Pharmacological interventions for preventing dry mouth and salivary gland dysfunction following radiotherapy

Summary of findings for the main comparison. Pilocarpine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy

Pilocarpine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy
Patient or population: patients receiving radiotherapy on its own or in addition to chemotherapy to the head and neck region Intervention: pilocarpine Comparison: no treatment/placebo
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with no treatment/placebo	Risk with pilocarpine	Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Xerostomia ‐ Up to and including 6 months postRT Studies used different ways of measuring the outcome and therefore we combined the studies using SMD	‐	SMD 0.35 lower (1.04 lower to 0.33 higher)	‐	126 (2 RCTs)	⊕⊝⊝⊝ VERY LOW¹	Insufficient evidence of a difference at this time point and also at the end of RT and 3 months postRT 1 of the 2 studies in this assessment showed inconsistent results when using an alternative way of measuring this outcome at the 6‐month time point. 2 further studies showed insufficient evidence of a difference, 1 at the end of RT and the other at 3 months postRT
Salivary flow rate (unstimulated) ‐ Up to and including 3 months postRT Studies used different ways of measuring the outcome and therefore we combined the studies using SMD	‐	MD 0.06 lower (0.23 lower to 0.11 higher)	‐	24 (1 RCT)	⊕⊝⊝⊝ VERY LOW²	Insufficient evidence of a difference at this time point and also at the end of RT Same results for stimulated salivary flow rates at end of RT, and 3, 6 and 12 months postRT Same results for a further study at the end of RT and 3 months postRT looking at whether or not stimulated and unstimulated salivary flow was > 0 g
Overall survival ‐ Up to and including 6 months postRT	724 per 1000	775 per 1000 (579 to 1000)	RR 1.07 (0.80 to 1.43)	60 (1 RCT)	⊕⊝⊝⊝ VERY LOW³	Insufficient evidence of a difference
Quality of life ‐ Up to and including 6 months postRT McMaster University Head and Neck Questionnaire (HNRQ). Score 1‐7, lower score = poorer quality of life	Control group mean was 5.3	MD 0.20 higher (0.19 lower to 0.59 higher)	‐	90 (1 RCT)	⊕⊝⊝⊝ VERY LOW³	Insufficient evidence of a difference at this time point and also at the end of RT and 3 months postRT
Adverse effects	Insufficient evidence of a difference between groups for any reported adverse event, apart from for sweating where data from 5 studies showed an increased risk associated with pilocarpine (RR 2.98, 95% CI 1.43 to 6.22; P = 0.004; I² = 0%; 389 participants; ⊕⊕⊝⊝ LOW⁴)
The risk in the intervention group (and its 95% CI) 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; RCT: randomised controlled trial; RR: risk ratio; SMD: standardised mean difference; RT: radiotherapy
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect
¹Downgraded by 1 level for risk of bias, 1 level for imprecision (small sample size and 95% CIs include both possibility of benefit and harm), and 1 level for inconsistency (I² = 68%). ²Downgraded by 1 level for risk of bias, and 2 levels for imprecision (single study with 12 participants per group and 95% CIs include both possibility of benefit and harm). ³Downgraded by 1 level for risk of bias, and 2 levels for imprecision (single study and 95% CIs include both possibility of benefit and harm). ⁴Downgraded by 1 level for risk of bias, and 1 level for imprecision (very wide 95% CIs).

Summary of findings 2. Amifostine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy

Amifostine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy
Patient or population: patients receiving radiotherapy on its own or in addition to chemotherapy to the head and neck region Intervention: amifostine Comparison: no treatment/placebo
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with no treatment/placebo	Risk with amifostine	Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Xerostomia (0‐4 scale ‐ grade 2 or above) ‐ 12 months postRT	418 per 1000	292 per 1000 (167 to 514)	RR 0.70 (0.40 to 1.23)	682 (7 studies)	⊕⊕⊝⊝ LOW¹	Insufficient evidence of a difference at this time point. However, both at the end of RT (RR 0.35, 95% CI 0.19 to 0.67; 3 studies, 119 participants) and up to 3 months postRT (RR 0.66, 95% CI 0.48 to 0.92; 5 studies, 687 participants), amifostine reduced the risk of developing grade ≥ 2 xerostomia
Salivary flow rate (mg/5 min) (unstimulated) ‐ 12 months postRT	Control group mean was 0.16	MD 0.32 higher (0.09 higher to 0.55 higher)	‐	27 (1 study)	⊕⊝⊝⊝ VERY LOW²	Amifostine led to increased unstimulated saliva flow both at 12 months postRT and at the end of RT, but there was insufficient evidence of a difference at 3 months postRT. This evidence was supported by a further study showing a benefit for amifostine at 12 months postRT when looking at incidence of producing > 0.1 g of saliva over 5 minutes (RR 1.45, 95% CI 1.13 to 1.86; 175 participants). A further study narratively reported no difference Insufficient evidence of a difference in stimulated saliva flow at any time point
Overall survival at 12 to 24 months postRT	450 per 1000**	531 per 1000 (383 to 747)	HR 1.18 (0.85 to 1.66)	271 (2 studies)	⊕⊝⊝⊝ VERY LOW³	Insufficient evidence to determine whether or not amifostine reduces overall survival, progression‐free survival, disease‐free survival or locoregional tumour control up to 24 months postRT
Quality of life (Patient Benefit Questionnaire) ‐ 12 months postRT 8 items each on a 10‐point scale where higher = better QoL	Control group mean was 6.66	MD 0.7 higher (0.2 higher to 1.2 higher)	‐	180 (1 study)	⊕⊝⊝⊝ VERY LOW²	Amifostine led to a small improvement in quality of life at 12 months postRT, but there was insufficient evidence of a difference at the end of RT and 3 months postRT A further study narratively reported no difference at end of RT and 6, 12, 18, and 24 months postRT
Adverse effects	Data from 5 studies showed an increased risk of vomiting associated with amifostine (RR 4.90, 95% CI 2.87 to 8.38; 601 participants; ⊕⊕⊝⊝ LOW⁴) Data from 3 studies showed an increased risk of hypotension associated with amifostine (RR 9.20, 95% CI 2.84 to 29.83; 376 participants; ⊕⊕⊝⊝ LOW⁴) Data from 4 studies showed an increased risk of nausea associated with amifostine (RR 2.60, 95% CI 1.81 to 3.74; 556 participants; ⊕⊕⊝⊝ LOW⁴) Data from 3 studies showed an increased risk of allergic response associated with amifostine (RR 7.51, 95% CI 1.40 to 40.39; 524 participants; ⊕⊕⊝⊝ LOW⁴) There was insufficient evidence of a difference between groups for any other adverse events
The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) 2014 5‐year overall survival rate of patients with head and neck squamous cell carcinoma (www.who.int/selection_medicines/committees/expert/20/applications/HeadNeck.pdf) CI:* confidence interval; HR: hazard ratio; MD: mean difference; QoL: quality of life; RR: risk ratio; RT: radiotherapy
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect
¹Downgraded by 1 level for risk of bias, and 1 level for inconsistency (I² = 83%). ²Downgraded by 1 level for risk of bias, and 2 levels for imprecision (single study and small sample size). ³Downgraded by 1 level for risk of bias, and 2 levels for imprecision (small sample size and 95% CIs include both possibility of benefit and harm). ⁴Downgraded by 1 level for risk of bias, and 1 level for imprecision (very wide 95% CIs).

See: Summary of findings for the main comparison Pilocarpine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy; Summary of findings 2 Amifostine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy; Summary of findings 3 Palifermin compared to placebo for preventing salivary gland dysfunction following radiotherapy

Summary of findings 3. Palifermin compared to placebo for preventing salivary gland dysfunction following radiotherapy

Palifermin compared to placebo for preventing salivary gland dysfunction following radiotherapy
Patient or population: patients receiving radiotherapy on its own or in addition to chemotherapy to the head and neck region Intervention: palifermin Comparison: placebo
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with placebo	Risk with palifermin
Xerostomia (0‐4 scale ‐ grade 2 or above) ‐ Up to and including 3 months postRT	727 per 1000	705 per 1000 (560 to 887)	RR 0.97 (0.77 to 1.22)	471 (3 studies)	⊕⊕⊝⊝ LOW¹	Insufficient evidence of a difference at this time point
Overall survival at 42 to 72 months from baseline	450 per 1000**	450 per 1000 (324 to 626)	HR 1.00 (0.72 to 1.39)	(3 studies)	⊕⊕⊕⊝ MODERATE²	Insufficient evidence to determine whether or not amifostine reduces overall survival and progression‐free survival up to 72 months
Adverse effects	There was insufficient evidence of patients in either group experiencing more or less adverse events
The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) 2014 5‐year overall survival rate of patients with head and neck squamous cell carcinoma (www.who.int/selection_medicines/committees/expert/20/applications/HeadNeck.pdf) CI:* confidence interval; HR: hazard ratio; RR: risk ratio; RT: radiotherapy
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect
¹Downgraded by 1 level for imprecision (95% CIs include both possibility of benefit and harm), and 1 level for inconsistency (I² = 76%). ²Downgraded by 1 level for imprecision (95% CIs include both possibility of benefit and harm).

Background

Description of the condition

Xerostomia (dry mouth) has been defined as the "subjective sensation of dryness" (Sreebny 1996), whilst salivary gland hypofunction has been defined as "any objectively demonstrable reduction in either whole and/or individual salivary gland flow rates" (Navazesh 1992). Xerostomia is usually the result of a decrease in the volume of saliva secreted. Indeed, healthy individuals complain of a dry mouth when their unstimulated whole salivary flow rate falls below 50% of their normal level (Dawes 1987). However, xerostomia may also occur without a reduction in salivary flow (Porter 2004), possibly resulting from a change in composition of saliva secreted (Pankhurst 1996). Thus, xerostomia may, or may not be associated with salivary gland hypofunction. Salivary gland dysfunction is an 'umbrella' term for the presence of either xerostomia, or salivary gland hypofunction.

Salivary gland dysfunction is an extremely common side effect of radiotherapy to the head and neck region (Guchelaar 1997). The total dose for a course of radiotherapy for head and neck cancer is 50 Gy (gray) to 70 Gy (Shiboski 2007). However, doses over 52 Gy will cause severe salivary gland dysfunction (Porter 2004). A major decrease in saliva flow develops within one week of starting radiotherapy, and continues to deteriorate throughout treatment, culminating in permanent salivary gland dysfunction (Shiboski 2007). Indeed, even a dose of 20 Gy is enough to permanently damage salivary flow if it is given as a single dose (Porter 2004). Salivary gland hypofunction is associated with a variety of oral problems in this group of people (e.g. oral discomfort, taste disturbance, difficulty chewing, difficulty swallowing, speech problems, dental caries, oral candidiasis, and other oral infections). Certainly salivary gland dysfunction is associated with a significant impairment of quality of life in this group of patients.

Description of the intervention

The literature discusses a wide range of pharmacological interventions for preventing radiation‐induced salivary gland dysfunction. Examples of these include.

Parasympathomimetic drugs (choline esters, cholinesterase inhibitors)

Parasympathomimetic drugs stimulate salivary secretion by stimulating the parasympathetic nervous system. The parasympathetic nervous system increases bodily secretions such as tears, gastric juices, mucus and saliva to defend the body and help digestion. The most widely used parasympathomimetic drug in this clinical situation is pilocarpine hydrochloride (a choline ester) and has been licensed in many countries for the treatment of radiation‐induced salivary gland dysfunction (Wiseman 1995). Other indirectly acting parasympathomimetics, for example bethanecol, are much more widely used in other contexts, but have also been used 'off‐licence' to treat this condition (Epstein 1994).

Parasympatholytic drugs

Parasympatholytic drugs have the opposite effect to parasympathomimetic drugs, their action is anticholinergic, i.e. they inhibit the secretion of saliva. Results from animal tests (Ahlner 1994) and a study by Rode et al (Rode 1999; Rode 2001) suggest that the inhibition of salivary secretion during radiotherapy might actually protect later damage of the salivary glands and improve salivation following the treatment.

Cytoprotective agents

Cytoprotective agents can be administered before, with, or after cancer therapy to reduce or prevent damage or toxicity to the normal cells and tissues without compromising therapeutic efficacy. Amifostine is a cytoprotective agent and has been shown to accumulate in the salivary glands (Takahashi 1986); there are reports that this might lead to a reduction in parotid parenchymal damage due to radiotherapy (Bohuslavizki 1998), and reduce the incidence of radiation‐induced xerostomia (Brizel 2000).

Why it is important to do this review

Salivary gland dysfunction is a significant and mostly permanent side effect of radiotherapy to the head and neck region that has numerous knock‐on effects, negatively affecting quality of life. Unfortunately, the evidence for prevention using pharmacological agents is weak and some guideline statements do not currently recommend any (Buglione 2016). Although there is a recently published Cochrane Review looking at parasympathomimetic drugs for treating radiation‐induced salivary gland dysfunction (Davies 2015), other drugs with different modes of action have the potential to be effective in this situation, and a broader review of prophylactic measures was needed.

Objectives

To assess the effects of pharmacological interventions for the prevention of radiation‐induced salivary gland dysfunction.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials of parallel design. Trials were included irrespective of language of publication or publication status.

Types of participants

We included participants of all ages, ethnic origin and gender scheduled to receive radiotherapy on its own or in addition to chemotherapy to the head and neck region. Participants could be outpatients or inpatients.

Types of interventions

Active agents

Any pharmacological agent prescribed prophylactically for salivary gland dysfunction prior to or during radiotherapy, by any route, any dose, and for any length of time. Radiation techniques were excluded.

Control groups

No preventative intervention, placebo, or another pharmacological preventative measure for salivary gland dysfunction.

Types of outcome measures

As radiotherapy‐induced salivary gland dysfunction is considered to be permanent, we were interested in long‐term treatment effects and only collected data starting from the end of radiotherapy, except in the case of adverse effects.

Primary outcomes

The primary outcome measure for the review is salivary gland dysfunction as indicated by either:

xerostomia, i.e. the subjective sensation of dryness of the mouth. It was anticipated that different investigators would use different scales to assess xerostomia, e.g. visual analogue scales, verbal rating scales;
salivary flow rates (stimulated or unstimulated).

Secondary outcomes

The secondary outcome measures of the review are:

adverse effects, e.g. sweating, lacrimation (excess tears, crying), rhinorrhoea (watery discharge from the nose), diarrhoea, nausea;
survival data (overall, disease‐free, progression‐free, locoregional control);
other oral signs/symptoms, e.g. oral discomfort/pain, dysgeusia (taste disturbance), dysmasesia (difficulty in chewing), dysphagia (difficulty in swallowing), dysphonia (difficulty in speaking);
quality of life, e.g. ability to sleep, work, speak;
patient satisfaction;
cost data.

Search methods for identification of studies

Electronic searches

Cochrane Oral Health's Information Specialist conducted systematic searches in the following databases for randomised controlled trials and controlled clinical trials. There were no publication year or publication status restrictions:

Cochrane Oral Health's Trials Register (searched 14 September 2016) (Appendix 1);
Cochrane Central Register of Controlled Trials (CENTRAL; 2016, Issue 8) in the Cochrane Library (searched 14 September 2016) (Appendix 2);
MEDLINE Ovid (1946 to 14 September 2016) (Appendix 3);
Embase Ovid (1980 to 14 September 2016) (Appendix 4);
CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature; 1937 to 14 September 2016) (Appendix 5);
LILACS BIREME Virtual Health Library (Latin American and Caribbean Health Science Information database; 1982 to 14 September 2016) (Appendix 6);
Zetoc Conference Proceedings (1993 to 14 September 2016) (Appendix 7);
OpenGrey (1997 to 14 September 2016) (Appendix 8).

Subject strategies were modelled on the search strategy designed for MEDLINE Ovid. Where appropriate, they were combined with subject strategy adaptations of the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials as described in the Cochrane Handbook for Systematic Reviews of Interventions Chapter 6 (Lefebvre 2011). The Embase subject search was linked to an adapted version of the Cochrane Crowd Project filter for identifying randomised controlled trials in Embase Ovid (see www.cochranelibrary.com/help/central‐creation‐details.html for information).

Language

The search attempted to identify all relevant studies irrespective of language. Articles in Chinese (Han 2010; He 2004; Hu 2005; Wang 1998) were translated and included in the review. An article in Spanish (Fuertes 2004) was translated and subsequently excluded.

Searching other resources

Ongoing studies

We searched the following trial registries for ongoing studies:

US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov; searched 14 September 2016) (Appendix 9);
World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch; searched 14 September 2016) (Appendix 10).

Reference list searching

The reference lists of review articles and standard clinical oncology textbooks were checked for additional studies. The reference lists of included studies were also checked for additional studies.

Handsearching

Only handsearching done as part of the Cochrane Worldwide Handsearching Programme and uploaded to CENTRAL was included.

Unpublished studies

Requests for information about unpublished studies/studies published in the 'grey literature' were sent to relevant pharmaceutical companies, relevant investigators, editors of radiotherapy journals, and relevant professional organisations.

Data collection and analysis

Selection of studies

The titles and abstracts of all records identified by the search strategy were scanned independently and in duplicate by two review authors. For both studies that appeared to meet the inclusion criteria, and studies that contained insufficient information in the title and abstract to determine eligibility, we obtained the full‐text report and two review authors independently assessed them to establish whether they met the inclusion criteria. Studies excluded at this or subsequent stages were entered in the table of excluded studies with the reasons for exclusion recorded. All disagreements were resolved by discussion.

Data extraction and management

Two review authors independently and in duplicate extracted data using specially designed data extraction forms. The data extraction forms were piloted on several papers and modified as required before use. The data extracted included.

Citation details: including year of publication, country of origin, setting and source of funding.
Details of participants: including demographic characteristics, cancer details (type, stage, location), radiation therapy and criteria for inclusion.
Details of intervention: including type, duration and method of administration.
Details of outcomes reported: including method of assessment (if measurement scales were used, details of whether the scale was validated were recorded).
Sample size calculation and trial registration.

Authors were contacted where possible for clarification and missing information.

Assessment of risk of bias in included studies

Two review authors independently assessed the risk of bias of each included study using the Cochrane domain‐based, two‐part tool as described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We contacted study authors for clarification or missing information where necessary and feasible. Disagreements were resolved through discussion, consulting a third review author to achieve consensus when necessary.

We completed a 'Risk of bias' table for each included study. For each domain of risk of bias, we described what was reported to have happened in the study. This information provided the rationale for our judgement of whether that domain was at low, high, or unclear risk of bias.

We assessed the following domains:

sequence generation (selection bias);
allocation concealment (selection bias);
blinding of participants and personnel (performance bias);
blinding of outcome assessment (detection bias);
incomplete outcome data (attrition bias);
selective outcome reporting (reporting bias);
other bias.

We categorised the overall risk of bias of individual studies as being at low, high, or unclear risk of bias according to the following criteria:

low risk of bias (plausible bias unlikely to seriously alter the results) if all domains were at low risk of bias;
high risk of bias (plausible bias that seriously weakens confidence in the results) if one or more domains were at high risk of bias; or
unclear risk of bias (plausible bias that raises some doubt about the results) if one or more domains were at unclear risk of bias.

We also presented the 'Risk of bias' summary graphically.

Measures of treatment effect

For continuous outcomes (e.g. xerostomia on a visual analogue scale) where studies use the same scale, we used the mean values and standard deviations (SDs) reported in the studies in order to express the estimate of effect as mean difference (MD) with 95% confidence interval (CI). Where different scales were used to measure the same outcome, we expressed the treatment effect as standardised mean difference (SMD) with 95% CI.

For dichotomous outcomes, the estimate of effect of an intervention is expressed as risk ratios (RR) together with 95% CIs.

Unit of analysis issues

The participant is the unit of analysis.

Dealing with missing data

We contacted the author(s) of included studies, where feasible, to identify missing data and details of any other outcomes that may have been measured but not reported. We would have used the methods described in Section 7.7.3 of the Cochrane Handbook for Systematic Reviews of Interventions to estimate missing SDs (Higgins 2011) if appropriate. We did not use any other statistical methods or perform any further imputation to account for missing data.

Assessment of heterogeneity

Before any pooling of data was conducted, for comparisons with two or more studies, clinical heterogeneity was assessed by examining the types of participants (e.g. cancer types), interventions (e.g. control group used, dose and mode of administration), and outcomes (e.g. stimulated salivary flow rates or quality of life questionnaires). Statistical heterogeneity was also assessed using a Chi² test, where a P value < 0.1 indicated statistically significant heterogeneity. We also quantified the heterogeneity using the I² statistic.

Assessment of reporting biases

Publication bias was to have been assessed for comparisons where at least 10 studies were included in a meta‐analysis. We would have used the recommendations on testing for funnel plot asymmetry (Egger 1997), as described in Section 10.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Data synthesis

Meta‐analyses were only undertaken where there were studies of similar comparisons reporting the same outcomes. We combined MDs for continuous data, and RRs for dichotomous data. Our general approach was to use a random‐effects model. Our preference for the more conservative random‐effects model is because statistical assessments can miss potentially important between‐study heterogeneity in small samples (Kontopantelis 2012).

We presented data not suitable for meta‐analysis in additional tables.

Subgroup analysis and investigation of heterogeneity

Where possible, subgroup analyses would have been performed according to cancer type and treatment plans for cancer, and age of participants (i.e. children under the age of 18 years).

Sensitivity analysis

Sensitivity analysis was to be undertaken on the primary outcomes by excluding studies at unclear and high risk of bias from the analyses and also excluding unpublished literature.

If any meta‐analyses had included studies with a large variation in sample size (for example several small studies and a single very large study), we would have undertaken a sensitivity analysis comparing the effect estimates from both random‐effects and fixed‐effect models. If these were different we would have reported on both analyses as part of the results section, and we would have considered possible interpretation.

Presentation of main results

We produced a 'Summary of findings' table for each comparison that included more than one study. We included data on: xerostomia, salivary flow rate, survival, quality of life and adverse events. We used GRADE methods (GRADE 2004), and the GRADEpro online tool for developing 'Summary of findings' tables (www.guidelinedevelopment.org). We assessed the quality of the body of evidence for each comparison and outcome by considering the overall risk of bias of the included studies, the directness of the evidence, the inconsistency of the results, the precision of the estimates, and the risk of publication bias. We described our level of certainty in the overall findings for each comparison/outcome in terms of high, moderate, low, very low.

Results

Description of studies

Results of the search

Electronic searches identified a total of 3536 titles and abstracts. A further study was identified by one of the review authors' knowledge of the topic area. After removal of duplicates, 2284 records were identified for screening. Following screening of these titles and abstracts by two review authors, 87 were identified as potentially relevant. Full papers were retrieved and authors of abstracts were written to in order to gain the full papers. Following a second screening of these studies, 46 were excluded for reasons described in the Excluded studies section and in the table of Characteristics of excluded studies. One study is ongoing and a further study is awaiting classification. Therefore, 39 studies met our eligibility criteria and were included in this review. This process is presented graphically in Figure 1.

Figure 1

Study flow diagram.

Included studies

Characteristics of trial setting, publication status and funding

Thirty‐nine trials were included; five were multinational (Brizel 2000; Brizel 2008; Buentzel 2006; Henke 2011; Le 2011), six were conducted in China (Han 2010; He 2004; Hu 2005; Lin 2014; Peng 2006; Wang 1998), four were conducted in Germany (Büntzel 1998; Büntzel 2010; Grötz 2001; Vacha 2003), four in the USA (Fisher 2003; Haddad 2009; Lozada‐Nur 1998; Valdez 1993), three in Canada (Duncan 2005; Gornitsky 2004; Warde 2002), three in Brazil (Jaguar 2015; Jham 2007; Pimentel 2014), two in Thailand (Sangthawan 2001; Veerasarn 2006), two in India (Patni 2004; Reshma 2012), two in the Netherlands (Burlage 2008; Jellema 2006), one in Croatia (Lajtman 2000), one in Turkey (Abacioglu 1997), one in Greece (Antonadou 2002), one in Iran (Haddad 2002), one in France (Bardet 2011), one in Spain (Lanzós 2010), one in Japan (Watanabe 2010), and one in Slovenia (Rode 1999).

All trials had a parallel‐group design. Ten trials had more than one published paper, with Büntzel 1998 publishing seven papers relating to the one trial. Abacioglu 1997 is an unpublished dissertation and data were gained from the authors of two trials following publication of their results as conference abstracts (Lozada‐Nur 1998; Patni 2004). Eighteen of the trials received external funding, six trials received internal or no funding and the funding source was not stated in 15 trials.

One trial is ongoing (NCT02430298) and will be considered for future updates.

Characteristics of the participants

All of the trials recruited adults scheduled to receive radiotherapy to the salivary glands for cancer. The majority of participants were male. The type of cancer was head and neck at different sites in 36 trials and nasopharyngeal in 3 trials (Han 2010; He 2004; Lozada‐Nur 1998).

Ten of the trials explicitly stated that chemotherapy was given as part of the treatment regimen in addition to radiotherapy for all patients (Antonadou 2002; Brizel 2008; Buentzel 2006; Büntzel 1998; Han 2010; Henke 2011; Le 2011; Peng 2006; Vacha 2003; Watanabe 2010). Chemotherapy was given to some patients of the following four trials: Bardet 2011; Gornitsky 2004; Haddad 2009; Lozada‐Nur 1998. The other trials either undertook no chemotherapy, or were unclear about whether any chemotherapy was given.

Four studies explicitly referred to neck dissection but varied in the clarity of reporting: two clearly reported the proportions in each group that had their submandibular glands removed (Burlage 2008; Vacha 2003); one only reported the proportion that had neck dissection in each group, but did not refer to salivary gland removal (Haddad 2009); and one only stated that participants were stratified by submandibular gland removal, but numbers of participants affected were not reported (Jellema 2006).

The 39 included studies randomised 3520 participants, ranging from 10 to 291.

The percentage of participants lost to follow‐up ranged from 0% to 38%.

Characteristics of the intervention

All of the trials provided a detailed description of the intervention including the dose and method of administration for the test and control groups. Twenty‐one trials included a placebo control group and 14 a 'no intervention' control group, the remaining four trials making head‐to‐head comparisons (Bardet 2011; Jellema 2006; Jham 2007; Watanabe 2010).

Pilocarpine hydrochloride was assessed in 12 trials at various dosages: Abacioglu 1997; Burlage 2008; Fisher 2003; Gornitsky 2004; Haddad 2002; Lajtman 2000; Lozada‐Nur 1998; Pimentel 2014; Rode 1999; Sangthawan 2001; Valdez 1993; Warde 2002.
Biperiden plus pilocarpine was assessed in one trial: Rode 1999.
Amifostine was assessed in 12 trials at various dosages: Antonadou 2002; Bardet 2011; Brizel 2000; Buentzel 2006; Büntzel 1998; Haddad 2009; He 2004; Jellema 2006; Patni 2004; Peng 2006; Vacha 2003; Veerasarn 2006.
Chinese medicine was assessed in five trials, all comparing different herbs: Han 2010; Hu 2005; Lin 2014; Reshma 2012; Wang 1998.
Palifermin was assessed in three trials: Brizel 2008; Henke 2011; Le 2011.
Bethanechol was assessed in two trials: Jaguar 2015; Jham 2007.
Artificial saliva was assessed in one trial: Jham 2007.
Selenium was assessed in one trial: Büntzel 2010.
Antiseptic mouthrinse was assessed in one trial: Lanzós 2010.
Antimicrobial lozenge was assessed in one trial: Duncan 2005.
Polaprezinc was assessed in one trial: Watanabe 2010.
Azulene oral rinse assessed in one trial: Watanabe 2010.
Venalot Depot (coumarin/ troxerutin) was assessed in one trial: Grötz 2001.

The length of follow‐up ranged from day 28 of the radiotherapy (RT) to 34 months: day 28/29 of RT (Pimentel 2014; Reshma 2012), end of RT (Abacioglu 1997; Hu 2005; Wang 1998), four weeks from start of RT (Lanzós 2010), four weeks after RT (Lin 2014; Grötz 2001), five weeks after RT (Gornitsky 2004), six weeks after RT (Büntzel 2010; Vacha 2003), seven weeks after RT (He 2004), two months after RT (Jham 2007), three months after RT (Brizel 2008; Han 2010; Jaguar 2015; Lozada‐Nur 1998), six months after RT (Duncan 2005; Fisher 2003; Haddad 2002; Sangthawan 2001; Warde 2002), 12 months after RT (Bardet 2011; Buentzel 2006; Büntzel 1998; Burlage 2008; Lajtman 2000; Rode 1999; Valdez 1993), 18 months after RT (Antonadou 2002), 24 months after RT (Brizel 2000; Henke 2011; Jellema 2006; Le 2011; Patni 2004; Veerasarn 2006), and 34 months after RT (Haddad 2009). Duration of follow‐up/timing of assessment was unclear in two studies (Peng 2006; Watanabe 2010).

Characteristics of outcome measures

The trials used a variety of assessment measures for salivary gland dysfunction. Ten trials included a subjective measure of salivary gland dysfunction, i.e. the patient was involved in the assessment through visual analogue scales (VAS) (Gornitsky 2004; Haddad 2002; Sangthawan 2001; Wang 1998), linear analogue scale (LASA) (Warde 2002), and modified patient questionnaires (Abacioglu 1997; He 2004; Jellema 2006; Lajtman 2000; Veerasarn 2006). One study reported 'acute' or 'chronic' dry mouth only (Peng 2006). Fifteen trials reported a clinical assessment of salivary gland dysfunction using various scales: RTOG/EORTC (Radiation Therapy Oncology Group/European Organisation for Research and Treatment of Cancer) scoring (Antonadou 2002; Brizel 2000; Buentzel 2006; Fisher 2003; Grötz 2001; He 2004; Jellema 2006; Patni 2004; Veerasarn 2006), NCI CTCAE (National Cancer Institute Common Terminology Criteria for Adverse Events) (Brizel 2008; Henke 2011; Le 2011), WHO (World Health Organization) grading/classification (Büntzel 1998), and Late Effects Normal Tissue Task Force (LENT)‐Subjective, Objective, Management, Analytic (SOMA) scales (Burlage 2008; Haddad 2002). Unstimulated or stimulated whole saliva secretion data or both were collected in ten trials (Abacioglu 1997; Brizel 2000; Buentzel 2006; Gornitsky 2004; He 2004; Lajtman 2000; Rode 1999; Valdez 1993; Veerasarn 2006; Wang 1998), and salivary gland scintigraphy was used in five trials (Fisher 2003; Grötz 2001; Lozada‐Nur 1998; Patni 2004; Veerasarn 2006).

Secondary outcomes were sporadically reported, using various scales. However, the majority of studies reported adverse events.

Excluded studies

Of the 87 trials that were identified as potentially eligible, 46 were excluded, with the main reason being the publication of an abstract only (17 publications), with insufficient information to allow thorough assessment: Bagga 2007; Borg 2007; Chambers 2005; Goyal 2007; Gu 2014; Kumarchandra 2010; Manoor 2014; Mitine 2000; Mix 2013; Norberg‐Spaak 1996; Norberg‐Spaak 1997; Park 2012; Park 2012a; Resubal 2011; Rudat 2005; Strnad 1997; Zale 1993.

Other reasons for exclusion were: not a randomised controlled trial or unclear if a randomised controlled trial; prevention of salivary gland dysfunction not the aim of study/not reported; radioactive iodine used rather than radiotherapy; study did not include head and neck cancer patients; the intervention was not a pharmacological agent.

Risk of bias in included studies

Allocation

Random sequence generation

Twenty of the included studies described an adequate method of random sequence generation and were assessed as at low risk of bias for this domain (Abacioglu 1997; Brizel 2000; Buentzel 2006; Burlage 2008; Gornitsky 2004; Haddad 2002; Haddad 2009; Henke 2011; Jaguar 2015; Jellema 2006; Jham 2007; Lanzós 2010; Le 2011; Lin 2014; Lozada‐Nur 1998; Pimentel 2014; Rode 1999; Sangthawan 2001; Valdez 1993; Veerasarn 2006). The remaining 19 studies stated that allocation was random but did not describe their methods and were therefore assessed as at unclear risk of bias for this domain.

Allocation concealment

Allocation concealment was clearly described in 16 of the included studies and they were assessed as being at low risk of bias for this domain (Abacioglu 1997; Brizel 2000; Buentzel 2006; Burlage 2008; Gornitsky 2004; Haddad 2002; Haddad 2009; Henke 2011; Lanzós 2010; Le 2011; Lozada‐Nur 1998; Pimentel 2014; Rode 1999; Sangthawan 2001; Valdez 1993; Veerasarn 2006). The remaining 23 did not describe any methods used to conceal the random sequence, and so were assessed as being at unclear risk of bias.

Blinding

Blinding of participants and personnel (performance bias)

Twenty‐one studies were placebo‐controlled and double‐blind, and were assessed at low risk of performance bias. In the remaining 18 studies, blinding of the patients and their caregivers to the allocated treatment was not possible because the active and control treatments were administered differently, the control group had no intervention at all, or the personnel administering or patients were not blinded to the intervention (Abacioglu 1997; Antonadou 2002; Bardet 2011; Brizel 2000; Büntzel 1998; Büntzel 2010; Haddad 2009; Han 2010; He 2004; Hu 2005; Jellema 2006; Jham 2007; Patni 2004; Peng 2006; Rode 1999; Vacha 2003; Veerasarn 2006; Watanabe 2010).

Blinding of outcome assessment (detection bias)

Twenty‐one studies were assessing the effect of the intervention versus a placebo where the assessor was also blinded and these have been assessed as at low risk of bias. A further study which was not placebo‐controlled was assessed at low risk of bias because the outcome assessment for salivary gland dysfunction was objective (Rode 1999). The remaining 17 studies were assessed as being at high risk of detection bias, as the assessor was not blinded, the intervention was assessed against no intervention, the administration of the drug was different in the intervention and control groups or the assessment of xerostomia was subjective (Abacioglu 1997; Antonadou 2002; Bardet 2011; Brizel 2000; Büntzel 1998; Büntzel 2010; Haddad 2009; Han 2010; He 2004; Hu 2005; Jellema 2006; Jham 2007; Patni 2004; Peng 2006; Vacha 2003; Veerasarn 2006; Watanabe 2010).

Incomplete outcome data

Twenty‐one studies had no or negligible attrition and were assessed as being low risk. Twelve studies were assessed to be at high risk of attrition bias, due to high dropout rates, no reasons given for dropouts or differential attrition between the groups, which could be linked to the intervention (Bardet 2011; Brizel 2008; Burlage 2008; Grötz 2001; Haddad 2002; Jellema 2006; Jham 2007; Lanzós 2010; Pimentel 2014; Vacha 2003; Veerasarn 2006; Warde 2002). For the six remaining studies, there was insufficient information to determine risk of attrition bias (Fisher 2003; Haddad 2009; Lajtman 2000; Lozada‐Nur 1998; Peng 2006; Sangthawan 2001).

Selective reporting

Nineteen of the included studies reported the outcomes specified in the methods section in full, including information about xerostomia and adverse effects (Abacioglu 1997; Antonadou 2002; Bardet 2011; Brizel 2000; Brizel 2008; Buentzel 2006; Büntzel 1998; Haddad 2002; Han 2010; He 2004; Henke 2011; Hu 2005; Jaguar 2015; Jham 2007; Le 2011; Lin 2014; Lozada‐Nur 1998; Veerasarn 2006; Warde 2002). One study was assessed to be at unclear risk of reporting bias (Peng 2006). The remaining 19 studies were assessed as at high risk of reporting bias as they did not report on adverse effects or xerostomia, did not report on all outcomes, only significant data were reported or data on individuals were not reported, and grouped data did not have the standard deviations.

Other potential sources of bias

We did not consider there to be any other issues arising from other potential sources in any of the studies and we therefore assessed them all as being at low risk of bias for this domain.

Overall risk of bias

Overall, three of the included studies (8%) were assessed as at low risk of bias for all domains (Buentzel 2006; Henke 2011; Le 2011), and four studies (10%) were assessed as being at unclear risk of bias for at least one domain (Brizel 2008; Jaguar 2015; Lin 2014; Lozada‐Nur 1998). The remaining 32 studies (82%) were at high risk of bias for at least one domain. Risk of bias can be viewed graphically in Figure 2 and Figure 3.

Figure 2

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

Figure 3

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

Effects of interventions

Pilocarpine versus no treatment/placebo

Pilocarpine hydrochloride, at various dosages, was assessed in 12 trials: Abacioglu 1997; Burlage 2008; Fisher 2003; Gornitsky 2004; Haddad 2002; Lajtman 2000; Lozada‐Nur 1998; Pimentel 2014; Rode 1999; Sangthawan 2001; Valdez 1993; Warde 2002. Over 900 participants were randomised to either pilocarpine or no treatment/placebo; 698 were evaluated (although number varied by outcome/timing of assessment). Eleven of the trials were judged to be at high risk of bias; one was at unclear risk (Lozada‐Nur 1998).

Xerostomia

Nine trials evaluated xerostomia, however, the method of assessment varied across studies.

Seven trials presented continuous data on xerostomia obtained by simple VAS or a composite based on a number of xerostomia‐focused questions (Abacioglu 1997; Burlage 2008; Gornitsky 2004; Haddad 2002; Lozada‐Nur 1998; Sangthawan 2001; Warde 2002). The trial by Burlage 2008 was unable to be included in any statistical pooling as data were presented by Gy dose, but the number receiving each dose is unclear. There was no evidence of a difference between treatment groups at end of radiotherapy (standardised mean difference (SMD) 0.20, 95% confidence interval (CI) ‐0.16 to 0.56; P = 0.27; 122 participants), up to three months postradiotherapy (SMD 0.02, 95% CI ‐0.33 to 0.37; P = 0.92; 125 participants), or up to six months postradiotherapy (SMD ‐0.35, 95% CI ‐1.04 to 0.33; P = 0.31; 126 participants) (Analysis 1.1). There was substantial statistical heterogeneity present for the six‐month data (I² = 68%, P = 0.08).

One trial used the LENT‐SOMA scale to provide an objective assessment of xerostomia (Haddad 2002). This single trial showed a statistically significant difference in favour of pilocarpine (mean difference (MD) ‐0.40, 95% CI ‐0.69 to ‐0.11; P = 0.006; 39 participants) at six months postradiotherapy (Analysis 1.2).

Two trials presented binary data on the number of participants with/without xerostomia (Lajtman 2000; Pimentel 2014). There was no evidence of a difference between treatment groups at the end of radiotherapy (risk ratio (RR) 0.60, 95% CI 0.18 to 2.02; P = 0.41; 11 participants) or at three months postradiotherapy (RR 1.00, 95% CI 0.92 to 1.08; P = 1.00; 48 participants) (Analysis 1.3).

Salivary flow rates

Eight studies presented continuous data for either stimulated or unstimulated salivary flow (e.g. ml/min or g) (Abacioglu 1997; Burlage 2008; Gornitsky 2004; Lajtman 2000; Lozada‐Nur 1998; Pimentel 2014; Rode 1999; Valdez 1993). The studies by Burlage 2008; Lajtman 2000; Pimentel 2014 and Rode 1999 were unable to be included in any statistical pooling due to insufficient reporting of data. There was no evidence of a difference between treatment groups for unstimulated or stimulated flow rates at any time point (Analysis 1.4; Analysis 1.5).

One study presented binary data on whether stimulated or unstimulated salivary flow was > 0 g (Fisher 2003). There is insufficient evidence to determine whether pilocarpine is beneficial for this outcome at any time point (Analysis 1.6; Analysis 1.7).

Survival

Only one trial reported on overall survival within the trial period (six months) (Haddad 2002). There was no evidence of a difference between treatment groups (RR 1.07, 95% CI 0.80 to 1.43; P = 0.66; 60 participants) (Analysis 1.8).

Quality of life and other oral related symptoms

There was insufficient evidence to determine whether or not pilocarpine improved quality of life measurements for global quality of life, quality of life (HNRQ), oral discomfort, eating difficulties and sleeping problems at the end of radiotherapy (Analysis 1.9; Additional Table 1). One trial (Gornitsky 2004) found an increased risk in speech difficulties at the end of radiotherapy in the pilocarpine group (MD 20.20, 95% CI 1.93 to 38.47; P = 0.03; 34 participants) when assessed using a VAS scale (0 to 100 mm) (Additional Table 1).

Table 1. Pilocarpine versus no treatment/placebo (other outcomes)

Outcome	Study ID	Time point	Pilocarpine	Control	Results	Comments
Oral related symptoms (other than salivary gland dysfunction/xerostomia)
Oral discomfort	Gornitsky 2004	End of radiotherapy	Mean 38.7 (SD 31.6) n = 16	Mean 56.7 (SD 26.7) n = 20	Mean difference ‐18.00 (95% CI ‐37.41 to 1.41), P = 0.07
Speech difficulties	Gornitsky 2004		Mean 57.5 (SD 26.8) n = 16	Mean 37.3 (SD 27.5) n = 18	Mean difference 20.20 (95% CI 1.93 to 38.47), P = 0.03
Eating difficulties	Gornitsky 2004		Mean 47.4 (SD 33.9) n = 15	Mean 61.8 (SD 25.4) n = 17	Mean difference ‐14.40 (95% CI ‐35.38 to 6.58), P = 0.18
Mucosal pain	Gornitsky 2004		Mean 38.8 (SD 33.9) n = 17	Mean 53.6 (SD 34.2) n = 19	Mean difference ‐14.80 (95% CI ‐37.07 to 7.47), P = 0.19
Oral complications	Pimentel 2014		1/5	4/6	RR 0.30 (95% CI 0.05 to 1.89), P = 0.20
Adverse events
Sweating	Abacioglu 1997		2/12	0/12	Random‐effects meta‐analysis of 5 studies: RR 2.98 (95% CI 1.43 to 6.22), P = 0.004 Heterogeneity: I² = 0%, P = 0.52
	Fisher 2003		18/118	5/114
	Gornitsky 2004		3/28	1/28
	Lozada‐Nur 1998		5/12	1/12
	Sangthawan 2001		1/29	2/24

Chilling	Abacioglu 1997		1/12	0/12	RR 3.00 (95% CI 0.13 to 67.06), P = 0.49

Nausea	Gornitsky 2004		7/28	5/28	Random‐effects meta‐analysis of 3 studies: RR 1.39 (95% CI 0.63 to 3.05), P = 0.41 Heterogeneity: I² = 0%, P = 0.93
	Haddad 2002		3/18	3/21
	Lozada‐Nur 1998		2/12	1/12

Vomiting	Fisher 2003		13/118	10/114	Random‐effects meta‐analysis of 3 studies: RR 1.28 (95% CI 0.70 to 2.35), P = 0.43 Heterogeneity: I² = 0%, P = 0.92
	Gornitsky 2004		6/28	5/28
	Lozada‐Nur 1998		2/12	1/12

Headache	Gornitsky 2004		2/28	3/28	RR 0.67 (95% CI 0.12 to 3.69), P = 0.64

Excessive lacrimation (tears)	Fisher 2003		3/118	0/114	Random‐effects meta‐analysis of 3 studies: RR 2.54 (95% CI 0.70 to 9.17), P = 0.15 Heterogeneity: I² = 0%, P = 0.71
	Haddad 2002		1/18	0/21
	Sangthawan 2001		4/25	2/22

Dysphasia	Lozada‐Nur 1998		3/12	2/12	RR 1.50 (95% CI 0.30 to 7.43), P = 0.62

Weakness	Fisher 2003		3/118	2/114	RR 1.45 (95% CI 0.25 to 8.51), P = 0.68

Nervous	Gornitsky 2004		0/28	1/28	Random‐effects meta‐analysis of 2 studies: RR 1.02 (95% CI 0.11 to 9.33), P = 0.99 Heterogeneity: I² = 0%, P = 0.33
Nervous	Lozada‐Nur 1998		1/12	0/12

Rhinitis	Fisher 2003		2/118	5/114	Random‐effects meta‐analysis of 3 studies: RR 0.87 (95% CI 0.41 to 1.86), P = 0.72 Heterogeneity: I² = 0%, P = 0.53
	Lozada‐Nur 1998		1/12	1/12
	Sangthawan 2001		8/29	6/24

Blurred vision	Lozada‐Nur 1998		1/12	0/12	RR 3.00 (95% CI 0.13 to 67.06), P = 0.49

Urinary frequency	Fisher 2003		7/118	5/114	Random‐effects meta‐analysis of 2 studies: RR 0.87 (95% CI 0.43 to 1.75), P = 0.70 Heterogeneity: I² = 0%, P = 0.32
Urinary frequency	Sangthawan 2001		6/25	8/22

Dizziness	Gornitsky 2004		0/28	2/28	Random‐effects meta‐analysis of 2 studies: RR 0.80 (95% CI 0.18 to 3.45), P = 0.76 Heterogeneity: I² = 13%, P = 0.28
Dizziness	Sangthawan 2001		4/25	3/22

Palpitation	Sangthawan 2001		0/25	4/22	RR 0.10 (95% CI 0.01 to 1.73), P = 0.11

Skin flushing	Fisher 2003		1/118	0/114	RR 2.90 (95% CI 0.12 to 70.44), P = 0.51

Motor tremors	Fisher 2003		2/118	1/114	RR 1.93 (95% CI 0.18 to 21.02), P = 0.59

Sleep problems	Gornitsky 2004	End of radiotherapy	Mean 37.3 (SD 36.4) n = 17	Mean 49.6 (SD 36.9) n = 19	Mean difference ‐12.30 (95% CI ‐36.27 to 11.67), P = 0.31

RTOG (grade 3; mucous membrane, pharynx and larynx)	Warde 2002					No statistically significant difference between treatment groups

CI = confidence interval; RR = risk ratio; RTOG = Radiation Therapy Oncology Group; SD = standard deviation.

There was insufficient evidence from one study (Gornitsky 2004) to determine whether or not pilocarpine improved oral mucosal pain at the end of radiotherapy (MD ‐14.80, 95% CI ‐37.07 to 7.47; P = 0.19; 36 participants) (Additional Table 1). The effect of pilocarpine on the treatment and prevention of mucositis has been assessed in more detail in separate Cochrane Reviews (Clarkson 2010; Worthington 2011).

Side effects

No evidence of a difference was found between treatment groups for any reported adverse event, apart from for sweating where data from five studies showed an increased risk associated with pilocarpine (RR 2.98, 95% CI 1.43 to 6.22; P = 0.004; 389 participants) (Additional Table 1). There was no observed statistical heterogeneity (I² = 0%, P = 0.52).

Cost

None of the included studies evaluating the effectiveness of pilocarpine reported cost data.

Biperiden plus pilocarpine versus no treatment

One trial, assessed at high risk of bias, compared biperiden and pilocarpine with no treatment (Rode 1999).

Xerostomia

No xerostomia data related to the effectiveness of biperiden and pilocarpine were reported.

Salivary flow rates

There was insufficient evidence, from a single trial of 60 participants (Rode 1999), to determine whether or not biperiden and pilocarpine reduced the unstimulated salivary flow rate between patients at the end of radiotherapy (Analysis 2.1).

Survival

Not reported.

Quality of life and other oral related symptoms

There was insufficient evidence to determine whether or not biperiden and pilocarpine reduces the risk of WHO grade 3+ dysphagia up to one year after radiotherapy (Additional Table 2). No further data on quality of life or other oral related symptoms were reported.

Table 2. Biperiden plus pilocarpine versus no treatment/placebo (other outcomes)

Outcome	Study ID	Time point	Pilocarpine	Control	Results	Comments
Dysphagia (WHO grade 3+)	Rode 1999	12 months after RT	1/30	4/30	RR 0.25 (95% CI 0.03 to 2.11), P = 0.20

CI = confidence interval; RR = risk ratio; RT = radiotherapy; WHO = World Health Organization.

Side effects

No data were reported on side effects.

Cost

No cost data related to the effectiveness of biperiden and pilocarpine were reported.

Amifostine versus no treatment/placebo

Eleven trials, one at low risk of bias (Buentzel 2006) and ten at high risk of bias (Antonadou 2002; Brizel 2000; Büntzel 1998; Haddad 2009; He 2004; Jellema 2006; Patni 2004; Peng 2006; Vacha 2003; Veerasarn 2006), randomised 1036 participants (887 analysed, although the number varied by outcome/timing of assessment) to amifostine or no treatment group/placebo. The trial by Jellema 2006 had three comparison groups: two different doses of amifostine and a 'no treatment' group. For the purpose of this comparison, the two amifostine groups were combined.

Xerostomia

Three studies (Büntzel 1998; He 2004; Veerasarn 2006) were combined in a meta‐analysis showing that amifostine reduced the risk of developing grade ≥ 2 xerostomia (on a 0 to 4 scale) at the end of radiotherapy (RR 0.35, 95% CI 0.19 to 0.67; P = 0.001; 119 participants) (Analysis 3.1).

Up to and including three months postradiotherapy, a smaller effect was observed in favour of amifostine in a meta‐analysis of five studies (Antonadou 2002; Brizel 2000; Buentzel 2006; Patni 2004; Veerasarn 2006) (RR 0.66, 95% CI 0.48 to 0.92; P = 0.01; 687 participants) (Analysis 3.1). However, there was substantial heterogeneity present (I² = 63%).

At 12 months postradiotherapy, there was insufficient evidence of a difference in the risk of grade ≥ 2 xerostomia (RR 0.70, 95% CI 0.40 to 1.23; P = 0.21; 682 participants analysed) (Antonadou 2002; Brizel 2000; Buentzel 2006; Büntzel 1998; Jellema 2006; Patni 2004; Veerasarn 2006) (Analysis 3.1). There was considerable heterogeneity present (I² = 83%).

Three further studies had no usable data: one failed to report the data by study group and reported that "For the end point xerostomia we are not able to demonstrate that amifostine had a positive effect, and there was no difference detected between the arms in terms of xerostomia, with 41% of patients reporting xerostomia of grade ≥ 2" (Haddad 2009); one failed to report the timing of assessment (Peng 2006); and one only reported xerostomia during radiotherapy (i.e. not at any of the time points we were interested in) (Vacha 2003).

Salivary flow rates

Unstimulated

There was inconsistent evidence regarding the effect of amifostine of unstimulated salivary flow rate. A greater salivary flow rate for those receiving amifostine was shown at the end of radiotherapy (MD 0.34, 95% CI 0.07 to 0.61; P = 0.01; 83 participants) (Analysis 3.2).

There was insufficient evidence of a difference from one study (Veerasarn 2006) up to and including three months postradiotherapy (MD 0.13, 95% CI ‐0.90 to 1.16; P = 0.8; 41 participants), but the same study showed a slight benefit in favour of amifostine at 12 months postradiotherapy (MD 0.32, 95% CI 0.09 to 0.55; P = 0.006; 27 participants) (Analysis 3.2). A further study (Brizel 2000) showed a benefit at 12 months postradiotherapy in favour of amifostine when looking at incidence of producing > 0.1 g of saliva over 5 minutes (RR 1.45, 95% CI 1.13 to 1.86; P = 0.004; 175 participants) (Analysis 3.3).

Haddad 2009 failed to report the salivary flow data by study group and simply reported that "No difference was observed between the 2 treatment arms."

Stimulated

There was insufficient evidence of a difference from one study (Veerasarn 2006) at the end of radiotherapy (MD ‐0.09, 95% CI ‐1.48 to 1.30; P = 0.90; 47 participants), up to and including three months postradiotherapy (MD 0.38, 95% CI ‐1.43 to 2.19; P = 0.68; 41 participants), or 12 months postradiotherapy (MD 0.82, 95% CI ‐0.47 to 2.11; P = 0.21; 27 participants) (Analysis 3.4). There was also insufficient evidence of a difference from one study (Brizel 2000), analysing 173 participants, when looking at incidence of producing > 0.1 g of saliva over 5 minutes at 12 months postradiotherapy (RR 1.12, 95% CI 0.89 to 1.41; P = 0.32) (Analysis 3.5).

Haddad 2009 failed to report the salivary flow data by study group and simply reported that "No difference was observed between the 2 treatment arms."

Survival

There was insufficient evidence to determine whether or not amifostine reduces overall survival, progression‐free survival, disease‐free survival or locoregional tumour control up to 24 months postradiotherapy.

Overall survival

There was insufficient evidence from a meta‐analysis of two studies (Brizel 2000; Buentzel 2006) to determine whether or not amifostine reduces overall survival at 12 to 24 months postradiotherapy (hazard ratio (HR) 1.18, 95% CI 0.85 to 1.66; P = 0.33; 271 participants) (Analysis 3.6). Two further studies found no difference in overall survival at two years or more postradiotherapy (Haddad 2009; Jellema 2006) (Analysis 3.7).

Progression‐free survival

There was insufficient evidence from a meta‐analysis of two studies (Brizel 2000; Buentzel 2006) to determine whether or not amifostine reduces progression‐free survival at 12 to 24 months postradiotherapy (HR 0.94, 95% CI 0.70 to 1.27; P = 0.70; 247 participants) (Analysis 3.8). A further study (Antonadou 2002) found no difference at 18 months postradiotherapy (RR 1.11, 95% CI 0.81 to 1.51; P = 0.52; 45 participants) (Analysis 3.9). This was supported by Haddad 2009 who reported "no differences noted" (Analysis 3.10).

Locoregional tumour control

There was insufficient evidence from a meta‐analysis of two studies (Brizel 2000; Buentzel 2006) to determine whether or not amifostine reduces locoregional tumour control at 12 to 24 months postradiotherapy (HR 0.90, 95% CI 0.74 to 1.11; P = 0.33; 279 participants) (Analysis 3.11). Three further studies reported narrative evidence to support this result (Haddad 2009; Jellema 2006; Patni 2004) (Analysis 3.12).

Disease‐free survival

There was insufficient evidence from one study (Patni 2004) to determine whether or not amifostine reduces disease‐free survival at 24 months postradiotherapy (RR 0.94, 95% CI 0.73 to 1.21; P = 0.64; 170 participants) (Analysis 3.13). Two studies reported narrative evidence to support this result (Patni 2004; Veerasarn 2006) (Analysis 3.14).

Quality of life and other oral related symptoms

There was insufficient evidence of a difference in quality of life from one study (Brizel 2000), both at the end of radiotherapy (MD 0.38, 95% CI ‐0.07 to 0.83; P = 0.1; 298 participants), and up to and including three months postradiotherapy (MD 0.52, 95% CI ‐0.02 to 1.06; P = 0.06; 233 participants). The same study showed a benefit in favour of amifostine at 12 months postradiotherapy (MD 0.70, 95% CI 0.20 to 1.20; P = 0.006; 180 participants) (Analysis 3.15). A further study reported no differences in quality of life but did not present data (Jellema 2006) (Additional Table 3).

Table 3. Amifostine versus no treatment/placebo (other outcomes)

Outcome	Study ID	Time point	Amifostine	Control	Results	Comments
Quality of life	Jellema 2006	Assessed at end of RT and 6, 12, 18 and 24 months after RT	No data	No data	"No significant differences between the 3 treatment arms"

Dysphagia (difficulty in swallowing) (0‐4 scale): grade 3 and above	Antonadou 2002	End of RT	14/22	23/23	Random‐effects meta‐analysis of 2 studies: RR 0.50 (95% CI 0.17 to 1.48); P = 0.21 Heterogeneity: I² = 40%, P = 0.20
	Büntzel 1998	End of RT	1/14	5/14
	Antonadou 2002	4 weeks after RT	2/22	3/23	RR 0.70 (95% CI 0.13 to 3.78); P = 0.68	By 8 weeks after RT, no participants had grade 3 or above dysphagia

Dysgeusia (taste disturbance) (0‐4 scale): grade 2 and above	Büntzel 1998	End of RT	3/14	14/14	RR 0.24 (95% CI 0.10 to 0.61); P = 0.003

Cost data (mean per patient supportive care costs)	Büntzel 1998	End of RT	USD 4401	USD 5873	P = 0.02

Vomiting	Antonadou 2002		1/22	0/23	Random‐effects meta‐analysis of 5 studies: RR 4.90 (95% CI 2.87 to 8.38); P < 0.00001 Heterogeneity: I² = 0%, P = 0.96
	Brizel 2000		55/150	11/153
	Buentzel 2006		8/66	2/64
	He 2004		1/17	0/15		"1 patient left due to gastrointestinal tract reaction/side effect, all other patients completed the treatment" "At the beginning of treatment, nausea and vomiting was obvious for amifostine group, but after treating with metoclopramide, there was no significant difference between 2 groups in gastrointestinal tract reaction/side effect"
	Jellema 2006		10/60	0/31
	Peng 2006		10/18			Data not reported in control group. Unknown if this was due to 0 events
	Veerasarn 2006		18/32			Data not reported in control group. Unknown if this was due to 0 events

Hypotension	Antonadou 2002		3/22	0/23	Random‐effects meta‐analysis of 3 studies: RR 9.20 (95% CI 2.84 to 29.83); P = 0.0002 Heterogeneity: I² = 0%, P = 0.88
	Brizel 2000		22/150	2/153
	Büntzel 1998		2/14	0/14
	Veerasarn 2006		5/32			Data not reported in control group. Unknown if this was due to 0 events

Nausea	Brizel 2000		66/150	25/153	Random‐effects meta‐analysis of 4 studies: RR 2.60 (95% CI 1.81 to 3.74); P < 0.00001 Heterogeneity: I² = 0%, P = 0.45
	Buentzel 2006		4/66	4/64
	He 2004		1/17	0/15		"1 patient left due to gastrointestinal tract reaction/side effect, all other patients completed the treatment" "At the beginning of treatment, nausea and vomiting was obvious for amifostine group, but after treating with metoclopramide, there was no significant difference between 2 groups in gastrointestinal tract reaction/side effect"
	Jellema 2006		23/60	3/31
	Peng 2006		10/18			Data not reported in control group. Unknown if this was due to 0 events
	Veerasarn 2006		20/32			Data not reported in control group. Unknown if this was due to 0 events

Allergic response	Brizel 2000		8/150	0/153	Random‐effects meta‐analysis of 3 studies: RR 7.51 (95% CI 1.40 to 40.39); P = 0.02 Heterogeneity: I² = 0%, P = 0.77
	Buentzel 2006		2/66	0/64
	Jellema 2006		4/60	0/31

Asthenia (weakness or lack of energy)	Buentzel 2006		3/66	1/64	RR 2.91 (95% CI 0.31 to 27.24); P = 0.35

Alopecia	Vacha 2003					Similar in both groups and increased continuously during the treatment

Skin toxicity	Vacha 2003					Similar in both groups and increased continuously during the treatment

Hot flush	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
Hot flush	Veerasarn 2006		17/32			Data not reported in control group. Unknown if this was due to 0 events

Somnolence (drowsiness)	Veerasarn 2006		18/32			Data not reported in control group. Unknown if this was due to 0 events

Sneezing	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
Sneezing	Veerasarn 2006		13/32			Data not reported in control group. Unknown if this was due to 0 events

Hiccup	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
Hiccup	Veerasarn 2006		10/32			Data not reported in control group. Unknown if this was due to 0 events

Dizziness	Peng 2006					"...dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"

Fatigue	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"

CI = confidence interval; RR = risk ratio; RT = radiotherapy; USD = US dollars.

Two of the 11 studies presented data on dysphagia (Antonadou 2002; Büntzel 1998). There was insufficient evidence of a difference in the risk of developing grade ≥ 3 dysphagia (on a 0 to 4 scale) at the end of radiotherapy (RR 0.50, 95% CI 0.17 to 1.48; P = 0.21; 73 participants) and up to and including three months postradiotherapy (four weeks after) (RR 0.70, 95% CI 0.13 to 3.78; P = 0.68; 45 participants) (Additional Table 3). In Antonadou 2002, no participants had grade 3 or above dysphagia by eight weeks after radiotherapy.

One study presented data on dysgeusia (Büntzel 1998). The study showed that amifostine reduced the risk of developing grade ≥ 2 dysgeusia (on a 0 to 4 scale) at the end of radiotherapy (RR 0.24, 95% CI 0.10 to 0.61; P = 0.003; 28 participants) (Additional Table 3).

Side effects

Adverse events were reported inconsistently across the 11 included trials. There was a higher risk of vomiting in the amifostine group (RR 4.90, 95% CI 2.87 to 8.38; P < 0.00001; five studies; 601 participants) (Antonadou 2002; Brizel 2000; Buentzel 2006; He 2004; Jellema 2006). Two further studies reported high rates of vomiting in the amifostine group but did not mention vomiting in the control group (Peng 2006; Veerasarn 2006). The risk of hypotension was higher in the amifostine group (RR 9.20, 95% CI 2.84 to 29.83; P = 0.0002; three studies; 376 participants) (Antonadou 2002; Brizel 2000; Büntzel 1998). Another study reported hypotension only in the amifostine group (Veerasarn 2006). The risk of nausea was higher in the amifostine group (RR 2.60, 95% CI 1.81 to 3.74; P < 0.00001; four studies; 556 participants) (Brizel 2000; Buentzel 2006; He 2004; Jellema 2006). Two further studies reported high rates of nausea in the amifostine group but did not mention nausea in the control group (Peng 2006; Veerasarn 2006). The risk of allergic response was higher in the amifostine group (RR 7.51, 95% CI 1.40 to 40.39; P = 0.02; three studies; 524 participants) (Brizel 2000; Buentzel 2006; Jellema 2006). There was insufficient evidence of a difference in asthenia (weakness/lack of energy) from one study (RR 2.91, 95% CI 0.31 to 27.24; P = 0.35; 130 participants) (Buentzel 2006). Other side effects (alopecia, skin toxicity, hot flush, drowsiness, sneezing, hiccupping, dizziness and fatigue) were reported either narratively or only for the amifostine group (Additional Table 3).

Cost

One study (Büntzel 1998), analysing 28 participants, reported economic cost data in a separate paper (Bennett 2001). In 2001, the mean per patient supportive care costs were lower in the amifostine group (USD 4401) than the control group (USD 5873) (P = 0.02) (Additional Table 3).

Amifostine (comparison of dosages)

One trial, at high risk of bias, compared two different amifostine regimens of 200 mg/m² either five or three times a week (Jellema 2006) (a third 'no treatment' group was not considered in this comparison).

Xerostomia

There was insufficient evidence to determine whether or not different amifostine dosages reduced the risk of developing grade ≥ 2 xerostomia (on a 0 to 4 scale) at 12 months postradiotherapy (RR 0.94, 95% CI 0.58 to 1.53; P = 0.80; 49 participants) (Analysis 4.1).

Salivary flow rates

No salivary flow rate data, related to the effectiveness of different doses of amifostine, were reported.

Survival

There is insufficient evidence reported on overall survival or locoregional tumour control (Analysis 4.2; Analysis 4.3).

Quality of life and other oral related symptoms

The paper reported "no significant differences between the three treatment arms" in quality of life assessed at the end of radiotherapy and 6, 12, 18 and 24 months after radiotherapy (Additional Table 4).

Table 4. Amifostine: comparison of different doses (other outcomes)

Outcome	Study ID	Time point	Amifostine 3 times weekly	Amifostine 5 times weekly	Results	Comments
Quality of life	Jellema 2006	Assessed at end of RT and 6, 12, 18 and 24 months after RT	No data	No data	"No significant differences between the 3 treatment arms"
Nausea	Jellema 2006		9/30	14/30	RR 0.64 (95% CI 0.33 to 1.25); P = 0.19
Vomiting	Jellema 2006		2/30	8/30	RR 0.25 (95% CI 0.06 to 1.08); P = 0.06
Allergic response	Jellema 2006		2/30	2/30	RR 1.00 (95% CI 0.15 to 6.64); P = 1

CI = confidence interval; RR = risk ratio; RT = radiotherapy.

Side effects

There was insufficient evidence of a difference in nausea (RR 0.64, 95% CI 0.33 to 1.25; P = 0.19; 60 participants), vomiting (RR 0.25, 95% CI 0.06 to 1.08; P = 0.06; 60 participants), or allergic response (RR 1.00, 95% CI 0.15 to 6.64; P = 1; 60 participants) (Additional Table 4).

Cost

No cost data related to the effectiveness of different doses of amifostine were reported.

Amifostine (intravenous versus subcutaneous)

One study, at high risk of bias, compared intravenous and subcutaneous delivery of amifostine (Bardet 2011).

Xerostomia

There was insufficient evidence to determine whether or not different methods of amifostine delivery reduced the risk of developing grade ≥ 2 xerostomia (on a 0 to 4 scale) up to and including three months postradiotherapy (RR 1.03, 95% CI 0.76 to 1.40; P = 0.86; 263 participants). There was a benefit in favour of amifostine at 12 months postradiotherapy (RR 0.61, 95% CI 0.42 to 0.88; P = 0.008; 127 participants) (Analysis 5.1).

Salivary flow rates

No salivary flow rate data, related to the effectiveness of different routes of administration of amifostine, were reported.

Survival

There was insufficient evidence of a difference in overall survival (HR 1.36, 95% CI 0.89 to 2.10; P = 0.16; Analysis 5.2) or locoregional tumour control (HR 1.34, 95% CI 0.76 to 2.36; P = 0.32; Analysis 5.3), both at 48 months postradiotherapy.

Quality of life and other oral related symptoms

No data on either quality of life or other oral related symptoms were reported.

Side effects

The single trial comparing intravenous and subcutaneous delivery of amifostine reported increased incidence of hypotension for intravenous delivery. Skin rash and local pain at injection site were worse for subcutaneous delivery (Additional Table 5). Results were inconclusive with regard to nausea/vomiting, fever, and asthenia (weakness/lack of energy).

Table 5. Amifostine: different routes of administration (other outcomes)

Outcome	Study ID	Time point	Intravenous	Subcutaneous	Results	Comments
Nausea/vomiting	Bardet 2011		29%	36%	P = 0.267
Hypotension	Bardet 2011		20%	8%	P = 0.007
Skin rash	Bardet 2011		10%	22%	P = 0.012
Local pain at injection site	Bardet 2011		0%	8%	P = 0.001
Fever	Bardet 2011		2%	0%	P = 0.256
Asthenia (weakness or lack of energy)	Bardet 2011		1%	6%	P = 0.054

Cost

No cost data related to the effectiveness of different routes of administration of amifostine were reported.

Chinese medicine versus no treatment/placebo

Five studies compared some form of Chinese medicine with no treatment/placebo. Four studies were assessed as being at high risk of bias (Han 2010; Hu 2005; Reshma 2012; Wang 1998); one study was assessed as being at unclear risk of bias (Lin 2014).

Xerostomia

Hu 2005 found that patients who received Shenqi Fanghon recipe had a reduced risk of xerostomia at the end of radiotherapy compared to those in the no treatment control group (RR 0.39, 95% CI 0.28 to 0.55; P < 0.00001; 140 participants) (Analysis 6.1). The paper was translated from Chinese, the methods were unclear.

Lin 2014, a trial of 71 participants, however, found no evidence of a difference for Tianwang Buxin Mini‐pills when compared with placebo when xerostomia was evaluated using both dichotomous data (Analysis 6.1) or continuous data (Analysis 6.2). Similarly, Han 2010, a trial of 95 participants, found no evidence of a difference for Jinlong capsules when compared with no intervention (Analysis 6.1).

Wang 1998, a trial of 50 participants, found a difference in favour of Chinese medicine in an assessment of xerostomia (VAS) against a no treatment group at the end of radiotherapy (P < 0.05). The results were graphically represented and the standard deviations were not available from the paper.

Salivary flow rates

One study showed an increase in stimulated salivary flow rate in favour of Chinese medicine when compared with no treatment at the end of radiotherapy (MD 0.09, 95% CI 0.03 to 0.15; P = 0.001; 50 participants) (Analysis 6.3). The paper was translated from Chinese, with the standard deviations being estimated (Wang 1998). Reshma 2012 mentioned salivary status but provided no data.

Survival

Hu 2005 evaluated overall survival (12 months postradiotherapy) but there was insufficient evidence to determine any effect (Analysis 6.4).

Quality of life and other oral related symptoms

Lin 2014 evaluated quality of life (at end of intervention and up to and including three months postradiotherapy) but there was insufficient evidence to determine any effect (Analysis 6.5). The same study showed insufficient evidence of a difference for other oral related symptoms (both at end of radiotherapy and one month after) (Additional Table 6). In Hu 2005 difficulty in mouth opening was worse in the control group (Additional Table 6).

Table 6. Chinese medicine (other outcomes)

Outcome	Study ID	Intervention	Time point	Study	Control	Results	Comments
Dysphasia (difficulty in swallowing) (score for EORTC‐H&N35 questionnaire: mean (SD))	Lin 2014	Chinese medicine (TWBXM)	End of RT	50.2 (26.3); n = 35	38.9 (25.9); n = 33	P = 0.07
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	30.2 (29.8); n = 23	26.7 (24.8); n = 21	P = 0.65
Dysgeusia (taste disturbance) (0 to 3 scale): grade 1 and above	Lin 2014	Chinese medicine (TWBXM)	End of RT	32/38	29/33	RR 0.96 (95% CI 0.79 to 1.16); P = 0.13
Speech difficulty (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	36.3 (26.7); n = 35	28.6 (26.2); n = 33	P = 0.23
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	27.4 (28.6); n = 23	22.7 (19.5); n = 21	P = 0.50
Difficulty in mouth opening (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	39.6 (28.2); n = 35	41.4 (27.7); n = 33	P = 0.79
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	32.2 (31.5); n = 23	33.3 (24.1); n = 21	P = 0.88
Difficulty in mouth opening (0 to 2 scale): grade 1 and above	Hu 2005	Chinese medicine (Shenqi Fanghou recipe)	"During the treatment"	22/70	52/70	RR 0.42 (95% CI 0.29 to 0.61); P < 0.001
Skin toxicity (0 to 3 scale): grade 1 and above	Lin 2014	Chinese medicine (TWBXM)	End of RT	35/38	30/33	RR 1.01 (95% CI 0.88 to 1.17); P = 0.82
Skin toxicity (0 to 4 scale): grade 1 and above	Hu 2005	Chinese medicine (Shenqi Fanghou recipe)	"During the treatment"	57/70	68/70	RR 0.84 (95% CI 0.74 to 0.94); P = 0.002
Skin toxicity (prevalence according to RTOG standards)	Han 2010	Chinese medicine (Jinlong capsule)		46.82%	58.32%		Quote: "toxicities during and after treatment were assessed" Comment: time point for assessment unclear; minor discrepancies in presented data
Nausea/vomiting (0 to 3 scale): grade 1 and above	Lin 2014	Chinese medicine (TWBXM)	End of RT	12/38	4/33	RR 2.61 (95% CI 0.93 to 7.30); P = 0.183
Hoarseness	Lin 2014	Chinese medicine (TWBXM)	End of RT	1/38	3/33	RR 0.29 (95% CI 0.03 to 2.65); P = 0.26
Fatigue (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	43.2 (26.2); n = 35	42.4 (23.0); n = 33	P = 0.88
Fatigue (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	31.2 (28.3); n = 23	36.4 (25.0); n = 21	P = 0.51
Pain (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	46.8 (23.2); n = 35	41.7 (27.4); n = 33	P = 0.40
Pain (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	35.9 (27.0); n = 23	40.9 (29.9); n = 21	P = 0.54
Pain (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	55.4 (25.1); n = 35	42.4 (20.5); n = 33	P = 0.02
Pain (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	31.6 (24.2); n = 23	37.8 (23.3); n = 21	P = 0.35
Dyspnea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	17.1 (23.1); n = 35	16.7 (20.7); n = 33	P = 0.93
Dyspnea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	20.5 (21.2); n = 23	13.6 (22.2); n = 21	P = 0.28
Insomnia (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	40.5 (25.0); n = 35	31.2 (25.3); n = 33	P = 0.13
Insomnia (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	30.8 (24.8); n = 23	31.8 (28.1); n = 21	P = 0.28
Appetite loss (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	45.0 (30.7); n = 35	45.8 (29.0); n = 33	P = 0.91
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	28.2 (26.1); n = 23	34.9 (30.0); n = 21	P = 0.42
Constipation (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	37.8 (27.4); n = 35	29.2 (20.3); n = 33	P = 0.15
Constipation (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	29.5 (30.3); n = 23	25.7 (20.4); n = 21	P = 0.63
Diarrhoea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	9.0 (15.0); n = 35	6.2 (13.2); n = 33	P = 0.42
Diarrhoea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	9.0 (15.1); n = 23	6.1 (16.7); n = 21	P = 0.53
Adverse effects	Han 2010	Chinese medicine (Jinlong capsule)		Leukopenia, nausea, vomiting, 1 participant had dizziness and blood pressure drop, 1 participant had skin rash	Not reported
Adverse effects	Hu 2005	Chinese medicine (Shenqi Fanghou recipe)	"During the treatment"	No adverse event	Not reported

CI = confidence interval; EORTC = European Organisation for Research and Treatment of Cancer; H&N = head and neck; RR = risk ratio; RT = radiotherapy; RTOG = Radiation Therapy Oncology Group; SD = standard deviation; TWBXM = Tianwang Buxin Mini‐pills.

Side effects

There was insufficient evidence of a difference for any side effects (at end of intervention and one month after) (Additional Table 6).

Cost

No cost data related to the effectiveness of Chinese medicine were reported.

Palifermin versus placebo

Three trials, two at low (Henke 2011; Le 2011) and one at unclear risk of bias (Brizel 2008), evaluated palifermin versus placebo.

Xerostomia

In a meta‐analysis of all three trials, there was insufficient evidence of a difference in the incidence of > grade 2 xerostomia up to three months postradiotherapy (RR 0.97, 95% CI 0.77 to 1.22; P = 0.78; 471 participants). There was considerable heterogeneity present (I² = 76%, P = 0.02) (Analysis 7.1). It should be noted that a large proportion of participants in Henke 2011 did not have assessments for xerostomia, but the intention‐to‐treat (ITT) rules stated that such participants would be assumed to have the outcome, and this may have had a substantial effect on the meta‐analysis result.

Xerostomia was measured up to 12 months in two studies but no data were reported (Henke 2011; Le 2011).

Salivary flow rates

None of the trials evaluating palifermin provided data on salivary flow rates.

Survival

All three trials reported data on overall and progression‐free survival at 42 to 72 months from baseline. There was insufficient evidence of a difference in both overall survival (HR 1.00, 95% CI 0.72 to 1.39; P = 0.99; Analysis 7.2) and progression‐free survival (HR 1.06, 95% CI 0.80 to 1.42; P = 0.67; Analysis 7.3).

Quality of life and other oral related symptoms

All three trials provided data for a meta‐analysis of dysphagia at three months postradiotherapy, with insufficient evidence of a difference (RR 1.32, 95% CI 0.55 to 3.13; P = 0.54) (Additional Table 7). There was also insufficient evidence of a difference in mouth and throat soreness at three months postradiotherapy in a meta‐analysis of two trials (MD ‐0.12, 95% CI ‐0.27 to 0.02; P = 0.10) (Additional Table 7).

Table 7. Palifermin versus placebo (other outcomes)

Outcome	Study ID	Time point	Palifermin	Placebo	Results
Oral related symptoms (other than salivary gland dysfunction/xerostomia)
Dysphagia	Le 2011	3 months postRT	29/94	19/91	Random‐effects meta‐analysis of 3 studies: RR 1.32 (95% CI 0.55 to 3.13); P = 0.54 Heterogeneity: I² = 94%, P < 0.00001
	Brizel 2008		61/64	31/32
	Henke 2011		32/92	20/93
Mouth and throat soreness ‐ 0 (no soreness) to 4 (extreme soreness) OMWQ‐HN scale	Le 2011	3 months postRT	n = 94, mean = 1.66, SD = 0.73	n = 94, mean = 1.86, SD = 0.65	Random‐effects meta‐analysis of 2 studies: mean difference ‐0.12 (95% CI ‐0.27 to 0.02); P = 0.10 Heterogeneity: I² = 13%, P = 0.28
	Henke 2011	3 months postRT	n = 92, mean = 1.52, SD = 0.69	n = 94, mean = 1.57, SD = 0.63
Adverse events
Nausea	Le 2011		47/94	42/91	Random‐effects meta‐analysis of 2 studies: RR 0.96 (95% CI 0.77 to 1.19); P = 0.69 Heterogeneity: I² = 28%, P = 0.24
Nausea	Brizel 2008		48/67	26/32
Fever	Brizel 2008		30/67	13/32	RR 1.10 (95% CI 0.67 to 1.81); P = 0.70
Constipation	Le 2011		31/94	24/91	Random‐effects meta‐analysis of 2 studies: RR 1.15 (95% CI 0.82 to 1.60); P = 0.42 Heterogeneity: I² = 0%, P = 0.57
Constipation	Brizel 2008		28/67	13/32
Diarrhoea	Brizel 2008		14/67	8/32	Random‐effects meta‐analysis of 2 studies: RR 1.28 (95% CI 0.49 to 3.36); P = 0.61 Heterogeneity: I² = 57%, P = 0.13
Diarrhoea	Henke 2011		11/92	5/93
Insomnia	Brizel 2008		12/67	4/32	Random‐effects meta‐analysis of 2 studies: RR 0.77 (95% CI 0.23 to 2.55); P = 0.67 Heterogeneity: I² = 63%, P = 0.10
Insomnia	Henke 2011		5/92	12/93
Dyspnea	Brizel 2008		9/67	1/32	RR 1.10 (95% CI 0.67 to 1.81); P = 0.70
Cough	Brizel 2008		8/67	5/32	RR 0.76 (95% CI 0.27 to 2.15); P = 0.61
Headache	Brizel 2008		8/67	2/32	Random‐effects meta‐analysis of 2 studies: RR 2.13 (95% CI 0.86 to 5.28); P = 0.10 Heterogeneity: I² = 0%, P = 0.86
Headache	Henke 2011		9/92	4/93
Decreased weight	Le 2011		29/94	27/91	Random‐effects meta‐analysis of 2 studies: RR 1.01 (95% CI 0.67 to 1.52); P = 0.96 Heterogeneity: I² = 0%, P = 0.73
Decreased weight	Brizel 2008		7/67	4/32
Dizziness	Brizel 2008		5/67	4/32	RR 0.60 (95% CI 0.17 to 2.07); P = 0.42
Anxiety	Brizel 2008		4/67	5/32	RR 0.38 (95% CI 0.11 to 1.33); P = 0.13
Hypomagnesemia	Brizel 2008		4/67	4/32	RR 0.48 (95% CI 0.13 to 1.79); P = 0.27
Vomiting	Le 2011		26/94	26/91	Random‐effects meta‐analysis of 2 studies: RR 0.98 (95% CI 0.72 to 1.33); P = 0.89 Heterogeneity: I² = 0%, P = 0.96
Vomiting	Brizel 2008		33/67	16/32
Radiation skin injury	Le 2011		25/94	13/91	RR 1.10 (95%CI 0.67 to 1.81); P = 0.70
Anaemia	Le 2011		21/94	34/91	Random‐effects meta‐analysis of 2 studies: RR 0.83 (95% CI 0.33 to 2.05); P = 0.68 Heterogeneity: I² = 54%, P = 0.14
Anaemia	Brizel 2008		10/67	3/32
Fatigue	Le 2011		21/94	20/91	Random‐effects meta‐analysis of 3 studies: RR 0.88 (95% CI 0.60 to 1.30); P = 0.52 Heterogeneity: I² = 2%, P = 0.36
	Henke 2011		7/92	14/93
	Brizel 2008		17/67	8/32
Leukopenia	Le 2011		21/94	12/91	Random‐effects meta‐analysis of 2 studies: RR 1.01 (95% CI 0.37 to 2.78); P = 0.98 Heterogeneity: I² = 79%, P = 0.03
Leukopenia	Henke 2011		12/92	20/93
Granulocytopenia	Brizel 2008		20/67	6/32	RR 1.59 (95% CI 0.47 to 5.39); P = 0.45
Pharyngolaryngeal pain	Le 2011		20/94	23/91	RR 0.84 (95% CI 0.50 to 1.42); P = 0.52
Hypokalemia	Le 2011		19/94	8/91	RR 2.04 (95% CI 0.98 to 4.28); P = 0.06
Pyrexia	Le 2011		16/94	19/91	RR 0.82 (95% CI 0.45 to 1.48); P = 0.50
Mucosal inflammation	Henke 2011		4/92	10/93	RR 0.40 (95% CI 0.13 to 1.24); P = 0.11
Asthenia	Henke 2011		13/92	7/93	RR 1.88 (95% CI 0.78 to 4.49); P = 0.16
Abdominal pain	Henke 2011		7/92	2/93	RR 3.54 (95% CI 0.75 to 16.58); P = 0.11
Back pain	Henke 2011		6/92	1/93	RR 6.07 (95% CI 0.74 to 49.40); P = 0.09
Febrile neutropenia	Henke 2011		1/92 Considered "serious adverse event"	0/93	RR 3.03 (95% CI 0.13 to 73.48); P = 0.50
Dehydration	Le 2011		13/94	19/91	Random‐effects meta‐analysis of 3 studies: RR 0.75 (95% CI 0.45 to 1.25); P = 0.27 Heterogeneity: I² = 30%, P = 0.24
	Henke 2011		6/92	13/93
	Brizel 2008		20/67	8/32

CI = confidence interval; OMWQ‐HN = Oral Mucositis Weekly Questionnaire ‐ Head and Neck Cancer; RR = risk ratio; RT = radiotherapy; SD = standard deviation.

Side effects

All three trials provided information on possible adverse events, sometimes reporting the same adverse event (Additional Table 7). There was no evidence of patients in either group experiencing more or less of these adverse events.

Cost

No cost data related to the effectiveness of palifermin were reported.

Bethanechol versus placebo

One study, at unclear risk of bias, compared bethanechol with placebo (Jaguar 2015).

Xerostomia

Bethanechol reduced the risk of developing grade ≥ 2 xerostomia (on a 0 to 3 scale) at the end of radiotherapy (RR 0.43, 95% CI 0.28 to 0.66; P = 0.0001; 84 participants). However, there was insufficient evidence of a difference up to and including three months postradiotherapy (RR 0.81, 95% CI 0.65 to 1.01; P = 0.06; 84 participants) (Analysis 8.1).

Salivary flow rates

Bethanechol increased unstimulated saliva flow (ml/min) at two months postradiotherapy (MD 0.19, 95% CI 0.06 to 0.32; P = 0.004; 97 participants) (Analysis 8.2).

There was insufficient evidence of a difference in stimulated saliva flow (ml/min) at two months postradiotherapy (MD 0.15, 95% CI ‐0.03 to 0.33; P = 0.11; 97 participants; Analysis 8.3).

Survival

No survival data were reported.

Quality of life and other oral related symptoms

No data on either quality of life or other oral related symptoms were reported.

Side effects

The study reported narratively that there were no statistical differences between the groups in bethanechol‐related toxicity and that "no patient experienced severe (grade 3) toxicity and no one dropped out of the study due to adverse effects" (Additional Table 8).

Table 8. Bethanechol versus placebo (other outcomes)

Outcome	Study ID	Results
Adverse effects	Jaguar 2015	No statistical difference between the groups in bethanechol‐related toxicity. Quote: "No patient experienced severe (grade 3) toxicity and no one dropped out of the study due to adverse effects"

Cost

No cost data related to the effectiveness of bethanecol versus placebo were reported.

Bethanechol versus artificial saliva

One study, at high risk of bias, compared bethanechol with artificial saliva (Jham 2007).

Xerostomia

There was insufficient evidence of a difference in having a dry mouth (yes/no) either at the end of radiotherapy (RR 0.63, 95% CI 0.30 to 1.29; P = 0.2; 36 participants) or at 8 to 40 weeks postradiotherapy (RR 0.56, 95% CI 0.30 to 1.05; P = 0.07; 30 participants) (Analysis 9.1).

Salivary flow rates

Bethanechol increased unstimulated saliva flow (ml/min) at the end of radiotherapy (MD 0.12, 95% CI 0.01 to 0.23; P = 0.03; 36 participants), but there was insufficient evidence of a difference at 8 to 40 weeks postradiotherapy (MD 0.07, 95% CI ‐0.02 to 0.16; P = 0.13; 33 participants) (Analysis 9.2).

There was insufficient evidence of a difference in stimulated saliva flow (ml/min) at the end of radiotherapy (MD 0.13, 95% CI ‐0.03 to 0.29; P = 0.12; 32 participants), but there was a benefit in favour of bethanechol at 8 to 40 weeks postradiotherapy (MD 0.21, 95% CI 0.01 to 0.41; P = 0.04; 29 participants) (Analysis 9.3).

Survival

There was insufficient evidence of a difference in overall survival at 40 weeks postradiotherapy (RR 1.59, 95% CI 0.43 to 5.84; P = 0.48; 43 participants; Analysis 9.4).

Quality of life and other oral related symptoms

No data on either quality of life or other oral related symptoms were reported.

Side effects

There were low rates of adverse effects (watering eyes, nervousness, frequent urination, sweating, warm face, cramps, diarrhoea, nausea) with insufficient evidence of any differences (Additional Table 9).

Table 9. Bethanechol versus artificial saliva (other outcomes)

Outcome	Study ID	Time point	Bethanechol	Artificial saliva	Results	Comments
Lacrimation (watering eyes)	Jham 2007	End of RT	3/22	0/21	RR 6.70 (95% CI 0.37 to 122.29); P = 0.2
Nervousness			3/22	0/21	RR 6.70 (95% CI 0.37 to 122.29); P = 0.2
Frequent urination			3/22	0/21	RR 6.70 (95% CI 0.37 to 122.29); P = 0.2
Sweating			2/22	0/21	RR 4.78 (95% CI 0.24 to 94.12); P = 0.3	"1 patient using bethanechol dropped out of the study due to excessive sweating (Grade 2 severity; National Cancer Institute Common Terminology Criteria for Adverse Events – NCI CTCAE, v 3"
Warm face			2/22	0/21	RR 4.78 (95% CI 0.24 to 94.12); P = 0.3
Cramps			1/22	0/21	RR 2.87 (95% CI 0.12 to 66.75); P = 0.51
Diarrhoea			1/22	0/21	RR 2.87 (95% CI 0.12 to 66.75); P = 0.51
Nausea			1/22	2/21	RR 0.48 (95% CI 0.05 to 4.88); P = 0.53

CI = confidence interval; RR = risk ratio; RT = radiotherapy.

Cost

No cost data related to the effectiveness of bethanechol versus artifical saliva were reported.

Selenium versus no intervention

Selenium was compared to no intervention in one trial assessed as at high risk of bias (Büntzel 2010).

Xerostomia

We were unable to use any of the data as bar charts of mean scores (baseline to postradiotherapy) were presented with no standard deviations. The results reported in the text indicated that there was no evidence that selenium reduced xerostomia (Analysis 10.1).

Salivary flow rates

No data on salivary flow rates were reported.

Survival

No survival data were reported.

Quality of life and other oral related symptoms

We were unable to use any of the data for loss of taste or dysphagia as bar charts of mean scores were presented with no standard deviations, but some information was reported in the text (Additional Table 10). There was insufficient evidence of any differences.

Table 10. Selenium versus no intervention (other outcomes)

Outcome	Study ID	Time point	Reported in text
Loss of taste	Büntzel 2010	6 weeks after end RT	"Ageusia was milder in the selenium group. But the difference was not significant"
Dysphagia	Büntzel 2010	6 weeks after end RT	"The only significant difference was observed at week 7, when the selenium group had developed a mean value of 1.533 versus 2.167 in the control group (P = 0.05)"
Adverse events	Büntzel 2010	6 weeks after end RT	"23 serious adverse events (SAEs) were seen in the selenium group, compared to 22 in the control group (P = 0.476). No statistically significant differences in toxicities were found using the 2‐tailed Fisher's exact test"

RT = radiotherapy.

Side effects

The text around adverse events implied that there was no evidence that selenium caused a higher number of events. We were unable to analyse the data as they were clustered (Additional Table 10).

Cost

No cost data related to the effectiveness of selenium were reported.

Antiseptic mouthrinse versus placebo

Antiseptic mouthrinse was assessed in one small trial at high risk of bias (Lanzós 2010).

Xerostomia

No data on xerostomia were reported.

Salivary flow rates

This outcome was only reported during radiotherapy (i.e. not at any of the time points we were interested in).

Survival

No survival data were reported.

Quality of life and other oral related symptoms

No quality of life data were reported, however, data on hyposialosis (drooling) were reported; there was insufficient evidence to determine whether or not antiseptic mouthrinse reduced or increased drooling in patients after four weeks (Additional Table 11).

Table 11. Antiseptic mouthrinse versus placebo (other outcomes)

Outcome

Study ID

Time point

Antiseptic rinse

Placebo

Results

Drooling

Lanzós 2010

4 weeks from baseline

Increased 6

No change or decreased 8

6/14

Increased 3

No change or decreased 7

3/10

RR 1.43 (95% CI 0.46 to 4.39); P = 0.53

Adverse events

Lanzós 2010

"No relevant adverse events were reported in any group"

CI = confidence interval; RR = risk ratio.

Side effects

The study reported "no relevant adverse events were reported in any group" (Additional Table 11).

Cost

No cost data related to the effectiveness of antiseptic mouthrinses were reported.

Antimicrobial lozenge versus placebo

Antimicrobial lozenge was assessed in one trial assessed as at high risk of bias (Duncan 2005).

Xerostomia

There was insufficient evidence that the antimicrobial lozenges reduced xerostomia up to and including three months postradiotherapy (RR 1.16, 95% CI 0.97 to 1.40; P = 0.11; 133 participants) (Analysis 11.1).

Salivary flow rates

No data on salivary flow rates were reported.

Survival

No survival data were reported.

Quality of life and other oral related symptoms

There was insufficient evidence of a difference between the groups for global quality of life (Analysis 11.2), mouth pain, sore/burning mouth, or throat pain (Additional Table 12).

Table 12. Antimicrobial lozenge versus placebo (other outcomes)

Outcome	Study ID	Time point	Antimicrobial lozenge	Placebo	Results
Mouth pain	Duncan 2005	Worse over 6 months	32/66	32/62	RR 0.94 (95% CI 0.66 to 1.33); P=0.72
Sore/burning mouth	Duncan 2005	Worse over 6 months	32/65	32/62	RR 0.95 (95% CI 0.68 to 1.35); P=0.79
Throat pain	Duncan 2005	Worse over 6 months	29/66	36/65	RR 0.79 (95% CI 0.56 to 1.12); P=0.19
Dryness in mouth	Duncan 2005	Worse over 6 months	55/66	46/65	RR 1.18 (95% CI 0.97 to 1.42); P=0.09
Nausea	Duncan 2005	Worse over 6 months	27/66	14/65	RR 1.90 (95% CI 1.10 to 3.28); P=0.02
Diarrhoea	Duncan 2005	Worse over 6 months	6/66	3/65	RR 1.97 (95% CI 0.51 to 7.54); P=0.32
Constipation	Duncan 2005	Worse over 6 months	24/66	26/65	RR 0.91 (95% CI 0.59 to 1.42); P=0.67

CI = confidence interval; RR = risk ratio.

Side effects

There is weak evidence that the antimicrobial lozenge may cause nausea, but insufficient evidence of a difference in dryness in the mouth, diarrhoea, or constipation (Additional Table 12).

Cost

No cost data related to the effectiveness of antimicrobial lozenges were reported.

Polaprezinc versus azulene oral rinse

One study at high risk of bias compared polaprezinc with azulene oral rinse (Watanabe 2010).

Xerostomia

There is some weak evidence that polaprezinc reduced severe xerostomia at the end of radiotherapy when compared with azulene rinse (RR 0.17, 95% CI 0.04 to 0.65; P = 0.009; 31 participants) (Analysis 12.1).

Salivary flow rates

No data on salivary flow rates were reported.

Survival

The study reported tumour response by RECIST (Response Evaluation Criteria In Solid Tumors) criteria for a specific group of patients only.

Quality of life and other oral related symptoms

There is some weak evidence that polaprezinc reduces severe oral pain and severe taste disturbance, however there is no evidence that polaprezinc helps patients to eat more when compared with azulene oral rinse (Additional Table 13).

Table 13. Polaprezinc versus azulene rinse (other outcomes)

Outcome	Study ID	Time point	Polaprezinc	Azulene rinse	Results
Pain > 2 (0‐3 scale)	Watanabe 2010	Over RT period	5/16	13/15	RR 0.36 (95% CI 0.17 to 0.77); P = 0.008
Taste disturbance > 2 (0‐3 scale)	Watanabe 2010	Over RT period	1/16	8/15	RR 0.12 (95% CI 0.02 to 0.83); P = 0.03
Disability of oral intake	Watanabe 2010	Over RT period	2/16	6/15	RR 0.31 (95% CI 0.07 to 1.31); P = 0.11

CI = confidence interval; RR = risk ratio; RT = radiotherapy.

Side effects

No data on adverse events were reported.

Cost

No cost data related to the effectiveness of polaprezinc or azulene were reported.

Venalot Depot (coumarin/troxerutin) versus placebo

One small trial, assessed as at high risk of bias, compared Venalot Depot with placebo (Grötz 2001).

Xerostomia

RTOG scores are reported however these scores were a composite of radiation side effects on different sites, so could not be used for assessing xerostomia.

Salivary flow rates

The sialometric data showed the reading dropped to an unmeasurable level in both groups so this was abandoned as a primary marker of efficacy and the protocol changed.

Survival

We were unable to use the data on locoregional control.

Quality of life and other oral related symptoms

No data on either quality of life or other oral related symptoms were reported.

Side effects

The study reported that "no adverse events could be attributed to the experimental medication" (Additional Table 14).

Table 14. Venalot Depot (coumarin/troxerutin) versus placebo

Outcome	Study ID	Results
Adverse events	Grötz 2001	"No adverse events could be attributed to the experimental medication"

Cost

No cost data related to the effectiveness of Venalot Depot were reported.

Sensitivity analysis

Risk of bias

There were too few studies at low risk of bias to carry out sensitivity analyses based on this factor.

Publication status

Four of the 39 studies were unpublished; Abacioglu 1997 was a dissertation, Lozada‐Nur 1998, Patni 2004 and Veerasarn 2006 were conference abstracts. The authors of these papers were contacted and their data were provided. Sensitivity analysis was undertaken to assess whether the inclusion of unpublished information had an effect on the results of the review.

Pilocarpine

Abacioglu 1997 and Lozada‐Nur 1998 assessed pilocarpine. Removing these unpublished trials from the results did not alter the findings of the review i.e. xerostomia and salivary flow rates at the end of radiotherapy and three months postradiotherapy were still not significant.

Amifostine

Patni 2004 and Veerasarn 2006 assessed amifostine.

Xerostomia

Removing Veerasarn 2006 from the analysis of xerostomia at the end of radiotherapy increases the effect estimate from RR 0.35 (95% CI 0.19 to 0.67; P = 0.001; 119 participants) to RR 0.22 (95% CI 0.09 to 0.53; P = 0.0008; 60 participants).

Removing Patni 2004 and Veerasarn 2006 from the analysis of xerostomia at three months postradiotherapy does not change the result but increases the uncertainty around the effect estimate from RR 0.66 (95% CI 0.48 to 0.92; P = 0.01; 687 participants) to RR 0.66 (95% CI 0.40 to 1.09; P = 0.1; 473 participants), thus including the possibility of harm associated with amifostine.

Removing Patni 2004 and Veerasarn 2006 from the analysis of xerostomia at 12 months postradiotherapy does not affect the result i.e. from RR 0.70 (95% CI 0.40 to 1.23; P = 0.21; 682 participants analysed) to RR 0.64 (95% CI 0.38 to 1.08; P = 0.09; 479 participants analysed).

Discussion

Summary of main results

A total of 39 trials were included in this review. We assessed the quality of the body of evidence for each outcome within a comparison (providing there was more than one study) using GRADE methodology (GRADE 2004).

Pilocarpine (summary of findings Table for the main comparison), compared with no treatment/placebo, was evaluated in 12 trials. There was no evidence of a difference in xerostomia between treatment groups at end of radiotherapy, three or six months. Similarly, there was also no evidence of a difference between treatment groups for salivary flow rates (stimulated or unstimulated) at any time point. There was insufficient evidence to determine the benefit of pilocarpine with regard to improving quality of life or increasing survival. There was no difference in reported adverse events, apart from sweating, where data from five studies showed an increased risk with pilocarpine. The body of evidence for each outcome was rated as very low quality, except for the adverse event of sweating which was low quality.

Amifostine (summary of findings Table 2), compared with no treatment/placebo, was evaluated in 11 studies. There is some (low‐quality) evidence that amifostine reduced the risk of developing grade ≥ 2 xerostomia (0 to 4 scale) at end of radiotherapy and, to a lesser extent, up to and including three months postradiotherapy. At 12 months postradiotherapy, there was insufficient evidence of a difference in the risk of grade ≥ 2 xerostomia. There was inconsistent (very low‐quality) evidence regarding the effect of amifostine on salivary flow rate. There was insufficient (very low‐quality) evidence to determine whether or not amifostine reduced overall survival, progression‐free survival, disease‐free survival or locoregional tumour control. Similarly, there was insufficient (very low‐quality) evidence to determine the benefit of amifostine in terms of quality of life. In general, adverse effects were poorly reported but there was (low‐quality) evidence that amifostine was associated with an increased risk of vomiting, hypotension, nausea and allergic response.

Palifermin (summary of findings Table 3), compared with placebo, was evaluated in three trials. There is insufficient (low‐quality) evidence to determine whether or not palifermin reduced the incidence of grade ≥ 2 xerostomia (0 to 4 scale) up to three months postradiotherapy. There was insufficient (moderate‐quality) evidence to determine the effect of palifermin on overall or progression‐free survival. There was no evidence of a difference in reported adverse effects.

All evidence from any remaining comparisons did not undergo formal GRADE assessment but was considered to be of very low quality.

Five trials (four at high risk of bias and one at unclear risk of bias) evaluated different forms of Chinese medicine. There is some evidence to suggest a benefit from Shenqi Fanghon recipe and an unspecified Chinese medicine at reducing xerostomia. Similarly, the unspecified Chinese medicine improved salivary flow rates. However, these findings were from single studies at high risk of bias. There was insufficient evidence to determine if any of the Chinese medicines had any effect on quality of life and survival.

Other interventions evaluated, for which there is currently insufficient evidence to draw conclusions were:

amifostine ‐ comparison of doses (single trial, at high risk of bias);
amifostine ‐ different routes of administration (single trial, at high risk of bias);
biperiden (single trial, at high risk of bias);
bethanechol (single trial, at unclear risk of bias);
bethanecol versus artificial saliva (single trial, at high risk of bias);
selenium (single trial, at high risk of bias);
antiseptic mouthrinse verus placebo (single trial, at high risk of bias);
antimicrobial lozenge versus placebo (single trial, at high risk of bias);
polaprezinc versus azulene oral rinse (single trial, at high risk of bias);
Venalot Depot versus placebo (single trial, at high risk of bias).

Overall completeness and applicability of evidence

Although we found 39 eligible studies that covered a wide range of interventions, the evidence found is not sufficient to highlight much promise in terms of effective preventative treatments for salivary gland dysfunction. This is because, despite there being a reasonable number of studies for three of the interventions (amifostine: 12; pilocarpine: 12; palifermin: 3), there was inconsistency in the way outcomes were reported and in the timing of outcome measurement. Most comparisons included only a single small study, the large majority being at high risk of bias. The most complete body of evidence was for amifostine and the outcome of incidence of moderate to severe xerostomia. Guideline statements point out the lack of an established pharmacological prophylaxis for salivary gland dysfunction and highlight the potential of radiotherapeutic techniques/precautions in reducing damage to the salivary glands (for example parotid‐sparing plans) (Buglione 2016). Therefore, for completeness, it may be sensible to also carry out a Cochrane Review of non‐pharmacological interventions, although we are not aware of many randomised controlled trials.

As mentioned, a wide range of interventions were assessed, but the studies were also conducted in both middle‐income and high‐income countries with no exclusion criteria in terms of the population included. Unfortunately, many studies did not include an objective measure of saliva flow to go with the more subjective measure of xerostomia. Furthermore, xerostomia was often measured differently between comparisons making it difficult to get an overall picture of the comparative effectiveness of the different interventions.

It was interesting that two of the three interventions for which we were able to carry out meta‐analyses (amifostine and palifermin) have shown promise in a Cochrane Review on the prevention of oral mucositis, another major side effect of cancer treatment (Worthington 2011). As the evidence for amifostine in the prevention of salivary gland dysfunction is promising, this could be beneficial for patients as they may require fewer medications. However, amifostine is not currently recommended in clinical practice guidelines due to high costs and its side effects (Buglione 2016). Furthermore, the evidence for its long‐term benefit is weak. There was insufficient evidence to support the use of palifermin in this review, but it is possible that the intention‐to‐treat rules in one of the three studies in the xerostomia meta‐analysis may have influenced the result (Henke 2011). Therefore, further studies assessing palifermin may be of interest. Worthington 2011 reported that there was no evidence that pilocarpine prevents oral mucositis.

Quality of the evidence

We included 39 studies that randomised 3520 participants; the number of participants analysed varied by outcome and time point.

Pilocarpine

We have very little confidence in the effect estimates for the outcomes of xerostomia, salivary flow rate, survival and quality of life (none of which showed a difference), mainly due to concerns regarding the risk of bias of the studies and imprecision of the results, but also due to inconsistency in the case of xerostomia. Further studies are likely to change the results. We had a little more confidence (although still limited) in the effect estimate for the adverse effect of sweating, which occurred more frequently in those receiving pilocarpine. Again it was risk of bias and imprecision which limited our confidence, and further studies would probably change the effect estimate. For more details see summary of findings Table for the main comparison.

Amifostine

Our confidence in the effect estimates for xerostomia was limited by concerns regarding the risk of bias and inconsistency in the results of the individual studies. New studies could change the results. We had a similar level of confidence in the results for the adverse effects of vomiting, hypotension, nausea and allergic response, which were all more frequent in those receiving amifostine. Risk of bias and imprecision were the factors affecting our confidence. We had very little confidence in the effect estimates for salivary flow rate, survival and quality of life due to risk of bias and imprecision. For more details see summary of findings Table 2.

Palifermin

Our confidence in the effect estimate for xerostomia was limited by concerns regarding imprecision and inconsistency in the results of the individual studies. We were moderately confident that palifermin did not compromise survival. For more details see summary of findings Table 3.

We did not formally assess the quality of the evidence for all other comparisons in this review, but it is all considered to be very low quality due to single small studies that are mostly at high risk of bias. Further studies would very likely change the effect estimates for all outcomes and time points within these comparisons.

Potential biases in the review process

Standard Cochrane methods were followed to avoid biases in the review process. However, we acknowledge that the decision to exclude data measured and reported during radiotherapy may be considered by some readers to be an arbitrary one. These data are potentially of interest and their exclusion may be thought of as a bias.

Agreements and disagreements with other studies or reviews

Although there are some systematic reviews on the treatment of salivary gland dysfunction caused by radiotherapy (Davies 2015; Mercadante 2017), we are not aware of any high quality systematic reviews on prevention.

Figure 1

Study flow diagram.

Figure 2

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

Figure 3

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

Analysis 1.1

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 1 Xerostomia.

Analysis 1.2

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 2 Xerostomia (LENT‐SOMA scale).

Analysis 1.3

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 3 Xerostomia.

Analysis 1.4

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 4 Salivary flow rate (unstimulated).

Analysis 1.5

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 5 Salivary flow rate (stimulated).

Analysis 1.6

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 6 Salivary flow rate (> 0 g) unstimulated.

Analysis 1.7

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 7 Salivary flow rate (> 0 g) stimulated.

Analysis 1.8

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 8 Overall survival.

Analysis 1.9

Comparison 1 Pilocarpine versus no treatment/placebo, Outcome 9 Quality of life.

Analysis 2.1

Comparison 2 Biperiden plus pilocarpine versus no treatment, Outcome 1 Salivary flow rate (unstimulated).

Analysis 2.2

Comparison 2 Biperiden plus pilocarpine versus no treatment, Outcome 2 Salivary flow rate (> 0 g) unstimulated.

Analysis 3.1

Comparison 3 Amifostine versus no treatment/placebo, Outcome 1 Xerostomia (0 to 4 scale ‐ grade 2 or above).

Analysis 3.2

Comparison 3 Amifostine versus no treatment/placebo, Outcome 2 Salivary flow rate (unstimulated).

Analysis 3.3

Comparison 3 Amifostine versus no treatment/placebo, Outcome 3 Salivary flow rate (unstimulated) ‐ incidence of > 0.1 g in 5 min.

Analysis 3.4

Comparison 3 Amifostine versus no treatment/placebo, Outcome 4 Salivary flow rate (stimulated).

Analysis 3.5

Comparison 3 Amifostine versus no treatment/placebo, Outcome 5 Salivary flow rate (stimulated) ‐ incidence of > 0.1 g in 5 min.

Analysis 3.6

Comparison 3 Amifostine versus no treatment/placebo, Outcome 6 Overall survival at 12 to 24 months postradiotherapy.


Study	Time point	Amifostine	Control	Comments
Haddad 2009	Median follow‐up 34 months after radiotherapy, minimum 26 months			"No differences noted"
Jellema 2006	24 months	3 times weekly = 84% 5 times weekly = 58%	70%	Reported narratively rather than as a risk ratio due to differing results in the amifostine arms

Analysis 3.7

Comparison 3 Amifostine versus no treatment/placebo, Outcome 7 Overall survival ‐ narrative data.

Analysis 3.8

Comparison 3 Amifostine versus no treatment/placebo, Outcome 8 Progression‐free survival at 12 to 24 months postradiotherapy.

Analysis 3.9

Comparison 3 Amifostine versus no treatment/placebo, Outcome 9 Progression‐free survival.


Study	Time point	Amifostine	Control	Comments
Haddad 2009	Median follow‐up 34 months after radiotherapy, minimum 26 months			"No differences noted"

Analysis 3.10

Comparison 3 Amifostine versus no treatment/placebo, Outcome 10 Progression‐free survival ‐ narrative data.

Analysis 3.11

Comparison 3 Amifostine versus no treatment/placebo, Outcome 11 Locoregional tumour control at 12 to 24 months postradiotherapy.


Study	Time point	Amifostine	Control	Comments
Haddad 2009	Median follow‐up 34 months after radiotherapy, minimum 26 months			"No differences noted"
Jellema 2006	24 months	3 times weekly = 67% 5 times weekly = 83%	79%	Reported narratively rather than as a risk ratio due to differing results in the amifostine arms
Patni 2004	24 month	No data	No data	"Amifostine does not alter the response or the survival"

Analysis 3.12

Comparison 3 Amifostine versus no treatment/placebo, Outcome 12 Locoregional tumour control ‐ narrative data.

Analysis 3.13

Comparison 3 Amifostine versus no treatment/placebo, Outcome 13 Disease‐free survival.


Study	Time point	Amifostine	Control	Comments
Patni 2004	24 months	No data	No data	"Amifostine does not alter the response or the survival"
Veerasarn 2006	24 months	No data	No data	"There was no statistical difference in 2‐year disease‐free survival"

Analysis 3.14

Comparison 3 Amifostine versus no treatment/placebo, Outcome 14 Disease‐free survival.

Analysis 3.15

Comparison 3 Amifostine versus no treatment/placebo, Outcome 15 Quality of life (Patient Benefit Questionnaire).

Analysis 4.1

Comparison 4 Amifostine (comparison of dosages), Outcome 1 Xerostomia (0 to 4 scale ‐ grade 2 or above).


Study	Time point	Amifostine 3 times weekly	Amifostine 5 times weekly	Comments
Jellema 2006	24 months	84%	58%

Analysis 4.2

Comparison 4 Amifostine (comparison of dosages), Outcome 2 Overall survival ‐ narrative data.


Study	Time point	Amifostine 3 times weekly	Amifostine 5 times weekly	Comments
Jellema 2006	24 months	67%	83%

Analysis 4.3

Comparison 4 Amifostine (comparison of dosages), Outcome 3 Locoregional tumour control ‐ narrative data.

Analysis 5.1

Comparison 5 Amifostine (intravenous versus subcutaneous), Outcome 1 Xerostomia (0 to 4 scale ‐ grade 2 or above).

Analysis 5.2

Comparison 5 Amifostine (intravenous versus subcutaneous), Outcome 2 Overall survival.

Analysis 5.3

Comparison 5 Amifostine (intravenous versus subcutaneous), Outcome 3 Locoregional tumour control.

Analysis 6.1

Comparison 6 Chinese medicine versus no treatment/placebo, Outcome 1 Xerostomia.

Analysis 6.2

Comparison 6 Chinese medicine versus no treatment/placebo, Outcome 2 Xerostomia.

Analysis 6.3

Comparison 6 Chinese medicine versus no treatment/placebo, Outcome 3 Salivary flow rate (stimulated).

Analysis 6.4

Comparison 6 Chinese medicine versus no treatment/placebo, Outcome 4 Overall survival (12 months postRT).

Analysis 6.5

Comparison 6 Chinese medicine versus no treatment/placebo, Outcome 5 Quality of life (EORTC‐C30).

Analysis 7.1

Comparison 7 Palifermin versus placebo, Outcome 1 Xerostomia (0 to 4 scale ‐ grade 2 or above).

Analysis 7.2

Comparison 7 Palifermin versus placebo, Outcome 2 Overall survival at 42 to 72 months from baseline.

Analysis 7.3

Comparison 7 Palifermin versus placebo, Outcome 3 Progression‐free survival at 42 to 72 months from baseline.

Analysis 8.1

Comparison 8 Bethanechol versus placebo, Outcome 1 Xerostomia (0 to 3 scale ‐ grade 2 or above).

Analysis 8.2

Comparison 8 Bethanechol versus placebo, Outcome 2 Salivary flow rate (unstimulated) ‐ ml/min.

Analysis 8.3

Comparison 8 Bethanechol versus placebo, Outcome 3 Salivary flow rate (stimulated) ‐ ml/min.

Analysis 9.1

Comparison 9 Bethanechol versus artificial saliva, Outcome 1 Xerostomia (dry mouth? yes/no).

Analysis 9.2

Comparison 9 Bethanechol versus artificial saliva, Outcome 2 Salivary flow rate (unstimulated) ‐ ml/min.

Analysis 9.3

Comparison 9 Bethanechol versus artificial saliva, Outcome 3 Salivary flow rate (stimulated) ‐ ml/min.

Analysis 9.4

Comparison 9 Bethanechol versus artificial saliva, Outcome 4 Overall survival.


Study
Büntzel 2010	"comparing the mean value of xerostomia, no statistically significant difference can be seen between the groups"

Analysis 10.1

Comparison 10 Selenium versus no selenium, Outcome 1 Xerostomia.

Analysis 11.1

Comparison 11 Antimicrobial lozenge versus placebo, Outcome 1 Xerostomia (QoL response for dryness).

Analysis 11.2

Comparison 11 Antimicrobial lozenge versus placebo, Outcome 2 Quality of life.

Analysis 12.1

Comparison 12 Polaprezinc versus azulene oral rinse, Outcome 1 Xerostomia (grade 2 or above).

Summary of findings for the main comparison. Pilocarpine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy

Pilocarpine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy
Patient or population: patients receiving radiotherapy on its own or in addition to chemotherapy to the head and neck region Intervention: pilocarpine Comparison: no treatment/placebo
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with no treatment/placebo	Risk with pilocarpine	Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Xerostomia ‐ Up to and including 6 months postRT Studies used different ways of measuring the outcome and therefore we combined the studies using SMD	‐	SMD 0.35 lower (1.04 lower to 0.33 higher)	‐	126 (2 RCTs)	⊕⊝⊝⊝ VERY LOW¹	Insufficient evidence of a difference at this time point and also at the end of RT and 3 months postRT 1 of the 2 studies in this assessment showed inconsistent results when using an alternative way of measuring this outcome at the 6‐month time point. 2 further studies showed insufficient evidence of a difference, 1 at the end of RT and the other at 3 months postRT
Salivary flow rate (unstimulated) ‐ Up to and including 3 months postRT Studies used different ways of measuring the outcome and therefore we combined the studies using SMD	‐	MD 0.06 lower (0.23 lower to 0.11 higher)	‐	24 (1 RCT)	⊕⊝⊝⊝ VERY LOW²	Insufficient evidence of a difference at this time point and also at the end of RT Same results for stimulated salivary flow rates at end of RT, and 3, 6 and 12 months postRT Same results for a further study at the end of RT and 3 months postRT looking at whether or not stimulated and unstimulated salivary flow was > 0 g
Overall survival ‐ Up to and including 6 months postRT	724 per 1000	775 per 1000 (579 to 1000)	RR 1.07 (0.80 to 1.43)	60 (1 RCT)	⊕⊝⊝⊝ VERY LOW³	Insufficient evidence of a difference
Quality of life ‐ Up to and including 6 months postRT McMaster University Head and Neck Questionnaire (HNRQ). Score 1‐7, lower score = poorer quality of life	Control group mean was 5.3	MD 0.20 higher (0.19 lower to 0.59 higher)	‐	90 (1 RCT)	⊕⊝⊝⊝ VERY LOW³	Insufficient evidence of a difference at this time point and also at the end of RT and 3 months postRT
Adverse effects	Insufficient evidence of a difference between groups for any reported adverse event, apart from for sweating where data from 5 studies showed an increased risk associated with pilocarpine (RR 2.98, 95% CI 1.43 to 6.22; P = 0.004; I² = 0%; 389 participants; ⊕⊕⊝⊝ LOW⁴)
The risk in the intervention group (and its 95% CI) 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; RCT: randomised controlled trial; RR: risk ratio; SMD: standardised mean difference; RT: radiotherapy
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect
¹Downgraded by 1 level for risk of bias, 1 level for imprecision (small sample size and 95% CIs include both possibility of benefit and harm), and 1 level for inconsistency (I² = 68%). ²Downgraded by 1 level for risk of bias, and 2 levels for imprecision (single study with 12 participants per group and 95% CIs include both possibility of benefit and harm). ³Downgraded by 1 level for risk of bias, and 2 levels for imprecision (single study and 95% CIs include both possibility of benefit and harm). ⁴Downgraded by 1 level for risk of bias, and 1 level for imprecision (very wide 95% CIs).

Summary of findings for the main comparison. Pilocarpine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy

Summary of findings 2. Amifostine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy

Amifostine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy
Patient or population: patients receiving radiotherapy on its own or in addition to chemotherapy to the head and neck region Intervention: amifostine Comparison: no treatment/placebo
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Outcomes	Risk with no treatment/placebo	Risk with amifostine	Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Xerostomia (0‐4 scale ‐ grade 2 or above) ‐ 12 months postRT	418 per 1000	292 per 1000 (167 to 514)	RR 0.70 (0.40 to 1.23)	682 (7 studies)	⊕⊕⊝⊝ LOW¹	Insufficient evidence of a difference at this time point. However, both at the end of RT (RR 0.35, 95% CI 0.19 to 0.67; 3 studies, 119 participants) and up to 3 months postRT (RR 0.66, 95% CI 0.48 to 0.92; 5 studies, 687 participants), amifostine reduced the risk of developing grade ≥ 2 xerostomia
Salivary flow rate (mg/5 min) (unstimulated) ‐ 12 months postRT	Control group mean was 0.16	MD 0.32 higher (0.09 higher to 0.55 higher)	‐	27 (1 study)	⊕⊝⊝⊝ VERY LOW²	Amifostine led to increased unstimulated saliva flow both at 12 months postRT and at the end of RT, but there was insufficient evidence of a difference at 3 months postRT. This evidence was supported by a further study showing a benefit for amifostine at 12 months postRT when looking at incidence of producing > 0.1 g of saliva over 5 minutes (RR 1.45, 95% CI 1.13 to 1.86; 175 participants). A further study narratively reported no difference Insufficient evidence of a difference in stimulated saliva flow at any time point
Overall survival at 12 to 24 months postRT	450 per 1000**	531 per 1000 (383 to 747)	HR 1.18 (0.85 to 1.66)	271 (2 studies)	⊕⊝⊝⊝ VERY LOW³	Insufficient evidence to determine whether or not amifostine reduces overall survival, progression‐free survival, disease‐free survival or locoregional tumour control up to 24 months postRT
Quality of life (Patient Benefit Questionnaire) ‐ 12 months postRT 8 items each on a 10‐point scale where higher = better QoL	Control group mean was 6.66	MD 0.7 higher (0.2 higher to 1.2 higher)	‐	180 (1 study)	⊕⊝⊝⊝ VERY LOW²	Amifostine led to a small improvement in quality of life at 12 months postRT, but there was insufficient evidence of a difference at the end of RT and 3 months postRT A further study narratively reported no difference at end of RT and 6, 12, 18, and 24 months postRT
Adverse effects	Data from 5 studies showed an increased risk of vomiting associated with amifostine (RR 4.90, 95% CI 2.87 to 8.38; 601 participants; ⊕⊕⊝⊝ LOW⁴) Data from 3 studies showed an increased risk of hypotension associated with amifostine (RR 9.20, 95% CI 2.84 to 29.83; 376 participants; ⊕⊕⊝⊝ LOW⁴) Data from 4 studies showed an increased risk of nausea associated with amifostine (RR 2.60, 95% CI 1.81 to 3.74; 556 participants; ⊕⊕⊝⊝ LOW⁴) Data from 3 studies showed an increased risk of allergic response associated with amifostine (RR 7.51, 95% CI 1.40 to 40.39; 524 participants; ⊕⊕⊝⊝ LOW⁴) There was insufficient evidence of a difference between groups for any other adverse events
The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) 2014 5‐year overall survival rate of patients with head and neck squamous cell carcinoma (www.who.int/selection_medicines/committees/expert/20/applications/HeadNeck.pdf) CI:* confidence interval; HR: hazard ratio; MD: mean difference; QoL: quality of life; RR: risk ratio; RT: radiotherapy
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect
¹Downgraded by 1 level for risk of bias, and 1 level for inconsistency (I² = 83%). ²Downgraded by 1 level for risk of bias, and 2 levels for imprecision (single study and small sample size). ³Downgraded by 1 level for risk of bias, and 2 levels for imprecision (small sample size and 95% CIs include both possibility of benefit and harm). ⁴Downgraded by 1 level for risk of bias, and 1 level for imprecision (very wide 95% CIs).

Summary of findings 2. Amifostine compared to no treatment/placebo for preventing salivary gland dysfunction following radiotherapy

Summary of findings 3. Palifermin compared to placebo for preventing salivary gland dysfunction following radiotherapy

Palifermin compared to placebo for preventing salivary gland dysfunction following radiotherapy
Patient or population: patients receiving radiotherapy on its own or in addition to chemotherapy to the head and neck region Intervention: palifermin Comparison: placebo
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with placebo	Risk with palifermin
Xerostomia (0‐4 scale ‐ grade 2 or above) ‐ Up to and including 3 months postRT	727 per 1000	705 per 1000 (560 to 887)	RR 0.97 (0.77 to 1.22)	471 (3 studies)	⊕⊕⊝⊝ LOW¹	Insufficient evidence of a difference at this time point
Overall survival at 42 to 72 months from baseline	450 per 1000**	450 per 1000 (324 to 626)	HR 1.00 (0.72 to 1.39)	(3 studies)	⊕⊕⊕⊝ MODERATE²	Insufficient evidence to determine whether or not amifostine reduces overall survival and progression‐free survival up to 72 months
Adverse effects	There was insufficient evidence of patients in either group experiencing more or less adverse events
The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) 2014 5‐year overall survival rate of patients with head and neck squamous cell carcinoma (www.who.int/selection_medicines/committees/expert/20/applications/HeadNeck.pdf) CI:* confidence interval; HR: hazard ratio; RR: risk ratio; RT: radiotherapy
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect
¹Downgraded by 1 level for imprecision (95% CIs include both possibility of benefit and harm), and 1 level for inconsistency (I² = 76%). ²Downgraded by 1 level for imprecision (95% CIs include both possibility of benefit and harm).

Summary of findings 3. Palifermin compared to placebo for preventing salivary gland dysfunction following radiotherapy

Table 1. Pilocarpine versus no treatment/placebo (other outcomes)

Outcome	Study ID	Time point	Pilocarpine	Control	Results	Comments
Oral related symptoms (other than salivary gland dysfunction/xerostomia)
Oral discomfort	Gornitsky 2004	End of radiotherapy	Mean 38.7 (SD 31.6) n = 16	Mean 56.7 (SD 26.7) n = 20	Mean difference ‐18.00 (95% CI ‐37.41 to 1.41), P = 0.07
Speech difficulties	Gornitsky 2004		Mean 57.5 (SD 26.8) n = 16	Mean 37.3 (SD 27.5) n = 18	Mean difference 20.20 (95% CI 1.93 to 38.47), P = 0.03
Eating difficulties	Gornitsky 2004		Mean 47.4 (SD 33.9) n = 15	Mean 61.8 (SD 25.4) n = 17	Mean difference ‐14.40 (95% CI ‐35.38 to 6.58), P = 0.18
Mucosal pain	Gornitsky 2004		Mean 38.8 (SD 33.9) n = 17	Mean 53.6 (SD 34.2) n = 19	Mean difference ‐14.80 (95% CI ‐37.07 to 7.47), P = 0.19
Oral complications	Pimentel 2014		1/5	4/6	RR 0.30 (95% CI 0.05 to 1.89), P = 0.20
Adverse events
Sweating	Abacioglu 1997		2/12	0/12	Random‐effects meta‐analysis of 5 studies: RR 2.98 (95% CI 1.43 to 6.22), P = 0.004 Heterogeneity: I² = 0%, P = 0.52
	Fisher 2003		18/118	5/114
	Gornitsky 2004		3/28	1/28
	Lozada‐Nur 1998		5/12	1/12
	Sangthawan 2001		1/29	2/24

Chilling	Abacioglu 1997		1/12	0/12	RR 3.00 (95% CI 0.13 to 67.06), P = 0.49

Nausea	Gornitsky 2004		7/28	5/28	Random‐effects meta‐analysis of 3 studies: RR 1.39 (95% CI 0.63 to 3.05), P = 0.41 Heterogeneity: I² = 0%, P = 0.93
	Haddad 2002		3/18	3/21
	Lozada‐Nur 1998		2/12	1/12

Vomiting	Fisher 2003		13/118	10/114	Random‐effects meta‐analysis of 3 studies: RR 1.28 (95% CI 0.70 to 2.35), P = 0.43 Heterogeneity: I² = 0%, P = 0.92
	Gornitsky 2004		6/28	5/28
	Lozada‐Nur 1998		2/12	1/12

Headache	Gornitsky 2004		2/28	3/28	RR 0.67 (95% CI 0.12 to 3.69), P = 0.64

Excessive lacrimation (tears)	Fisher 2003		3/118	0/114	Random‐effects meta‐analysis of 3 studies: RR 2.54 (95% CI 0.70 to 9.17), P = 0.15 Heterogeneity: I² = 0%, P = 0.71
	Haddad 2002		1/18	0/21
	Sangthawan 2001		4/25	2/22

Dysphasia	Lozada‐Nur 1998		3/12	2/12	RR 1.50 (95% CI 0.30 to 7.43), P = 0.62

Weakness	Fisher 2003		3/118	2/114	RR 1.45 (95% CI 0.25 to 8.51), P = 0.68

Nervous	Gornitsky 2004		0/28	1/28	Random‐effects meta‐analysis of 2 studies: RR 1.02 (95% CI 0.11 to 9.33), P = 0.99 Heterogeneity: I² = 0%, P = 0.33
Nervous	Lozada‐Nur 1998		1/12	0/12

Rhinitis	Fisher 2003		2/118	5/114	Random‐effects meta‐analysis of 3 studies: RR 0.87 (95% CI 0.41 to 1.86), P = 0.72 Heterogeneity: I² = 0%, P = 0.53
	Lozada‐Nur 1998		1/12	1/12
	Sangthawan 2001		8/29	6/24

Blurred vision	Lozada‐Nur 1998		1/12	0/12	RR 3.00 (95% CI 0.13 to 67.06), P = 0.49

Urinary frequency	Fisher 2003		7/118	5/114	Random‐effects meta‐analysis of 2 studies: RR 0.87 (95% CI 0.43 to 1.75), P = 0.70 Heterogeneity: I² = 0%, P = 0.32
Urinary frequency	Sangthawan 2001		6/25	8/22

Dizziness	Gornitsky 2004		0/28	2/28	Random‐effects meta‐analysis of 2 studies: RR 0.80 (95% CI 0.18 to 3.45), P = 0.76 Heterogeneity: I² = 13%, P = 0.28
Dizziness	Sangthawan 2001		4/25	3/22

Palpitation	Sangthawan 2001		0/25	4/22	RR 0.10 (95% CI 0.01 to 1.73), P = 0.11

Skin flushing	Fisher 2003		1/118	0/114	RR 2.90 (95% CI 0.12 to 70.44), P = 0.51

Motor tremors	Fisher 2003		2/118	1/114	RR 1.93 (95% CI 0.18 to 21.02), P = 0.59

Sleep problems	Gornitsky 2004	End of radiotherapy	Mean 37.3 (SD 36.4) n = 17	Mean 49.6 (SD 36.9) n = 19	Mean difference ‐12.30 (95% CI ‐36.27 to 11.67), P = 0.31

RTOG (grade 3; mucous membrane, pharynx and larynx)	Warde 2002					No statistically significant difference between treatment groups
CI = confidence interval; RR = risk ratio; RTOG = Radiation Therapy Oncology Group; SD = standard deviation.

Table 1. Pilocarpine versus no treatment/placebo (other outcomes)

Table 2. Biperiden plus pilocarpine versus no treatment/placebo (other outcomes)

Outcome	Study ID	Time point	Pilocarpine	Control	Results	Comments
Dysphagia (WHO grade 3+)	Rode 1999	12 months after RT	1/30	4/30	RR 0.25 (95% CI 0.03 to 2.11), P = 0.20
CI = confidence interval; RR = risk ratio; RT = radiotherapy; WHO = World Health Organization.

Table 2. Biperiden plus pilocarpine versus no treatment/placebo (other outcomes)

Table 3. Amifostine versus no treatment/placebo (other outcomes)

Outcome	Study ID	Time point	Amifostine	Control	Results	Comments
Quality of life	Jellema 2006	Assessed at end of RT and 6, 12, 18 and 24 months after RT	No data	No data	"No significant differences between the 3 treatment arms"

Dysphagia (difficulty in swallowing) (0‐4 scale): grade 3 and above	Antonadou 2002	End of RT	14/22	23/23	Random‐effects meta‐analysis of 2 studies: RR 0.50 (95% CI 0.17 to 1.48); P = 0.21 Heterogeneity: I² = 40%, P = 0.20
	Büntzel 1998	End of RT	1/14	5/14
	Antonadou 2002	4 weeks after RT	2/22	3/23	RR 0.70 (95% CI 0.13 to 3.78); P = 0.68	By 8 weeks after RT, no participants had grade 3 or above dysphagia

Dysgeusia (taste disturbance) (0‐4 scale): grade 2 and above	Büntzel 1998	End of RT	3/14	14/14	RR 0.24 (95% CI 0.10 to 0.61); P = 0.003

Cost data (mean per patient supportive care costs)	Büntzel 1998	End of RT	USD 4401	USD 5873	P = 0.02

Vomiting	Antonadou 2002		1/22	0/23	Random‐effects meta‐analysis of 5 studies: RR 4.90 (95% CI 2.87 to 8.38); P < 0.00001 Heterogeneity: I² = 0%, P = 0.96
	Brizel 2000		55/150	11/153
	Buentzel 2006		8/66	2/64
	He 2004		1/17	0/15		"1 patient left due to gastrointestinal tract reaction/side effect, all other patients completed the treatment" "At the beginning of treatment, nausea and vomiting was obvious for amifostine group, but after treating with metoclopramide, there was no significant difference between 2 groups in gastrointestinal tract reaction/side effect"
	Jellema 2006		10/60	0/31
	Peng 2006		10/18			Data not reported in control group. Unknown if this was due to 0 events
	Veerasarn 2006		18/32			Data not reported in control group. Unknown if this was due to 0 events

Hypotension	Antonadou 2002		3/22	0/23	Random‐effects meta‐analysis of 3 studies: RR 9.20 (95% CI 2.84 to 29.83); P = 0.0002 Heterogeneity: I² = 0%, P = 0.88
	Brizel 2000		22/150	2/153
	Büntzel 1998		2/14	0/14
	Veerasarn 2006		5/32			Data not reported in control group. Unknown if this was due to 0 events

Nausea	Brizel 2000		66/150	25/153	Random‐effects meta‐analysis of 4 studies: RR 2.60 (95% CI 1.81 to 3.74); P < 0.00001 Heterogeneity: I² = 0%, P = 0.45
	Buentzel 2006		4/66	4/64
	He 2004		1/17	0/15		"1 patient left due to gastrointestinal tract reaction/side effect, all other patients completed the treatment" "At the beginning of treatment, nausea and vomiting was obvious for amifostine group, but after treating with metoclopramide, there was no significant difference between 2 groups in gastrointestinal tract reaction/side effect"
	Jellema 2006		23/60	3/31
	Peng 2006		10/18			Data not reported in control group. Unknown if this was due to 0 events
	Veerasarn 2006		20/32			Data not reported in control group. Unknown if this was due to 0 events

Allergic response	Brizel 2000		8/150	0/153	Random‐effects meta‐analysis of 3 studies: RR 7.51 (95% CI 1.40 to 40.39); P = 0.02 Heterogeneity: I² = 0%, P = 0.77
	Buentzel 2006		2/66	0/64
	Jellema 2006		4/60	0/31

Asthenia (weakness or lack of energy)	Buentzel 2006		3/66	1/64	RR 2.91 (95% CI 0.31 to 27.24); P = 0.35

Alopecia	Vacha 2003					Similar in both groups and increased continuously during the treatment

Skin toxicity	Vacha 2003					Similar in both groups and increased continuously during the treatment

Hot flush	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
Hot flush	Veerasarn 2006		17/32			Data not reported in control group. Unknown if this was due to 0 events

Somnolence (drowsiness)	Veerasarn 2006		18/32			Data not reported in control group. Unknown if this was due to 0 events

Sneezing	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
Sneezing	Veerasarn 2006		13/32			Data not reported in control group. Unknown if this was due to 0 events

Hiccup	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
Hiccup	Veerasarn 2006		10/32			Data not reported in control group. Unknown if this was due to 0 events

Dizziness	Peng 2006					"...dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"

Fatigue	Peng 2006					"..dizziness, fatigue, hiccup, sneezing, facial flush all in less than 5% of the patients"
CI = confidence interval; RR = risk ratio; RT = radiotherapy; USD = US dollars.

Table 3. Amifostine versus no treatment/placebo (other outcomes)

Table 4. Amifostine: comparison of different doses (other outcomes)

Outcome	Study ID	Time point	Amifostine 3 times weekly	Amifostine 5 times weekly	Results	Comments
Quality of life	Jellema 2006	Assessed at end of RT and 6, 12, 18 and 24 months after RT	No data	No data	"No significant differences between the 3 treatment arms"
Nausea	Jellema 2006		9/30	14/30	RR 0.64 (95% CI 0.33 to 1.25); P = 0.19
Vomiting	Jellema 2006		2/30	8/30	RR 0.25 (95% CI 0.06 to 1.08); P = 0.06
Allergic response	Jellema 2006		2/30	2/30	RR 1.00 (95% CI 0.15 to 6.64); P = 1
CI = confidence interval; RR = risk ratio; RT = radiotherapy.

Table 4. Amifostine: comparison of different doses (other outcomes)

Table 5. Amifostine: different routes of administration (other outcomes)

Outcome	Study ID	Time point	Intravenous	Subcutaneous	Results	Comments
Nausea/vomiting	Bardet 2011		29%	36%	P = 0.267
Hypotension	Bardet 2011		20%	8%	P = 0.007
Skin rash	Bardet 2011		10%	22%	P = 0.012
Local pain at injection site	Bardet 2011		0%	8%	P = 0.001
Fever	Bardet 2011		2%	0%	P = 0.256
Asthenia (weakness or lack of energy)	Bardet 2011		1%	6%	P = 0.054

Table 5. Amifostine: different routes of administration (other outcomes)

Table 6. Chinese medicine (other outcomes)

Outcome	Study ID	Intervention	Time point	Study	Control	Results	Comments
Dysphasia (difficulty in swallowing) (score for EORTC‐H&N35 questionnaire: mean (SD))	Lin 2014	Chinese medicine (TWBXM)	End of RT	50.2 (26.3); n = 35	38.9 (25.9); n = 33	P = 0.07
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	30.2 (29.8); n = 23	26.7 (24.8); n = 21	P = 0.65
Dysgeusia (taste disturbance) (0 to 3 scale): grade 1 and above	Lin 2014	Chinese medicine (TWBXM)	End of RT	32/38	29/33	RR 0.96 (95% CI 0.79 to 1.16); P = 0.13
Speech difficulty (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	36.3 (26.7); n = 35	28.6 (26.2); n = 33	P = 0.23
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	27.4 (28.6); n = 23	22.7 (19.5); n = 21	P = 0.50
Difficulty in mouth opening (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	39.6 (28.2); n = 35	41.4 (27.7); n = 33	P = 0.79
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	32.2 (31.5); n = 23	33.3 (24.1); n = 21	P = 0.88
Difficulty in mouth opening (0 to 2 scale): grade 1 and above	Hu 2005	Chinese medicine (Shenqi Fanghou recipe)	"During the treatment"	22/70	52/70	RR 0.42 (95% CI 0.29 to 0.61); P < 0.001
Skin toxicity (0 to 3 scale): grade 1 and above	Lin 2014	Chinese medicine (TWBXM)	End of RT	35/38	30/33	RR 1.01 (95% CI 0.88 to 1.17); P = 0.82
Skin toxicity (0 to 4 scale): grade 1 and above	Hu 2005	Chinese medicine (Shenqi Fanghou recipe)	"During the treatment"	57/70	68/70	RR 0.84 (95% CI 0.74 to 0.94); P = 0.002
Skin toxicity (prevalence according to RTOG standards)	Han 2010	Chinese medicine (Jinlong capsule)		46.82%	58.32%		Quote: "toxicities during and after treatment were assessed" Comment: time point for assessment unclear; minor discrepancies in presented data
Nausea/vomiting (0 to 3 scale): grade 1 and above	Lin 2014	Chinese medicine (TWBXM)	End of RT	12/38	4/33	RR 2.61 (95% CI 0.93 to 7.30); P = 0.183
Hoarseness	Lin 2014	Chinese medicine (TWBXM)	End of RT	1/38	3/33	RR 0.29 (95% CI 0.03 to 2.65); P = 0.26
Fatigue (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	43.2 (26.2); n = 35	42.4 (23.0); n = 33	P = 0.88
Fatigue (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	31.2 (28.3); n = 23	36.4 (25.0); n = 21	P = 0.51
Pain (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	46.8 (23.2); n = 35	41.7 (27.4); n = 33	P = 0.40
Pain (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	35.9 (27.0); n = 23	40.9 (29.9); n = 21	P = 0.54
Pain (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	55.4 (25.1); n = 35	42.4 (20.5); n = 33	P = 0.02
Pain (mean (SD) score for EORTC‐H&N35 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	31.6 (24.2); n = 23	37.8 (23.3); n = 21	P = 0.35
Dyspnea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	17.1 (23.1); n = 35	16.7 (20.7); n = 33	P = 0.93
Dyspnea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	20.5 (21.2); n = 23	13.6 (22.2); n = 21	P = 0.28
Insomnia (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	40.5 (25.0); n = 35	31.2 (25.3); n = 33	P = 0.13
Insomnia (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	30.8 (24.8); n = 23	31.8 (28.1); n = 21	P = 0.28
Appetite loss (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	45.0 (30.7); n = 35	45.8 (29.0); n = 33	P = 0.91
	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	28.2 (26.1); n = 23	34.9 (30.0); n = 21	P = 0.42
Constipation (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	37.8 (27.4); n = 35	29.2 (20.3); n = 33	P = 0.15
Constipation (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	29.5 (30.3); n = 23	25.7 (20.4); n = 21	P = 0.63
Diarrhoea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	End of RT	9.0 (15.0); n = 35	6.2 (13.2); n = 33	P = 0.42
Diarrhoea (mean (SD) score for EORTC‐C30 questionnaire)	Lin 2014	Chinese medicine (TWBXM)	1 month after RT	9.0 (15.1); n = 23	6.1 (16.7); n = 21	P = 0.53
Adverse effects	Han 2010	Chinese medicine (Jinlong capsule)		Leukopenia, nausea, vomiting, 1 participant had dizziness and blood pressure drop, 1 participant had skin rash	Not reported
Adverse effects	Hu 2005	Chinese medicine (Shenqi Fanghou recipe)	"During the treatment"	No adverse event	Not reported
CI = confidence interval; EORTC = European Organisation for Research and Treatment of Cancer; H&N = head and neck; RR = risk ratio; RT = radiotherapy; RTOG = Radiation Therapy Oncology Group; SD = standard deviation; TWBXM = Tianwang Buxin Mini‐pills.

Table 6. Chinese medicine (other outcomes)

Table 7. Palifermin versus placebo (other outcomes)

Outcome	Study ID	Time point	Palifermin	Placebo	Results
Oral related symptoms (other than salivary gland dysfunction/xerostomia)
Dysphagia	Le 2011	3 months postRT	29/94	19/91	Random‐effects meta‐analysis of 3 studies: RR 1.32 (95% CI 0.55 to 3.13); P = 0.54 Heterogeneity: I² = 94%, P < 0.00001
	Brizel 2008		61/64	31/32
	Henke 2011		32/92	20/93
Mouth and throat soreness ‐ 0 (no soreness) to 4 (extreme soreness) OMWQ‐HN scale	Le 2011	3 months postRT	n = 94, mean = 1.66, SD = 0.73	n = 94, mean = 1.86, SD = 0.65	Random‐effects meta‐analysis of 2 studies: mean difference ‐0.12 (95% CI ‐0.27 to 0.02); P = 0.10 Heterogeneity: I² = 13%, P = 0.28
	Henke 2011	3 months postRT	n = 92, mean = 1.52, SD = 0.69	n = 94, mean = 1.57, SD = 0.63
Adverse events
Nausea	Le 2011		47/94	42/91	Random‐effects meta‐analysis of 2 studies: RR 0.96 (95% CI 0.77 to 1.19); P = 0.69 Heterogeneity: I² = 28%, P = 0.24
Nausea	Brizel 2008		48/67	26/32
Fever	Brizel 2008		30/67	13/32	RR 1.10 (95% CI 0.67 to 1.81); P = 0.70
Constipation	Le 2011		31/94	24/91	Random‐effects meta‐analysis of 2 studies: RR 1.15 (95% CI 0.82 to 1.60); P = 0.42 Heterogeneity: I² = 0%, P = 0.57
Constipation	Brizel 2008		28/67	13/32
Diarrhoea	Brizel 2008		14/67	8/32	Random‐effects meta‐analysis of 2 studies: RR 1.28 (95% CI 0.49 to 3.36); P = 0.61 Heterogeneity: I² = 57%, P = 0.13
Diarrhoea	Henke 2011		11/92	5/93
Insomnia	Brizel 2008		12/67	4/32	Random‐effects meta‐analysis of 2 studies: RR 0.77 (95% CI 0.23 to 2.55); P = 0.67 Heterogeneity: I² = 63%, P = 0.10
Insomnia	Henke 2011		5/92	12/93
Dyspnea	Brizel 2008		9/67	1/32	RR 1.10 (95% CI 0.67 to 1.81); P = 0.70
Cough	Brizel 2008		8/67	5/32	RR 0.76 (95% CI 0.27 to 2.15); P = 0.61
Headache	Brizel 2008		8/67	2/32	Random‐effects meta‐analysis of 2 studies: RR 2.13 (95% CI 0.86 to 5.28); P = 0.10 Heterogeneity: I² = 0%, P = 0.86
Headache	Henke 2011		9/92	4/93
Decreased weight	Le 2011		29/94	27/91	Random‐effects meta‐analysis of 2 studies: RR 1.01 (95% CI 0.67 to 1.52); P = 0.96 Heterogeneity: I² = 0%, P = 0.73
Decreased weight	Brizel 2008		7/67	4/32
Dizziness	Brizel 2008		5/67	4/32	RR 0.60 (95% CI 0.17 to 2.07); P = 0.42
Anxiety	Brizel 2008		4/67	5/32	RR 0.38 (95% CI 0.11 to 1.33); P = 0.13
Hypomagnesemia	Brizel 2008		4/67	4/32	RR 0.48 (95% CI 0.13 to 1.79); P = 0.27
Vomiting	Le 2011		26/94	26/91	Random‐effects meta‐analysis of 2 studies: RR 0.98 (95% CI 0.72 to 1.33); P = 0.89 Heterogeneity: I² = 0%, P = 0.96
Vomiting	Brizel 2008		33/67	16/32
Radiation skin injury	Le 2011		25/94	13/91	RR 1.10 (95%CI 0.67 to 1.81); P = 0.70
Anaemia	Le 2011		21/94	34/91	Random‐effects meta‐analysis of 2 studies: RR 0.83 (95% CI 0.33 to 2.05); P = 0.68 Heterogeneity: I² = 54%, P = 0.14
Anaemia	Brizel 2008		10/67	3/32
Fatigue	Le 2011		21/94	20/91	Random‐effects meta‐analysis of 3 studies: RR 0.88 (95% CI 0.60 to 1.30); P = 0.52 Heterogeneity: I² = 2%, P = 0.36
	Henke 2011		7/92	14/93
	Brizel 2008		17/67	8/32
Leukopenia	Le 2011		21/94	12/91	Random‐effects meta‐analysis of 2 studies: RR 1.01 (95% CI 0.37 to 2.78); P = 0.98 Heterogeneity: I² = 79%, P = 0.03
Leukopenia	Henke 2011		12/92	20/93
Granulocytopenia	Brizel 2008		20/67	6/32	RR 1.59 (95% CI 0.47 to 5.39); P = 0.45
Pharyngolaryngeal pain	Le 2011		20/94	23/91	RR 0.84 (95% CI 0.50 to 1.42); P = 0.52
Hypokalemia	Le 2011		19/94	8/91	RR 2.04 (95% CI 0.98 to 4.28); P = 0.06
Pyrexia	Le 2011		16/94	19/91	RR 0.82 (95% CI 0.45 to 1.48); P = 0.50
Mucosal inflammation	Henke 2011		4/92	10/93	RR 0.40 (95% CI 0.13 to 1.24); P = 0.11
Asthenia	Henke 2011		13/92	7/93	RR 1.88 (95% CI 0.78 to 4.49); P = 0.16
Abdominal pain	Henke 2011		7/92	2/93	RR 3.54 (95% CI 0.75 to 16.58); P = 0.11
Back pain	Henke 2011		6/92	1/93	RR 6.07 (95% CI 0.74 to 49.40); P = 0.09
Febrile neutropenia	Henke 2011		1/92 Considered "serious adverse event"	0/93	RR 3.03 (95% CI 0.13 to 73.48); P = 0.50
Dehydration	Le 2011		13/94	19/91	Random‐effects meta‐analysis of 3 studies: RR 0.75 (95% CI 0.45 to 1.25); P = 0.27 Heterogeneity: I² = 30%, P = 0.24
	Henke 2011		6/92	13/93
	Brizel 2008		20/67	8/32
CI = confidence interval; OMWQ‐HN = Oral Mucositis Weekly Questionnaire ‐ Head and Neck Cancer; RR = risk ratio; RT = radiotherapy; SD = standard deviation.

Table 7. Palifermin versus placebo (other outcomes)

Table 8. Bethanechol versus placebo (other outcomes)

Outcome	Study ID	Results
Adverse effects	Jaguar 2015	No statistical difference between the groups in bethanechol‐related toxicity. Quote: "No patient experienced severe (grade 3) toxicity and no one dropped out of the study due to adverse effects"

Table 8. Bethanechol versus placebo (other outcomes)

Table 9. Bethanechol versus artificial saliva (other outcomes)

Outcome	Study ID	Time point	Bethanechol	Artificial saliva	Results	Comments
Lacrimation (watering eyes)	Jham 2007	End of RT	3/22	0/21	RR 6.70 (95% CI 0.37 to 122.29); P = 0.2
Nervousness			3/22	0/21	RR 6.70 (95% CI 0.37 to 122.29); P = 0.2
Frequent urination			3/22	0/21	RR 6.70 (95% CI 0.37 to 122.29); P = 0.2
Sweating			2/22	0/21	RR 4.78 (95% CI 0.24 to 94.12); P = 0.3	"1 patient using bethanechol dropped out of the study due to excessive sweating (Grade 2 severity; National Cancer Institute Common Terminology Criteria for Adverse Events – NCI CTCAE, v 3"
Warm face			2/22	0/21	RR 4.78 (95% CI 0.24 to 94.12); P = 0.3
Cramps			1/22	0/21	RR 2.87 (95% CI 0.12 to 66.75); P = 0.51
Diarrhoea			1/22	0/21	RR 2.87 (95% CI 0.12 to 66.75); P = 0.51
Nausea			1/22	2/21	RR 0.48 (95% CI 0.05 to 4.88); P = 0.53
CI = confidence interval; RR = risk ratio; RT = radiotherapy.

Table 9. Bethanechol versus artificial saliva (other outcomes)

Table 10. Selenium versus no intervention (other outcomes)

Outcome	Study ID	Time point	Reported in text
Loss of taste	Büntzel 2010	6 weeks after end RT	"Ageusia was milder in the selenium group. But the difference was not significant"
Dysphagia	Büntzel 2010	6 weeks after end RT	"The only significant difference was observed at week 7, when the selenium group had developed a mean value of 1.533 versus 2.167 in the control group (P = 0.05)"
Adverse events	Büntzel 2010	6 weeks after end RT	"23 serious adverse events (SAEs) were seen in the selenium group, compared to 22 in the control group (P = 0.476). No statistically significant differences in toxicities were found using the 2‐tailed Fisher's exact test"
RT = radiotherapy.

Table 10. Selenium versus no intervention (other outcomes)

Table 11. Antiseptic mouthrinse versus placebo (other outcomes)

Outcome	Study ID	Time point	Antiseptic rinse	Placebo	Results
Drooling	Lanzós 2010	4 weeks from baseline	Increased 6 No change or decreased 8 6/14	Increased 3 No change or decreased 7 3/10	RR 1.43 (95% CI 0.46 to 4.39); P = 0.53
Adverse events	Lanzós 2010				"No relevant adverse events were reported in any group"
CI = confidence interval; RR = risk ratio.

Table 11. Antiseptic mouthrinse versus placebo (other outcomes)

Table 12. Antimicrobial lozenge versus placebo (other outcomes)

Outcome	Study ID	Time point	Antimicrobial lozenge	Placebo	Results
Mouth pain	Duncan 2005	Worse over 6 months	32/66	32/62	RR 0.94 (95% CI 0.66 to 1.33); P=0.72
Sore/burning mouth	Duncan 2005	Worse over 6 months	32/65	32/62	RR 0.95 (95% CI 0.68 to 1.35); P=0.79
Throat pain	Duncan 2005	Worse over 6 months	29/66	36/65	RR 0.79 (95% CI 0.56 to 1.12); P=0.19
Dryness in mouth	Duncan 2005	Worse over 6 months	55/66	46/65	RR 1.18 (95% CI 0.97 to 1.42); P=0.09
Nausea	Duncan 2005	Worse over 6 months	27/66	14/65	RR 1.90 (95% CI 1.10 to 3.28); P=0.02
Diarrhoea	Duncan 2005	Worse over 6 months	6/66	3/65	RR 1.97 (95% CI 0.51 to 7.54); P=0.32
Constipation	Duncan 2005	Worse over 6 months	24/66	26/65	RR 0.91 (95% CI 0.59 to 1.42); P=0.67
CI = confidence interval; RR = risk ratio.

Table 12. Antimicrobial lozenge versus placebo (other outcomes)

Table 13. Polaprezinc versus azulene rinse (other outcomes)

Outcome	Study ID	Time point	Polaprezinc	Azulene rinse	Results
Pain > 2 (0‐3 scale)	Watanabe 2010	Over RT period	5/16	13/15	RR 0.36 (95% CI 0.17 to 0.77); P = 0.008
Taste disturbance > 2 (0‐3 scale)	Watanabe 2010	Over RT period	1/16	8/15	RR 0.12 (95% CI 0.02 to 0.83); P = 0.03
Disability of oral intake	Watanabe 2010	Over RT period	2/16	6/15	RR 0.31 (95% CI 0.07 to 1.31); P = 0.11
CI = confidence interval; RR = risk ratio; RT = radiotherapy.

Table 13. Polaprezinc versus azulene rinse (other outcomes)

Table 14. Venalot Depot (coumarin/troxerutin) versus placebo

Outcome	Study ID	Results
Adverse events	Grötz 2001	"No adverse events could be attributed to the experimental medication"

Table 14. Venalot Depot (coumarin/troxerutin) versus placebo

Comparison 1. Pilocarpine versus no treatment/placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia Show forest plot	6		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only

1.1 End of radiotherapy	4	122	Std. Mean Difference (IV, Random, 95% CI)	0.20 [‐0.16, 0.56]
1.2 Up to and including 3 months postradiotherapy	3	125	Std. Mean Difference (IV, Random, 95% CI)	0.02 [‐0.33, 0.37]
1.3 Up to and including 6 months postradiotherapy	2	126	Std. Mean Difference (IV, Random, 95% CI)	‐0.35 [‐1.04, 0.33]
2 Xerostomia (LENT‐SOMA scale) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

2.1 Up to and including 6 months postradiotherapy	1		Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
3 Xerostomia Show forest plot	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

3.1 End of radiotherapy	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
3.2 Up to and including 3 months postradiotherapy	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
4 Salivary flow rate (unstimulated) Show forest plot	3		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only

4.1 End of radiotherapy	3	76	Std. Mean Difference (IV, Random, 95% CI)	0.24 [‐0.24, 0.72]
4.2 Up to and including 3 months postradiotherapy	1	24	Std. Mean Difference (IV, Random, 95% CI)	‐0.27 [‐1.07, 0.54]
5 Salivary flow rate (stimulated) Show forest plot	3		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only

5.1 End of radiotherapy	2	58	Std. Mean Difference (IV, Random, 95% CI)	0.08 [‐0.44, 0.59]
5.2 Up to and including 3 months postradiotherapy	1	24	Std. Mean Difference (IV, Random, 95% CI)	0.40 [‐0.41, 1.21]
5.3 Up to and including 6 months postradiotherapy	1	9	Std. Mean Difference (IV, Random, 95% CI)	0.52 [‐0.84, 1.87]
5.4 Up to and including 12 months postradiotherapy	1	9	Std. Mean Difference (IV, Random, 95% CI)	0.53 [‐0.83, 1.88]
6 Salivary flow rate (> 0 g) unstimulated Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

6.1 End of radiotherapy	1	154	Risk Ratio (M‐H, Random, 95% CI)	1.90 [0.98, 3.69]
6.2 Up to and including 3 months postradiotherapy	1	152	Risk Ratio (M‐H, Random, 95% CI)	2.0 [0.86, 4.68]
7 Salivary flow rate (> 0 g) stimulated Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

7.1 End of radiotherapy	1	138	Risk Ratio (M‐H, Random, 95% CI)	1.87 [0.77, 4.52]
7.2 Up to and including 3 months postradiotherapy	1	139	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.23, 2.11]
8 Overall survival Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

9 Quality of life Show forest plot	2		Mean Difference (IV, Random, 95% CI)	Totals not selected

9.1 End of radiotherapy	1		Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
9.2 Up to and including 3 months postradiotherapy	1		Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
9.3 Up to and including 6 months postradiotherapy	1		Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]

Comparison 1. Pilocarpine versus no treatment/placebo

Comparison 2. Biperiden plus pilocarpine versus no treatment

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Salivary flow rate (unstimulated) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

1.1 End of radiotherapy	1	40	Mean Difference (IV, Random, 95% CI)	0.02 [‐0.08, 0.12]
1.2 Up to and including 3 months postradiotherapy	1	17	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2 Salivary flow rate (> 0 g) unstimulated Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

2.1 End of radiotherapy	1	60	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.77, 1.58]
2.2 Up to and including 3 months postradiotherapy	1	60	Risk Ratio (M‐H, Random, 95% CI)	7.50 [1.88, 29.99]
2.3 Up to and including 6 months postradiotherapy	1	60	Risk Ratio (M‐H, Random, 95% CI)	29.00 [1.81, 465.07]
2.4 Up to and including 12 months postradiotherapy	1	60	Risk Ratio (M‐H, Random, 95% CI)	35.00 [2.20, 556.71]

Comparison 2. Biperiden plus pilocarpine versus no treatment

Comparison 3. Amifostine versus no treatment/placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (0 to 4 scale ‐ grade 2 or above) Show forest plot	8		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 End of radiotherapy	3	119	Risk Ratio (M‐H, Random, 95% CI)	0.35 [0.19, 0.67]
1.2 Up to and including 3 months postradiotherapy	5	687	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.48, 0.92]
1.3 12 months postradiotherapy	7	682	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.40, 1.23]
2 Salivary flow rate (unstimulated) Show forest plot	2		Mean Difference (IV, Random, 95% CI)	Subtotals only

2.1 End of radiotherapy	2	83	Mean Difference (IV, Random, 95% CI)	0.34 [0.07, 0.61]
2.2 Up to and including 3 months postradiotherapy	1	41	Mean Difference (IV, Random, 95% CI)	0.13 [‐0.90, 1.16]
2.3 12 months postradiotherapy	1	27	Mean Difference (IV, Random, 95% CI)	0.32 [0.09, 0.55]
3 Salivary flow rate (unstimulated) ‐ incidence of > 0.1 g in 5 min Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

3.1 12 months postradiotherapy	1	175	Risk Ratio (M‐H, Random, 95% CI)	1.45 [1.13, 1.86]
4 Salivary flow rate (stimulated) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

4.1 End of radiotherapy	1	47	Mean Difference (IV, Random, 95% CI)	‐0.09 [‐1.48, 1.30]
4.2 Up to and including 3 months postradiotherapy	1	41	Mean Difference (IV, Random, 95% CI)	0.38 [‐1.43, 2.19]
4.3 12 months postradiotherapy	1	27	Mean Difference (IV, Random, 95% CI)	0.82 [‐0.47, 2.11]
5 Salivary flow rate (stimulated) ‐ incidence of > 0.1 g in 5 min Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

5.1 12 months postradiotherapy	1	173	Risk Ratio (M‐H, Random, 95% CI)	1.12 [0.89, 1.41]
6 Overall survival at 12 to 24 months postradiotherapy Show forest plot	2	271	Hazard ratio (Random, 95% CI)	1.18 [0.85, 1.66]

7 Overall survival ‐ narrative data Show forest plot			Other data	No numeric data

8 Progression‐free survival at 12 to 24 months postradiotherapy Show forest plot	2	247	Hazard ratio (Random, 95% CI)	0.94 [0.70, 1.27]

9 Progression‐free survival Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

9.1 18 months postradiotherapy	1	45	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.81, 1.51]
10 Progression‐free survival ‐ narrative data Show forest plot			Other data	No numeric data

11 Locoregional tumour control at 12 to 24 months postradiotherapy Show forest plot	2	279	Hazard ratio (Random, 95% CI)	0.90 [0.74, 1.11]

12 Locoregional tumour control ‐ narrative data Show forest plot			Other data	No numeric data

13 Disease‐free survival Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

13.1 24 months postradiotherapy	1	170	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.73, 1.21]
14 Disease‐free survival Show forest plot			Other data	No numeric data

15 Quality of life (Patient Benefit Questionnaire) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

15.1 End of radiotherapy	1	298	Mean Difference (IV, Random, 95% CI)	0.38 [‐0.07, 0.83]
15.2 Up to and including 3 months postradiotherapy	1	233	Mean Difference (IV, Random, 95% CI)	0.52 [‐0.02, 1.06]
15.3 12 months postradiotherapy	1	180	Mean Difference (IV, Random, 95% CI)	0.70 [0.20, 1.20]

Comparison 3. Amifostine versus no treatment/placebo

Comparison 4. Amifostine (comparison of dosages)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (0 to 4 scale ‐ grade 2 or above) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 12 months postradiotherapy	1	49	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.58, 1.53]
2 Overall survival ‐ narrative data Show forest plot			Other data	No numeric data

3 Locoregional tumour control ‐ narrative data Show forest plot			Other data	No numeric data

Comparison 4. Amifostine (comparison of dosages)

Comparison 5. Amifostine (intravenous versus subcutaneous)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (0 to 4 scale ‐ grade 2 or above) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 Up to and including 3 months postradiotherapy	1	263	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.76, 1.40]
1.2 12 months postradiotherapy	1	127	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.42, 0.88]
2 Overall survival Show forest plot	1		Hazard Ratio (Random, 95% CI)	Subtotals only

2.1 48 months after radiotherapy	1		Hazard Ratio (Random, 95% CI)	1.36 [0.89, 2.10]
3 Locoregional tumour control Show forest plot	1		Hazard Ratio (Random, 95% CI)	Subtotals only

3.1 48 months after radiotherapy	1		Hazard Ratio (Random, 95% CI)	1.34 [0.76, 2.36]

Comparison 5. Amifostine (intravenous versus subcutaneous)

Comparison 6. Chinese medicine versus no treatment/placebo

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

1.1 End of radiotherapy: Shenqi Fanghou recipe versus no intervention	1	140	Risk Ratio (M‐H, Random, 95% CI)	0.39 [0.28, 0.55]
1.2 End of radiotherapy: TWBXM versus placebo	1	71	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.78, 1.03]
1.3 Up to and including 3 months postradiotherapy: Jinlong capsules versus no intervention	1	95	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.59, 1.36]
2 Xerostomia Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

2.1 End of radiotherapy: TWBXM versus placebo	1	68	Mean Difference (IV, Random, 95% CI)	‐2.41 [‐16.19, 11.37]
2.2 Up to and including 3 months postradiotherapy: TWBXM versus placebo	1	44	Mean Difference (IV, Random, 95% CI)	‐0.10 [‐17.21, 17.01]
3 Salivary flow rate (stimulated) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

3.1 End of radiotherapy: Chinese medicine versus no intervention	1	50	Mean Difference (IV, Random, 95% CI)	0.09 [0.03, 0.15]
4 Overall survival (12 months postRT) Show forest plot	1	78	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.84, 1.30]

5 Quality of life (EORTC‐C30) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

5.1 End of radiotherapy: TWBXM versus placebo	1	68	Mean Difference (IV, Random, 95% CI)	2.39 [‐8.74, 13.52]
5.2 Up to and including 3 months postradiotherapy: TWBXM versus placebo	1	44	Mean Difference (IV, Random, 95% CI)	1.93 [‐13.04, 16.90]

Comparison 6. Chinese medicine versus no treatment/placebo

Comparison 7. Palifermin versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (0 to 4 scale ‐ grade 2 or above) Show forest plot	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 Up to and including 3 months postRT	3	471	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.77, 1.22]
2 Overall survival at 42 to 72 months from baseline Show forest plot	3		Hazard Ratio (Random, 95% CI)	1.00 [0.72, 1.39]

3 Progression‐free survival at 42 to 72 months from baseline Show forest plot	3		Hazard Ratio (Random, 95% CI)	1.06 [0.80, 1.42]

Comparison 7. Palifermin versus placebo

Comparison 8. Bethanechol versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (0 to 3 scale ‐ grade 2 or above) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 End of radiotherapy	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.28, 0.66]
1.2 Up to and including 3 months postradiotherapy	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.65, 1.01]
2 Salivary flow rate (unstimulated) ‐ ml/min Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

2.1 2 months postradiotherapy	1	97	Mean Difference (IV, Random, 95% CI)	0.19 [0.06, 0.32]
3 Salivary flow rate (stimulated) ‐ ml/min Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

3.1 2 months postradiotherapy	1	97	Mean Difference (IV, Random, 95% CI)	0.15 [‐0.03, 0.33]

Comparison 8. Bethanechol versus placebo

Comparison 9. Bethanechol versus artificial saliva

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (dry mouth? yes/no) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 End of radiotherapy	1	36	Risk Ratio (M‐H, Random, 95% CI)	0.63 [0.30, 1.29]
1.2 8 to 40 weeks postradiotherapy	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.56 [0.30, 1.05]
2 Salivary flow rate (unstimulated) ‐ ml/min Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

2.1 End of radiotherapy	1	36	Mean Difference (IV, Random, 95% CI)	0.12 [0.01, 0.23]
2.2 8 to 40 weeks postradiotherapy	1	33	Mean Difference (IV, Random, 95% CI)	0.07 [‐0.02, 0.16]
3 Salivary flow rate (stimulated) ‐ ml/min Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Subtotals only

3.1 End of radiotherapy	1	32	Mean Difference (IV, Random, 95% CI)	0.13 [‐0.03, 0.29]
3.2 8 to 40 weeks postradiotherapy	1	29	Mean Difference (IV, Random, 95% CI)	0.21 [0.01, 0.41]
4 Overall survival Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

4.1 40 weeks postradiotherapy	1	43	Risk Ratio (M‐H, Random, 95% CI)	1.59 [0.43, 5.84]

Comparison 9. Bethanechol versus artificial saliva

Comparison 10. Selenium versus no selenium

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia Show forest plot			Other data	No numeric data

Comparison 10. Selenium versus no selenium

Comparison 11. Antimicrobial lozenge versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (QoL response for dryness) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 Up to and including 3 months postradiotherapy	1	133	Risk Ratio (M‐H, Random, 95% CI)	1.16 [0.97, 1.40]
2 Quality of life Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

2.1 Up to and including 3 months postradiotherapy (change score over 6 months)	1	131	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.65, 1.50]

Comparison 11. Antimicrobial lozenge versus placebo

Comparison 12. Polaprezinc versus azulene oral rinse

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Xerostomia (grade 2 or above) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

1.1 End of radiotherapy	1	31	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.04, 0.65]

Comparison 12. Polaprezinc versus azulene oral rinse