Types of studies

We will include only randomised controlled trials (RCTs) of the effect of shunting in idiopathic normal pressure hydrocephalus (iNPH).

People may be reluctant to participate in clinical trials in which they are at risk of not being shunted. Therefore, we anticipate that we will encounter trials with a one‐arm, cross‐over design, in which half of the cases will have the intervention initially, and the other half will have the intervention at a later time; such studies may be blinded or unblinded (Saper 2017). In trials of this design, there is an early post‐randomisation time period when there are parallel intervention and control groups, which may be used to investigate the effect of the intervention. 

Types of participants

To be eligible for this study, participants must have at least one symptom of the Adams–Hakim triad: gait apraxia, dementia, or urinary incontinence, and an Evan's index of > 0.3 on cranial imaging. Participants must be at least 60 years of age, and have a normal cerebrospinal fluid (CSF) opening pressure.

We will exclude participants with potential secondary causes of normal pressure hydrocephalus (NPH), such as previous head trauma, meningitis, or subarachnoid haemorrhage.

These criteria are consistent with the Japanese Society of Normal Pressure Hydrocephalus Guidelines (Nakajima 2021), except that we will define elevated opening pressure as > 24.5 cm of water, which is consistent with international diagnostic guidelines in idiopathic normal pressure hydrocephalus (iNPH), and normal CSF reference ranges used in other neurological fields (Mollan 2018; Relkin 2005a). We will include studies in which participants may have had only one of Hakim's triad of symptoms, to ensure we do not exclude studies conducted before current diagnostic guidelines became available, but we will use sensitivity analysis to determine the effect of including these studies (Ishikawa 2004; Relkin 2005a). Similarly, we will not restrict participants to those who have a tight, high convexity and enlarged sylvian fissures on cranial imaging, to ensure we include any RCTs conducted before these iNPH‐specific imaging features were identified.

Types of interventions

Experimental interventions will be any permanent CSF shunting technique for the treatment of iNPH, including ventriculoperitoneal (VP) shunt, lumboperitoneal (LP) shunt, ventriculoatrial (VA) shunt, ventriculopleural (VPl) shunt, or third ventriculostomy.

Comparator interventions will be no CSF shunting, or the insertion of a shunt, but with the programmable valve not yet activated to a draining position (placebo shunt).

Types of outcome measures

Broadly, the outcome categories are:

• Gait function

• Cognitive function

• Urinary function

• Disability

• Quality of life

• Adverse events

We will assess non‐adverse outcomes after CSF shunting in the short‐term (less than six months), as we anticipate that there will be short‐term control data. Due to the preferred cross‐over trial design in the field of study, we anticipate that there will be no control data for long‐term outcomes; we will provide descriptive analysis of any non‐controlled long‐term data available (longer than six months).

We consider gait function the primary efficacy outcome, along with disability and quality of life, because it is usually pivotal in the decision to conduct CSF shunting.

Electronic searches

We will search ALOIS (alois.medsci.ox.ac.uk) ‐ the Cochrane Dementia and Cognitive Improvement Group’s Specialised Register. ALOIS is maintained by the Cochrane Dementia and Cognitive Improvement Group's Information Specialists, and contains dementia and cognitive improvement studies identified from the following:

• Monthly searches of a number of major healthcare databases: MEDLINE Ovid, Embase Ovid, CINAHL EBSCO, PsycINFO Ovid, and LILACS (Latin American and Caribbean Health Science Information database) BIREME

• Monthly searches of the US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov) and the World Health Organization International Clinical Trials Registry Platform (ICTRP; apps.who.int/trialsearch)

• Monthly search of the Cochrane Library’s Central Register of Controlled trials (CENTRAL)

• Six‐monthly searches of ISI Web of Science Core Collection

We will run additional separate searches of: MEDLINE Ovid, Embase Ovid, CINAHL EBSCO, PsycINFO Ovid, Web of Science Core Collection (ISI Web of Science), ClinicalTrials.gov, and ICTRP to ensure we retrieve the most current results. The MEDLINE search strategy is in Appendix 1.

Searching other resources

We will review citation indices for the identified relevant articles

Selection of studies

Two review authors will independently examine titles and abstracts of citations obtained from the searches, and exclude any clearly ineligible or duplicate articles, using Covidence (Covidence 2021). Following the initial screening, we will independently assess the full‐text article for inclusion in the review, using pre‐defined inclusion and exclusion criteria. A third review author will arbitrate disagreements, to create consensus. We will identify and record reasons for exclusion of the ineligible studies. We will record the selection process and create a PRISMA flow diagram (Moher 2009). The review authors will not be blind to trial authors, institution, or journal.

Data extraction and management

We will pilot the extraction process on at least one study in the review. We will use Covidence to manage study selection, and risk of bias assessment (Covidence 2021).

Two review authors will independently extract study characteristics and outcome data from included studies. Study characteristics will include: study design, setting, characteristics of participants (e.g. gender, age, ethnicity, disease severity, number of Hakim's triad of symptoms needed to fulfil NPH criteria for the study), randomisation, eligibility criteria, intervention details, type of control, the outcomes assessed, source of study funding, and investigator conflicts of interests. We will resolve disagreements by consensus, involving a third author (CC) where necessary.

For each outcome of interest, we will extract mean scores and standard deviations. When continuous outcome measures have been reported using different scales for the same construct, the lead review author (CC) will derive standardised mean differences, as outlined in the Cochrane Handbook for Systematic Reviews of Interventions, Chapter 6.6.2.1 (Higgins 2021).

We will export data from Covidence to Review Manager Web software (RevMan Web 2022).

Assessment of risk of bias in included studies

Two review authors will independently assess risk of bias in studies, using the revised risk of bias in randomised trials (RoB 2) tool (Higgins 2021a; Sterne 2019). Any disagreements will be resolved by discussion with a third author to reach consensus agreement.

We will assess risk of bias for each study outcome, using the following Cochrane RoB 2 criteria:

• Bias arising from the randomisation process

• Bias due to deviations from intended interventions

• Bias due to missing outcome data

• Bias in measurement of the outcome

• Bias in selection of the reported result

For each domain, we will answer a series of signalling questions with yes, probably yes, no information, probably no, or no, to determine the risk of bias (low risk, some concerns, and high risk). We will include text alongside the judgements to provide supporting information for our decisions.

We will decide the risk of bias for each outcome in our summary of findings tables (e.g. gait speed (short‐term)) by its performance in each domain: if we judge one domain to have some concerns, we will take this judgement for the whole outcome. We will summarise the risk of bias in traffic lights on the forest plots.

Measures of treatment effect

We will calculate effect estimates with 95% confidence intervals (CIs), using time point scores for each trial outcome. If outcomes are measured with ordinal rating scales; provided these contained a reasonably large number of categories (more than 10), we will treat the data as continuous, arising from a normal distribution. When outcomes are measured with a single continuous scale, the measure of treatment effect will be the mean difference (MD). When the same outcome is measured on different scales, we will use a standardised mean difference (SMD). We will use 'Guiding rules' for interpreting SMDs (or Cohen’s effect sizes), as outlined in Chapter 15.5.3.1 in the Cochrane Handbook for Systematic Reviews of Interventions, where a difference of 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect (Schünemann 2021a).

Unit of analysis issues

We anticipate that there will be several unit of analysis issues to contend with in this field of study:

Dealing with missing data

We will contact study investigators to obtain missing outcome or baseline characteristic data, when needed.

When the trials report change from baseline data only, we will compare this with time point data from the other studies identified in our search.

Assessment of heterogeneity

In addition to a visual inspection of the forest plots, we will assess statistical heterogeneity using a standard Chi² statistic and the associated l² statistic. Consistent with recommendations from Deeks 2021, we will deem heterogeneity to be present when the Chi² statistic is significant at the P = 0.1 level, or when l² suggests that more than 40% of the variability in the effect estimate is due to heterogeneity.

Assessment of reporting biases

We will assess in‐trial reporting bias as part of our risk of bias assessment, by assessing whether outcomes reported in the methods section are reported in the results for the included studies. If our searches identify trial protocols, clinical trial registrations, or abstracts, indicating the existence of unpublished studies, we will attempt to determine the status of any unpublished studies by contacting the investigators.

Data synthesis

We will undertake meta‐analysis using Review Manager Web when the results from more than one RCT are available, and we consider the trials sufficiently similar (RevMan Web 2022). We will use fixed‐effect meta‐analyses as we anticipate potential small‐study effects, and we anticipate studies having very similar designs.

Subgroup analysis and investigation of heterogeneity

We anticipate identifying only a few studies, with relatively low sample sizes, and thus, do not intend to undertake subgroup analysis.

Sensitivity analysis

If we have sufficient studies, we plan to undertake these sensitivity analyses:

• remove studies with the highest risk of bias from the analysis for the major outcomes

• remove studies that used no‐shunt comparator groups (as opposed to placebo‐shunt comparator groups)

• remove studies in which the majority of participants had only one of the Hakim‐Adam's triad of symptoms.

• If we consider that there may be important diversity between studies, we will re‐run the analyses using a random‐effects model, to test the robustness of findings with the meta‐analytic model used.

Summary of findings and assessment of the certainty of the evidence

We will assess the certainty of evidence for each outcome, using the evidence grading system developed by the GRADE collaboration, as described in Chapter 14 of Cochrane Handbook for Systemic Reviews of Interventions (Schünemann 2021).

Two authors will independently apply the GRADE approach, and assess the certainty of evidence at high, moderate, low, or very low. We will discuss the certainty of evidence ratings for each outcome with the other members of the review team. A third review author will judge any disagreement, to reach consensus for final decisions on the ratings.

We will take the following factors into account when deciding whether or not to downgrade the certainty of evidence in relation to each outcome:

• Risk of bias

• Inconsistency of results

• Indirectness of evidence

• Imprecision of results

• Publication bias

Since we will only include RCTs, we will start with high certainty for each outcome. We will downgrade certainty of the evidence by one level if we consider there is a serious limitation in relation to a particular factor, or by two levels if we consider there to be a very serious limitation. We will explain our reason(s) for downgrading in the footnotes.

We will generate a summary of findings tables using GRADEpro GDT software (GRADEpro GDT). The summary of findings table will compare CSF shunt to no shunt or a placebo shunt (using randomised parallel‐group data only) for the following outcomes;

• Gait speed (short‐term)

• Qualitative gait function (short‐term)

• Patient disability (short‐term)

• Quality of life (QoL; short‐term)

• Adverse events (short‐term)

• Adverse events (long‐term)

• Cognitive function (short‐term)