi 2005; Rodriguez‐Merchan 1997).

Although there were fewer complications relating to the median nerve, such as carpal tunnel syndrome, reported in the pinning group, the evidence is very low‐quality downgraded two levels for serious risk of bias and two levels for very serious imprecision: 3/161 versus 11/202; RR 0.53, 95% CI 0.19 to 1.46; 5 studies.

Secondary outcomes

Return to work or former activities

All 43 participants, who were all heavy labour workers in Delgado 2009 returned to the same work activity. Two people, both in the nonsurgical treatment group, out of 11 former labourers in Rodriguez‐Merchan 1997, did not return to their former work (Analysis 1.4). The evidence for this outcome is very low quality, downgraded two levels for very serious risk of bias and two levels for very serious imprecision. Verhulst 1990 (number of participants to which this applied unknown, maximum 130) reported duration of incapacity to work did not differ between the two groups at the 5% significance level. Wong 2010, which measured functional status relating to employment as part of the Mayo Wrist Score, found no difference (reported P = 0.914) between the two groups in this item at final (mean 19.5 months) follow‐up; it was clear that the majority of the 60 participants had 'returned to regular employment' or equivalent.

Grip strength

Grip strength was reported in seven trials but no pooling of data was possible (Azzopardi 2005; Delgado 2009; Rodriguez‐Merchan 1997; Shankar 1992; Stoffelen 1998; Wong 2010; Zyluk 2007). Zyluk 2007 reported medians and ranges for global grip strength and global grip strength relative to the other wrist at 1.5, 3 and 6 months for 60 participants; see Analysis 1.5; Analysis 1.6. Zyluk 2007 indicated that the difference in grip strength, which favoured the pinning group, was not statistically significant at 1.5 months (median grip strength 7.2 kg versus 6.0 kg; relative grip strength 31% versus 27%) but was at six months (median grip strength 18.2 kg versus 16.0 kg; relative grip strength 72% versus 61%); very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision). Results at six months also favoured the pinning group in Shankar 1992 which reported on the numbers of participants with under half grip strength at six months: 0/23 versus 11/22; RR 0.04, 95% CI 0.00 to 0.67; very low‐quality evidence downgraded two levels for very serious risk of bias, one level for serious imprecision and one level for indirectness in terms of the outcome measure (Analysis 1.7). At 12 months, grip strength presented as a percentage of the grip strength of the unaffected arm, was similar in Azzopardi 2005 (mean 77%, SD 21% versus 72%, SD 17%, reported P = 0.54) and greater in Rodriguez‐Merchan 1997 (85% versus 65%; P not reported). Both Delgado 2009 (43 participants) and Stoffelen 1998 (98 participants) reported similar grip strength results in the two groups at 12 months without providing data. Wong 2010 found no significant differences at follow‐up averaging 19.5 months: MD ‐0.50 kg; CI ‐3.90 to 2.90; very low‐quality evidence downgraded one levels for serious risk of bias and two levels for very serious imprecision (Analysis 1.8).

Range of motion

At six months, all components of wrist movement were significantly better in the pinning group in Gupta 1999 (Analysis 1.9), as were the components of wrist movement in Shankar 1992 (statistically significant differences were reported for extension and ulnar deviation). In contrast, Venkatesh 2016 found no differences between the two groups; the differences in the means for all six components of wrist motion ranged from 1 to 3.5 degrees. Zyluk 2007 reported there were no statistically significant differences between the two groups in flexion and extension range of motion and pronation and supination range of motion at 6 weeks, 3 and 6 months; Analysis 1.10 showed the available median and range data for 6 weeks and 6 months. At 12 months, Azzopardi 2005 reported no statistically significant differences in the components of wrist movement, except for ulnar deviation (93% versus 76% of normal; reported P = 0.009). The overall range of wrist motion at 12 months was reported to be better in the pinning group in Rodriguez‐Merchan 1997 (80% versus 60% of normal). Delgado 2009 stated that flexion‐extension arc loss compared with the contralateral wrist was less in the nonsurgical group at 3 and 6 months, but similar in both groups at 12 months; no supporting data were available in the abstract report. There were no differences in range of motion between treatment groups at longer follow‐up in Wong 2010 (mean 19.5 months) (Analysis 1.11). Verhulst 1990 also reported that range of motion along the three axes showed no statistically significant differences between the two groups at two years. Overall, the quality of the evidence for range of motion was very low at short‐, medium‐ and long‐term follow‐up; this reflected the serious limitation in the data including downgrading one level for serious risk of bias and two levels for very serious imprecision.

Health‐related quality of life

The evidence for this outcome, reported by two trials, is very low‐quality, being downgraded one level for serious risk of bias and two levels for very serious imprecision. Azzopardi 2005 found no significant differences between the two groups in the physical or mental score domains of the Short‐Form 36 at four months (Analysis 1.12). Wong 2010 reported no between‐group differences in WHOQoL‐BREF scores, either relating to a single question on the participant's view of their quality of life (Analysis 1.13) or for four domains (physiological, psychological, social, environment), at an average of 19.5 months follow‐up. Azzopardi 2005 reported there were no significant differences in either unilateral (reported P = 0.43) or bilateral (reported P = 0.74) activities of daily living scores at 12 months.

Pain

There were no statistically significant differences between the two groups in pain scores at one year (0.7 versus 1.2 on a 0 (no pain) to 10 (worst pain) point scale; reported P = 0.16) in Azzopardi 2005 or in the numbers of participants with occasional pain at long‐term follow‐up in Rodriguez‐Merchan 1997 (2/20 versus 4/20; RR 0.50, 95% CI 0.10 to 2.43; very low‐quality evidence downgraded one level for serious risk of bias and two levels for very serious imprecision; Analysis 1.14). Although Delgado 2009 implied comparable pain scores at 12 months (2.9 in the surgery group versus 2.1 in the nonsurgical treatment group), these data were incomplete and there were no details of the pain scale used. Wong 2010, which measured pain as part of the Mayo Wrist Score, found no difference (reported P = 0.873) between the two groups in this item at final (mean 19.5 months) follow‐up; the mean scores implied that the majority appeared to have mild or occasional pain over the past four weeks.

Composite and mainly physician‐reported functional scoring systems

The composite and mainly physician‐reported functional scoring systems used by five of the trials (Gupta 1999; Rodriguez‐Merchan 1997; Shankar 1992, Venkatesh 2016; Zyluk 2007) were all derived from Gartland and Werley's scheme (Gartland 1951), which also rates deformity and various complications. Fewer participants had fair or poor outcome in the pinning group based on these measures: 29/128 versus 52/127; RR 0.53, 95% CI 0.37 to 0.78; 255 participants, 5 studies; I² = 13%; very low‐quality evidence downgraded two levels for very serious risk of bias, one level for serious imprecision and one level for serious indirectness given nonvalidated categorisation of the score (Analysis 1.18). Stoffelen 1998 and Wong 2010 used the Cooney or Mayo Wrist Score, which included pain, functional status (employment), range of motion and grip strength (range 0 to 100, higher scores mean better outcome). There was minimal difference between the two groups in the numbers with a fair or poor outcome in Stoffelen 1998; very low‐quality evidence for the same reasons as above (Analysis 1.18). Wong 2010 reported the Mayo wrist score at final follow‐up (mean 19.5 months; range 13 to 24 months) and also found no difference between the treatment groups; very low‐quality evidence downgraded one level for risk of serious bias and two levels for very serious imprecision (Analysis 1.19).

Satisfaction

Mardani 2011 reported on patient‐reported satisfaction at 12 weeks based on criteria set in the Saito score. They found more participants in the surgery group had an excellent outcome (no pain, no disability and no limitation of motion): 93/99 versus 81/99, RR 1.15, 95% CI 1.03 to 1.28; very low‐quality evidence downgraded two levels for very serious risk of bias; two levels for serious imprecision and one level for serious indirectness relating to the interpretation of the outcome (Analysis 1.15). None reported a poor outcome. Wong 2010 found no difference between the two groups in patient satisfaction at 13 to 24 months; very low‐quality evidence downgraded one level for serious risk of bias; two levels for serious imprecision and one level for serious indirectness relating to the crude categorisation of the outcome measure (Analysis 1.16).

Radiographic outcomes

Four trials (Azzopardi 2005; Gupta 1999; Venkatesh 2016; Wong 2010) presented complete results for dorsal angulation, radial angulation and radial length and two trials (Venkatesh 2016; Wong 2010) for ulnar variance. Although we considered it is useful to present anatomical results, we decided against pooling these data because of the likely differences in their measurement. The data suggest better anatomical results in favour of the pinning group for all four measures but the differences in radiographic measurements recorded in most cases were small, may reflect measurement error, and are of uncertain clinical significance. Mardani 2011 and Rodriguez‐Merchan 1997 did not report on measurable radiographic parameters. Healing of all fractures was confirmed in Gupta 1999 and Shankar 1992 and is likely in the other trials. Loss of reduction and need for secondary procedures due to loss of position is captured under complications, as detailed above (Analysis 1.3).

Recording of anatomical results in the remaining trials was incomplete. Delgado 2009 reported that "radiographic correction" was anatomical in 38% of cases in the nonsurgical group and 80% in the pinning group. Losses in radial angulation, radial length and dorsal angulation from the reduced position were reported as being significantly greater in the nonsurgical treatment group in Shankar 1992. In contrast, Stoffelen 1998 did not report superior anatomical results for the Kapandji's pinning group. Whilst radial shift (lateral) was less in the pinning group (mean values: 0.5 mm versus 1.5 mm), radial shortening was greater (mean values: 2 mm versus < 1 mm). Stoffelen 1998 claimed that these differences were statistically significant. Stoffelen 1998 made no specific mention of the apparently greater, though indicated as not statistically different, difference in dorsal angulation between the two groups. Zyluk 2007 reported comparable results for the comparison groups in dorsal angulation, radial length and radial inclination at final follow‐up at six months.

Resource outcomes

None of the trials reported on economic outcomes. Mardani 2011 reported a mean number of postoperation visits was 4.4 in the pinning group and 3.6 in the cast‐only group. However, this was out of five scheduled visits and so can be considered more a measure of compliance.

Any method of percutaneous pinning versus any other method of percutaneous pinning

Kapandji intrafocal pinning (two or three wires) versus trans‐styloid fixation (two wires)

In Lenoble 1995, Kapandji fixation followed by immediate mobilisation was compared with trans‐styloid fixation followed by around six weeks of plaster cast immobilisation in 120 people. However, results were presented only for the 96 participants available for follow‐up at 24 months. As well as not reporting a PROM for function, return to prior activities, quality of life, or patient satisfaction, Lenoble 1995 did not provide standard deviations for continuous outcomes such as grip strength. There was very low‐quality evidence for all reported outcomes; this was downgraded two levels for very serious risk of bias and one level for serious imprecision, reflecting the small sample size and low number of events.

Lenoble 1995 did not report the numbers of participants who had one or more complications. There was a higher number of reported complications in the Kapandji group, but all of the confidence intervals for the individual complications crossed the line of no effect (Analysis 2.1). In particular, there were more Kapandji group participants with persistent symptoms related to the superficial radial nerve (8/54 versus 3/42; RR 2.07, 95% CI 0.59 to 7.34); an identical result applied for bone‐scan confirmed RSD. Lenoble 1995 suggested that the radial nerve complications were related to Kirschner wire removal; these are closer to the sensory branches of the radial nerve in Kapandji pinning. The redisplaced fracture in the trans‐styloid fixation group was treated with an external fixator. There were no cases of tendon or vascular injury or median nerve dysfunction. All four cases of pin‐track infection were superficial.

Lenoble 1995 reported there was no statistically significant difference between the two groups in either grip or pinch strength, both presented as a percentage of the uninjured wrist (one‐year results: mean grip strength 84% versus 83%; mean pinch strength 89% versus 87%). More participants of the Kapandji group were reported to suffer pain but the difference between the groups in the mean visual analogue scale values was not statistically significant (two‐year results: 7.6 versus 6.9 on a 0 (no pain) to 100 (unbearable pain) point scale). Wrist mobility was reported to be significantly better in the Kapandji group up until the removal of the K‐wires at six to eight weeks (all values as percentage of uninjured arm, flexion: 59% versus 46%; extension: 58% versus 44%; radial deviation: 60% versus 38%; ulnar deviation: 61% versus 52%; pronation: 80% versus 62%; supination: 76% versus 64%) but not subsequently. Lenoble 1995 considered that better early range of motion results for the Kapandji group related to the early mobilisation of this group and occurred at the cost of increased pain.

Similar long‐term results for anatomical outcomes in the two groups were evident from graphs presented in the trial report. The mean ulnar variance, which was positive in both groups, was approximately one millimetre greater in the Kapandji group. One of the six over‐reduced (by 15 degrees in the anterior direction) fractures in the Kapandji group developed a more extreme anterior tilt whilst the two over‐reduced fractures in the trans‐styloid group did not worsen. The four cases of DISI (dorsal intercalated segment instability) and seven cases of VISI (volar intercalated segment instability) were reported as being "evenly" distributed through the two groups.

Modified Kapandji (dorsal Kapandji wires and trans‐styloid fixation) pinning versus Kapandji intrafocal pinning

This comparison was made in Gravier 2006, a quasi‐randomised trial that reported only on radiological outcomes in 78 participants at a follow‐up of 45 days. Although the authors claimed final follow‐up was a minimum of six months, these results were not reported. The available complete radiographic data (dorsal and radial angulation) are presented in Analysis 3.1; this is very low‐quality evidence, downgraded two levels for very serious risk of bias and two levels for very serious imprecision, reflecting the small sample size and wide confidence intervals. Although the authors reported on "variance radio‐ulnaire", this was in degrees whereas ulnar variance is typically reported in millimetres; as we are uncertain what this measurement represents, we have not presented these data. Gravier 2006 also reported percentage loss of anatomical reduction at 45 days but did not provide sufficient data for us to evaluate this intermediate outcome.

Kapandji intrafocal pinning (three wires) versus Py's isoelastic pinning (two wires)

Table 4 provides details of the two trials that compared Kapandji intrafocal versus Py's isoelastic pinning in 207 people with dorsally displaced distal radius fractures (Fikry 1998; Saddiki 2012). As well as demographic differences between the two trials, it is notable that Saddiki 2012 included both intra‐ or extra‐articular dorsally displaced distal radius fractures, while Fikry 1998 excluded "comminuted" fractures.

Table 4. Kapandji intrafocal pinning (3 wires) versus Py's isoelastic pinning (2 wires)

Trial

Participants

Fracture type^a

Kapandji pinning

Py's pinning

Cast use and wire removal

Fikry 1998

110, demographics and results only for 88 who completed follow‐up: 25% female, mean age 34 years, range 18‐66

Dorsally displaced distal radius fractures, with or without separation of ulnar styloid fragment. Comminuted fractures were excluded. Classification not given

3 K‐wires inserted at fracture site. Image intensifier used

2 K‐wires inserted through radial epiphysis, across fracture and along medullary canal up to radial head. Image intensification not used: grinding of wire against the cortical bone used to determine wire placement

Wires cut below the skin, hidden ends covered by a small bore drainage tube. Forearm plaster cast for 4 weeks; received instruction for shoulder and finger exercises. Pins removed at around 8 weeks under local or regional anaesthesia

Saddiki 2012

97, 86% female, mean age 63 years

Extra‐articular fractures: associated with fracture of the ulnar styloid process = Gerard‐Marchand (GM) fracture, not associated with fracture of the ulnar styloid process = Pouteau‐Colles (PC) fracture.

Intra‐articular fractures: simple: fractures with postero‐medial fragment (PMF), complex: T‐shaped intra‐articular fractures (sagittal and/or frontal)^b .

3 K‐wires used to perform intrafocal pinning. After intraoperative reduction by close manipulation, the first K‐wire was introduced laterally while the two others were inserted postero‐laterally and postero‐medially, at a minimum 40º angle relative to the vertical axis of the radius.

2 K‐wires featuring a spatulated tip for easier progression along the medullary canal of the radius were introduced into the epiphysis of the distal radius after reduction by close manipulation, the first one being inserted from the tip of the radial styloid and the second one from inside the Lister’s tubercle. The entry point of each K‐wire was radiographically controlled. Retrograde pinning of the radius was then performed up to the sub‐chondral bone of the radial head.

Plaster splint for 3‐weeks immobilisation, then a removable splint in all cases allowing the start of daily self‐rehabilitation exercises. Pin removal after 6 weeks postoperatively

^a Available information reported
^b Results reported per fracture category (incomplete)

Fikry 1998 compared Kapandji intrafocal pinning with Py's isoelastic pinning in 110 people but presented results only for the 88 participants available at 27 months follow‐up. Saddiki 2012 reported on 97 participants of whom 64 (66%) were available at final follow‐up of 12 months. The results for the two trials are presented together in the analyses but pooling was possible for individual complications only. Notably, both trials recorded components of the Jakim radio‐clinical score (Jakim 1991), but data for pooling were not available from Saddiki 2012, which reported incomplete data by fracture subgroup. Additionally, there were contradictions between text and table for the Jakim 1991 scores in the report of Fikry 1998: thus, we have presented two sets of values in the analyses showing mean Jakim scores: the first set of values are those in the report; in the second set, the results for the two groups are reversed, based on the assumption that the rows in the table were presented in the reverse order. Neither trial reported return to former activities, quality of life, separate pain measures or patient satisfaction.

There is very low‐quality evidence, downgraded two levels for very serious risk of bias and one level for serious imprecision reflecting a wide confidence interval, from Saddiki 2012 of no or very little difference between the two groups in function assessed using the DASH (0 to 100; higher scores mean worse function): MD 4.80, 95% CI ‐5.40 to 15.00; 64 participants (Analysis 4.1). Given the estimated condition‐specific minimum detectable change for DASH is 14, these results indicated no clinically important difference between the two groups (Kleinlugtenbelt 2018). A similar finding of no between‐group difference applied to the subjective (pain and function) scores, rated with the Jakim scoring system, for Fikry 1998 as shown in Analysis 4.2.

Neither trial presented the numbers of participants who had one or more complications and clear unit of analysis issues, as exemplified in Saddiki 2012 which implied that pin migration had been sometimes associated with superficial infection and prevented the estimation of this outcome from the sum of the individual complications. Nonetheless, there were more complications after Kapandji intrafocal pinning with very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision (Analysis 4.3). In particular, there were more Kapandji group participants with secondary displacement (8/73 versus 3/79, RR 2.9, 95% CI 0.81 to 10.46) and wire displacement (10/73 versus 0/79, RR 11.87, 95% CI 1.57 to 89.61). Similar numbers of participants suffered tendon rupture or RSD in the two groups. Fikry 1998 reported all participants with ruptured tendons had successful secondary tendon transfer, and all those with RSD required prolonged rehabilitation. Saddiki 2012 did not report on the subsequent treatment of these complications but reported a remedial operation for severe malunion in one person, with group not stated. Fikry 1998 suggested that the excess of secondary fracture displacements and wire displacements in the Kapandji group was linked to posterior comminution of the original fracture. The iatrogenic fracture in the Py's pinning group of Fikry 1998 was successfully treated by a plate.

Data unavailability and discrepancies means that we are very uncertain of the relative effects of the two pinning methods on clinical outcomes. Mobility and grip strength were reported as part of the objective category (mobility, grip strength and deformity) of the Jakim scoring system in Fikry 1998, which found higher scores of uncertain clinical relevance in the Kapapandji pinning group (according to the table) or in the Py pinning group (according to their conclusions). The clinical importance of the 3‐point difference shown in Analysis 4.4 is unknown. Saddiki 2012 reported there was no statistically significant between‐group difference in grip strength at one year: 75.6% versus 79.8% of the healthy side, reported P > 0.05. Range of motion results were reported only by type of fracture in Saddiki 2012, which did not provide denominators.

Fikry 1998 reported the overall (combined function and radiological results) Jakim scores (0 to 100; best outcome), which favoured either the Kapandji or the Py pinning group; again, the clinical importance of the 6‐point difference is not known (Analysis 4.5). Fikry 1998 also graded overall outcome into four categories (excellent, good, fair, poor). There was very low‐quality evidence, downgraded two levels for very serious risk of bias and two levels for very serious imprecision of little between‐group difference, in unsatisfactory outcome, based on the numbers of participants with fair or poor overall outcome: 7/42 versus 6/46; RR 1.28, 95% CI 0.47 to 3.50 (Analysis 4.6).

Anatomical results including dorsal and radial angulation, radial length, articulation of the radio‐ulnar joint, incongruity ("step‐off") and signs of arthritis were scored according to Jakim's scoring system (Jakim 1991). Fikry 1998 claimed a superior result for the Py's pinning group, but again the mean values for the radiological section presented in the report showed the converse (Analysis 4.8). As noted above, more secondary displacement of fractures occurred in the Kapandji group (Analysis 4.3). Saddiki 2012 found small but unimportant differences between the two groups in radial and dorsal angulation at one year (Analysis 4.8).

Modified Kapandji intrafocal pinning (three wires) versus Willenegger pinning (two wires)

In Strohm 2004, modified Kapandji intrafocal pinning was compared with Willenegger pinning (two wires in the radial styloid) in 100 participants. The length of follow‐up was variable with a mean of 10 months (range six to 20). Nineteen participants were lost to follow‐up (two died) and only the remaining 81 were included by the authors in their analysis. Strohm 2004 did not report function using a PROM nor separate results for return to function, mobility, grip strength, quality of life, pain, satisfaction or radiographic findings. There was very low‐quality evidence for all reported outcomes; this was downgraded two levels for very serious risk of bias and one level for serious imprecision, reflecting the small sample size and low number of events. Additionally, the quality of the evidence for the Martini score was downgraded one level for serious indirectness in view of the difficulties in interpreting this nonvalidated composite measure.

Strohm 2004 reported that the modified Kapandji group had significantly higher modified Martini scores (this composite score included pain, subjective assessment, strength, work and sports, wrist and forearm mobility, radiological results and complications) at a median of 10 months follow‐up: 34 versus 28 on a scale 0 (worst) to 38 (best); reported P < 0.005). Strohm 2004 reported that 14 participants incurred complications but did not present separate data for the two groups. There were no significant differences between the two groups in the numbers of people with individual complications (Analysis 5.1). Two participants of the Willenegger group had an undefined "conversion procedure". Both cases of RSD resolved after treatment; and both cases of carpal tunnel syndrome were operated on. Strohm 2004 reported there was no significant difference between the two groups in operating times, but that imaging time was significantly less in the modified Kapandji group (89.5 versus 156 seconds; reported P < 0.004).

"Spring‐loaded intramedullary" pinning (not Py) versus "traditional transcortical" pinning (three wires used for both)

Shannon 2003 compared three wire pinning inserted either using "a novel spring‐loaded intramedullary technique" or in "the traditional transcortical fashion" in 46 participants, who were followed up for a minimum of six weeks. Return to prior activities, clinical outcomes, quality of life, pain or patient satisfaction were not reported. Based on an undefined "Wrist Fracture Score", Shannon 2003 reported only that functional outcome "was similar in both groups at 6 weeks". The only reported complications were two early fixation failures in the "spring‐loaded group", both in class AO C3 fractures: 2/24 versus 0/20; RR 4.60, 95% CI 0.23 to 90.84; very low‐quality evidence, downgraded two levels for very serious risk of bias and two levels for very serious imprecision (Analysis 6.1). Radiographic parameters (mean dorsal angle and mean loss of radial length) at six weeks were incompletely reported and could not be analysed.

Any technique or type of material or device used for percutaneous pinning versus any other technique or type of material or device at surgery or postoperatively

Biodegradable pins or wires versus metal pins or wires

Table 5 provides brief details of the two trials that compared biodegradable pins with Kirschner wires in 70 people with extra‐articular or intra‐articular distal radial fractures (Casteleyn 1992; Korner 1999). Despite the differences in the pinning techniques used in the trials, the results for the two trials are presented together in the analyses but no pooling has been performed. Both trials were incompletely reported and neither recorded PROMS for function, return to former activities, clinical outcomes, quality of life, pain or patient satisfaction. Length of follow‐up was one year in Casteleyn 1992 with function data unavailable for 1 of 15 and range of motion or anatomical data for 2 of 15 participants from each comparison group at final follow‐up. Korner 1999 assessed 17 of 19 participants in the biodegradable pins group and 19 of 21 in the Kirschner wire group at a median of 25.4 months.

Table 5. Biodegradable pins or wires versus metal pins or wires; participant characteristics and interventions

Trial	Participants	Fracture type^a	Type of surgery	Biodegradable pinning	Metal pinning	Cast use and wire removal
Casteleyn 1992	30, 77% female; mean age 61, range 22 to 85 years	Extra‐ and intra‐articular; Frykman I, II, V, VI (involving the DRUJ)^b	Closed reduction using finger traps. All fractures pinned percutaneously (stab incisions) using Kapandji's iintrafocal technique	Holes predrilled; 2 PGA (polyglycodic acid ‐ biodegradable) rods inserted and gently tapped home	Two 2 mm Kirscher‐wires inserted into fracture gap (one dorsally and the other laterally) and drilled into the opposite cortex of the radial shaft. Then cut to be under the skin	No cast applied. Wire removal at 6 weeks
Korner 1999	60, no other details	Extra‐ and intra‐articular; mention of AO A2, A3 and B1^c	Mention of Willenegger: probably 2 pins placed through radial styloid and across fracture	Biodegradable pins	Kirschner wires	No details

^a Available information reported
^b DRUJ: distal radial ulnar joint
^c AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)

In both trials, there was an excess of complications in the biodegradable pinning group (see Analysis 7.1); very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision. Difficulties during pin insertion in Korner 1999 (8/19 versus 0/21; RR 18.70, 95% CI 1.15 to 303.59) contributed to the longer duration of the operation in the biodegradable pinning group (mean duration: 40.5 versus 23.5 minutes). Secondary surgery was performed for the two tendon ruptures (Kirschner wiring group) and for one of the two cases of carpal tunnel syndrome; both of the latter were attributed to swelling brought on by the biodegradable rod. Additionally, one person in the biodegradable pinning group in Korner 1999 had a secondary displacement requiring revision. Painful scars, sinus (a cavity in bone or other tissue) formation and severe osteolytic reactions (9/15 versus 0/15; RR 19.00, 95% CI 1.20 to 299.63) were also related to the degradation and resorption of the biodegradable material in Casteleyn 1992. A histological examination carried out in one participant showed a nonspecific foreign‐body reaction with abundant giant cells. Most of the severe osteolytic reactions were most obvious around three to six months postoperatively, but had resolved or regressed by one year. Casteleyn 1992 pointed out that the main putative advantage of biodegradable implants, namely, that they do not need to be removed, was offset by the "management cost and inconvenience of the late inflammatory reactions".

Casteleyn 1992 assessed functional ability using an own physician‐reported scoring system that was not described in detail. Functional recovery was reported to have been quicker in the Kirschner wiring group of Casteleyn 1992, although the eventual difference at one year between the two groups (based on a 5‐point scale) was reported as not being statistically significant, and no participant had any restriction in activities of daily living (very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious indirectness).

Range of motion was satisfactory in both groups; with no statistically significant difference between them (see Analysis 7.2), with very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision. Korner 1999 did not report whether the greater loss in range of motion in the biodegradable pinning group was statistically significant: mean loss compared to contralateral wrist: 30 versus 20 degrees (extension/flexion); 35 versus 30 degrees (pronation/supination).

Final anatomical results were considered satisfactory in both groups in Casteleyn 1992, which found no statistically significant differences between the two groups (seeAnalysis 7.3).

Pinning with buried wires versus exposed percutaneous wires

Table 6 provides brief details of the three trials that evaluated the technique of burying wires versus leaving them exposed in 168 people with isolated distal radius fractures requiring Kirchner wire fixation (Hargreaves 2004; Murphy 2008; Waheed 2004). Details of antibiotic administration and wound care were provided only in Hargreaves 2004 (see Characteristics of included studies). Length of follow‐up was at around six weeks, coinciding with the return to hospital for wire removal in all three trials. None of the studies reported on participant flow, although it is probable that all participants were followed up. All three trials were incompletely reported and none recorded function, return to former activities, clinical outcomes or quality of life.

Table 6. Buried versus exposed percutaneous wires: participants and interventions

Trial	Participants	Fracture type^a	Type of surgery	Buried wiring	Exposed wiring	Cast use and wire removal
Hargreaves 2004	56, 45% female; mean age 35, range 7 to 81 years	54 closed and 2 open	Closed reduction and 2 crossed wires. Open reduction in 9 cases	Buried deep to the skin; closed with sutures. Wire ends bent over	Exposed, skin released round wire. Wire ends bent over	Forearm cast applied. Removal at 6 weeks
Murphy 2008	60, no other details	Displaced	No details	Buried	Exposed	No details. Probably removed at 6 weeks
Waheed 2004	52; 73% female; mean age 57, range 19 to 84 years	No details	No details	Buried	Wires left protruding	Cast and wires removed mean 5.8 weeks at outpatients

^a Available information reported

The available data for complications are presented in Analysis 8.1, with very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision. No pooling was possible for overall or individual outcomes. Hargreaves 2004 found more wires were infected in the exposed wires group (2/27 versus 10/29; RR 0.21, 95% CI 0.05 to 0.89); Murphy 2008 reported slightly more superficial infection in the exposed group at two weeks but not at six weeks; and Waheed 2004 reported there was no difference in infection between the two groups (reported P = 0.15). Although Waheed 2004 did not state to which group(s) the four infections occurred, exploratory analysis based on the P value indicated that it was likely that all four probably occurred in the exposed wires group. Hargreaves 2004 reported that all five wires removed early were in the exposed wires group. Murphy 2008 reported one case of superficial radial nerve damage that followed removal of a buried wire; and confirmed the absence of tendon injury.

More intensive treatment for removing buried wires were reported in Hargreaves 2004 (need for removal of wires in theatre: day surgery versus clinic) and Waheed 2004 (need for local anaesthesia): 44/52 versus 6/56; RR 7.36, 95% CI 3.54 to 15.31; 108 participants; 2 studies; low‐quality evidence downgraded two levels for serious risk of bias (Analysis 8.2).

Waheed 2004 reported, without providing data, no between‐group differences in pain during fixation or in patient satisfaction.

Any type or duration of postoperative immobilisation versus any other type or duration of immobilisation including none

Cast immobilisation of wrist in dorsiflexion versus palmar flexion after intrafocal pinning

Snow 2007 evaluated the effect of immobilising the wrist in 30 degrees dorsiflexion versus 30 degrees palmar flexion after closed manipulation and percutaneous Kirschner wire fixation in 63 participants whose fracture had redisplaced within 14 days. Snow 2007 did not report on the number of participants assessed at final follow‐up at 17 weeks; it is probable though all were present except for three participants excluded post‐randomisation because of bilateral fractures. This trial did not report a PROM for function, return to prior activities, quality of life, pain or patient satisfaction. There was very low‐quality evidence for all reported outcomes; this was downgraded two levels for very serious risk of bias and one level for serious imprecision, reflecting the small sample size and wide confidence intervals.

The available data for complications are presented in Analysis 9.1. The only notable difference between the two groups was the greater number of referrals for physiotherapy in the palmar flexion group (2/27 versus 10/33; RR 0.24, 95% CI 0.06 to 1.02). There were high rates of malunion (defined as > 10 degrees of dorsal angulation, > 15 degrees of volar tilt or > 3 mm of radial shortening compared with the opposite side) in both groups: 5/33 versus 4/27; RR 0.98, 95% CI 0.29 to 3.29. One participant with malunion in the palmar flexion group had distal radial osteotomy at 17 months. Overall, there was very low‐quality evidence of no or minimal between‐group differences at 6 or 17 weeks follow‐up in grip strength (Analysis 9.2), pinch strength (Analysis 9.3), and range of motion (Analysis 9.4), all which were reported as percentages of the contra‐lateral side; or the time taken in seconds to perform the Roylan nine‐peg dexterity test (Analysis 9.5). There was no evidence of differences between the two groups in the four reported radiographic outcomes (dorsal angle, radial angle, radial height and length) at 17 weeks as shown in Analysis 9.6.

Duration of postoperative immobilisation after percutaneous pinning

Table 7 provides brief details of the three trials that compared one week plaster cast immobilisation versus either four weeks in 50 people in Soleiman pour 2011 or six weeks in 120 people in Allain 1999 and Milliez 1992. Length of follow‐up was one year in Allain 1999 and three months for Milliez 1992 and Soleiman pour 2011. None of the trials reported a PROM for function or quality of life. Data for pooling were available for a few individual complications. Given the clear clinical heterogeneity, particularly in terms of the different methods of percutaneous pinning (trans‐styloid fixation in Allain 1999, Kapandji intrafocal pinning in Milliez 1992 and undetermined method in Soleiman pour 2011), and differences in duration and type of immobilisation, we pooled results for these outcomes on an exploratory basis.

Table 7. Duration of postoperative immobilisation after percutaneous pinning: participants and interventions

Trial	Participants	Fracture type^a	Type of surgery	Short‐term immobilisation	Longer‐term immobilisation	Immobilisation type
Allain 1999	60, 75% female; mean age 55, range 18‐87 years	AO types A2, A3, C1, C2^b	Trans‐styloid fixation: 2 wires removed at 45 days	1 week	6 weeks	Short arm circular cast
Milliez 1992	60, 73% female, age 55, range 19‐91 years	Closed displaced intra‐ and extra‐articular	Kapandji intrafocal pinning, 3 wires removed at 6 weeks	1 week	6 weeks	Plaster splint
Soleiman pour 2011	50, 46% female; no details on age	displaced metaphyseal distal radius fractures and non‐comminuted intra‐articular fractures	Not specified, Kapandji mentioned	1 week	4 weeks	Not specified

^a Available information reported
^b AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)

The available data for complications are presented in Analysis 10.1. Although overall complications were more frequently present in the early mobilisation groups of each trial, unit of analysis issues precluded meta‐analysis of this outcome. There is very low‐quality evidence, downgraded two levels for very serious risk of bias and two levels for very serious imprecision reflecting very few events, for little or no difference between the two groups in the incidence of individual complications: displaced wires or pins (6 cases in all); treatment failure (3 cases); superficial pin‐track infection (4 cases); tendon rupture (6 cases); nerve damage (4 cases); RSD (4 cases); and non‐union (3 cases) (see Analysis 10.1). In Allain 1999, surgery was necessary for the one person in the 6‐week group who had two tendon ruptures, and an anterior approach for removal of a migrated wire was required in another person in the same group. Three of the four reported cases of persistent radial nerve paraesthesia or hypoaesthesia appeared only after Kirschner wire removal. In Milliez 1992, all reported complications occurred in the early mobilisation group with the exception of one case of RSD. One of the five instances of wire displacement in the early mobilisation group in Milliez 1992 was associated with a tendon rupture, one resulted from poor technique (wiring was done by "junior" operators) and two from osteoporotic bone. In Milliez 1992, the wrists of three people with "treatment failure" were immobilised (one due to the need for tendon repair and two due to the risk of secondary displacement). Soleiman pour 2011 reported no cases of pin displacement, similar rates of pin site infection and non‐union, being the only study to report any cases of the latter.

Allain 1999 reported no significant between‐group differences in the mean duration of sick leave (which was seven weeks) for the 39 of 60 trial participants who worked, in the numbers reporting functional discomfort during domestic chores (14 of 60), or in the numbers of "sportsmen" with residual discomfort (3 of 22). One participant from the 6‐week group (who developed RSD) was the only one of 19 working people who had not returned to work by three months in Milliez 1992. Without providing numerical data for the 60 participants followed up, Soleiman pour 2011 stated that participants of the shorter immobilisation group returned to their daily activities "faster", with no differences at three months.

Very few data were available for presentation in the analyses of the clinical outcomes of grip strength and range of movement. At final follow‐up, one year in Allain 1999 and three months in Milliez 1992, the differences in grip strength, which favoured the early mobilisation group in both trials, were reported not to be statistically significant: 25 kg versus 21 kg in Allain 1999; and 55% versus 44% of the normal side in Milliez 1992. At one year follow‐up, the mean values for the six parameters of range of movement in Allain 1999 (60 participants) were between 0 and 4 degrees higher in the early mobilisation group; only the difference in ulnar deviation was reported as being statistically significant (39 versus 37 degrees; reported P = 0.03). The clinical importance of these small differences is uncertain. Milliez 1992, with 60 participants, found no statistically significant differences in flexion (74% versus 77% of the normal side) or extension (72% versus 74% of the normal side), although the results were marginally better in the group immobilised for six weeks; pronation and supination were reported as unrestricted. Soleiman pour 2011 reported there was no significant difference in "range of motion improvement" between the two groups at final follow‐up at three months but provided data only for six weeks (1.5 months) that illustrate a faster restoration of range of movement in the early mobilisation group (Analysis 10.2).

In Milliez 1992, as calculated from percentages in the trial report, fewer people in the early mobilisation groups had some degree of pain at final follow‐up (4/27 versus 10/30; RR 0.44, 95% CI 0.16 to 1.25; very low‐quality evidence downgraded two levels for very serious risk of bias and two levels for imprecision (Analysis 10.3)). Allain 1999 reported no statistically significant difference between groups in the mean pain scores (13 versus 12.5 (15 = no pain)) and use of analgesics.

Based on data from Milliez 1992, there was very low‐quality evidence of no or little between‐group difference in the numbers of people who were disappointed with their outcome: 3/30 versus 1/30; RR 3.00, 95% CI 0.33 to 27.23 (Analysis 10.4).

Similar radiological results in the two treatment groups were reported by Allain 1999 and Milliez 1992. As it is likely that the unusually small "standard deviations" provided in Allain 1999 were standard errors, the results for dorsal and radial angulation and radial length, with standard deviations calculated from the supposed standard errors, are shown in Analysis 10.5; these results are compatible with the claim of no statistical significance. The three radiological parameters listed in Milliez 1992 are defined in Characteristics of included studies. However, the results at three months seem to be for dorsal angulation (volar tilt: 6.9 versus 6.5 degrees), radial inclination (21.4 versus 21.3 degrees) and ulnar variance (1.4 mm versus 1.1 mm).

Discussion

This review covers one of the most commonly used surgical interventions for one of the most commonly sustained adult fractures. This review update doubled the number of trials from 13 in 2007 to 26. These involved a total of 1946 generally female and older adults with dorsally displaced and potentially or evidently unstable distal radial fractures.

Summary of main results

The 10 different treatment comparisons made across 26 trials are summarised below.

Percutaneous pinning versus plaster cast immobilisation alone

This comparison was tested in 11 heterogeneous trials involving 917 participants in dorsally displaced distal radius fractures. The majority of participants were female in six trials and male in two trials. As shown in Table 2, the trials employed different pinning techniques and durations of immobilisation. Nine trials (776 participants) applied across‐fracture pinning, Stoffelen 1998 used Kapandji's triple intrafocal technique with early mobilisation in 98 participants, and Delgado 2009 (43 participants) provided no details of the wiring technique. The anaesthetic/analgesic technique in the nonsurgical arm also varied, with four trials employing general anaesthesia. Two trials specified that casting was performed under fluoroscopic control; the latter was not reported in the remaining nine trials.

The main evidence for this comparison is summarised in summary of findings Table for the main comparison. All quality of the evidence for all outcomes was rated at very low, invariably reflecting serious risk of bias and imprecision, which indicates our uncertainly in the results. Thus, we are uncertain if percutaneous pinning compared with plaster cast alone makes any difference to patient‐reported function, measured using the DASH, at six weeks or six months (incomplete data from one trial, 60 participants). There is no evidence available for patient‐reported function in the long term, that is over 12 months follow‐up.

There are no data available for overall numbers of participants incurring complications or complications requiring substantive treatment. The evidence for individual complications is also incomplete. We considered the evidence for four common individual complications requiring substantive treatment: treated redisplacement, pin tract infection, Complex Regional Pain Syndrome Type 1 (CRPS‐1), and persistent finger stiffness. Redisplacement resulting in secondary treatment, reported in six studies, occurred on average in 12% (range 3.3% to 75%) of participants treated nonsurgically in cast alone, whereas pin tract infection, requiring antibiotics and, often, early wire removal, reported in seven studies, occurred on average in 7.7% (range 0% to 15%) of participants treated by percutaneous pinning. There was one reported case of persistent deep infection.

We are uncertain whether pinning affects the incidence of RSD (or CRPS‐1); although reported in four studies, the results were dominated by the very high (28%) incidence of reported short‐term RSD in one study. Although two studies found finger stiffness after cast removal was less common after surgery (20% versus 36%), the severity and treatment implications were not stated. Reported events for other outcomes were generally few and reported in single trials only. These were mainly surgery‐related. The exception was complications relating to the median nerve, such as carpal tunnel syndrome, which were fewer in the pinning group but overall uncommon, amounting to 14 cases in 383 participants (3.7%) reported in five studies.

Based on incomplete data or qualitative statements from only four studies, we are uncertain of the effects of pinning on grip strength at 12 months. We are uncertain if percutaneous pinning compared with plaster cast alone makes any difference to patient‐reported quality of life at four months; this was assessed in one study only.

Similar uncertainties, and the underlying incompleteness, apply to the evidence for other outcomes reported in one or more of the 11 studies: return to work; grip strength at different follow‐up times; range of movement at different follow‐up times; pain; composite and mainly physician‐reported functional scoring systems such as that of Gartland and Werley; patient satisfaction; and final radiographic outcomes.

Any method of percutaneous pinning versus any other method of percutaneous pinning

Kapandji intrafocal pinning (two or three wires) versus trans‐styloid fixation (two wires)

This comparison was tested in one study of 120 participants that reported results for the 96 people available for follow‐up. Those in the Kapandji pinning group were mobilised immediately whereas those in the trans‐styloid group had cast immobilisation for around six weeks. No PROM outcome for function, return to former activities, quality of life or patient satisfaction was recorded and all clinical and anatomical outcomes were incompletely reported. The number of participants with one or more complications was not available. The very low‐quality evidence and few events from one trial only means that we are uncertain of the finding of a slightly higher number of cases of superficial radial nerve symptoms and of bone‐scan confirmed RSD in the Kapandji pinning group. There is very low‐quality evidence of no difference between the two pinning methods at 12 months in grip and pinch strength, in pain or in radiological outcomes.

Kapandji intrafocal pinning versus modified Kapandji (dorsal Kapandji wires and trans‐styloid fixation)

This comparison was tested in one study(Gravier 2006) of 85 participants, that reported only on limited radiographic outcomes at 45 days post‐surgery. Any reported differences in measured dorsal and radial angulation were small and not clinically meaningful. This very low‐quality evidence gives no basis for drawing conclusions on the relative effects of the two modes of fixation.

Kapandji intrafocal pinning (three wires) versus Py's isoelastic pinning (two wires)

This comparison was tested in two poorly reported studies, with marked differences in their population characteristics. Outcome data were available for a maximum of 152 participants of the 207 recruited into the two trials. Neither trial reported return to former activities, pain, quality of life or patient satisfaction and pooled data were available only for a few individual complications. There is very low‐quality evidence of no difference between the two groups in function at 12 months measured via the DASH or at an average of 27 months measured via the subjective component of the Jakim score. Although the number of participants with one or more complications could not be determined, there is very low‐quality evidence of a higher incidence of some complications, such as secondary displacement and wire migration, after Kapandji pinning. There is incomplete and, for one trial, contradictory evidence on clinical outcomes, primarily grip strength and range of motion.

Modified Kapandji intrafocal pinning (three wires) versus Willenegger pinning (two wires)

This comparison was tested in one incompletely reported study for which data for presentation in the analyses were available only for individual complications in 81 participants. There is very low‐certainly evidence of little difference between the two pinning methods in individual complications. The incomplete data for a nonvalidated composite score hinders interpretation of the reported superior mean score in the modified Kapandji pinning group at final follow‐up, based on very low‐quality evidence.

"Spring‐loaded intramedullary" pinning (not Py) versus "traditional transcortical" pinning (three wires used for both)

This comparison was tested in one study involving 46 participants that was reported in a conference abstract only. We reluctantly included this study as it met the inclusion criteria but, in particular, there was incomplete information on the interventions and outcome measurement, and the usable data for the few reported outcomes were limited to two early fixation failures (8.3%) in the spring‐loaded intramedullary pinning group. Aside from the very low quality of the available evidence, there is nothing that we could conclude about this comparison.

Any technique or type of material or device used for percutaneous pinning versus any other technique or type of material or device at surgery or postoperatively

Biodegradable pins or wires versus metal pins or wires

The use of biodegradable pins was tested in two small trials using two very different pinning techniques. There was insufficient evidence to detect a difference in anatomical and clinical outcomes between the two groups and neither trial recorded PROMs for function, return to former activities, clinical outcomes, quality of life, pain or patient satisfaction. Although there is very low quality of evidence, what is apparent are the extra demands at surgery of insertion of biodegradable pins, and the excess of complications associated with biodegradable material. Thus, the recommendation, expressed by both studies, not to use biodegradable pins for these fractures (in adults) seems justified. This conclusion could be have been moderated by further evidence from a multicentre trial involving 115 participants in 'Studies awaiting assessment' (Russe 2000); however, further information about this study, which was last reported 19 years ago in a series of abstracts, is unlikely to now emerge.

Pinning with buried wires versus exposed percutaneous wires

The technique of burying wires versus leaving them exposed was tested in three small poorly‐reported trials involving 168 participants. Our attempts to obtain more information, including the full reports of Murphy 2008; and Waheed 2004, were unsuccessful. None of the trials reported on function, clinical outcomes, return to former activities or quality of life. We are uncertain of the finding of a lower risk of superficial infection with buried wires reported in one trial; and possible in the other two trials (very low‐quality evidence). There is low‐quality evidence that burying of wires may be associated with a higher risk of requiring more invasive treatment for wire removal, such as it being performed in an operating theatre rather than a clinic, or requiring the use of local anaesthetic generally in an operating theatre.

Any type or duration of postoperative immobilisation versus any other type or duration of immobilisation including none

Cast immobilisation in wrist dorsiflexion versus wrist palmar flexion after intrafocal pinning

This comparison was tested in one study reporting limited results for 60 people whose fracture had redisplaced within 14 days and treated via intrafocal pinning. Snow 2007 did not report on function, return to prior activities, quality of life, pain or patient satisfaction. This study found more participants were referred for physiotherapy in the palmar flexion group (very low‐quality evidence), but otherwise there is very low‐quality evidence of little or no differences between the two groups for other complications. The same applied to clinical outcomes (grip and pinch strength, range of wrist motion, time to complete a dexterity test) at 6 and 17 weeks and radiographic outcomes at 17 weeks. Overall, we are uncertain of the findings of this incompletely reported trial; this uncertainty means that no recommendation for or against placing the wrist in dorsiflexion can be made.

Duration of postoperative immobilisation after percutaneous pinning

Three small incompletely‐reported trials, two of which were quasi‐randomised, compared cast immobilisation for one week (early mobilisation) versus four or six weeks after percutaneous pinning in 170 people. While the pinning technique was undefined in one trial, the other two trials used very different pinning techniques, respectively trans‐styloid fixation and Kapandji intrafocal fixation, an important source of clinical heterogeneity. There are insufficient data for subgroup analysis to check this, and indeed only data for individual complications were available for pooling, which was done on an exploratory basis in light of the clinical heterogeneity. None of the three trials reported a PROM for function or quality of life.

The very low‐quality evidence means we are uncertain of the potential finding of more complications following early mobilisation. The latter is based primarily on a higher number of individual complications only, in Milliez 1992, where there were five cases (17% of 30) each of displaced wires and of tendon rupture, and three cases (10%) of treatment failure in the early mobilisation group after Kapandji pinning. There is very low‐quality evidence of no clinically important between‐group differences in grip strength or range of movement at final follow‐up (3 or 12 months). There is incomplete and very low‐quality evidence of little or no between‐group difference in return to work or former activities. There is very low‐quality evidence of little or no between‐group difference in pain, participant dissatisfaction with outcome or anatomical results at final follow‐up.

Any method or timing of pin or wire removal versus any other method or timing of pin or wire removal including no removal

There are no included trials on this topic.

Overall completeness and applicability of evidence

Completeness of the evidence

This updated review includes 26 trials that recruited 1946 people with distal radius fractures. This is double the number of trials and just over double the number of participants since the first version of the review (Handoll 2007). Despite this increase, the incompleteness of the available evidence in terms of outcomes reported and available data remains noteworthy. In particular, there were minimal data available for the six included studies reported in conference abstracts only; one of these provided the sole evidence for one comparison (Shannon 2003). We were unsuccessful in obtaining further information on these six trials and, overall, our requests for additional unpublished information were successful for two trials only (Gupta 1999; Zyluk 2007). The included trials tested 10 treatment comparisons, with 11 trials comparing pinning with nonsurgical treatment alone in 917 people. The maximum number of participants in any pooled analysis for this review was 591 (64% of 917) for redisplacement requiring secondary treatment for the pinning versus nonsurgical treatment comparison. Notably, only 3 of 26 trials reported on a validated patient‐reported outcome measure of function (DASH) and, of these, only Saddiki 2012 (88 participants) provided data that could be presented in a forest plot (Analysis 4.1).

Results of our formal assessment of applicability

To inform consideration of applicability of the evidence from individual trials, we provide quite extensive details in the Characteristics of included studies on the study populations and interventions; as well as providing summaries of these in additional tables by comparison. Additionally, Table 8 shows our assessments for each trial of four aspects of relevance to ascertaining external validity: definition of the study population, description of the interventions, definition of the main outcome measures and length of follow‐up. Incomplete descriptions of study inclusion criteria (15 trials) and interventions (8 trials) are clearly unhelpful. It is notable that two trials where the study population was not clearly defined ('No') and five of the six trials where the interventions were not sufficiently described ('No') were for trials that were only reported in conference abstracts. The main outcome measures were sufficiently described in only nine trials. In eight trials, only three of which were reported in conference abstracts only, the description of the main outcomes was considered wholly insufficient ('No') to be certain what was being measured. The eight trials that had only short‐term follow‐up, thus within three months of randomisation, were considered not to have appropriate timing of outcome measurement. This is particularly the case for five trials that reported outcomes at around six weeks, corresponding to the time of wire removal (Gravier 2006; Hargreaves 2004; Murphy 2008; Shannon 2003; Waheed 2004).

Table 8. Assessment of items relating to applicability of trial findings

Study ID	Clearly defined study population?	Interventions sufficiently described?	Main outcomes sufficiently described?	Appropriate timing of outcome measurement? (Yes = ≥ 12 months)
Allain 1999	Yes	Yes	Partial: Inadequate definition of discomfort during activity; and CRPS‐1^c	Yes: 12 months
Azzopardi 2005	Yes	Yes	Yes	Yes: 12 months
Casteleyn 1992	Partial: Insufficient information on exclusions	Yes	Partial: Poor description of functional grading tool	Yes: 12 months
Delgado 2009^a	Partial: “Heavy‐labour workers with unstable fractures of distal radius.” Only description of participants	No: No details of type of wiring or cast or duration of immobilisation	Partial: Some outcomes types reported, not defined	Yes: 12 months
Fikry 1998	Yes: However, a young population.	Yes	No: Not defined, just reported.	Yes: 20 months minimum
Gravier 2006	Partial: Insufficient information on displacement criteria for surgery.	Yes	No: Radiological outcomes only; lack of clarity over one outcome	No: 45 days (data for)
Gupta 1999	Partial: Colles radius fracture in the skeletally mature	Yes	Partial: Inadequate description of assessment of outcomes, including complications. Crude categorisation of functional scores	Partial: 6 months
Hargreaves 2004	Partial: No clear information on indications for surgery.	Yes	Yes	No: 6 weeks
Korner 1999^a	No: Distal radius fracture only.	No: Composition of bioabsorbable pins not given; no detail on modification to Wellenegger technique	No: Not defined, just reported. Inadequate descriptions of redisplacement and arthrotic findings	Partial: although median follow‐up was 25.4 months, some participants may not have been followed up for 12 months.
Lenoble 1995	Yes	Yes	Yes	Yes: 24 months
Mardani 2011	Partial: Not explicitly dorsally displaced fractures though implied by the study context i.e. reference to dorsal tilt and dorsal comminution	No: Inadequate information about the pin configuration, no broad description of the technique or approach used. Also no descriptors of cast positions or type or mode of application, bar “short” and "long arm cast”. No information as to if or when intraoperative imaging was used. Care provider was the same surgeon in both cases, assuming the lead surgeon but not clear	Partial: Primary outcome not determined. Lack of definition of finger stiffness. Loss to follow‐up taken to be a positive outcome, when 5 visits were scheduled for all participants. Failure of treatment (need for further intervention) not treated as an outcome parameter	No: 3 months
Milliez 1992	Yes	Partial: Not enough detail on early mobilisation	Partial: Inadequate description of CRPS‐1 and crude measure of pain.	No: 3 months
Murphy 2008^a	Partial: “Patients with a distal radial fracture managed with percutaneous wire fixation and casting only.” Minimal information available. By deduction, probably adults	No: Interventions were only mentioned and not described. Care providers not defined	Partial: Outcome assessment appeared limited to K‐wire‐related complications – some detail provided on assessment	No: 6 weeks
Rodriguez‐Merchan 1997	Partial: Other than fracture type, no exclusion criteria although age range provided.	Yes	Partial: Inadequate description of assessment of clinical union, radiographic measures, and complications	Yes: 12 months
Saddiki 2012	Yes	Yes	Yes	Yes: 12 months
Shankar 1992	Partial: Intra‐articular fractures but limited description of population.	Yes	Yes: Descriptions given of outcome measures, although crude measurement of subjective outcomes within composite outcome measure	Partial: 6 months
Shannon 2003^a	No: “Patients with unstable fractures of the distal radius were entered into the study.” No report of exclusion criteria or description of fracture type	No: Minimal information on intervention technique	No: No definition of outcome measures	No: 6 weeks
Snow 2007	Yes	Yes	Yes: Main outcome criteria (radiographic) were clearly defined	Partial: 17 weeks (4.25 months)
Soleiman pour 2011	Partial: No indication of participant age. No indication of type of pinning	Partial: No detail on method(s) of pinning nor immobilisation; no identification of care providers	Partial: Insufficient definitions provided for outcomes; but only part translation obtained	No: 3 months
Stoffelen 1998	Yes	Yes	Yes: Descriptions given of outcome measures, although crude measurement of subjective outcomes within composite outcome measure	Yes: 12 months
Strohm 2004	Partial: Colles type and AO^b classification but no exclusions	Yes	Partial: Inadequate description of assessment of outcomes, including complications	Partial: range 6 to 20 months
Venkatesh 2016	Partial: Not explicitly dorsally displaced fractures though implied by the study context i.e. reference to Colles' fracture, application of Colles' cast and restoration of volar tilt	Yes	Yes	Partial: 6 months
Verhulst 1990^a	Partial: Colles' radius fracture only	Partial: No information on postoperative care or type of immobilisation	Partial: Inadequate descriptions of outcome assessment	Yes: 2 years
Waheed 2004^a	Partial: “Patients with a distal radial fracture managed with percutaneous wire fixation and casting only.”	No: Intervention not described	No: No definition of outcome measures	No: Most likely up to 6 weeks
Wong 2010	Yes	Yes	Yes	Yes: minimum 13 months
Zyluk 2007	Partial: Inclusion criteria inadequately defined: “displaced isolated fractures”. Exclusion criteria listed: “… excluded people who did not meet the prerequisites, i.e. they had a fracture without displacement, open, with associated injuries of other structures, or did not agree to participate in the study”.	Yes	No: "The rate of secondary displacement was considered a primary outcome measure.” Secondary displacement not defined, some participants with secondary displacement excluded from analysis (if they required further surgery within 2 weeks) whereas others were analysed (if treated operatively and did not require further surgery).	Partial: 6 months

^a Reported in conference abstracts only
^b AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)
^c CRPS‐1: Complex Regional Pain Syndrome type 1

Applicability of the evidence

Population

Where stated, the predominantly female sex of most trial populations indicates these were generally representative of the people who sustain these injuries in industrialised societies. Some exceptions include those trials which included more male participants, such as Delgado 2009 which focused on heavy labour workers in Spain. Assessment of applicability based on the fracture population was less straightforward, although it seems likely that the fractures were predominantly dorsally displaced. The variety of fracture classification systems, with associated issues of reliability and validity, complicates this area (Jupiter 1997). For example, the two fracture classifications used by trials in this review (the AO and Frykman) place different emphases on various fracture patterns and anatomical components. Studies have revealed unsatisfactory interobserver reliability and intraobserver reproducibility for both classification systems (Andersen 1996; Kreder 1996b), and neither was useful for predicting clinical outcome (Flinkkila 1998). Four trials in this review stipulated criteria for anatomical displacement of the fracture for trial entry. However, Kreder 1996a found quite broad margins of error ("tolerance limits") for anatomical measurements in general. This, along with other factors, such as variations in anatomical reference points, again hinders treatment comparison, both in terms of defining the population and when used as criteria for determining secondary displacement. A recent Delphi study including 43 expert surgeons indicated some agreement on radiological thresholds for surgical intervention; these were greater in people over 75 years (Johnson 2019a). However, pre‐injury function was the most important factor influencing the surgeon's decision, which is harder to assess in this review.

Fracture instability was the inherent or explicit criterion for many of the included trials, but it is not established how best to predict this. A study of 4024 patients concluded that patient age, metaphyseal comminution of the fracture and ulnar variance were the most important factors in predicting instability of distal radial fractures (MacKenney 2006). This unvalidated prognostic model is too complex for use in routine clinical practice (Downing 2008). Ultimately, a direct association between radiological parameters and patient‐rated outcome has yet to be established (Downing 2008; Johnson 2019b).

Interventions

Where details are provided, the variation in the interventions for the different comparisons is evident from the Additional Tables. For the pinning versus cast‐only comparison, this variation is notable not only for the primary interventions but also the co‐interventions, where described. While we anticipate most interventions are available and in use somewhere in the world, likely exceptions in terms of general applicability are the spring‐loaded intramedullary pinning technique tested in Snow 2007, the pinning technique described in Wong 2010, and the use of biodegradable pins or wires. Usually cast immobilisation was applied to both groups for comparable durations, around five to eight weeks, of the trials in this comparison. Notable exceptions were the pinning group of Wong 2010, where there was immediate mobilisation of the pinning group with a removable palmar splint used for resting, and Stoffelen 1998 where cast immobilisation was one week in the pinning group compared with six weeks in the nonsurgical group.

Outcomes

Although some improvement was notable in more recently conducted trials, overall there was inadequate assessment of outcome, particularly of patient‐reported function and long‐term outcome. Our restructuring of Types of outcome measures in this update gives prominence to validated patient‐rated assessment of function instruments such as the Patient‐Rated Wrist Evaluation (PRWE) and the Disability of the Arm, Shoulder, and Hand questionnaire (DASH) (MacDermid 2000). This helped highlight the fact that only three trials reported this outcome. Trials, however, continue to report nonvalidated and composite outcome measures, such as those based on the Gartland and Werley scoring system (Gartland 1951), which combine aspects of function, pain, deformity and complications. These are particularly crude indicators of outcome and considerable caution is needed in their interpretation, even more so when the scores have been categorised.

Lastly, only two trials reported on health‐related quality of life.

Quality of the evidence

Where data were available, the quality of the evidence for all outcomes in all comparisons was either low or, most often, very low.

We downgraded all evidence for risk of bias, which we considered either serious or very serious. In particular, this reflects the susceptibility to performance and detection biases for most outcomes where blinding to the allocated intervention is not possible. Other common sources of bias were selection bias (allocation concealment was deemed secure in just one trial) and selective reporting bias.

For some outcomes, we downgraded the quality for indirectness. This was always in relation to outcome assessment or presentation, such as inappropriate categorisation of outcomes.

We did not downgrade specifically for inconsistency. This reflected on the lack of data for pooling of most outcomes and the fact that the quality of the evidence had already been downgraded to very low in relation to risk of bias and imprecision.

Downgrading for imprecision was common. This reflected wide confidence intervals but also the problems relating to small sample sizes and typically small numbers of events. The need to be wary of the results from small single‐centre trials was often behind a decision to downgrade for imprecision. However, we did not apply this indiscriminately but considered the results of single trials in the context of properties of the measure used, including distinguishing between binary and continuous measures, and if there was other corroborative evidence.

We did not downgrade for publication bias. Constructing funnel plots to explore the possibility of publication bias was not viable, given the few trials. This does not, however, mean that we can discount the possibility of publication bias. The fact that nearly a quarter of the included trials were incompletely published as conference abstracts also points to significant concerns about publication bias.

Potential biases in the review process

Before we started our review update, we revised our protocol methods, including restructuring Types of outcome measures and identifying the main outcomes for presentation in 'Summary of Findings' tables, in accordance with the contemporary standards. We have noted the main changes in methods from the last version of the review under Differences between protocol and review.

Although we avoided bias by prespecifying our main outcomes, the selection of these, which is formally but not rigidly set at a maximum of seven, was a source of extensive discussion. Missing a key outcome of primary importance to people incurring these fractures is of concern. Potentially, one such outcome is return to former activities. However, this outcome is problematic because it can be measured in various ways, such as return to former work, duration of sick leave, and return and time to return to former daily activity. Additionally, the activity in question may only apply to a subgroup of participants or be influenced by different factors and circumstance such as, for employment, insurance and workers' compensation schemes, statutory sick‐pay and transport. We thus decided against including this as a main outcome.

Our search for trials was comprehensive, and screening and study selection were performed systematically and according to protocol. As shown by our identification of six unpublished trials, the possibility of other unpublished trials, such as conference abstracts, remains. However, based on our experiences of failing to obtain more data on unpublished trials, we anticipate the potential contribution of these, if included, to the evidence is likely to be very limited and thus we do not think this is an important source of bias. Our attempts to obtain further data and information on published trials were also largely unsuccessful. It is inevitable then that our appraisal of the evidence is one that reflects the quality of the reporting and not necessarily the actual conduct of the included trials.

Whether from the lack of information on or clear differences in trial characteristics, some of our decisions to place trials under the same comparison and then, where outcomes were in common, pool data can be open to question. There were insufficient data for subgroup analyses to examine decisions based on key characteristics. As well as being cautious in our interpretation of the evidence, we have provided tables of key characteristics of participants and interventions for each comparison that allow the reader to judge the key areas of clinical heterogeneity.

We were mindful of potential unit of analysis problems and took a cautious approach particularly when presenting data for the numbers of participants with any complication. The possibility that participants could have more than one complication is illustrated by Strohm 2004, which reported that 14 of the 81 participants at follow‐up incurred complications but provided data for individual complications by group; these totaled 29 complications.

Lastly, GRADE is a blunt instrument and the downgrading by whole levels can rapidly result in a very low‐quality rating depicting "uncertainty about the estimate". This indeed applied to the evidence available for our largest comparison (pinning versus cast‐only). Although it is possible that our judgements were too severe, it is also the case that definitive evidence was not available for any of the comparisons.

Agreements and disagreements with other studies or reviews

In Why it is important to do this review, we pointed to the findings of DRAFFT in support for updating this review on percutaneous pinning: "Contrary to the existing literature, and against the rapidly increasing use of locking plate fixation, this trial found no difference in functional outcome in patients with dorsally displaced fractures of the distal radius treated with Kirschner wires or volar locking plates. Kirschner wire fixation, however, is cheaper and quicker to perform." (DRAFFT 2014). DRAFFT was aimed at dorsally displaced distal radius fractures in adults whose fractures, including simple articular fractures, could be reduced by closed manipulation. Based on English Hospital Episode Statistics data, Costa 2016 reported that in the five years prior to DRAFFT, 75% of patients were treated with plate fixation versus 12% with percutaneous pinning. Subsequent to the trial publication, the reported figures were 48% having plate fixation and 42% having K‐wire fixation; the proportion of other procedures stayed the same.

DRAFFT was a pragmatic trial where the size and number of wires, the insertion technique and the configuration of wires was left to the discretion of the surgeon as per their normal practice. This reliance on methods which the surgeon is familiar and experienced with is appropriate where, as seen in our review, there is a lack of evidence to inform the selection of any particular pinning technique. DRAFFT 2014 provided information on number of wires used (predominantly two or three); wire size (almost all 1.6 mm); and the technique: Kapandji (27%); 'interfragmentary' (38%) and mixed techniques (35%). The latter data illustrate that both Kapandji and transfixation approaches are still distinct methods employed in UK practice, but also that surgeons often also use combination techniques.

We do not think it serves any purpose to compare this unique review with other systematic reviews covering the management of fractures of the distal radius. A summary of the characteristics of 41 systematic reviews published up to May 2017 on treatments for fractures of the distal radius provided in Table 1 of Belloti 2019, confirms that our review is the only one looking at all aspects of percutaneous pinning. Mellstrand Navarro 2019, a more recently published broad intervention systematic review that confines its focus to the 'elderly' (study populations of mean age 60 years), included only two trials comparing pinning versus plaster cast alone (Azzopardi 2005; Wong 2010).

$Example of a dorsally displaced "Colles'" distal radius fracture$

Figure 1

Example of a dorsally displaced "Colles'" distal radius fracture

Figure 2

Study flow diagram for updated review

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 4

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.


Study	Pinning	Cast only	Statistically significant?
At 1.5 months
Zyluk 2007	Median, range, no. 85, 30‐125, 30	Median, range, no. 97, 50‐142, 30	No
At 3 months
Zyluk 2007	Median, range, no. 67, 41‐101, 30	Median, range, no. 84, 40‐128, 30	No
At 6 months
Zyluk 2007	Median, range, no. 48, 32‐94, 30	Median, range, no. 64, 40‐119, 30	Yes

Analysis 1.1

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 1 DASH scores (30 to 150; worst disability).

Analysis 1.2

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 2 DASH (30 to 150; worst outcome) at 6 months: reduced function indicated by > 60 score.

Analysis 1.3

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 3 Complications.

Analysis 1.4

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 4 Non return to work (labourers).


Study	Pinning	Cast only	Statistically significant?
At 1.5 months
Zyluk 2007	Median, range, no. 7.2 kg, 0‐56 kg, 30	Median, range, no. 6.0 kg, 2‐15 kg, 30	No
At 3 months
Zyluk 2007	Median, range, no. 15.0 kg, 10‐46 kg, 30	Median, range, no. 10.0 kg, 2‐20 kg, 30	Yes
At 6 months
Zyluk 2007	Median, range, no. 18.2 kg, 12‐48 kg, 30	Median, range, no. 16.0 kg, 6‐26 kg, 30	Yes

Analysis 1.5

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 5 Global grip strength (kg).


Study	Pinning	Cast only	Statistically significant?
At 1.5 months
Zyluk 2007	Median, range, no. 31%, 0‐86%, 30	Median, range, no. 27%, 10‐55%, 30	No
At 3 months
Zyluk 2007	Median, range, no. 61%, 41‐88%, 30	Median, range, no. 41%, 7‐86%, 30	Yes
At 6 months
Zyluk 2007	Median, range, no. 72%, 57‐94%, 30	Median, range, no. 61%, 30‐124%, 30	Yes

Analysis 1.6

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 6 Global grip strength [probably] relative to other side (%).

Analysis 1.7

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 7 Under half grip strength at 6 months.

Analysis 1.8

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 8 Grip strength at 13 to 24 months (kg).

Analysis 1.9

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 9 Range of movement at 6 months.


Study	Pinning	Cast only	Statistically significant?
Flexion and extension at 1.5 months
Zyluk 2007	Median, range, no. 88º, 64º to 126º, 30	Median, range, no. 81º, 17º to 130º, 30	No
Flexion and extension at 6 months
Zyluk 2007	Median, range, no. 143º, 109º to 190º, 30	Median, range, no. 133º, 81º to 170º, 30	No
Pronation and supination at 1.5 months
Zyluk 2007	Median, range, no. 110º, 11º to 164º, 30	Median, range, no. 117º, 9º to 160º, 30	No
Pronation and supination: at 6 months
Zyluk 2007	Median, range, no. 142º, 84º to 164º, 30	Median, range, no. 138º, 61º to 175º, 30	No

Analysis 1.10

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 10 Range of motion (data from Zyluk 2007).

Analysis 1.11

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 11 Range of movement at 13 to 24 months.

Analysis 1.12

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 12 Short Form‐36.

Analysis 1.13

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 13 Overall quality of life (1 very poor to 5 very good) at 13 to 24 months.

Analysis 1.14

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 14 Pain (occasional).

Analysis 1.15

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 15 Patient satisfaction at 12 wks: excellent result (no pain, disability or motion limitation).

Analysis 1.16

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 16 Patient satisfaction at 13 to 24 months (1 to 4, lower score equals higher satisfaction).

Analysis 1.17

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 17 Deformity (clinical and radiological).

Analysis 1.18

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 18 Composite score functional grading: fair or poor.

Analysis 1.19

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 19 Mayo wrist score at 13 to 24 months (0 to 100; worst outcome).

Analysis 1.20

Comparison 1 Percutaneous pinning versus plaster cast, Outcome 20 Anatomical measurements.

Analysis 2.1

Comparison 2 Kapandji intrafocal pinning (2 or 3 wires) versus trans‐styloid fixation (2 wires), Outcome 1 Complications.

Analysis 3.1

Comparison 3 Modified Kapandji (dorsal Kapandji wires and trans‐styloid fixation) versus Kapandji intrafocal pinning, Outcome 1 Anatomical measurements (at 45 days).

Analysis 4.1

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 1 DASH score at one year (0: worst disability to 100: no disability).

Analysis 4.2

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 2 Subjective results: pain and function (normal = 30 points); from Jakim score.

Analysis 4.3

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 3 Complications.

Analysis 4.4

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 4 Objective results: mobility, grip strength, deformity (normal = 30 points); from Jakim score.

Analysis 4.5

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 5 Overall results: radiological, subjective and objective (normal = 100 points); from Jakim score.

Analysis 4.6

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 6 Overall outcome grades.

Analysis 4.7

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 7 Radiological results (normal = 40 points); from Jakim score.

Analysis 4.8

Comparison 4 Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires), Outcome 8 Anatomical measurements at one year.

Analysis 5.1

Comparison 5 Modified Kapandji pinning (3 wires) versus Willenegger pinning (2 wires), Outcome 1 Complications.

Analysis 6.1

Comparison 6 “Spring‐loaded intramedullary" pinning (not Py) versus transcortical pinning, Outcome 1 Complications.

Analysis 7.1

Comparison 7 Biodegradeable pins versus Kirschner wires, Outcome 1 Complications.

Analysis 7.2

Comparison 7 Biodegradeable pins versus Kirschner wires, Outcome 2 Range of movement.

Analysis 7.3

Comparison 7 Biodegradeable pins versus Kirschner wires, Outcome 3 Anatomical measurements.

Analysis 8.1

Comparison 8 Buried wires versus exposed percutaneous wires, Outcome 1 Complications.

Analysis 8.2

Comparison 8 Buried wires versus exposed percutaneous wires, Outcome 2 More invasive treatment for wire removal.

Analysis 9.1

Comparison 9 Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning), Outcome 1 Complications.

Analysis 9.2

Comparison 9 Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning), Outcome 2 Grip strength (as percentage of other side minus 30% for nondominant side).

Analysis 9.3

Comparison 9 Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning), Outcome 3 Pinch strength (as percentage of contralateral minus 30% for nondominant side)(.

Analysis 9.4

Comparison 9 Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning), Outcome 4 Range of motion (as percentage of contralateral normal side).

Analysis 9.5

Comparison 9 Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning), Outcome 5 Rolyan nine hole peg test of dexterity (seconds).

Analysis 9.6

Comparison 9 Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning), Outcome 6 Anatomical measurements at week 17.

Analysis 10.1

Comparison 10 Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery, Outcome 1 Complications.

Analysis 10.2

Comparison 10 Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery, Outcome 2 Range of movement (at 1.5 months; interim results).

Analysis 10.3

Comparison 10 Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery, Outcome 3 Pain at 3 months: usually or during effort.

Analysis 10.4

Comparison 10 Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery, Outcome 4 Patient dissatisfaction with outcome.

Analysis 10.5

Comparison 10 Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery, Outcome 5 Anatomical measurements (at 1 year): standard errors conversion.

Summary of findings for the main comparison. Summary of findings: percutaneous pinning versus cast or brace immobilisation only

Percutaneous pinning compared with cast or brace immobilisation only for treating distal radius fractures in adults
Patient or population: adults with dorsally displaced distal radius fractures Settings: emergency department, hospital Intervention: reduction and percutaneous pinning, usually supplemented by cast immobilisation Comparison: reduction and cast (or brace) immobilisation alone
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	cast only	percutaneous pinning
Short‐term patient‐reported hand/wrist/upper limb function. DASH scores (30 to 150; worst disability) At 1.5 months	Median 97 (study result)	Median 85 (study result)	Not available	60 (1 study)	⊕⊝⊝⊝ very low^a	There are very limited data for this outcome. The study reported the difference was not statistically significant.
Medium‐term patient‐reported hand/wrist/upper limb function. DASH scores (30 to 150; worst disability) At 6 months	Median 84	Median 67	Not available	60 (1 study)	⊕⊝⊝⊝ very low^a	There are very limited data for this outcome. The study reported the difference was statistically significant. However, it may not be clinically important.^b Another study (43 participants) reported similar functional results in the two groups, probably at 12 months.
Long‐term patient‐reported hand/wrist/upper limb function. Over 12 months follow‐up	See comment	See comment	‐	‐	‐	This was not reported in any of the 11 studies (917 participants) for this comparison.
Number of people incurring one or more complications requiring substantive treatment	See comment	See comment	‐	799 (9 studies)	⊕⊝⊝⊝ very low^c	Data are available for the most common individual complications in this category: treated redisplacement, pin tract infection, CRPS type 1 and persistent finger stiffness.^d
Overall number of people incurring one or more complications	See comment	See comment	‐	799 (9 studies)	⊕⊝⊝⊝ very low^c	As well as those complications listed above, data were available for other surgery‐related complications (e.g. K‐wire migration) and other complications occurring in both groups, particularly those related to the median nerve (e.g. carpal tunnel syndrome).^e
Grip strength At 12 months	See comment	See comment	‐	238 (4 studies)	⊕⊝⊝⊝ very low^f	None of the four studies provided data for pooling. Results were reported as similar in the two groups in three trials and favouring pinning in the fourth trial.
Medium‐term quality of life measured via the Short Form 36 ‐ Physical score (0: worst to 100: best health) At 4 months	Mean SF‐36 in the plaster cast group was 38.2	Mean SF‐36 in the intervention group was 4.00 higher (1.59 lower to 9.59 higher)		54 (1 study)	⊕⊝⊝⊝ very low^g	A similar lack of between‐group difference was found for mental scores (MD 0.62, 95% CI ‐5.32 to 6.56).
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval;MD: Mean difference; RR: Risk Ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
^a Very low‐quality evidence downgraded two levels for very serious risk of bias and one level for serious imprecision. ^b Based on a scoring scheme of 0 to 100 (worst outcome), a minimal clinically important difference (MCID) of 15 is recommended in DASH/QuickDASH; this translates to 18 for a scale of 30 to 150; which is greater than the 16 difference in median scores at six months. ^c All the evidence for individual complications was considered very low quality, downgraded two levels for very serious risk of bias and one or two levels for serious or very serious imprecision. ^d Redisplacement resulting in secondary treatment only occurred in the nonsurgical treatment group: 38/316 (12%); range 3.3% to 75% (reported in 6 studies). Pin tract infection requiring antibiotics and, often, early wire removal only occurred in the pinning group: 22/285 (7.7%); range 0% to 15% (reported in 7 studies). There was one case of persistent deep infection. Incidence of RSD (nowadays, complex regional pain syndrome type 1), reported in 4 studies, was similar in the two groups: 13/123 (10.6%) versus 17/125 (13.6%); this was dominated by the results for short‐term RSD in one study: 12/48 versus 15/50. Persistent finger stiffness was less common after pinning versus cast alone after cast removal or at 12 weeks: 29/147 versus 53/149; RR 0.52, 95% CI 0.36 to 0.76; 2 studies; the treatment implications of this were not stated. ^e Reported events for other outcomes were generally few and reported in single trials only. These were mainly surgery‐related. The exception was complications relating to the median nerve, such as carpal tunnel syndrome, which occurred in both groups but were uncommon overall, amounting to 14 cases in 383 participants (3.7%) reported in five studies. ^f The quality of the evidence was very low given the qualitative and incomplete nature of the reported results and the very serious risk of bias. ^g Very low quality, downgraded one level for serious risk of bias and two levels for very serious imprecision.

Summary of findings for the main comparison. Summary of findings: percutaneous pinning versus cast or brace immobilisation only

Table 1. Definition of radiological parameters

Parameter	Definition (Radiographic view)	Approximate normative values
Dorsal angulation (dorsal or volar or palmar tilt)	Angle between a) the line which connects the most distal points of the dorsal and volar cortical rims of the radius and b) the line drawn perpendicular to the longitudinal axis of the radius Lateral view	11–12 degrees, where positive values indicate a palmar angulation and negative values indicate angulation dorsally beyond neutral
Radial length	Distance between a) a line drawn at the tip of the radial styloid process, perpendicular to the longitudinal axis of the radius and b) a second perpendicular line at the level of the distal articular surface of the ulnar head Posteroanterior (PA) view	11‐12 mm
Radial angle or radial inclination	Angle between a) the line drawn from the tip of the radial styloid process to the ulnar corner of the articular surface of the distal end of the radius and b) the line drawn perpendicular to the longitudinal axis of the radius. Posteroanterior (PA) view.	22‐23 degrees
Ulnar variance	Vertical distance between a) a line drawn parallel to the proximal surface of the lunate facet of the distal radius and b) a line parallel to the articular surface of the ulnar head. Posteroanterior (PA) view obtained with the wrist in neutral forearm rotation, the elbow flexed 90° and the shoulder abducted 90°	Varies in different populations, with a mean usually neutral to slightly negative (e.g. to ‐1 mm)

Table 1. Definition of radiological parameters

Table 2. Percutaneous pinning versus nonsurgical treatment: participants & interventions

Trial	Participants	Fracture type and classification	Timing of surgery and reduction method	Fixation	Nonsurgical treatment
Azzopardi 2005	57; 89% female (of 54); mean age 71.5 years	Unstable (dorsal angulation 20+ degrees) extra‐articular fractures. AO type A3^a; Frykman I and II	Timing not stated. Closed reduction	Crossed pins, one from the radial styloid and the other from the dorso‐ulnar side of the distal fragment into the distal radial shaft. Cast immobilisation for 5 weeks	Closed reduction under general anaesthesia and fluoroscopy with three‐point fixation obtained in a "well‐moulded" short‐arm cast for 5 weeks
Delgado 2009	43; 21% female; mean age 40 years	Unstable fractures of distal radius (intra‐articular fracture mentioned in abstract report), no further details	Timing not stated. Closed reduction	No details on wiring type or duration of immobilisation	Closed reduction. No details on cast type or duration of immobilisation
Gupta 1999	50; 74% female; mean age 56 years	Colles'. All extra‐articular. Frykman I and II	Timing not stated. Closed reduction	Crossed pins, one from the radial styloid and the other from the dorso‐ulnar side of the distal fragment into the distal radial shaft. Cast immobilisation for 6 weeks	Closed reduction under traction and plaster cast for 6 weeks (change in position and new plaster at 3 weeks)
Mardani 2011	198; 44% female; mean age 50.8 years	Displaced distal radius fracture with congruous joint with less than 2 mm joint gap [described as Fernandez classification type 1]. Dorsal displacement can be inferred by content of study report.	Timing not stated. Closed reduction	Percutaneous pinning with smooth 1.5 mm or 2 mm pin and immobilised with short arm cast. Cast immobilisation for 6 or 8 weeks	Closed reduction under general anaesthesia and long arm cast applied by the same orthopaedist. Cast immobilisation for 6 or 8 weeks
Rodriguez‐Merchan 1997	40; 73% female; mean age 57 years	Comminuted unstable fractures. All intra‐articular. Displaced (10+ degrees dorsal angulation/3+ mm radial shortening). Frykman III to VIII	Probably next day after presentation at hospital. Closed reduction	Crossed pins, two pins from the radial side and one from the ulnar side of the distal fragment into the distal shaft. Cast immobilisation for 7 weeks.	Closed reduction under local anaesthesia and plaster cast for 7 weeks
Shankar 1992	45; 88% female; age range 17‐88 years	Comminuted Colles', Frykman IV to VIII intra‐articular fractures	Timing not stated. Closed reduction	Two pins inserted obliquely from the radial side across the inferior radio‐ulnar joint and into the ulnar medial cortex. Pins incorporated into plaster. Cast immobilisation for 5 to 6 weeks	Closed reduction under general anaesthesia and image intensifier control plaster with cast for 5 to 6 weeks
Stoffelen 1998	98; % female unknown (discrepancies between trial reports but fewer females); mean age 58 years	Colles'. Extra‐articular fractures with dorsal displacement. Frykman I and II	Timing not stated. Probably closed reduction	Triple intrafocal Kapandji pinning: 3 pins, inserted at the fracture site and driven into the radial shaft, act as buttresses to the distal fragment. Cast immobilisation for 1 week	Closed reduction. Above‐elbow plaster cast for 3 weeks; below‐elbow for 3 weeks
Venkatesh 2016	70; % female unknown; mean age 47.75 years	Extra‐articular fractures of distal radius (AO types 23‐A2, 23‐A3^a). Dorsal displacement can be inferred by content of study report and treatment in Colles' cast	Not older than 2 weeks. Closed reduction	5 mm K‐wire through the radial styloid process piercing the far medial cortex of proximal fragment. An additional K‐wire was passed through the ulnar side of the radius engaging the opposite cortex when deemed necessary by the operating surgeon. Below elbow cast in neutral for 6 weeks	Closed reduction under general anaesthesia. Below elbow cast maintaining the wrist in palmar flexion and ulnar deviation (Colles' cast) for 6 weeks
Verhulst 1990	130; % female unknown; "elderly"	Colles'	Timing not stated. Closed reduction	Single Kirschner wire (pin) inserted through radial styloid, across fracture and into the distal radial shaft. Cast immobilisation for 4 weeks (mean)	Closed reduction and plaster cast for 5 weeks (mean)
Wong 2010	62; 82% (of 60) female; mean age 70.5 years	Unstable (dorsal angulation > 20 degrees and radial shortening > 5 mm) extra‐articular fracture of the distal radius	Fracture less than 2 weeks old. Closed reduction	Bier's block; 3 percutaneous K‐wires ("tripod" construct described in report); removable palmar splint for resting purpose; immediate gentle mobilisation under instructions of a physiotherapist was allowed; wires were removed "after fracture consolidation"	Closed reduction under haematoma block, Jones's methods used for reduction, below elbow plaster. No details on duration of immobilisation
Zyluk 2007	124; 82% (of 60) female; mean age 61 years	Displaced distal radius fracture. AO A2, A3, B1, B2, C1, C2^a and Frykman I‐VII	Timing not stated. Closed reduction	Regional block; percutaneous fixation with 2 to 4 k‐wires introduced via the radial styloid so that they pierce the opposite cortex. In some cases, one K‐wire was introduced "from the base of the arm". Wires were buried and a plaster immobilising "only the wrist" was applied. Wires removed under local anaesthesia in clinic at 6 weeks	Closed reduction under local anaesthesia and immobilisation in a "Vienna" type above elbow plaster/backslab, shortened to below elbow at 2 weeks. Mobilised at 5 weeks
^a AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)

Table 2. Percutaneous pinning versus nonsurgical treatment: participants & interventions

Table 3. Percutaneous pinning versus nonsurgical treatment: length and completeness of follow‐up

Study ID^a	Length of follow‐up in months	Completeness of follow‐up
Mardani 2011	3	No participant flow diagram. Some loss, as the authors described mean postoperative visits of 4.4 for group 1 and 3.6 for group 2 out of a possible 5.
Gupta 1999	6	No losses (reported)
Shankar 1992	6	No losses (inferred)
Venkatesh 2016	6	10 participants lost to follow‐up (5 from each group) were excluded from the analysis.
Zyluk 2007	6	64 (31 from the pinning and 33 from the nonsurgical group)
Azzopardi 2005	12	9 (3 deaths and 6 lost to follow‐up). No information on which group they were from
Delgado 2009	12	Not known
Rodriguez‐Merchan 1997	12	No losses (inferred)
Stoffelen 1998	12	No losses (inferred)
Verhulst 1990	24	Not known
Wong 2010	19.5 (13 to 24)	2 participants were excluded from the analysis; these were two deaths from heart disease "which was not related to postoperative complications"; hence they appeared to be from the surgery group.
^a Ordered by length of follow‐up

Table 3. Percutaneous pinning versus nonsurgical treatment: length and completeness of follow‐up

Table 4. Kapandji intrafocal pinning (3 wires) versus Py's isoelastic pinning (2 wires)

Trial	Participants	Fracture type^a	Kapandji pinning	Py's pinning	Cast use and wire removal
Fikry 1998	110, demographics and results only for 88 who completed follow‐up: 25% female, mean age 34 years, range 18‐66	Dorsally displaced distal radius fractures, with or without separation of ulnar styloid fragment. Comminuted fractures were excluded. Classification not given	3 K‐wires inserted at fracture site. Image intensifier used	2 K‐wires inserted through radial epiphysis, across fracture and along medullary canal up to radial head. Image intensification not used: grinding of wire against the cortical bone used to determine wire placement	Wires cut below the skin, hidden ends covered by a small bore drainage tube. Forearm plaster cast for 4 weeks; received instruction for shoulder and finger exercises. Pins removed at around 8 weeks under local or regional anaesthesia
Saddiki 2012	97, 86% female, mean age 63 years	Extra‐articular fractures: associated with fracture of the ulnar styloid process = Gerard‐Marchand (GM) fracture, not associated with fracture of the ulnar styloid process = Pouteau‐Colles (PC) fracture. Intra‐articular fractures: simple: fractures with postero‐medial fragment (PMF), complex: T‐shaped intra‐articular fractures (sagittal and/or frontal)^b .	3 K‐wires used to perform intrafocal pinning. After intraoperative reduction by close manipulation, the first K‐wire was introduced laterally while the two others were inserted postero‐laterally and postero‐medially, at a minimum 40º angle relative to the vertical axis of the radius.	2 K‐wires featuring a spatulated tip for easier progression along the medullary canal of the radius were introduced into the epiphysis of the distal radius after reduction by close manipulation, the first one being inserted from the tip of the radial styloid and the second one from inside the Lister’s tubercle. The entry point of each K‐wire was radiographically controlled. Retrograde pinning of the radius was then performed up to the sub‐chondral bone of the radial head.	Plaster splint for 3‐weeks immobilisation, then a removable splint in all cases allowing the start of daily self‐rehabilitation exercises. Pin removal after 6 weeks postoperatively
^a Available information reported ^b Results reported per fracture category (incomplete)

Table 4. Kapandji intrafocal pinning (3 wires) versus Py's isoelastic pinning (2 wires)

Table 5. Biodegradable pins or wires versus metal pins or wires; participant characteristics and interventions

Trial	Participants	Fracture type^a	Type of surgery	Biodegradable pinning	Metal pinning	Cast use and wire removal
Casteleyn 1992	30, 77% female; mean age 61, range 22 to 85 years	Extra‐ and intra‐articular; Frykman I, II, V, VI (involving the DRUJ)^b	Closed reduction using finger traps. All fractures pinned percutaneously (stab incisions) using Kapandji's iintrafocal technique	Holes predrilled; 2 PGA (polyglycodic acid ‐ biodegradable) rods inserted and gently tapped home	Two 2 mm Kirscher‐wires inserted into fracture gap (one dorsally and the other laterally) and drilled into the opposite cortex of the radial shaft. Then cut to be under the skin	No cast applied. Wire removal at 6 weeks
Korner 1999	60, no other details	Extra‐ and intra‐articular; mention of AO A2, A3 and B1^c	Mention of Willenegger: probably 2 pins placed through radial styloid and across fracture	Biodegradable pins	Kirschner wires	No details
^a Available information reported ^b DRUJ: distal radial ulnar joint ^c AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)

Table 5. Biodegradable pins or wires versus metal pins or wires; participant characteristics and interventions

Table 6. Buried versus exposed percutaneous wires: participants and interventions

Trial	Participants	Fracture type^a	Type of surgery	Buried wiring	Exposed wiring	Cast use and wire removal
Hargreaves 2004	56, 45% female; mean age 35, range 7 to 81 years	54 closed and 2 open	Closed reduction and 2 crossed wires. Open reduction in 9 cases	Buried deep to the skin; closed with sutures. Wire ends bent over	Exposed, skin released round wire. Wire ends bent over	Forearm cast applied. Removal at 6 weeks
Murphy 2008	60, no other details	Displaced	No details	Buried	Exposed	No details. Probably removed at 6 weeks
Waheed 2004	52; 73% female; mean age 57, range 19 to 84 years	No details	No details	Buried	Wires left protruding	Cast and wires removed mean 5.8 weeks at outpatients
^a Available information reported

Table 6. Buried versus exposed percutaneous wires: participants and interventions

Table 7. Duration of postoperative immobilisation after percutaneous pinning: participants and interventions

Trial	Participants	Fracture type^a	Type of surgery	Short‐term immobilisation	Longer‐term immobilisation	Immobilisation type
Allain 1999	60, 75% female; mean age 55, range 18‐87 years	AO types A2, A3, C1, C2^b	Trans‐styloid fixation: 2 wires removed at 45 days	1 week	6 weeks	Short arm circular cast
Milliez 1992	60, 73% female, age 55, range 19‐91 years	Closed displaced intra‐ and extra‐articular	Kapandji intrafocal pinning, 3 wires removed at 6 weeks	1 week	6 weeks	Plaster splint
Soleiman pour 2011	50, 46% female; no details on age	displaced metaphyseal distal radius fractures and non‐comminuted intra‐articular fractures	Not specified, Kapandji mentioned	1 week	4 weeks	Not specified
^a Available information reported ^b AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF)

Table 7. Duration of postoperative immobilisation after percutaneous pinning: participants and interventions

Table 8. Assessment of items relating to applicability of trial findings

Study ID	Clearly defined study population?	Interventions sufficiently described?	Main outcomes sufficiently described?	Appropriate timing of outcome measurement? (Yes = ≥ 12 months)
Allain 1999	Yes	Yes	Partial: Inadequate definition of discomfort during activity; and CRPS‐1^c	Yes: 12 months
Azzopardi 2005	Yes	Yes	Yes	Yes: 12 months
Casteleyn 1992	Partial: Insufficient information on exclusions	Yes	Partial: Poor description of functional grading tool	Yes: 12 months
Delgado 2009^a	Partial: “Heavy‐labour workers with unstable fractures of distal radius.” Only description of participants	No: No details of type of wiring or cast or duration of immobilisation	Partial: Some outcomes types reported, not defined	Yes: 12 months
Fikry 1998	Yes: However, a young population.	Yes	No: Not defined, just reported.	Yes: 20 months minimum
Gravier 2006	Partial: Insufficient information on displacement criteria for surgery.	Yes	No: Radiological outcomes only; lack of clarity over one outcome	No: 45 days (data for)
Gupta 1999	Partial: Colles radius fracture in the skeletally mature	Yes	Partial: Inadequate description of assessment of outcomes, including complications. Crude categorisation of functional scores	Partial: 6 months
Hargreaves 2004	Partial: No clear information on indications for surgery.	Yes	Yes	No: 6 weeks
Korner 1999^a	No: Distal radius fracture only.	No: Composition of bioabsorbable pins not given; no detail on modification to Wellenegger technique	No: Not defined, just reported. Inadequate descriptions of redisplacement and arthrotic findings	Partial: although median follow‐up was 25.4 months, some participants may not have been followed up for 12 months.
Lenoble 1995	Yes	Yes	Yes	Yes: 24 months
Mardani 2011	Partial: Not explicitly dorsally displaced fractures though implied by the study context i.e. reference to dorsal tilt and dorsal comminution	No: Inadequate information about the pin configuration, no broad description of the technique or approach used. Also no descriptors of cast positions or type or mode of application, bar “short” and "long arm cast”. No information as to if or when intraoperative imaging was used. Care provider was the same surgeon in both cases, assuming the lead surgeon but not clear	Partial: Primary outcome not determined. Lack of definition of finger stiffness. Loss to follow‐up taken to be a positive outcome, when 5 visits were scheduled for all participants. Failure of treatment (need for further intervention) not treated as an outcome parameter	No: 3 months
Milliez 1992	Yes	Partial: Not enough detail on early mobilisation	Partial: Inadequate description of CRPS‐1 and crude measure of pain.	No: 3 months
Murphy 2008^a	Partial: “Patients with a distal radial fracture managed with percutaneous wire fixation and casting only.” Minimal information available. By deduction, probably adults	No: Interventions were only mentioned and not described. Care providers not defined	Partial: Outcome assessment appeared limited to K‐wire‐related complications – some detail provided on assessment	No: 6 weeks
Rodriguez‐Merchan 1997	Partial: Other than fracture type, no exclusion criteria although age range provided.	Yes	Partial: Inadequate description of assessment of clinical union, radiographic measures, and complications	Yes: 12 months
Saddiki 2012	Yes	Yes	Yes	Yes: 12 months
Shankar 1992	Partial: Intra‐articular fractures but limited description of population.	Yes	Yes: Descriptions given of outcome measures, although crude measurement of subjective outcomes within composite outcome measure	Partial: 6 months
Shannon 2003^a	No: “Patients with unstable fractures of the distal radius were entered into the study.” No report of exclusion criteria or description of fracture type	No: Minimal information on intervention technique	No: No definition of outcome measures	No: 6 weeks
Snow 2007	Yes	Yes	Yes: Main outcome criteria (radiographic) were clearly defined	Partial: 17 weeks (4.25 months)
Soleiman pour 2011	Partial: No indication of participant age. No indication of type of pinning	Partial: No detail on method(s) of pinning nor immobilisation; no identification of care providers	Partial: Insufficient definitions provided for outcomes; but only part translation obtained	No: 3 months
Stoffelen 1998	Yes	Yes	Yes: Descriptions given of outcome measures, although crude measurement of subjective outcomes within composite outcome measure	Yes: 12 months
Strohm 2004	Partial: Colles type and AO^b classification but no exclusions	Yes	Partial: Inadequate description of assessment of outcomes, including complications	Partial: range 6 to 20 months
Venkatesh 2016	Partial: Not explicitly dorsally displaced fractures though implied by the study context i.e. reference to Colles' fracture, application of Colles' cast and restoration of volar tilt	Yes	Yes	Partial: 6 months
Verhulst 1990^a	Partial: Colles' radius fracture only	Partial: No information on postoperative care or type of immobilisation	Partial: Inadequate descriptions of outcome assessment	Yes: 2 years
Waheed 2004^a	Partial: “Patients with a distal radial fracture managed with percutaneous wire fixation and casting only.”	No: Intervention not described	No: No definition of outcome measures	No: Most likely up to 6 weeks
Wong 2010	Yes	Yes	Yes	Yes: minimum 13 months
Zyluk 2007	Partial: Inclusion criteria inadequately defined: “displaced isolated fractures”. Exclusion criteria listed: “… excluded people who did not meet the prerequisites, i.e. they had a fracture without displacement, open, with associated injuries of other structures, or did not agree to participate in the study”.	Yes	No: "The rate of secondary displacement was considered a primary outcome measure.” Secondary displacement not defined, some participants with secondary displacement excluded from analysis (if they required further surgery within 2 weeks) whereas others were analysed (if treated operatively and did not require further surgery).	Partial: 6 months
^a Reported in conference abstracts only ^b AO: Arbeitsgemeinschaft fur Osteosynthesefragen/Association for the Study of Internal Fixation (or ASIF) ^c CRPS‐1: Complex Regional Pain Syndrome type 1

Table 8. Assessment of items relating to applicability of trial findings

Comparison 1. Percutaneous pinning versus plaster cast

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 DASH scores (30 to 150; worst disability) Show forest plot			Other data	No numeric data

1.1 At 1.5 months			Other data	No numeric data
1.2 At 3 months			Other data	No numeric data
1.3 At 6 months			Other data	No numeric data
2 DASH (30 to 150; worst outcome) at 6 months: reduced function indicated by > 60 score Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3 Complications Show forest plot	9		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

3.1 Redisplacement	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.18, 1.65]
3.2 Redisplacement resulting in secondary treatment (reduction or reduction and K‐wire fixation)	6	591	Risk Ratio (M‐H, Fixed, 95% CI)	0.09 [0.03, 0.27]
3.3 K‐wire migration	2	180	Risk Ratio (M‐H, Fixed, 95% CI)	8.68 [1.19, 63.23]
3.4 K‐wire extrusion	1	50	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.5 K‐wire removal due to metal sensitivity	1	45	Risk Ratio (M‐H, Fixed, 95% CI)	2.88 [0.12, 67.03]
3.6 Pin track infection	7	571	Risk Ratio (M‐H, Fixed, 95% CI)	8.33 [2.54, 27.28]
3.7 Stab wound infection	1	130	Risk Ratio (M‐H, Fixed, 95% CI)	13.09 [0.69, 247.92]
3.8 Deep infection, joint infection, osteomyelitis	2	164	Risk Ratio (M‐H, Fixed, 95% CI)	3.10 [0.13, 74.58]
3.9 Tendon injury/rupture	2	94	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.10 Median nerve compression/neuropathy/contusion/CTS	5	363	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.19, 1.46]
3.11 Superfical radial nerve injury	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	13.53 [0.78, 233.82]
3.12 Reflex sympathetic dystrophy/Sudeck's atrophy	4	248	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.43, 1.48]
3.13 Persistent reflex sympathetic dystrophy	1	98	Risk Ratio (M‐H, Fixed, 95% CI)	2.08 [0.40, 10.85]
3.14 Shoulder hand syndrome	2	95	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.01, 7.45]
3.15 Finger stiffness	2	296	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.36, 0.76]
4 Non return to work (labourers) Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

5 Global grip strength (kg) Show forest plot			Other data	No numeric data

5.1 At 1.5 months			Other data	No numeric data
5.2 At 3 months			Other data	No numeric data
5.3 At 6 months			Other data	No numeric data
6 Global grip strength [probably] relative to other side (%) Show forest plot			Other data	No numeric data

6.1 At 1.5 months			Other data	No numeric data
6.2 At 3 months			Other data	No numeric data
6.3 At 6 months			Other data	No numeric data
7 Under half grip strength at 6 months Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

8 Grip strength at 13 to 24 months (kg) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

9 Range of movement at 6 months Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

9.1 Flexion (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.2 Extension (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.3 Radial deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.4 Ulnar deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.5 Supination (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.6 Pronation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
10 Range of motion (data from Zyluk 2007) Show forest plot			Other data	No numeric data

10.1 Flexion and extension at 1.5 months			Other data	No numeric data
10.2 Flexion and extension at 6 months			Other data	No numeric data
10.3 Pronation and supination at 1.5 months			Other data	No numeric data
10.4 Pronation and supination: at 6 months			Other data	No numeric data
11 Range of movement at 13 to 24 months Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

11.1 Flexion (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11.2 Extension (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11.3 Radial deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11.4 Ulnar deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11.5 Supination (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11.6 Pronation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
12 Short Form‐36 Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

12.1 Physical score at 4 months (0: worst to 100: best health)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
12.2 Mental score at 4 months (0: worst to 100: best health)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
13 Overall quality of life (1 very poor to 5 very good) at 13 to 24 months Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

14 Pain (occasional) Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

15 Patient satisfaction at 12 wks: excellent result (no pain, disability or motion limitation) Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

16 Patient satisfaction at 13 to 24 months (1 to 4, lower score equals higher satisfaction) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

17 Deformity (clinical and radiological) Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

17.1 Angulated malunion	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
17.2 Articular incongruity	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
17.3 Articular step off > 2 mm	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
17.4 Residual deformity ‐ prominence of ulnar styloid	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
18 Composite score functional grading: fair or poor Show forest plot	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

18.1 Score included deformity and complications	5	255	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.37, 0.78]
18.2 Score confined to pain and function	1	98	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.49, 1.89]
19 Mayo wrist score at 13 to 24 months (0 to 100; worst outcome) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

20 Anatomical measurements Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

20.1 Dorsal angulation (degrees)	4		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
20.2 Radial angulation (degrees)	4		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
20.3 Radial length (mm)	4		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
20.4 Ulnar variance (mm)	2		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 1. Percutaneous pinning versus plaster cast

Comparison 2. Kapandji intrafocal pinning (2 or 3 wires) versus trans‐styloid fixation (2 wires)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Redisplacement requiring secondary treatment	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Superficial pin‐track infection	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Tendon complications	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Superficial radial nerve ‐ symptoms	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.5 Median nerve dysfunction	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.6 Confirmed reflex sympathetic dystrophy (bone scan)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 2. Kapandji intrafocal pinning (2 or 3 wires) versus trans‐styloid fixation (2 wires)

Comparison 3. Modified Kapandji (dorsal Kapandji wires and trans‐styloid fixation) versus Kapandji intrafocal pinning

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Anatomical measurements (at 45 days) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.1 Dorsal angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Radial angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 3. Modified Kapandji (dorsal Kapandji wires and trans‐styloid fixation) versus Kapandji intrafocal pinning

Comparison 4. Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 DASH score at one year (0: worst disability to 100: no disability) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

2 Subjective results: pain and function (normal = 30 points); from Jakim score Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

2.1 In paper	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Reversed results	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Complications Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

3.1 Secondary displacement (early or reported as malunion)	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	2.90 [0.81, 10.46]
3.2 Fracture caused by pinning	1	88	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.02, 8.71]
3.3 Wire displacement	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	11.87 [1.57, 89.61]
3.4 Superficial infection	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	6.47 [0.80, 52.46]
3.5 Tendon rupture	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	3.29 [0.36, 30.38]
3.6 Reflex sympathetic dystrophy (Syndrome algoneurodystrophique)	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.39, 1.97]
3.7 Superficial radial nerve paraesthesia	1	64	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.16, 7.10]
3.8 Lunate (avascular) necrosis	1	64	Risk Ratio (M‐H, Fixed, 95% CI)	3.19 [0.13, 75.43]
4 Objective results: mobility, grip strength, deformity (normal = 30 points); from Jakim score Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

4.1 In paper	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 Reversed results	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Overall results: radiological, subjective and objective (normal = 100 points); from Jakim score Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 In paper	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 Reversed results	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Overall outcome grades Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

6.1 Fair or poor	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 Radiological results (normal = 40 points); from Jakim score Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

7.1 In paper	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.2 Reversed results	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
8 Anatomical measurements at one year Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

8.1 Dorsal angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.2 Radial angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 4. Kapandji intrafocal pinning (3 wires) versus Py isoelastic pinning (2 wires)

Comparison 5. Modified Kapandji pinning (3 wires) versus Willenegger pinning (2 wires)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 "Conversion procedure" (secondary treatment)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Wire migration	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Nerve irritation	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Carpal tunnel syndrome	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.5 Tendon injury or rupture	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.6 Reflex sympathetic dystrophy	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.7 Signs of swelling	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 5. Modified Kapandji pinning (3 wires) versus Willenegger pinning (2 wires)

Comparison 6. “Spring‐loaded intramedullary" pinning (not Py) versus transcortical pinning

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Early fixation failure	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 6. “Spring‐loaded intramedullary" pinning (not Py) versus transcortical pinning

Comparison 7. Biodegradeable pins versus Kirschner wires

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Operational difficulties in pin or wire insertion	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Secondary displacement requiring revision	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Superficial wound infection	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Tendon rupture	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.5 Carpal tunnel syndrome	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.6 Reflex sympathetic dystrophy	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.7 Secondary surgery	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.8 Painful scars	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.9 Sinus formation	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.10 Severe osteolytic reactions	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.11 Arthrosis (radiological signs)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Range of movement Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

2.1 Flexion (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Extension (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Radial deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.4 Ulnar deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.5 Supination (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.6 Pronation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Anatomical measurements Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

3.1 Dorsal angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 Radial angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.3 Radio‐ulnar index (mm)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 7. Biodegradeable pins versus Kirschner wires

Comparison 8. Buried wires versus exposed percutaneous wires

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Infected wire(s)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Wires removed early	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Superficial radial nerve symptoms	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Tendon damage (EPL)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 More invasive treatment for wire removal Show forest plot	2	108	Risk Ratio (M‐H, Fixed, 95% CI)	7.36 [3.54, 15.31]

2.1 Removal of wires in theatre	1	56	Risk Ratio (M‐H, Fixed, 95% CI)	6.80 [2.27, 20.42]
2.2 Local anaesthesia required	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	7.85 [2.94, 20.94]

Comparison 8. Buried wires versus exposed percutaneous wires

Comparison 9. Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Superficial pin‐site infection	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Deep infection	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Reflex sympathetic dystrophy	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Malunion	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.5 Need for physiotherapy	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.6 Secondary surgery (distal radial osteotomy)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Grip strength (as percentage of other side minus 30% for nondominant side) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

2.1 at 8 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 at 17 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Pinch strength (as percentage of contralateral minus 30% for nondominant side)( Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

3.1 at 8 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 at 17 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Range of motion (as percentage of contralateral normal side) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

4.1 Flexion at 8 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 Flexion at 17 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.3 Extension at 8 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.4 Extension at 17 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Rolyan nine hole peg test of dexterity (seconds) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 at 8 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 at 17 weeks	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Anatomical measurements at week 17 Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

6.1 Dorsal angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 Radial angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.3 Radial height (mm)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.4 Radial length (mm)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 9. Immobilisation of wrist in dorsiflexion versus palmar flexion (intrafocal pinning)

Comparison 10. Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Complications Show forest plot	3		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.1 Displaced wires or pins	3	170	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.62, 14.53]
1.2 Treatment failure (change of treatment)	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	7.0 [0.38, 129.93]
1.3 Superficial pin‐track infection	2	110	Risk Ratio (M‐H, Fixed, 95% CI)	2.33 [0.36, 15.18]
1.4 Tendon rupture	2	120	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.62, 14.53]
1.5 Persistent radial nerve paraesthesia or hypoaesthesia	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.04, 3.03]
1.6 Reflex sympathetic dystrophy	2	120	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.14, 6.93]
1.7 Non‐union	1	50	Risk Ratio (M‐H, Fixed, 95% CI)	2.0 [0.19, 20.67]
2 Range of movement (at 1.5 months; interim results) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

2.1 Flexion (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Extension (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Radial deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.4 Ulnar deviation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.5 Pronation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Pain at 3 months: usually or during effort Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

4 Patient dissatisfaction with outcome Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

5 Anatomical measurements (at 1 year): standard errors conversion Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 Dorsal angulation ‐ volar tilt (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 Radial angulation (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.3 Radial length (mm)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 10. Early (after 1 week cast immobilisation) versus later (after 4 to 6 weeks) mobilisation post surgery