Interventions for preventing oral mucositis for patients with cancer receiving treatment

Helen V Worthington; Janet E Clarkson; Gemma Bryan; Susan Furness; Anne-Marie Glenny; Anne Littlewood; Martin G McCabe; Stefan Meyer; Tasneem Khalid; Philip Riley

doi:10.1002/14651858.CD000978.pub5

. 2011 Apr 13;2011(4):CD000978. doi: 10.1002/14651858.CD000978.pub5

Interventions for preventing oral mucositis for patients with cancer receiving treatment

Helen V Worthington ^1,^✉, Janet E Clarkson ¹, Gemma Bryan ², Susan Furness ³, Anne-Marie Glenny ⁴, Anne Littlewood ¹, Martin G McCabe ⁵, Stefan Meyer ⁶, Tasneem Khalid ⁷, Philip Riley ¹

Editor: Cochrane Oral Health Group

PMCID: PMC7032547 PMID: 21491378

Abstract

Background

Treatment of cancer is increasingly more effective but is associated with short and long term side effects. Oral side effects remain a major source of illness despite the use of a variety of agents to prevent them. One of these side effects is oral mucositis (mouth ulcers).

Objectives

To evaluate the effectiveness of prophylactic agents for oral mucositis in patients with cancer receiving treatment, compared with other potentially active interventions, placebo or no treatment.

Search methods

Electronic searches of Cochrane Oral Health Group and PaPaS Trials Registers (to 16 February 2011), CENTRAL (The Cochrane Library 2011, Issue 1), MEDLINE via OVID (1950 to 16 February 2011), EMBASE via OVID (1980 to 16 February 2011), CINAHL via EBSCO (1980 to 16 February 2011), CANCERLIT via PubMed (1950 to 16 February 2011), OpenSIGLE (1980 to 2005) and LILACS via the Virtual Health Library (1980 to 16 February 2011) were undertaken. Reference lists from relevant articles were searched and the authors of eligible trials were contacted to identify trials and obtain additional information.

Selection criteria

Randomised controlled trials of interventions to prevent oral mucositis in patients receiving treatment for cancer.

Data collection and analysis

Information regarding methods, participants, interventions, outcome measures, results and risk of bias were independently extracted, in duplicate, by two review authors. Authors were contacted for further details where these were unclear. The Cochrane Collaboration statistical guidelines were followed and risk ratios calculated using random‐effects models.

Main results

A total of 131 studies with 10,514 randomised participants are now included. Overall only 8% of these studies were assessed as being at low risk of bias. Ten interventions, where there was more than one trial in the meta‐analysis, showed some statistically significant evidence of a benefit (albeit sometimes weak) for either preventing or reducing the severity of mucositis, compared to either a placebo or no treatment. These ten interventions were: aloe vera, amifostine, cryotherapy, granulocyte‐colony stimulating factor (G‐CSF), intravenous glutamine, honey, keratinocyte growth factor, laser, polymixin/tobramycin/amphotericin (PTA) antibiotic pastille/paste and sucralfate.

Authors' conclusions

Ten interventions were found to have some benefit with regard to preventing or reducing the severity of mucositis associated with cancer treatment. The strength of the evidence was variable and implications for practice include consideration that benefits may be specific for certain cancer types and treatment. There is a need for further well designed, and conducted trials with sufficient numbers of participants to perform subgroup analyses by type of disease and chemotherapeutic agent.

Plain language summary

Interventions for preventing oral mucositis for patients with cancer receiving treatment

Treatment for cancer (including bone marrow transplant) can cause oral mucositis (severe ulcers in the mouth). This painful condition can cause difficulties in eating, drinking and swallowing, and may also be associated with infections which may require the patient to stay longer in hospital. Different strategies are used to try and prevent this condition, and the review of trials found that some of these are effective. Two interventions, cryotherapy (ice chips) and keratinocyte growth factor (palifermin®) showed some benefit in preventing mucositis. Sucralfate is effective in reducing the severity of mucositis, and a further seven interventions, aloe vera, amifostine, intravenous glutamine, granulocyte‐colony stimulating factor (G‐CSF), honey, laser and antibiotic lozenges containing polymixin/tobramycin/amphotericin (PTA) showed weaker evidence of benefit. These were evaluated in patients with different types of cancer, undergoing different types of cancer treatment. Benefits may be restricted to the disease and treatment combinations evaluated.

Summary of findings

Background

Description of the condition

Treatment for malignancies with cytotoxic chemotherapy or radiotherapy or both are becoming increasingly effective but are associated with short and long term side effects. Among the clinically important acute side effects is the disruption in the function and integrity of the mouth. The consequences of this include severe ulceration (mucositis) and fungal infection of the mouth (oral candidiasis, thrush). These disease and treatment induced complications may also produce oral discomfort or pain, poor nutrition, delays in cancer treatment, increased hospital stays and costs and, in some patients, life threatening infection (septicaemia).

Description of the intervention

Oral complications remain a major source of illness despite the use of a plethora of prophylactic agents, many of which are not evidence based. Individual cancer centres use different mouth care regimens to prevent mucositis (Glenny 2004), frequently with scant evidence of efficacy. Given the costs to patients and their clinicians of mucositis‐related morbidity it is surprising that a uniform approach to mucositis prevention, backed by a strong evidence base, is lacking. There are variations in usage between cancer centres in terms of the mouth care regimen used. Compliance with recommended use of product is variable and there are conflicting reports of the effectiveness of prophylactic agents. The qualitative and quantitative benefits, side effects and costs of oral therapies are of importance to the cancer teams responsible for the treatment of patients.

There have been several traditional reviews published and most of these present a general discussion for both chemotherapy and radiotherapy induced oral side effects (Andreassen 2003; Chang 2003; De Pauw 1997; Denning 1992; Duncan 2003; Lortholary 1997; Savarese 2003; Stevens 1995; Symonds 1998; Verdi 1993; White 1993). The conclusions drawn and recommendations made vary from advocating a particular therapy to recommending oral care procedures that have not been systematically investigated. Three systematic reviews have focused on the prevention of oral mucositis in patients with cancer (Kowanko 1998; Stokman 2006; Sunderland 2001). Kowanko and colleagues concluded that for most strategies reviewed there was insufficient evidence to draw any conclusions regarding their effectiveness (Kowanko 1998). Sunderland and colleagues focused exclusively on patients with head and neck cancer. Their main analysis combined all the interventions in one meta‐analysis and found a beneficial effect of prophylactic interventions (Sunderland 2001). The most recent published review (Stokman 2006) considered eight different interventions and found four that showed a statistically significant effect in preventing the development or severity of mucositis.

Since the previous update of this review the most significant development in this field has been the publication of a review of evidence and the development of clinical practice guidelines by the Mucositis Study Section of the Multinational Association of Supportive Care in Cancer and the International Society for Oral Oncology (Keefe 2007; Rubenstein 2004; Sonis 2004). The development of clinical practice guidelines highlights the fact that preventing and managing mucositis in cancer patients worldwide is imperative. However, given the large number of clinical trials identified in this systematic review, it is important that researchers work co‐operatively to maintain up to date systematic reviews, which can then be used to provide the evidence base for clinical guidelines, which can be adapted at the local level to take account of policy and resources.

A previous version of this Cochrane review looked at the use of prophylactic agents for the prevention of oral mucositis and oral candidiasis in patients with cancer treated by chemotherapy (Clarkson 2000). The review concluded that there was some evidence that using ice chips during the chemotherapy treatment was effective in preventing mucositis. The review was updated in 2003 (Clarkson 2003a) and this update broadened the oral mucositis part of the initial review and looked at the prevention of oral mucositis in patients receiving any treatment for cancer, including patients with all types of cancer, as well as head and neck cancer, and including comparisons between any interventions for prevention. A second review update was carried out in 2006 (Worthington 2006) and reviews on the prevention of oral candidiasis have also been published in The Cochrane Library (Clarkson 2007a; Worthington 2004a). The third prevention review update was carried out in 2007 (Worthington 2007).These reviews form part of a series of Cochrane reviews on the prevention and treatment of oral mucositis, oral candidiasis (Clarkson 2007b; Worthington 2007a); xerostomia (Tavender 2004) and herpes simplex in patients receiving cancer therapy (Glenny 2009).

How the intervention might work

Mucositis presents a particular challenge due to its complex biological nature and the range of interventions tested have included mucosal surface protectants, anti‐inflammatory formulations, antimicrobials, growth factors and a plethora of other miscellaneous agents.

Why it is important to do this review

The Cochrane Oral Health Group undertook an extensive prioritisation exercise in 2014 to identify a core portfolio of titles that were the most clinically important ones to maintain on the Cochrane Library (Worthington 2015). This review was identified as a priority title by the oral medicine expert panel (Cochrane OHG priority review portfolio).

Objectives

To evaluate the effectiveness of interventions (which may include placebo or no treatment) for the prevention of oral mucositis in patients with cancer receiving radiotherapy, chemotherapy or targeted therapies.

We also investigated the following secondary outcomes for benefits or harms provided there were three or more trials or one trial with more than 100 participants, otherwise we recorded outcomes reported:

Oral hygiene measures
Relief of pain/use of analgesia
Duration or severity of dysphagia
Use of parenteral nutrition or feeding tube
Incidence of systemic infection or use of antibiotics
Febrile episodes
Blood changes
Treatment interruption
Days of stay in hospital
Toxicity (nausea/vomiting/constipation/diarrhoea)
Toxicity ‐ skin changes
Toxicity ‐ unspecific
Xerostomia
Cost of care
Patient quality of life
Death
Weight loss/gain
Caloric intake by oral nutrition
Eating/drinking difficulty
Overall health
Recurrence of cancer.

The following subgroup analyses were proposed:

Cancer type (leukaemia, head and neck, other solid tumours and mixed)
Cancer treatment (specific, for example 5‐fluorouracil (5‐FU))
Age group (adults, children or both).

Methods

Criteria for considering studies for this review

Types of studies

Only randomised controlled trials (RCTs) were eligible for inclusion in this review, including both cross‐over and parallel group studies.

Types of participants

Anyone with cancer who received radiotherapy, chemotherapy or targeted therapies.

Types of interventions

(This review did not include studies of different cancer treatments when the primary outcome was survival or cure with mucositis as a toxicity.) Active agents: any agent prescribed as prophylaxis for oral mucositis. Control: may be placebo or no treatment, or another active intervention.

Types of outcome measures

Primary outcome

Mucositis (at all levels of severity) (ulcers). Mucositis measured on a 0 to 4 point scale (none to severe) was used and this was dichotomised as any mucositis (0 versus 1+), moderate plus severe mucositis (0 to 1 versus 2+), severe mucositis (0 to 2 versus 3+).
Trials where mucositis is measured using a scale which individually grades multiple components of oral health such as teeth or plaque (Eilers 1988), are excluded. It is acknowledged that these oral effects are important, but it is not possible to separate the 'mucositis only' score, from the total score, so this review will not include studies which evaluate mucositis using such component scores.

Secondary outcomes

Oral hygiene measures
Relief of pain/use of analgesia
Duration or severity of dysphagia
Use of parenteral nutrition or feeding tube
Incidence of systemic infection or use of antibiotics
Febrile episodes
Blood changes
Treatment interruption
Days of stay in hospital
Toxicity (nausea/vomiting/constipation/diarrhoea)
Toxicity ‐ skin changes
Toxicity ‐ unspecific
Xerostomia
Cost of care
Patient quality of life
Death
Weight loss/gain
Caloric intake by oral nutrition
Eating/drinking difficulty
Overall health
Recurrence of cancer.

Search methods for identification of studies

This review is part of a series of four reviews on the prevention and treatment of oral candidiasis and oral mucositis in patients with cancer, and the same search strategies were used for all four reviews.

The searches attempted to identify all relevant trials irrespective of language. Papers not in English were translated by members of The Cochrane Collaboration. Sensitive search strategies were developed for each database using a combination of free text and MeSH terms. The MEDLINE and CANCERLIT searches combined the subject search with the Cochrane Highly Sensitive Search Strategy (CHSSS) for identifying randomised trials in MEDLINE: sensitivity maximising version (2009 revision) as referenced in Chapter 6.4.11.1 and detailed in boxes 6.4.a and 6.4.c of the Cochrane Handbook for Systematic Reviews of Interventions version 5.0.2 (updated September 2009) (Higgins 2009). The EMBASE and CINAHL searches were combined with sensitive search strategies developed by the Cochrane Oral Health Group for identifying randomised controlled trials (RCTs). The LILACs subject search was combined with the Brazilian Cochrane Centre search strategy for identifying RCTs in LILACS.

Electronic searches

The databases searched were:

Cochrane Oral Health Group Trials Register (to 16 February 2011) (Appendix 1)
Cochrane Pain, Palliative and Supportive Care (PaPaS) Group Trials Register (to 16 February 2011) (Appendix 1)
Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2011, Issue 1) (Appendix 2)
MEDLINE via OVID (1950 to 16 February 2011) (Appendix 3)
EMBASE via OVID (1980 to 16 February 2011) (Appendix 4)
CANCERLIT via PubMed (1950 to 16 February 2011) (Appendix 5)
OpenSIGLE (1980 to 2005) (Appendix 6)
LILACS via The Virtual Health Library (to 16 February 2011) (Appendix 7)
CINAHL via EBSCO (1980 to 16 February 2011) (Appendix 8).

Only handsearching carried out by The Cochrane Collaboration was included in the search (see master list www.cochrane.org).

The controlled trials database (www.controlled-trials.com) was also searched to identify ongoing and completed trials and to contact trialists for further information about these trials.

The review will be updated every 2 years using the Cochrane Oral Health Group Trials Register, CENTRAL, MEDLINE, EMBASE, CINAHL, CANCERLIT and LILACS. The search of OpenSIGLE was discontinued as this database ceased being updated in 2005.

Searching other resources

The reference list of related review articles and all articles obtained were checked for further trials. Authors of trial reports and specialists in the field known to the review authors were written to concerning further published and unpublished trials.

Data collection and analysis

Selection of studies

The titles and abstracts (when available) of all reports identified through the electronic searches were scanned independently by two review authors. For studies appearing to meet the inclusion criteria, or for which there were insufficient data in the title and abstract to make a clear decision, the full report was obtained. The full reports, containing names of the authors, institutions, journal of publication and results, obtained from all the electronic and other methods of searching were assessed independently by two authors with expertise in this content area, to establish whether the studies met the inclusion criteria or not. Disagreements were resolved by discussion. Where resolution was not possible, a third review author was consulted. All studies meeting the inclusion criteria then underwent validity assessment and data extraction. Studies rejected at this or subsequent stages were recorded in the 'Characteristics of excluded studies' table, and reasons for exclusion recorded.

Data extraction and management

Data were extracted by two review authors independently using specially designed data extraction forms. The characteristics of the trial participants, interventions and outcomes in the included trials are presented in the study tables. Mucositis may be dichotomised at different levels of severity. In order to maximise the availability of similar outcome data we recorded the number of patients in each category of mucositis. We planned to form three dichotomies of mucositis: absent versus present (0 versus 1+), mild versus moderate/severe (0 to 1 versus 2+) and moderate versus severe (0 to 2 versus 3+). Pain was assessed on visual analogue scales (0 to 100), the means and standard deviations for each group were recorded. The duration of trials and timing of assessments were recorded in order to make a decision about which to include for commonality. We also recorded the country where the trial was conducted and whether a dentist was involved in the investigation. Some of the authors were contacted for clarification or for further information.

Assessment of risk of bias in included studies

For any relevant studies identified, two review authors independently graded the relevant trials following the domain‐based evaluation described in the Cochrane Handbook for Systematic Reviews of Interventions 5.0.2 (updated September 2009) (http://archie.cochrane.org/sections/documents/view?document=8D67D37282E26AA201AA9B4D54ED9EBC&format=REVMAN#REF‐Higgins‐2009#REF‐Higgins‐2009). The review authors then compared evaluations, discussed and resolved any disagreements and reported their assessments in a risk of bias in included studies table in Review Manager (RevMan) software.

An assessment of the overall risk of bias involved the consideration of the relative importance of different domains and studies were to be categorised as low, high or unclear risk of bias.

The review authors were to assess the following domains as 'Yes' (i.e. low risk of bias), 'Unclear' (i.e. uncertain risk of bias) or 'No' (i.e. high risk of bias):

adequate sequence generation;
allocation concealment;
blinding (of participants, carers and outcome assessors);
incomplete outcome data addressed;
free of selective outcome reporting;
free of other bias.

These risk of bias assessments were then for each study across all domains. Adequate allocation concealment and blinding of outcome assessors were designated as key domains for this assessment.

Overall risk of bias was categorised according to the following:

Low risk of bias (plausible bias unlikely to seriously alter the results) for all key domains;
Unclear risk of bias (plausible bias that raises some doubt about the results) if one or more key domains were assessed as unclear; or
High risk of bias (plausible bias that seriously weakens confidence in the results) if one or more key domains were assessed to be at high risk of bias.

Measures of treatment effect

For dichotomous outcomes, the estimate of effect of an intervention was expressed as risk ratios (RR) together with 95% confidence intervals (CIs). For continuous outcomes, mean differences and standard deviations were used to summarise the data for each group using mean differences and 95% CIs. Appropriate data were extracted from the cross‐over studies and the generic inverse variance method was used to enter this into RevMan.

Unit of analysis issues

The patient was the unit of analysis in all trials.

Dealing with missing data

All authors were contacted to retrieve missing data from authors of trials.

The analysis will generally include only the available data (ignoring missing data) however methods for estimating missing standard deviations in Section 7.7.3 of the Cochrane Handbook for Systematic Reviews of Interventions 5.0.2 (Higgins 2009) were to be used. Otherwise we do not intend to undertake any imputations nor to use statistical methods to allow for missing data.

Assessment of heterogeneity

The significance of any discrepancies in the estimates of the treatment effects from the different trials was to be assessed by means of Cochran's test for heterogeneity and heterogeneity would have been considered significant if P < 0.1 (Higgins 2009).

The I² statistic, which describes the percentage total variation across studies that is due to heterogeneity rather than chance, was used to quantify heterogeneity with I² over 50% being considered substantial heterogeneity (Higgins 2009 Section 9.5.2). In order to assist in the readers in the interpretation of heterogeneity we would still have included the pooled meta‐analysis in the forest plot because the I² and Chi² statistics are helpful. With substantial heterogeneity pooling the data may not be appropriate and this will be considered in the results for each intervention.

Assessment of reporting biases

If there had been sufficient numbers of trials (more than 10) in any meta‐analysis, publication bias would have been assessed according to the recommendations on testing for funnel plot asymmetry (Egger 1997) as described in the Cochrane Handbook for Systematic Reviews of Interventions 5.0.2 (Higgins 2009). If asymmetry were identified we would have examined possible causes.

Data synthesis

A meta‐analysis would have only be conducted if there were studies of similar comparisons reporting the same mucositis outcome measures. Single studies would not be entered into forest plots. Risk ratios were to be combined for dichotomous data, and mean differences for continuous data, using random‐effect models provided there were more than three studies in the meta‐analysis.

It is possible to conduct cross‐over trials in this area as patients may be receiving several chemotherapy sessions, any mucositis completely healing in the periods between the sessions. The treatment effects from cross‐over trials were combined with those from parallel group trials where appropriate, using the data from both periods of the cross‐over studies (Elbourne 2002). The generic inverse variance method incorporated in RevMan was used for all analyses. Where data for the cross‐tabulation of pairs were not available, all possible paired comparisons for each study were calculated, giving rise to the same risk ratios values with different confidence intervals. The widest confidence interval was used in the analysis.

Subgroup analysis and investigation of heterogeneity

Clinical heterogeneity was to be assessed by examining the types of participants and interventions for all outcomes in each study. We proposed a priori to conduct subgroup analyses for different cancer types (leukaemia, head and neck, other solid tumours and mixed), cancer treatment (for example 5‐fluorouracil (5‐FU)) and age groups (children, adults and both). However, there were insufficient trials by intervention type to do this.

Sensitivity analysis

It was planned to undertake sensitivity analyses to examine the effect of the study quality assessment on the overall estimates of effect. In addition, the effect of including unpublished literature on the review's findings was also to be examined. There were too few trials to undertake these analyses.

Summary of findings and assessment of the certainty of the evidence

We described the results for the main outcome in 'Summary of findings' tables and assessed the certainty of the evidence for each comparison and the main outcome. We used GRADE criteria to judge the evidence as having high, moderate, low or very low certainty.

Results

Description of studies

Results of the search

Three hundred and eighty‐three reports of trials were initially identified as eligible according to the defined inclusion criteria for this review with regard to study design, participants, interventions and outcomes. From this list, the total number of included trials was 131, there were 10 duplicate reports and 176 studies were excluded. Due to changes in the inclusion criteria for studies in this review, some previously excluded studies are being reconsidered for inclusion in the next update. Information about these studies is in Characteristics of studies awaiting classification.

Included studies

Characteristics of the trial settings and investigators

Of the 131 included trials 124 were designed as parallel group studies and seven as cross‐over studies (Anderson 1998; Chi 1995; Dozono 1989; Jebb 1994; Loprinzi 1990; Mahood 1991; Pfeiffer 1990). None of the published reports of cross‐over studies reported the 'paired' data in an appropriate form to be used in a meta‐analysis. All the authors were contacted and replies were received supplying data for three studies (Anderson 1998; Loprinzi 1990; Mahood 1991). Data from the other cross‐over studies were extracted as outlined in the methods section.

Of the 131 included trials all included data on assessment of mucositis. Ninety‐two (70%) of the 131 included trials were conducted at a single site. Thirty‐two of these trials were conducted in Europe, 20 in the USA, seven in China, five in India, four in Canada, three in Brazil and Japan, two in Iran, Israel, Thailand, Taiwan and Turkey, and one in each of the following countries: Argentina, Egypt, Hong Kong, Malaysia, Mexico, South Africa, South Korea, and Uruguay. Thirty‐six trials were multicentre studies. Ten of these were conducted in the USA (Blazar 2006; Crawford 1999; Dodd 1996; Freytes 2004; Giles 2004; Hanson 1995; Mahood 1991; Meropol 2003; Scarantino 2006; Spielberger 2004), three in the USA and Canada (Epstein 2001; Foote 1994; Nemunaitis 1995), two in the USA and European countries (Brizel 2000; Buentzel 2006), two in European countries alone (Castagna 2001; Vokurka 2005), one in the USA, Canada and Europe (Trotti 2004), one in the USA and Australia (Rosen 2006) and one in the USA, Australia and Canada (Brizel 2008). Two studies were conducted in Australia (Spencer 2005; Veness 2006), Canada (Bjarnason 2009; El Sayed 2002), France (Bensadoun 1999 Gandemer 2007), Iran (Abbasi‐Nazari 2007; Motallebnejad 2008) and Taiwan (Chi 1995; Lin 2006), and one in each of the following countries: China (Tu 1998), Germany (Dorr 2007), Italy (Gori 2007), South Korea (Wu 2009), Russia (Peterson 2009) and Thailand (Veerasarn 2006). It was unclear whether the remaining three studies were conducted at a single site or were multisite studies (Bubley 1989; Lievens 1998; Qin 2007).

Eighty‐one studies (62%) provided funding information. Thirty‐seven (28%) trials were funded by the pharmaceutical industry. Twenty‐three studies received government funding. Whilst an additional nine trials were funded by a combination of funding sources (Anderson 1998; Blazar 2006; Cerchietti 2006; Dodd 1996; Ferretti 1988; Foote 1994; Makkonen 1994; Sornsuvit 2008; van der Lelie 2001). Five studies were funded by charities (Bjarnason 2009; Dickson 2000; Franzen 1995; Lilleby 2006; Lockhart 2005). Four trials reported that they received university funding (Biswal 2003; Madan 2008; Motallebnejad 2008; Yuen 2001). Three studies were funded by other sources (Gori 2007; Lin 2006; Oberbaum 2001). The remaining 50 (38%) studies reported either none, or insufficient information about funding sources. A dentist was involved in 26 (20%) of the trials.

Characteristics of the participants

One hundred and fourteen (87%) of the included trials recruited only adult patients, 13 included both adults and children (with a difference in age as large as 1 to 70 years) and three trials were conducted solely on paediatric patients (Cruz 2007; Gandemer 2007; Shenep 1988), the age group being unclear in one trial (Mahood 1991). The type of cancer for which patients were being treated was exclusively head and neck cancer in 69 trials (53%), leukaemia in 13 trials, solid tumours in 22 trials and a combination of haematological and solid tumours in 21 trials, the cancer type being unclear in six trials. The radiotherapy and/or chemotherapy regimen was described in most of the trials though the chemotherapeutic agents were not always described in full detail. Twenty‐nine trials included patients who were undergoing a bone marrow transplant. The chemotherapy regimen included 5‐fluorouracil (5‐FU) in 25 trials. In 13 of these trials the patients had solid tumours, in six trials patients had head and neck cancer, in two trials patients had mixed cancers and in four trials the cancer type was unclear. It was not always clear if the dose was in a bolus or continuous form. Trials in which patients received radiotherapy generally gave information about the total and daily or weekly dose. Total radiotherapy for head and neck cancer was generally 60 to 74 Gy and the Karnofsky performance > 60.

Characteristics of the interventions

All of the 131 trials provided a clear description of the interventions including the dose and method of administration for the test and control groups. The dosage of the test agents varied for similar products. Thirty‐six trials compared an active intervention with no treatment and 87 trials used a placebo control. In some trials the placebo was matched in taste and appearance to the active intervention and in others the following interventions were described as placebo: water, albumin, glycine, sugar solution, polycal, saline. Two trials included in the no treatment control group tested different oral care protocols and in each case one group received limited oral hygiene (usual care) (Borowski 1994; Shieh 1997). A further three trials included in the placebo control group included both a placebo control and a direct comparison (Freytes 2004 ‐ two different doses of keratinocyte growth factor, Madan 2008 ‐ chlorhexidine versus povidone iodine versus salt & soda, and Sorensen 2008 chlorhexidine mouthwash versus ice chips).

A further two trials (Huang 2003; Wang 2002) compared Chinese medicine to a control described as Dobell's solution, which is "a solution of sodium borate, sodium bicarbonate, phenol, and glycerol, used as a wash for mucous membranes" (Merck Index 2010). Two trials compared patients receiving radiotherapy in the morning with patients receiving radiotherapy in the afternoon (Bjarnason 2009; Goyal 2009). Another four trials compared two active interventions: ice chips sucked for different time periods (30 minutes versus 60 minutes) (Rocke 1993), granulocyte/macrophage colony‐stimulating factor (GM‐CSF) versus sucralfate (Saarilahti 2002), chlorhexidine versus laser (Arun Maiya 2006) and polaprezinc versus azulene (Watanabe 2010).

The interventions for the 131 studies assessing oral mucositis were:

aciclovir (Bubley 1989)
allopurinol mouthrinse (Abbasi‐Nazari 2007; Dozono 1989; Loprinzi 1990; Panahi 2009)
aloe vera (Puataweepong 2009; Su 2004;)
amifostine (Antonadou 2002; Bourhis 2000; Brizel 2000; Buentzel 2006; Buntzel 1998; Haddad 2009; Hartmann 2001; Koukourakis 2000; Spencer 2005; Vacha 2003; Veerasarn 2006)
antibiotic pastille or paste (El Sayed 2002)
antibiotic systemic (clarithromycin) (Yuen 2001)
azulene (Watanabe 2010*)
benzydamine (Epstein 1989; Epstein 2001; Kazemian 2009; Prada 1987)
beta carotene (Mills 1988)
chamomile (Fidler 1996)
chewing gum (Gandemer 2007)
Chinese herbs (details of herbs used are given in Characteristics of included studies table) (Huang 2003*; Wang 2002*)
chlorhexidine (Arun Maiya 2006*; Dodd 1996; Ferretti 1988; Foote 1994; Madan 2008*; McGaw 1985; Pitten 2003; Sorensen 2008; Spijkervet 1989; Wahlin 1989)
cryotherapy (Cascinu 1994; Gori 2007; Lilleby 2006; Mahood 1991; Rocke 1993*; Svanberg 2007; Sorensen 2008*)
dental stent (Qin 2007)
epidermal growth factor (Wu 2009)
glutamine (Anderson 1998; Cerchietti 2006; Choi 2007; Dickson 2000; He 2008; Huang 2000; Jebb 1994; Li 2006; Okuno 1999; Sornsuvit 2008)
granulocyte colony‐stimulating factor (G‐CSF) (Crawford 1999; Katano 1995; Schneider 1999; Su 2006)
granulocyte/macrophage colony‐stimulating factor (GM‐CSF) (Cartee 1995; Chi 1995; Dazzi 2003; Ifrah 1999; Makkonen 2000; McAleese 2006; Nemunaitis 1995; Saarilahti 2002*; van der Lelie 2001)
histamine gel (Elad 2006)
honey (Biswal 2003; Motallebnejad 2008; Rashad 2008)
hydrolytic enzymes (details of enzymes used are given in Characteristics of included studies table) (Dorr 2007; Gujral 2001; Kaul 1999)
indigo wood root (You 2009)
intestinal trefoil factor (Peterson 2009)
iseganan (Giles 2004; Trotti 2004)
keratinocyte growth factor (GF) (Blazar 2006; Brizel 2008; Freytes 2004*; Meropol 2003; Rosen 2006; Spielberger 2004; Vadhan‐Raj 2010)
laser (Antunes 2007; Arun Maiya 2006*; Bensadoun 1999; Chor 2010; Cruz 2007; Schubert 2007)
non‐steroidal anti‐inflammatory drug (Pillsbury 1986)
oral care (Borowski 1994; Shieh 1997)
pentoxifylline (Attal 1993)
pilocarpine (Lockhart 2005; Scarantino 2006)
PTA (polymixin/tobramycin/amphotericin) lozenges/paste (Stokman 2003; Symonds 1996; Wijers 2001)
polaprezinc (Watanabe 2010*)
povidone iodine (Arun Maiya 2006; Madan 2008*; Rahn 1997; Vokurka 2005)
prednisone (Leborgne 1997)
propantheline anticholinergic (Ahmed 1993)
prostaglandin (Duenas 1996; Hanson 1995; Labar 1993; Pillsbury 1986; Veness 2006)
radiation: morning versus evening (Bjarnason 2009*, Goyal 2009*)
shenqi‐fanghou (Hu 2005)
superoxide dismutase (SOD) (Tu 1998)
sucralfate (Carter 1999; Castagna 2001; Cengiz 1999; Epstein 1994; Evensen 2001; Franzen 1995; Lievens 1998; Makkonen 1994; Nottage 2003; Pfeiffer 1990; Saarilahti 2002*; Scherlacher 1990; Shenep 1988)
traumeel (Oberbaum 2001)
yangygin‐humo decoctalion (Dai 2009)
zinc sulphate (Ertekin 2004; Lin 2006).

* studies with two or more different active treatments.

Characteristics of outcome measures

Mucositis

All trials used a graded scale to record the severity of mucositis. Most described the index used or referred to published criteria, mainly World Health Organization (WHO) or European Organization for Research and Treatment of Cancer (EORTC). Scales were similar to the 5‐point WHO scale ranging from 0 (normal) to 4 (severe). The categories initially relate to visible changes in the mucosa and gradually record pain and inability to eat solid foods. The duration of the trials varied from a few days up to a year after treatment. The interval during which mucositis was recorded varied from 5 to 90 days or until the end of the radiotherapy, or the leukocyte count was above 8000 mm³. Several studies presented data at different time points, with the median time point being 28 days. The nearest assessment to 28 days was used for all studies.

Secondary outcomes

There was little consistency on the other outcome measures reported.

Oral hygiene measures (Biswal 2003; Cruz 2007; Dodd 1996; El Sayed 2002; Elad 2006; Ertekin 2004; Evensen 2001; Ferretti 1988; Foote 1994; Gandemer 2007; Kazemian 2009; Leborgne 1997; Lockhart 2005; Makkonen 1994; Makkonen 2000; McGaw 1985; Rahn 1997; Rashad 2008; Scherlacher 1990; Spijkervet 1989; Stokman 2003; Symonds 1996; Vadhan‐Raj 2010; Vokurka 2005; Wahlin 1989; Wijers 2001; Wu 2009).
Relief of pain/use of analgesia (morphine) (Antunes 2007; Arun Maiya 2006; Attal 1993; Bensadoun 1999; Blazar 2006; Brizel 2008; Carter 1999; Castagna 2001; Cengiz 1999; Cerchietti 2006; Cruz 2007; Dazzi 2003; Dorr 2007; El Sayed 2002; Epstein 1989; Epstein 1994; Epstein 2001; Ertekin 2004; Ferretti 1988; Franzen 1995; Freytes 2004; Gandemer 2007; Giles 2004; Hanson 1995; Lilleby 2006; Lockhart 2005; Makkonen 2000; Meropol 2003; Nottage 2003; Oberbaum 2001; Peterson 2009; Pfeiffer 1990; Prada 1987; Puataweepong 2009; Rosen 2006; Saarilahti 2002; Schubert 2007; Shenep 1988; Shieh 1997; Spencer 2005; Spielberger 2004; Su 2004; Svanberg 2007; Trotti 2004; Tu 1998; Vadhan‐Raj 2010; van der Lelie 2001; Veness 2006; Vokurka 2005; Watanabe 2010; Wijers 2001; Wu 2009).
Duration or severity of dysphagia (Antonadou 2002; Bensadoun 1999; Bjarnason 2009; Bourhis 2000; Brizel 2008; Buntzel 1998; Castagna 2001; Cengiz 1999; Choi 2007; Dorr 2007; El Sayed 2002; Elad 2006; Epstein 1994; Fidler 1996; Franzen 1995; Giles 2004; Goyal 2009; Gujral 2001; Haddad 2009; Hartmann 2001; Kaul 1999; Lievens 1998; Lilleby 2006; Lockhart 2005; McAleese 2006; Oberbaum 2001; Prada 1987; Rosen 2006; Scarantino 2006; Scherlacher 1990; Spielberger 2004; Su 2006; Symonds 1996; Trotti 2004; Vadhan‐Raj 2010; Veerasarn 2006; You 2009).
Use of parenteral nutrition or feeding tube (Bjarnason 2009; Bourhis 2000; Brizel 2008; Carter 1999; Cerchietti 2006; Cruz 2007; Dickson 2000; El Sayed 2002; Evensen 2001; Foote 1994; Franzen 1995; Gandemer 2007; Haddad 2009; Hanson 1995; Hartmann 2001; Leborgne 1997; Lilleby 2006; Lockhart 2005; Rashad 2008; Saarilahti 2002; Shenep 1988; Spencer 2005; Spielberger 2004; Stokman 2003; Su 2006; Symonds 1996; Trotti 2004; Vadhan‐Raj 2010; van der Lelie 2001; Yuen 2001).
Incidence of systemic infection or use of antibiotics (Antunes 2007; Attal 1993; Blazar 2006; Borowski 1994; Brizel 2008; Bubley 1989; Buntzel 1998; Castagna 2001; Cerchietti 2006; Crawford 1999; Cruz 2007; Duenas 1996; El Sayed 2002; Ertekin 2004; Ferretti 1988; Freytes 2004; Gandemer 2007; Hanson 1995; Hartmann 2001; Ifrah 1999; Jebb 1994; Labar 1993; McGaw 1985; Pitten 2003; Puataweepong 2009; Shenep 1988; Sornsuvit 2008; Spencer 2005; Spielberger 2004; Su 2004; Symonds 1996; Trotti 2004; Tu 1998; Vadhan‐Raj 2010; van der Lelie 2001; Vokurka 2005; Wahlin 1989; Yuen 2001).
Febrile episodes (Ahmed 1993; Anderson 1998; Attal 1993; Borowski 1994; Brizel 2000; Chi 1995; Chor 2010; Crawford 1999; Duenas 1996; Ferretti 1988; Freytes 2004; Ifrah 1999; Katano 1995; Labar 1993; McGaw 1985; Nemunaitis 1995; Pitten 2003; Shenep 1988; Spencer 2005; Spielberger 2004; van der Lelie 2001; Vokurka 2005; Wahlin 1989; Yuen 2001).
Blood changes (Ahmed 1993; Antonadou 2002; Attal 1993; Blazar 2006; Brizel 2000; Buentzel 2006; Buntzel 1998; Cartee 1995; Cascinu 1994; Cerchietti 2006; Chi 1995; Crawford 1999; Dazzi 2003; Dickson 2000; Dorr 2007; Duenas 1996; Elad 2006; Epstein 2001; Ertekin 2004; Ferretti 1988; Franzen 1995; Freytes 2004; Hartmann 2001; Huang 2003; Ifrah 1999; Katano 1995; Labar 1993; Li 2006; Lilleby 2006; Lin 2006; Mahood 1991; Makkonen 2000; Meropol 2003; Nemunaitis 1995; Pitten 2003; Rocke 1993; Rosen 2006; Saarilahti 2002; Schneider 1999; Shenep 1988; Sorensen 2008; Sornsuvit 2008; Spencer 2005; Spielberger 2004; Svanberg 2007; van der Lelie 2001; Veerasarn 2006; Vokurka 2005; Wahlin 1989; Wu 2009; You 2009).
Treatment interruption (Antonadou 2002; Biswal 2003; Bjarnason 2009; Bourhis 2000; Brizel 2000; Brizel 2008; Carter 1999; Dazzi 2003; El Sayed 2002; Foote 1994; Franzen 1995; Haddad 2009; Huang 2003; Ifrah 1999; Koukourakis 2000; Leborgne 1997; Makkonen 1994; Makkonen 2000; Pfeiffer 1990; Puataweepong 2009; Saarilahti 2002; Trotti 2004; van der Lelie 2001; Vadhan‐Raj 2010; Veness 2006; Wu 2009; You 2009).
Days of stay in hospital (Antonadou 2002; Attal 1993; Cerchietti 2006; Chor 2010; Dickson 2000; Duenas 1996; Ifrah 1999; Lilleby 2006; McGaw 1985; Saarilahti 2002; Sornsuvit 2008; van der Lelie 2001).
Toxicity ‐ nausea/vomiting/constipation/diarrhoea (Antonadou 2002; Blazar 2006; Bourhis 2000; Brizel 2000; Brizel 2008; Bubley 1989; Buentzel 2006; Cascinu 1994; Castagna 2001; Cengiz 1999; Dickson 2000; Duenas 1996; El Sayed 2002; Elad 2006; Epstein 1994; Epstein 2001; Ertekin 2004; Fidler 1996; Freytes 2004; Gandemer 2007; Giles 2004; Gujral 2001; Haddad 2009; Hartmann 2001; He 2008; Kazemian 2009; Labar 1993; Li 2006; Lievens 1998; Lockhart 2005; Mahood 1991; Meropol 2003; Nottage 2003; Oberbaum 2001; Okuno 1999; Peterson 2009; Pfeiffer 1990; Rocke 1993; Rosen 2006; Scarantino 2006; Shenep 1988; Sornsuvit 2008; Spencer 2005; Trotti 2004; Tu 1998; Vadhan‐Raj 2010; Yuen 2001).
Toxicity ‐ skin changes (Antonadou 2002; Blazar 2006; Bourhis 2000; Buntzel 1998; Choi 2007; Dickson 2000; Dorr 2007; El Sayed 2002; Evensen 2001; Giles 2004; Goyal 2009; Gujral 2001; Haddad 2009; Kaul 1999; Lievens 1998; Lin 2006; Meropol 2003; Rosen 2006; Scarantino 2006; Shenep 1988; Spielberger 2004; Tu 1998; Vacha 2003; Yuen 2001).
Toxicity ‐ unspecific (Buentzel 2006; Cerchietti 2006; Chi 1995; Duenas 1996; El Sayed 2002; Epstein 1994; Fidler 1996; Freytes 2004; Giles 2004; Gujral 2001; Mahood 1991; Makkonen 1994; Makkonen 2000; Okuno 1999; Puataweepong 2009; Spielberger 2004; Su 2006; Tu 1998).
Xerostomia (Antonadou 2002; Bourhis 2000; Brizel 2000; Brizel 2008; Buentzel 2006; Buntzel 1998; Castagna 2001; Cengiz 1999; Dazzi 2003; Elad 2006; Epstein 1994; Epstein 2001; Goyal 2009; Hartmann 2001; Koukourakis 2000; Lockhart 2005; Makkonen 2000; McAleese 2006; Meropol 2003; Nottage 2003; Oberbaum 2001; Saarilahti 2002; Scarantino 2006; Spencer 2005; Vacha 2003; Veerasarn 2006; Watanabe 2010).
Cost (Bourhis 2000; Brizel 2000; Buentzel 2006; Buntzel 1998; Choi 2007; Dodd 1996; Haddad 2009; Hartmann 2001; Hu 2005; Huang 2000; Koukourakis 2000; Makkonen 2000; McAleese 2006; Nemunaitis 1995; Nottage 2003; Sornsuvit 2008; Su 2006; Yuen 2001).
Patient quality of life (Bjarnason 2009; Brizel 2000; McAleese 2006; Nottage 2003; Scarantino 2006; Spielberger 2004; Veness 2006).
Death (Ahmed 1993; Attal 1993; Bjarnason 2009; Blazar 2006; Brizel 2008; Dickson 2000; Dodd 1996; Epstein 2001; Ertekin 2004; Ferretti 1988; Giles 2004; Gujral 2001; Kazemian 2009; Leborgne 1997; Madan 2008; Mills 1988; Oberbaum 2001; Rahn 1997; Rosen 2006; Schubert 2007; Spielberger 2004).
Weight loss/gain (Antonadou 2002; Biswal 2003; Bjarnason 2009; Brizel 2000; Buentzel 2006; Buntzel 1998; Carter 1999; Cengiz 1999; Cerchietti 2006; El Sayed 2002; Elad 2006; Ertekin 2004; Foote 1994; Freytes 2004; Haddad 2009; Hanson 1995; He 2008; Hu 2005; Huang 2000; Koukourakis 2000; Leborgne 1997; Lievens 1998; Lilleby 2006; Lin 2006; Makkonen 2000; Motallebnejad 2008; Nottage 2003; Pillsbury 1986; Puataweepong 2009; Qin 2007; Shenep 1988; Sornsuvit 2008; Stokman 2003; Su 2004; Su 2006; Symonds 1996; Trotti 2004; Vacha 2003; Veerasarn 2006; Veness 2006; Wu 2009; You 2009).
Caloric intake by oral nutrition (Castagna 2001; Cruz 2007; Dickson 2000; Freytes 2004; Hartmann 2001; He 2008; Lilleby 2006; Pfeiffer 1990; Shenep 1988; Spencer 2005; Watanabe 2010).
Eating/drinking difficulty (Anderson 1998; Carter 1999; Cengiz 1999; Cerchietti 2006; Dickson 2000; El Sayed 2002; Evensen 2001; Franzen 1995; Freytes 2004; Jebb 1994; Lilleby 2006; Lockhart 2005; Nottage 2003; Oberbaum 2001; Pfeiffer 1990; Prada 1987; Rashad 2008; Rosen 2006; Scarantino 2006; Shenep 1988; Sornsuvit 2008; Spielberger 2004; Stokman 2003; Symonds 1996; Vadhan‐Raj 2010).
Overall health (Antonadou 2002; Bourhis 2000; Brizel 2000; Buentzel 2006; Elad 2006; Ertekin 2004; Haddad 2009; Ifrah 1999; Jebb 1994; Lilleby 2006; Makkonen 2000; McAleese 2006; Nemunaitis 1995; Rosen 2006; Shenep 1988; Su 2004; Su 2006; Wu 2009).
Recurrence of cancer (Ahmed 1993; Attal 1993; Bjarnason 2009; Blazar 2006; Brizel 2000; Brizel 2008; Cerchietti 2006; Chi 1995; Dickson 2000; Duenas 1996; Goyal 2009; Gujral 2001; Leborgne 1997; Li 2006; Makkonen 2000; Mills 1988; Okuno 1999; Pillsbury 1986; Rosen 2006; Saarilahti 2002; Schneider 1999; Spielberger 2004; Vadhan‐Raj 2010; Watanabe 2010).

Excluded studies

See Characteristics of excluded studies table for further information on each excluded study. In summary, studies were excluded for the following reasons.

Abstracts only available, insufficient information to include study in the review (36 studies: Antonadou 1998; Buentzel 1999; Castro 2009; Clarke 2001; Collova 2004; Colombat 1995; Costa 1999; Gabison 1995; Goldberg 2003; Gordon 1993; Harris 1995; He 2004; Kante 1995; Klocke 2006; Lavendag 1998; Le 2008; Leong 1995; Lozada 1998; Marcial 1994; Merte 1999; Papas 1984; Pouli 1999; Radmard 2002; Robustelli 1999; Schwerkoske 1999; Sharma 2009; Shea 2007; Shidfar 2008; Spadaro 1991; Spielberger 2001; Throuvalas 1995; Valcárcel 1997; Vesole 1999; Villar 2009; Vitello 2000; Wagner 2002).
Comparing different cancer treatments including radiotherapy regimens (29 trial reports: Andersen 1987; Ardizzoni 2002; Awada 2002; Awwad 2002; Bensadoun 2006; Bentzen 2001; Bleehen 1996; Bourhis 2006; Calais 2000; Cassidy 2002; Cunningham 1995; Damon 2004; De Boer 2002; Denham 1999; Dobrowsky 1998; Doroshow 1987; Erkisi 1996; Erlichman 1988; Ezzat 2005; Falcone 2001; Giles 2003a; Giles 2003b; Gladkov 2007; Lee 1989; Levi 1997; Mahmoud 1996; Pyrhonen 1995; Rabinovitch 2006; Rocci 2005).
Not a randomised controlled trial (64 study reports: Aisa 2005; Altmann 1999; Arora 2008; Awada 2004; Baydar 2005; Calais 2004; Cheng 2001; Cheng 2002; Colella 2010; Costa 2003; Dreicer 1997; Edelman 1998; Eisen 2003; El‐Sayed 2002a; Fahlke 1999; Fay 1994; Foncuberta 2001; Gandara 1997; Gutierrez 1996; Horsley 2007; Hu 2003; Hunter 2007; Inagaki 2006; Ito 2002; Johnson 2002; Ju 2009; Karacetin 2004; Khouri 2009; Kuriakose 2002; Labbate 2003; Luglié 2002; Maddocks‐Jennings 2009; Madero 1999; Malaker 1991; Mantovani 2003; Martin 2006; Matejka 1990; Mills 1995; Mori 2006; Nicolatou‐Galitis 2006; Okutomi 2000; Papadeas 2007; Penpattanagul 2007; Peters 1993; Phillips 2002; Putwatana 2009; Sato 1997; Sato 2006; Schuster 2008; Shabanloei 2009; Simoes 2009; Stokman 2004; Thieblemont 2002; Tiemann 2006; Toubai 2003; Uchiyama 2005; Wang 2002a; Ward 2007; Weiss 1990; Whelan 2002; Whelan 2004; Wollina 2002; Wymenga 1999; Yokomizo 2004).
Multicomponent oral assessment instruments (which included voice, teeth etc) ‐ specifically excluded in methods section of review (16 trial reports: Aquino 2005; Cowen 1997; Dudjak 1987; Epstein 1992; Etiz 2000; Feber 1995; Feber 1996; Grotz 2001; Jebb 1995; Kenny 1990; McIlroy 1996; Nikoletti 2005; Piccirillo 2003; Pytlik 2002; Rothwell 1990; Verdi 1995).
Some mucositis present at baseline (12 trial reports: Anderson 1998b; Barasch 1995; Djuric 2006; Ferretti 1990; Genot‐Klastersky 2008; Kuhn 2009; Lanzos 2010; Loo 2010; Masucci 2005; Prada 1985; Ryu 2007; Valcarcel 2002).
Mucositis prevention not purpose of study (one trial report: Jham 2007).
Data presented as episodes rather than patients, where patients were re‐entered into the study, so data not independent (seven trial reports: Abramoff 2008; Awidi 2001; Hickey 1982; Karthaus 1998; Lorusso 2003; Rojas 2001; van Zaanen 1994).
Major change to protocol half way through study, blinded and unblinded patient data combined (one trial report: Okuno 1997).
Unclear if mucositis present at baseline (one trial report: Cheng 2006).
Unclear if randomised (four trial reports: Apaydin 1996; Howell 1983; Teshima 1986; Zanin 2010).
No mucositis data ‐ study stopped early because preset stooping rule triggered (two trial reports: Antin 2002; Rades 2004).
No clear mucositis data presented and unable to contact authors (two trial reports: Niibe 1985; Rutkauskas 1993).
Design flaw ‐ confounded interventions (one trial report: Papas 2003).

Risk of bias in included studies

Eleven studies were assessed at low overall risk of bias (8%) (Cartee 1995 Dazzi 2003 Foote 1994; Madan 2008; Oberbaum 2001; Pitten 2003; Saarilahti 2002; Schneider 1999; Shenep 1988; Stokman 2003; Su 2006), 82 (63%) were described as unclear and the remaining 38 studies (29%) were defined as being at high overall risk of bias (Antonadou 2002; Antunes 2007; Biswal 2003; Bjarnason 2009; Borowski 1994; Bourhis 2000; Brizel 2000; Buntzel 1998; Cascinu 1994; Chi 1995; Choi 2007; Dai 2009; Dozono 1989; Gandemer 2007; Gori 2007; Gujral 2001; Haddad 2009; Hartmann 2001; Huang 2000; Katano 1995; Kaul 1999; Koukourakis 2000; Lilleby 2006; Makkonen 2000; Mills 1988; Rahn 1997; Rashad 2008; Rocke 1993; Shieh 1997; Spencer 2005; Svanberg 2007; Vacha 2003; Vadhan‐Raj 2010; Veerasarn 2006; Wahlin 1989; Watanabe 2010; Wu 2009; Yuen 2001). See Figure 1 and Figure 2.

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

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

Allocation

Random sequence generation

Twenty‐seven studies (21%) were deemed to have adequate sequence generation, and therefore were classified as being at low risk of bias for this domain. Thirteen of these studies employed computer‐based sequence generation; while, four studies employed minimization. Four studies did not provide enough information about the randomisation process; however it was the opinion of the assessors that the setting of these trials made adequate randomisation likely. These studies were conducted at the Dana Faber cancer institute (Haddad 2009), the Memorial Sloan Kettering Cancer Centre (Su 2004), the Duke Centre (Cartee 1995) and the Finish cancer registry (Makkonen 2000). One study used biased coin randomisation (Su 2004). Of the remaining five studies, three used a table of random numbers (Huang 2003 Koukourakis 2000; Pitten 2003), and two studies provided limited information but made reference to appropriate literature concerning randomisation (Brizel 2000 Shieh 1997). Those studies considered to use an inappropriate method of randomisation were excluded; therefore no studies were given a decision of no for this category. The remaining 104 studies (79%) were judged as 'unclear'. The majority of these unclear studies gave no more information than that they were 'randomised'. Four studies stated that they employed the "closed envelope" method of randomisation. However, no information was provided about whether these envelopes were shuffled prior to the patient being randomised. They were therefore classified as "unclear".

Allocation concealment

Nineteen studies (14%) employed adequate methods of allocation concealment and were therefore classified as being at low risk of bias. Central randomisation was mentioned in 15 studies, with eight studies employing pharmacy controlled randomisation, six studies communicating by telephone, and one study by fax (Gandemer 2007). Two studies employed sequentially numbered drug containers which were identical in appearance (Foote 1994 Madan 2008). Two studies (1.5%) used open number tables without concealment and were therefore deemed to be at high risk of bias (Huang 2003 Koukourakis 2000). The remaining 111 studies were classified as unclear.

Blinding

Blinding was assessed for three different groups: patients, carers and outcome assessors.

Carer blinding: Nineteen studies (15%) described some method of blinding and were therefore deemed to be at low risk of bias for carer blinding. Forty‐five studies (34%) were classified as being at high risk of bias, as no blinding was employed. Sixty‐seven (51%) studies were classified as 'unclear'.

Patient blinding: Forty‐five studies were classified as being at high risk of bias for patient blinding. The majority of these were studies which employed no blinding, however one study was described by its authors as double blind, but then went on to state that a patient withdrew from the study because they were not allocated the intervention of interest (Wu 2009). The assessors were concerned that this suggested a failure in the blinding of patients in this study, and therefore decided to characterise the study at high risk of bias for all three blinding categories. Seventy‐five studies (58%) were deemed to be at low risk of bias for patient blinding. Eleven studies (8%) were classified as unclear. Of these studies, four were deemed unclear as they employed the use of a placebo control, and therefore blinding could not be discounted, while three studies were assessed for risk of bias from a data collection sheet provided by a translator without any additional information. Of the remaining four 'unclear' studies, one study which compared povidone‐iodine to saline was described as 'blind' to patients, however, this was considered by the assessors to be an inappropriate control as presumably the iodine solution would differ in colour from the saline (Vokurka 2005). Another which investigated zinc in head and neck patients receiving a mix of radiotherapy and chemoradiotherapy (Ertekin 2004), was classified as unclear for two reasons: firstly, because the study authors described the need for a double blind study in the introduction, and then failed to provide any information about blinding in the remainder of the text, and secondly, because the authors used empty capsules as the control, and the assessors were concerned that this would be noticeable to the patients. The third study (Vadhan‐Raj 2010) was deemed unclear because the authors stated that adverse events associated with the intervention (keratinocyte growth factor) may have affected the integrity of the blinding. The remaining trial (Ahmed 1993) only stated that "trial drugs were administered blind", without any additional information.

Outcome assessor blinding: Seventy‐seven studies (59%) were deemed to describe any method of outcome assessor blinding adequately and were considered to be at low risk of bias. Sixteen studies (12%) were classified as unclear and 38 studies (29%) were given a decision of 'no' and were therefore considered to be at high risk of bias in this category. In a subanalysis of those studies providing blinding information, only 19 studies gave specific information regarding the blinding of an outcome assessor. The remaining 58 studies were only described as "double blind" by the authors.

Incomplete outcome data

One hundred and six studies (82%) were considered to be at low risk of bias for this category. Seventeen studies (13%) were given a decision of unclear and eight were considered to be at high risk of bias. These seven studies experienced a high rate of drop out.

Selective reporting

Forty‐nine (37%) studies were deemed to be free of selective reporting for mucositis grade, which was determined prior to assessment as the outcome of interest for this category. These studies were therefore deemed to be at a low risk of bias. The remaining 82 studies were classified as unclear. These studies tended to only provide subsets of data for severe mucositis (grade > 2) rather than all the information of interest. No studies were given a decision of 'no', and consequently classified at high risk of bias, as studies which did not provide mucositis information for at least one of the dichotomies of interest could not be included in the review.

Other potential sources of bias

Thirty‐five studies (27%) were deemed to be at high risk of bias in the final 'other' category. A baseline imbalance was reported by 11 studies. Three studies reported gender imbalances (Abbasi‐Nazari 2007 Makkonen 1994; Puataweepong 2009), while three studies reported age imbalances (Bensadoun 1999; Ifrah 1999; Makkonen 1994). Two or more baseline imbalances were reported by four studies (Bensadoun 1999; Ifrah 1999 Makkonen 1994; Puataweepong 2009). Puataweepong 2009 reported baseline imbalances in both patient gender (P = 0.03) and previous surgery (P = 0.04). Meanwhile, in the Ifrah 1999 study, patients randomised to receive GM‐CSF in the intervention arm of the study, were older (P = 0.04) and more likely to have the Philadelphia chromosomal re‐arrangement (P = 0.026). Baseline imbalances in age and gender were reported by Makkonen and colleagues (Makkonen 1994). Bensadoun and colleagues reported imbalances in the number of patients receiving supplementary application of laser to the neck, which was hypothesised to exert a distant beneficial effect. In this study patients in the intervention group also tended to be older. However, no P values were presented by the authors for this imbalance (Bensadoun 1999). Risk of bias was assessed for eight studies from a data collection form completed by a translator. Loprinzi and colleagues initially aimed to recruit 120 patients into their allopurinol study, however, the power calculation was re‐run after 77 patients and as the results were found to favour the intervention, the study was terminated and the data published (Loprinzi 1990). In the Duenas and colleagues study (Duenas 1996), an interim analysis conducted in the 16 patients recruited into the study showed a significant different in favour of the placebo, and the authors therefore decided to cease recruitment. Epstein and colleagues also report the results of an interim analysis, in this case a trial of 33 patients which compared sucralfate to placebo. This trial was terminated after an interim analysis suggested that the impact of sucralfate on mucositis prevention was minimal (Epstein 1994).

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5

Summary of findings 1. Cryotherapy versus no treatment for preventing oral mucositis for patients with cancer receiving treatment.

Cryotherapy versus no treatment for preventing oral mucositis for patients with cancer receiving treatment
Patient or population: preventing oral mucositis for patients with cancer receiving treatment Settings: Intervention: Cryotherapy versus no treatment
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Cryotherapy versus no treatment
Mucositis (any) 0‐4 scale Follow‐up: median 28 days	Low risk population¹		RR 0.74  (0.57 to 0.95)	472 (5 studies)	⊕⊕⊝⊝ low^2,3
	600 per 1000	444 per 1000 (342 to 570)
	High risk population¹
	950 per 1000	703 per 1000 (541 to 902)
Mucositis (severe) 0‐4 scale Follow‐up: median 28 days	Low risk population⁴		RR 0.36  (0.17 to 0.77)	472 (5 studies)	⊕⊕⊝⊝ low^2,5
	300 per 1000	108 per 1000 (51 to 231)
	High risk population⁴
	650 per 1000	234 per 1000 (111 to 501)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Amifostine versus placebo/no treatment for preventing oral mucositis for patients with cancer receiving treatment
Patient or population: preventing oral mucositis for patients with cancer receiving treatment Settings: Intervention: Amifostine versus placebo/no treatment
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Amifostine versus placebo/no treatment
Mucositis (any) 0‐4 scale Follow‐up: median 28 days	Low risk population¹		RR 0.95  (0.91 to 0.99)	430 (3 studies)	⊕⊕⊕⊝ moderate²
	600 per 1000	570 per 1000 (546 to 594)
	High risk population¹
	950 per 1000	902 per 1000 (865 to 941)
Mucositis (severe) 0‐4 scale Follow‐up: median 28 days	Low risk population³		RR 0.68  (0.45 to 1.03)	845 (9 studies)	⊕⊝⊝⊝ very low^4,5,6
	300 per 1000	204 per 1000 (135 to 309)
	High risk population³
	650 per 1000	442 per 1000 (292 to 669)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Keratinocyte GF versus placebo for preventing oral mucositis for patients with cancer receiving treatment
Patient or population: patients with preventing oral mucositis for patients with cancer receiving treatment Settings: Intervention: Keratinocyte GF versus placebo
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Keratinocyte GF versus placebo
Mucositis (any) 0‐4 scale Follow‐up: median 28 days	Low risk population¹		RR 0.82  (0.71 to 0.94)	160 (2 studies)	⊕⊕⊝⊝ low^2,3
	600 per 1000	492 per 1000 (426 to 564)
	High risk population¹
	950 per 1000	779 per 1000 (674 to 893)
Mucositis (severe) 0‐4 scale Follow‐up: median 28 days	Low risk population⁴		RR 0.72  (0.58 to 0.9)	559 (6 studies)	⊕⊕⊝⊝ low^5,6
	300 per 1000	216 per 1000 (174 to 270)
	High risk population⁴
	650 per 1000	468 per 1000 (377 to 585)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Sucralfate versus placebo/usual care for preventing oral mucositis for patients with cancer receiving treatment
Patient or population: preventing oral mucositis for patients with cancer receiving treatment Settings: Intervention: Sucralfate versus placebo/usual care
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Sucralfate versus placebo/usual care
Mucositis (any) 0‐4 scale Follow‐up: median 28 days	Low risk population¹		RR 0.98  (0.88 to 1.1)	222 (3 studies)	⊕⊕⊕⊝ moderate²
	600 per 1000	588 per 1000 (528 to 660)
	High risk population¹
	950 per 1000	931 per 1000 (836 to 1000)
Mucositis (severe) 0‐4 scale Follow‐up: median 28 days	Low risk population³		RR 0.67  (0.48 to 0.92)	428 (7 studies)	⊕⊕⊕⊝ moderate⁴
	300 per 1000	201 per 1000 (144 to 276)
	High risk population³
	650 per 1000	435 per 1000 (312 to 598)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Chlorhexidine versus placebo/no treatment for preventing oral mucositis for patients with cancer receiving treatment
Patient or population: preventing oral mucositis for patients with cancer receiving treatment Settings: Intervention: Chlorhexidine versus placebo/no treatment
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Chlorhexidine versus placebo/no treatment
Mucositis (any) 0‐4 scale Follow‐up: median 28 days	Low risk population¹		RR 0.76  (0.47 to 1.24)	454 (4 studies)	⊕⊕⊝⊝ low^2,3
	600 per 1000	456 per 1000 (282 to 744)
	High risk population¹
	950 per 1000	722 per 1000 (446 to 1000)
Mucositis (severe) 0‐4 scale Follow‐up: median 28 days	Low risk population⁴		RR 0.82  (0.54 to 1.23)	244 (4 studies)	⊕⊕⊝⊝ low^5,6
	300 per 1000	246 per 1000 (162 to 369)
	High risk population⁴
	650 per 1000	533 per 1000 (351 to 800)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

		Experimental		Control		RR (95%CI)	P value
	Mucositis category	Events/ Mean (SD)	Total	Events/ Mean (SD)	Total	RR (95%CI)	P value
Aciclovir versus placebo Bubley 1989	Any	18	27	15	30	1.33 (0.85, 2.08)	0.21
Aloe vera versus placebo	Any Puataweepong 2009	29	30	31	31	0.97 (0.88, 1.06)	0.46
Aloe vera versus placebo	Severe Su 2004	1	30	8	31	0.13 (0.02, 0.97)	0.05
BCoG antibiotic pastilles El Sayed 2002	Any	62	69	60	68	1.02 (0.91, 1.15)	0.76
	Moderate plus severe	50	69	53	68	0.93 (0.77, 1.13)	0.46
	Severe	31	69	34	68	0.90 (0.63, 1.28)	0.55
Benzydamine versus placebo	Any Prada 1987	12	19	16	17	0.67 (0.47, 0.97)	0.03
Benzydamine versus placebo	Severe Kazemian 2009	17	39	33	42	0.55 (0.38, 0.82)	0.003
Beta carotene versus no treatment control	Severe Mills 1988	3	10	8	10	0.38 (0.14, 1.02)	0.05
Camomile versus placebo Fidler 1996	Any	33	82	37	82	0.89 (0.62, 1.27)	0.53
	Moderate plus severe	12	82	19	82	0.63 (0.33, 1.22)	0.17
	Severe	8	82	7	82	1.14 (0.43, 3.01)	0.79
Chewing gum versus no chewing gum Gandemer 2007	Any	48	70	55	70	0.87 (0.71, 1.07)	0.18
Chewing gum versus no chewing gum Gandemer 2007	Severe	36	70	31	70	1.16 (0.82, 1.64)	0.40
Clarithromycin (systemic antibiotic) versus no treatment Yuen 2001	Moderate plus severe	18	35	26	35	0.69 (0.48, 1.01)	0.06
Dental stent versus no treatment control Qin 2007	Moderate plus severe	14	19	21	24	0.84 (0.62, 1.15)	0.27
Dental stent versus no treatment control Qin 2007	Severe	5	19	9	24	0.70 (0.28, 1.75)	0.45
Epidermal growth factor versus placebo Wu 2009	Moderate plus severe	32	76	17	27	0.67 (0.45, 0.99)	0.04
G‐CSF versus placebo Su 2006	Moderate plus severe	11	19	15	21	0.81 (0.51, 1.30)	0.38
Histamine gel versus placebo Elad 2006	Any	17	20	12	19	1.35 (0.91, 1.99)	0.14
Histamine gel versus placebo Elad 2006	Severe	2	20	2	19	0.95 (0.15, 6.08)	0.96
Hydrolytic enzymes (papain, trypsin, chymotrypsin, pancreatin, bromelain) versus no treatment Kaul 1999	Severe	0	25	2	25	0.20 (0.01, 3.97)	0.29
Indomethacin versus placebo Pillsbury 1986	Moderate plus severe	10	10	8	8	1.00 (0.82, 1.23)	1.00
Indigo wood root versus saline (placebo) You 2009	Moderate to severe	8	11	9	9	0.75 (0.50, 1.10)	0.14
Indigo wood root versus saline (placebo) You 2009	Severe	1	11	6	9	0.14 (0.02, 0.93)	0.04
Intestinal trefoil factor versus placebo Peterson 2009	Any	23	66	22	33	0.52 (0.35, 0.79)	0.002
	Moderate plus severe	7	66	16	33	0.22 (0.10, 0.48)	0.0001
	Severe	1	66	0	33	1.52 (0.06, 36.39)	0.80
Oral care protocol versus none	Any Shieh 1997	12	20	10	10	0.62 (0.43, 0.91)	0.01
	Moderate plus severe Borowski 1994	64	75	70	75	0.91 (0.82, 1.02)	0.12
	Severe Borowski 1994	49	75	58	75	0.84 (0.69, 1.04)	0.11
Pentoxifylline versus no treatment Attal 1993	Moderate plus severe	30	70	30	70	1.00 (0.68, 1.47)	1.00
Pilocarpine versus placebo Lockhart 2005	Severe	12	18	12	16	0.89 (0.58, 1.37)	0.59
Prednisone versus placebo Leborgne 1997	Any	28	32	29	34	1.03 (0.85, 1.24)	0.79
	Moderate plus severe	16	32	24	34	0.71 (0.47, 1.07)	0.10
	Severe	3	32	5	34	0.64 (0.17, 2.45)	0.51
Prostaglandin versus placebo (Veness 2006)	Moderate plus severe	36	42	37	41	0.95 (0.81, 1.11)	0.53
Propantheline versus placebo Ahmed 1993	Any	2	6	5	6	0.40 (0.12, 1.31)	0.13
PTA antibiotic paste (Wijers 2001)	Moderate plus severe	26	39	28	38	0.90 (0.68, 1.21)	0.5
Shenqi‐fanghou versus no treatment Hu 2005	Any	60	70	67	70	0.90 (0.80, 1.00)	0.04
	Moderate plus severe	30	70	59	70	0.51 (0.38, 0.68)	0.00001
	Severe	10	70	50	70	0.20 (0.11, 0.36)	0.00001
Superoxide dismutase versus placebo Tu 1998	Any	59	119	17	40	1.17 (0.78, 1.75)	0.45
Sucralfate mouthwash plus gel on skin versus placebo mouthwash plus gel on skin (Evensen 2001)	Any	30	30	28	30	1.07 (0.96, 1.20)	0.24
	Moderate plus severe	29	30	24	30	1.21 (1.00, 1.46)	0.05
	Severe	26	30	23	30	1.13 (0.89, 1.44)	0.32
Traumeel versus placebo Oberbaum 2001	Any	10	15	14	15	0.71 (0.49, 1.05)	0.08
Zinc sulphate versus placebo Ertekin 2004	Any	13	15	12	12	0.88 (0.69, 1.11)	0.28
	Moderate plus severe	5	15	12	12	0.36 (0.18, 0.71)	0.003
	Severe	0	15	8	12	0.05 (0.00, 0.75)	0.03

Intervention	Study	Text
Amifostine versus no treatment	Vacha 2003	CTC mucositis index used. Results presented as weekly means for both arms with standard deviations. Text indicated statistically significant difference in favour of amifostine at 2 weeks but does not mention overall result.
Benzydamine versus placebo	Epstein 1989	Signs of mucositis were recorded by area of involvement, severity of inflammation, severity of ulceration and maximum size of ulceration for each region of the oral cavity. Results in Table 3 indicate borderline statistically significant differences in favour of benzydamine. Maximum size of ulcerations (P = 0.04); total area of ulcerations (P = 0.05); average area of mucositis (P = 0.050).
Benzydamine versus placebo	Epstein 2001	Area under the curve of mean mucositis cores presented for different radiotherapy intervals. Overall there was a statistically significant difference in favour of benzydamine (P = 0.006), Table 4.
Chlorhexidine versus placebo	McGaw 1985	Hickey (0‐3 scale) index for mucositis used over 4‐week period. During the third and fourth weeks the average mucositis scores were significantly higher in the control group.
Chlorhexidine versus povidone‐iodine, salt/soda versus water	Madan 2008	After 4 weeks there was a statistically significant decrease in mean mucositis scores in each of the active treatment groups compared to placebo.
Cryotherapy versus no treatment control	Svanberg 2007	"The results demonstrated that oral cryotherapy alleviated the development of mucositis and oral pain, which resulted in a reduction in the number of days of iv opioids for patients treated with autologous BMT."
GM‐CSF versus no treatment control	Chi 1995	Cross‐over study showing period effect but indicating GM‐CSF significantly prevents mucositis (P < 0.001).
Laser versus placebo light treatment	Bensadoun 1999	Parallel group study mucositis measured on 0‐4 scale. Mean calculated for each patient over 7 weeks. Quote "the mean grade of mucositis during radiotherapy was 2.1 +/ 0.26 for the group without laser and 1.7 +/‐ 0.26 for the group with laser (P = 0.01)."
Laser versus sham laser treatment	Schubert 2007	OMI appropriate index (Schubert 1992). Quote: "Figure 1 shows the mean OMI over time by treatment group. The placebo patient scores are higher on average than the laser patient scores at nearly every time point, signifying more severe mucositis over the course of the study.." The authors then present day 11 data and statistical test for that day (P = 0.06). "The peak severity of mucositis that generally occurs during the second week of transplant was reduced in the 650 nm laser group." The results of the overall burden over time in Table 2 showed the differences in the unadjusted model to be non‐significant. Only one difference comparing low‐level laser with placebo was significant in the adjusted model (P = 0.03).
Prostaglandin versus placebo	Hanson 1995	Data from 2 centres reported separately. Overall ANOVA for 1 centre showed no significant difference. The other centre found statistically significant differences for weeks 4 and 5 with less mucositis in the intervention group (P < 0.05).
Sucralfate versus placebo	Epstein 1994	Signs of mucositis were recorded by area of involvement, severity of inflammation, severity of ulceration and maximum size of ulceration for each region of the oral cavity. No statistically significant difference was seen in mucositis ulceration or the composite mucositis score (Table 3). Total mucositis score: Placebo (n = 17) mean = 18.7 +/‐ 21.3 Sucralfate (n = 16) mean = 22.3 +/‐ 31.2 Total mucosal ulceration: Placebo (n = 17) mean = 15.3 +/‐ 17.6 Sucralfate (n = 16) mean = 19.4 +/‐ 19.8.
Sucralfate versus placebo	Lievens 1998	Mucositis scores on 0‐6 ECOG scale. Graph (fig 1) displays mean mucositis scores for each week 1‐7. Quote: "Comparing the time course of the mean scores for . . .mucositis. . . no statistically significant differences between the two treatment arms were observed." Quote: "At 5 week when the mucosal reaction tends to be most severe a clear but not statistically significant advantage is seen for sucralfate as opposed to placebo."
Zinc versus placebo	Lin 2006	RTOG mucositis index used. Results presented as graphs. Quote: "This study no significant difference was found in the improvement of radiation mucositis and dermatitis during the 2 weeks between the patients with zinc supplement and those without."

Author	Type of keratinocyte	Dose of keratinocyte	Schedule	Number and gender of patients	Cancer type	Treatment
Blazar 2006	Palifermin	40 µg/kg or 60 µg/kg	Cohort 1: patients randomised to placebo or palifermin at either 40 µg (8 patients, total dose 240 µg) or 60 µg (10 patients, total dose 360 µg) per day for 3 days before conditioning (days ‐11 to ‐9) and for 3 days after transplant (days 0, 1 and 2). Cohort 2: 14 patients received palifermin at 60 µg/day, for 3 days before conditioning (day ‐11 to ‐9) and then for 6 days after transplant (days 0‐2 and then days 7‐9) (9 doses total). Total dose of palifermin received was 540 µg. Cohort 3: 37 patients received palifermin at 60 µg/day for 3 days before conditioning (day ‐11 to ‐9) and then for 9 days after transplant (days 0‐2, 7‐9 and 14‐16) (12 doses total). Total dose of palifermin received was 720 µg	100 (58M/ 42F)	Leukaemia, lymphoma, myelodysplastic syndrome	Allogeneic stem cell transplant. Cyclophosphamide 60 mg/kg per day, TBI total dose = 13.2 Gy (fractionated as 165 Gy twice daily for 4 days) or busulfan 1 mg/kg per dose given 4 times daily for 4 days then cyclophosphamide 60 mg/kg per day for 2 days
Brizel 2008	Palifermin	60 µg (67 patients)	1 dose administered weekly on the Friday before the first week of chemotherapy (then continued each Friday for 7 consecutive weeks). 2 additional doses given weeks 8 and 9	99 (82M/17F)	Head and neck	Chemotherapy (cisplatin 20 mg/m²/d IV bolus and fluorouracil 1000 mg/m²/d continuous infusion) administered for first 4 days of the first and fifth weeks of radiotherapy. Radiotherapy (daily fractions of 2 Gy until 70 Gy) or hyperfractionated radiotherapy (1.25 Gy twice daily until 72 Gy)
Freytes 2004	Repifermin	25 µg/kg or 50 µg/kg	Cohort 1: 25 µg/kg repifermin. Cohort 2: 50 µg/kg repifermin	42 (31M/11F)	Haematologic malignancies or lymphoma	Autologous stem cell transplant. Cyclophosphamide, etoposide and carmustine, or melphalan monotherapy, or melphalan combination, or cyclophosphamide + TBI or thiotepa + TBI or cyclophosphamide + busulfan
Meropol 2003	Palifermin	1 µg/kg/d or 10 µg/kg/d or 20 µg/kg/d or 40 µg/kg/d or 60 µg/kg/d or 80 µg/kg/d	All cohorts received palifermin on days 1 to 3 of each cycle. Cohort 1: 1 µg/kg/d palifermin Cohort 2: 10 µg/kg/d palifermin Cohort 3: 20 µg/kg/d palifermin Cohort 4: 40 µg/kg/d palifermin Cohort 4: 60 µg/kg/d palifermin Cohort 5: 80 µg/kg/d palifermin. 27 placebo patients (randomised 1:1 in cohort 1 and 2:1 in all other cohorts)	81(47M/34F)	Metastatic colorectal cancer	Leucovorin 20 mg/m² by IV followed immediately by 425 mg/m² for 5 consecutive days on days 4 to 8 of each 28 day cycle
Rosen 2006	Palifermin	40 µg/kg per day (28 patients)	3 consecutive days before chemotherapy	64 (42M/22F)	Solid tumours (colon and rectum)	Chemotherapy (5‐FU 425 mg/m²/day IV for 5 days, leucovorin 20 mg/m²/day for 5 days)
Spielberger 2004	Palifermin	60 µg/kg per day	3 consecutive days (starting 3 days before TBI) and 3 consecutive doses after transplantation (day 0, day 1, day 2)	212 (131M/81F)	Lymphoma, leukaemia and multiple myeloma	Autologous stem cell transplant. TBI (total 1200 Gy) chemotherapy included etoposide (60 mg/kg) and cyclophosphamide (100 mg/kg)
Vadhan‐Raj 2010	Palifermin	180 µg/kg	1 dose 3 days before chemotherapy	48 (25M/23F)	Sarcoma	Chemotherapy (doxorubicin 90 mg/m² over 72 hours, ifosfamide 10 m/m² 3 hour infusion for 4 days)

Date	Event	Description
31 January 2013	Amended	Corrected error in summary of findings table 1
8 March 2011	New citation required and conclusions have changed	Results and conclusions changed and five summary of findings tables added.
8 March 2011	New search has been performed	Updated search.
10 November 2010	New search has been performed	42 new studies and 14 new interventions added. Risk of bias assessments incorporated.
10 November 2010	New citation required and conclusions have changed	Nine interventions found to be beneficial for the prevention of mucositis. Conclusions changed.
16 June 2008	Amended	Converted to new review format.
21 August 2007	New citation required and conclusions have changed	Substantive amendment. This substantial update with a search only 14 months after the previous one includes 18 new included studies, bringing the total of number of studies up to 89. There are four new interventions included, bringing the total number of interventions to 33.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Mucositis (any)	4		Risk Ratio (IV, Random, 95% CI)	0.77 [0.50, 1.19]
1.2 Mucositis (moderate plus severe)	2	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.50, 0.86]
1.3 Mucositis (severe)	2	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.63, 1.04]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Mucositis (any)	3	430	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.91, 0.99]
3.2 Mucositis (moderate plus severe)	6	757	Risk Ratio (M‐H, Random, 95% CI)	0.75 [0.58, 0.96]
3.3 Mucositis (severe)	9	845	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.45, 1.03]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Mucositis (any)	4	454	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.47, 1.24]
4.2 Mucositis (moderate plus severe)	3	233	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.72, 1.21]
4.3 Mucositis (severe)	4	244	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.54, 1.23]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
5.1 Mucositis (any)	5	Risk Ratio (IV, Random, 95% CI)	0.74 [0.57, 0.95]
5.2 Mucositis (moderate plus severe)	5	Risk Ratio (IV, Random, 95% CI)	0.53 [0.31, 0.91]
5.3 Mucositis (severe)	5	Risk Ratio (IV, Random, 95% CI)	0.36 [0.17, 0.77]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
6.1 Mucositis (any)	6	Risk Ratio (IV, Random, 95% CI)	Subtotals only
6.1.1 Oral suspension/supplementation	5	Risk Ratio (IV, Random, 95% CI)	0.78 [0.57, 1.08]
6.1.2 IV supplementation	1	Risk Ratio (IV, Random, 95% CI)	0.33 [0.04, 2.63]
6.2 Mucositis (moderate plus severe)	6	Risk Ratio (IV, Random, 95% CI)	Subtotals only
6.2.1 Oral suspension/supplementation	5	Risk Ratio (IV, Random, 95% CI)	0.88 [0.69, 1.12]
6.2.2 IV supplementation	1	Risk Ratio (IV, Random, 95% CI)	0.33 [0.04, 2.63]
6.3 Mucositis (severe)	8	Risk Ratio (IV, Random, 95% CI)	Subtotals only
6.3.1 Oral suspension/supplementation	5	Risk Ratio (IV, Random, 95% CI)	0.69 [0.37, 1.29]
6.3.2 IV supplementation	3	Risk Ratio (IV, Random, 95% CI)	0.25 [0.10, 0.62]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Mucositis (any)	4	263	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.48, 1.23]
7.2 Mucositis (severe)	2	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.15, 0.86]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Mucositis (any)	2	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.84, 1.04]
8.2 Mucositis (moderate plus severe)	2	138	Risk Ratio (M‐H, Fixed, 95% CI)	0.89 [0.76, 1.05]
8.3 Mucositis (severe)	6	373	Risk Ratio (M‐H, Random, 95% CI)	0.73 [0.39, 1.40]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Mucositis (any)	3	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.56, 0.88]
9.2 Mucositis (moderate plus severe)	2	80	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.31, 0.74]
9.3 Mucositis (severe)	2	80	Risk Ratio (M‐H, Fixed, 95% CI)	0.26 [0.13, 0.52]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
10.1 Mucositis (any)	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
10.2 Mucositis (moderate plus severe)	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 Mucositis (moderate plus severe)	2	926	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.89, 1.03]
11.2 Mucositis (severe)	2	926	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.89, 1.10]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
12.1 Mucositis (any)	2	160	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.71, 0.94]
12.2 Mucositis (moderate plus severe)	7	640	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.62, 0.89]
12.3 Mucositis (severe)	6	559	Risk Ratio (M‐H, Random, 95% CI)	0.72 [0.58, 0.90]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
13.1 Mucositis (any)	3	131	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.71, 1.17]
13.2 Mucositis (moderate plus severe)	2	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.64 [0.38, 1.08]
13.3 Mucositis (severe)	2	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.20 [0.06, 0.62]

PERMALINK

Interventions for preventing oral mucositis for patients with cancer receiving treatment

Helen V Worthington

Janet E Clarkson

Gemma Bryan

Susan Furness

Anne-Marie Glenny

Anne Littlewood

Martin G McCabe

Stefan Meyer

Tasneem Khalid

Philip Riley

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcome

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

Results

Description of studies

Results of the search

Included studies

Characteristics of the trial settings and investigators

Characteristics of the participants

Characteristics of the interventions

Mucositis

Secondary outcomes

Excluded studies

Risk of bias in included studies

1.

2.

Allocation

Random sequence generation

Allocation concealment

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings 1. Cryotherapy versus no treatment for preventing oral mucositis for patients with cancer receiving treatment.

Summary of findings 2. Amifostine versus placebo/no treatment for preventing oral mucositis for patients with cancer receiving treatment.

Summary of findings 3. Keratinocyte GF versus placebo for preventing oral mucositis for patients with cancer receiving treatment.

Summary of findings 4. Sucralfate versus placebo/usual care for preventing oral mucositis for patients with cancer receiving treatment.

Summary of findings 5. Chlorhexidine versus placebo/no treatment for preventing oral mucositis for patients with cancer receiving treatment.

Interventions with more than one trial

Allopurinol versus placebo/no treatment (Analysis 1.1; Analysis 1.2; Analysis 1.3)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
14.1 Mucositis (any)	2	276	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [1.00, 1.10]
14.2 Mucositis (moderate plus severe)	2	276	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.94, 1.16]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
15.1 Mucositis (any)	2	172	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.76, 1.08]
15.2 Mucositis (moderate plus severe)	2	172	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.62, 1.10]
15.3 Mucositis (severe)	2	172	Risk Ratio (M‐H, Fixed, 95% CI)	0.65 [0.40, 1.06]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
16.1 Mucositis (any)	3	159	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.94, 1.12]
16.2 Mucositis (severe)	2	143	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.72, 1.43]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
17.1 Mucositis (any)	2	298	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.78, 0.96]
17.2 Mucositis (severe)	2	135	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.66, 1.09]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
19.1 Mucositis (any)	3	222	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.88, 1.10]
19.2 Mucositis (moderate plus severe)	4	164	Risk Ratio (M‐H, Random, 95% CI)	0.75 [0.54, 1.04]
19.3 Mucositis (severe)	7		Risk Ratio (IV, Random, 95% CI)	0.67 [0.48, 0.92]

*Study characteristics*
Methods	Randomised, parallel group multicentre study conducted in Iran. Unclear if dentist involved in the study. Drop outs: unclear. Duration: 42 days.
Participants	Adults with cancers of the oral cavity, nasopharynx or hypopharynx treated with radiotherapy or chemoradiotherapy (Cisplatine). Data presented for 24 patients. No dates for start and finish of recruitment.
Interventions	2 groups: placebo versus allopurinol mouthrinse. 10 mL of solution rinsed three times a day for three minutes. Solution then discarded.
Outcomes	Authors state that mucositis was graded weekly using the WHO instrument. However, the instrument reproduced in the publication is not the WHO instrument. Other reported outcomes: None.
Notes	Funding source: pharmaceutical.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "A dynamic randomization procedure was utilized to divide patients to receive allopurinol mouthwash (treatment group) or placebo mouthwashes (control group)". Comment: random component not described.
Allocation concealment (selection bias)	Unclear risk	Insufficient information to determine 'yes' or 'no'.
Blinding (performance bias and detection bias) Carers	Unclear risk	Insufficient information to determine 'yes' or 'no'.
Blinding (performance bias and detection bias) Patients	Low risk	Quote: "double blind". Comment: probably done.
Blinding (performance bias and detection bias) Outcome assessors	Low risk	Quote: "double blind". Comment: probably done.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "Patients who experienced hypersensitivity reaction or serious side effects were excluded from the study. Also patients who complained about pain or other signs attributed to severe mucositis were excluded from the study." Comment: 24 patients included in analysis. However, it is unclear how many patients were recruited.
Selective reporting (reporting bias)	Unclear risk	Data presented for 24 patients for grade of mucositis by grade of mucositis by week for each arm of the study (figures 1 and 2). However, it is unclear how many patients were initially recruited.
Other bias	High risk	Significant gender imbalance between the arms of the study (P = 0.028). Authors state that they used to WHO instrument. However, this is not the instrument presented in table 1.

*Study characteristics*
Methods	Randomised, parallel group study conducted in USA. Unclear whether dentist was involved in study.
Participants	Adults with haematological malignancies prior to BMT after conditioning with etoposide. 12 enrolled and completed.
Interventions	2 groups, placebo versus propantheline (30 mg every 6 hours during infusion and 12 hours after, for total of 6 doses).
Outcomes	Mucositis graded with reference to previous publication. Data presented as number of patients developing mucositis in both groups. Assessment used: day 3. Other reported outcomes: blood counts febrile episodes, survival, tumour response.
Notes	All patients received conditioning regimen of etoposide, cyclophosphamide and carmustine, together with acyclovir, and nystatin or clotrimazole. Funding source: unclear.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "patients were randomly assigned". Comment: random component not described.
Allocation concealment (selection bias)	Unclear risk	Insufficient information to determine 'yes' or 'no'.
Blinding (performance bias and detection bias) Carers	Unclear risk	Insufficient information to determine 'yes' or 'no'.
Blinding (performance bias and detection bias) Patients	Unclear risk	Quote: "Trial drugs were administered blind". Comment: unclear who was blind.
Blinding (performance bias and detection bias) Outcome assessors	Unclear risk	Quote: "Assessment of mucositis severity was performed by two independent observers". Comment: unclear if they were blinded to allocation.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	12 patients randomised. Authors do not give information about drop outs. Unclear how many patients in each arm. Unclear how many patients included in results.
Selective reporting (reporting bias)	Unclear risk	Brief mucositis data presented in text (enough for use in mucositis absent versus present dichotomy).
Other bias	Low risk	Study appears to be free of other sources of bias.

*Study characteristics*
Methods	Randomised, cross‐over study conducted in USA. Unclear if dentist was involved in study. Duration 14 days. Recruitment May 1993 to April 1996.
Participants	Children and adults with solid cancer (sarcoma/blastoma) who have previously had chemotherapy and experienced mucositis. 24 patients eligible and enrolled, 13 completed.
Interventions	2 groups, glycine control (described as placebo) versus glutamine (4 ml/M² twice daily swish and swallow) for 14 days.
Outcomes	Mucositis (patient's description on 0‐4 scale). Grade >= 2 painful mucositis which altered food intake. Assessment used: day 14. Other reported outcomes: none.
Notes	Funding source: charity/university.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were assigned randomly to two courses of glutamine and two courses of glycine". Comment: random component not described.
Allocation concealment (selection bias)	Unclear risk	Insufficient information to determine ‘yes’ or ‘no’.
Blinding (performance bias and detection bias) Carers	Low risk	Quote: "In addition to the patients, the nurses and oncologists involved in the care of these patients also were blinded". Comment: probably done.
Blinding (performance bias and detection bias) Patients	Low risk	Quote: "double blind". Comment: probably done.
Blinding (performance bias and detection bias) Outcome assessors	Low risk	Quote: "double blind". Comment: probably done.
Incomplete outcome data (attrition bias) All outcomes	High risk	24 patients randomised. Paired outcome data available for 13 patients. Authors give full reasons for attrition/exclusion. Potential for overall estimate to be reversed if excluded patients were included.
Selective reporting (reporting bias)	Unclear risk	Data presented for total days of mucositis compared to patient age and days of mucositis > grade 2 by chemotherapy regime.
Other bias	Low risk	Study appears to be free of other sources of bias.

*Study characteristics*
Methods	Randomised, parallel multisite study conducted in France. Clear information about withdrawals: 0. Dentist involved in study. 30 patients recruited between September 1994 and March 1998. Duration: 5 consecutive days each week during the 7 weeks of radiotherapy.
Participants	Adults with head and neck cancers receiving radiotherapy (at least 65 Gy in total).
Interventions	2 groups, sham laser versus laser (wavelength: 632.8 nm; power: 60 mW in Nice and Mareilles, 25 mW in Reims).
Outcomes	Mucositis assessed weekly using the WHO scale. Other outcome measures: pain, ability to swallow, incidence and duration of treatment gaps.
Notes	Funding source: unclear.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "patients were assigned to either laser treatment or sham‐treatment by computer blocked randomisation, 15 in each arm”. Comment: computer generated randomisation.
Allocation concealment (selection bias)	Unclear risk	Insufficient information to determine 'yes' or 'no'.
Blinding (performance bias and detection bias) Carers	Low risk	Quote: "this operator was the only person to know whether or not the patients was being sham treated, and did not participate in the evaluation and scoring of mucositis". Comment: unlikely that nursing staff would know of allocation.
Blinding (performance bias and detection bias) Patients	Low risk	Quote: "double blind". Comment: probably done.
Blinding (performance bias and detection bias) Outcome assessors	Low risk	Quote: "Objective assessment of degree of mucositis was recorded weekly by a physician blinded to the type of treatment the patient received". Comment: probably done.
Incomplete outcome data (attrition bias) All outcomes	Low risk	30 patients randomised. No missing outcome data.
Selective reporting (reporting bias)	Unclear risk	Data presented for mean grade of mucositis by arm by week of treatment (table 3/ fig 3) and distribution of mucositis grades, in number of weeks, for both treatment weeks.
Other bias	Unclear risk	Possible imbalance of groups at baseline laser group older (mean 62.7 vs 58.1). Also 12/15 of laser group vs. 6/15 of placebo group received laser treatment to skin of neck ("possible distant laser effect”). Different power of laser used at one study site (Relms).

*Study characteristics*
Methods	Randomised parallel group multisite study conducted in Canada. Clear information about withdrawals. Dentist involved in study. Recruitment of patients took place between August 1999 and November 2004.
Participants	Adults with head and neck cancers receiving radiotherapy (50‐70 Gy).
Interventions	2 groups, radiotherapy in the morning (8am to 10am) versus radiotherapy in the evening (4pm to 6pm).
Outcomes	Mucositis assessed weekly using the OMAS instrument. Other reported outcomes: disease reoccurrence, quality of life, compliance with treatment, death, smoking status, weight loss.
Notes	Funding source: National Cancer Institute of Canada.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "A minimization procedure was used to randomize patients". Quote: "patients were stratified by treatment centre, pretreatment smoking behaviour...tobacco use questionnaire, and planned total radiation dose". Comment: minimization.
Allocation concealment (selection bias)	Unclear risk	Insufficient information to determine 'yes' or 'no'.
Blinding (performance bias and detection bias) Carers	High risk	Morning versus evening radiation. No apparent blinding.
Blinding (performance bias and detection bias) Patients	High risk	Morning versus evening radiation. No apparent blinding.
Blinding (performance bias and detection bias) Outcome assessors	High risk	Quote: "the oncologists performing the evaluations were aware of the treatment arm for each patient".
Incomplete outcome data (attrition bias) All outcomes	Low risk	216 patients recruited. 7 patients deemed ineligible (4/108 arm A, 3/108 arm B), 3 patients in arm A did not receive radiotherapy (3/108 arm A), 1 patient in arm B did not have mucositis assessment. 205 patients included in final analysis.
Selective reporting (reporting bias)	Unclear risk	Data presented for number of patients with grade 3 or greater mucositis, median interval to mucositis and median grade of mucositis (all RTOG). Data collected using OMAS instrument also presented (table 3).
Other bias	Low risk	Study appears to be free from other sources of bias.