Nutritional interventions for reducing gastrointestinal toxicity in adults undergoing radical pelvic radiotherapy

  • Review
  • Intervention

Authors


Abstract

Background

Across the developed world, an estimated 150,000 to 300,000 people are treated annually with pelvic radiotherapy and 80% will develop gastrointestinal (GI) symptoms during treatment. Acute GI symptoms are associated with a greater risk of chronic, often debilitating, GI symptoms. Up to one-third of patients are malnourished before pelvic radiotherapy and up to four-fifths of patients lose weight during treatment. Malnutrition is linked to a higher risk of GI toxicity, which can lead to breaks in radiotherapy and early cessation of chemotherapy, thus compromising the efficacy of the primary cancer treatment.

Objectives

To assess the effects of nutritional interventions for reducing GI toxicity in adults undergoing radical pelvic radiotherapy.

Search methods

We searched the Cochrane Gynaecological Cancer Group's Trials Register, the Cochrane Central Register of Controlled Trials (CENTRAL), Issue 4, 2012, MEDLINE and EMBASE to May 2012. We handsearched the citation lists of included studies and previous systematic reviews identified to identify further relevant trials.

Selection criteria

We included studies if they were randomised controlled trials (RCTs) or non-randomised studies with concurrent comparison groups including quasi-randomised trials, cluster RCTs, non-randomised trials, prospective and retrospective cohort studies, and case series of 30 or more patients. We only included studies if they assessed the effect of a nutritional intervention in adults aged 18 years or over undergoing radical pelvic radiotherapy as part of anticancer treatment for a primary pelvic malignancy. We excluded patients with stomas and a previous history of inflammatory bowel disease. Nutritional support interventions could be provided at any stage before or during pelvic radiotherapy and included dietary counselling; dietary modification of fibre, lactose or fat; supplementary foods or drinks or fortified foods; standard oral nutrition supplements including polymeric-, peptide- or amino acid-based supplements and those where novel substrates have been added; enteral tube feeds; or parenteral nutrition (partial or total). We excluded probiotics, prebiotics and synbiotics.

Data collection and analysis

Two review authors independently assessed trial quality and extracted data. We contacted study authors to obtain missing data. We assessed bias for each of the included studies using the bias assessment tables in the Cochrane software Review Manager5. We performed meta-analysis, when indicated, using the Mantel-Haenszel fixed-effect method or inverse variance fixed-effect method displayed with heterogeneity. We undertook meta-analyses on trials evaluating dietary modification against standard treatment for diarrhoea at the end of radiotherapy and for change in weight from baseline to end of radiotherapy.

Main results

The searches identified 7558 titles, and we excluded 7513 during title and abstract searches. We reviewed 45 papers in full, and excluded 39. We identified four studies on handsearching of the references, which, along with the six eligible papers from the database search, led to 10 studies being included. Four studies, three of which were RCTs and one prospective study, investigated the effect of elemental diet on GI symptoms; one RCT investigated the effect of dietary modification and elemental diet; and five RCTs investigated dietary modification. Studies were varied in terms of risk of bias. Data were dichotomised for presence and absence of diarrhoea at the end of radiotherapy for four trials evaluating dietary modification comprising modified fat, lactose, fibre or combinations of these dietary changes. A reduction in diarrhoea was demonstrated with nutritional intervention (risk ratio (RR) 0.66; 95% confidence interval (CI) 0.51 to 0.87, four studies, 413 participants, moderate quality of evidence) with low heterogeneity (Chi2 = 3.50, I2 = 14%). Two trials evaluating dietary modification on weight change (comparing baseline and end of radiotherapy) showed no difference between intervention or control (mean difference (MD) -0.57 kg; 95% CI -1.22 to 0.09) with low heterogeneity (Chi2 = 1.41, I2 = 29%). Generally adverse effects were poorly reported in included studies. Elemental diet in particular was poorly tolerated. GI symptoms or toxicity > 6 months after radiotherapy was not reported in included studies

Authors' conclusions

There have been benefits demonstrated with dietary modification during pelvic radiotherapy to reduce diarrhoea. Those diets included single interventions or combinations of modified fat, lactose-restriction, fat-restriction and fibre supplementation. We were unable to meta-analyse elemental diet, as data were not available. We considered some of the studies to be at high risk of bias. There have been recent advances in novel, more targeted radiotherapy techniques, such that the findings of older studies need to be interpreted with caution. In addition, there were problems with compliance and palatability with some of the interventions, particularly elemental diet, which limits its usefulness in clinical practice.

Résumé scientifique

Interventions nutritionnelles pour réduire la toxicité gastro-intestinale chez les adultes subissant une radiothérapie pelvienne radicale

Contexte

Parmi les pays développés, on estime que 150 000 à 300 000 patients sont traités annuellement par radiothérapie au niveau de la région pelvienne et 80% développent des symptômes gastro-intestinaux (GI) pendant le traitement. Les symptômes GI aigus sont associés à un risque accru de symptômes GI chroniques, souvent invalidant. Jusqu' à un tiers des patients sont dénutris avant la radiothérapie au niveau de la région pelvienne et jusqu' à quatre cinquième des patients perdent du poids pendant la période de traitement. La dénutrition est associée à un risque plus élevé de toxicité GI, ce qui peut conduire à effectuer une radiothérapie par intervalles et un arrêt précoce de la chimiothérapie, compromettant par conséquent l'efficacité du traitement du cancer primaire.

Objectifs

Évaluer les effets des interventions nutritionnelles pour réduire la toxicité GI chez les adultes subissant une radiothérapie radicale au niveau de la région pelvienne.

Stratégie de recherche documentaire

Nous avons effectué des recherches dans le registre des essais cliniques du groupe Cochrane sur les cancers gynécologiques, le registre Cochrane des essais contrôlés (CENTRAL), numéro 4, 2012, MEDLINE et EMBASE jusqu' en mai 2012. Nous avons effectué une recherche manuelle dans les bibliographies des études incluses et des revues systématiques précédentes afin d'identifier d'autres essais pertinents.

Critères de sélection

Nous avons inclus les études s’il s’agissait d’essais contrôlés randomisés (ECR) ou des études non randomisées avec des groupes de comparaison concurrents, y compris les essais quasi-randomisés, les ECR en cluster, les essais non randomisés, les études de cohortes prospectives et rétrospectives et les séries de cas de 30 patients ou plus. Nous avons uniquement inclus les études si elles évaluaient l'effet d'une intervention nutritionnelle chez des adultes âgés de 18 ans ou plus subissant une radiothérapie pelvienne radicale dans le cadre du traitement anticancéreux pour une malignité pelvienne primaire. Nous avons exclu les patients atteints de stomies et ayant des antécédents d'une maladie intestinale inflammatoire. Les interventions de soutien nutritionnel pouvaient être administrées à n'importe quel stade avant ou pendant la radiothérapie au niveau de la région pelvienne et incluaient les conseils diététiques ; les modification en fibres alimentaires, en lactose ou en matières grasses; les aliments et les boissons complémentaires ou les aliments enrichis ; les suppléments nutritifs standards par voie orale, incluant les suppléments à base d’acide polymérique, peptides ou aminés et ceux avec l’ajout de nouveaux substrats; une alimentation par sonde entérale ou une nutrition parentérale (partielle ou totale). Nous avons exclu les probiotiques, les prébiotiques et les symbiotiques.

Recueil et analyse des données

Deux auteurs de la revue ont indépendamment évalué la qualité des essais et extrait les données. Nous avons contacté les auteurs des études afin d'obtenir des données manquantes. Nous avons évalué le biais pour chacune des études incluses à l'aide de l'évaluation des tableaux de biais dans le logiciel Cochrane Review Manager5. Nous avons effectué une méta-analyse, lorsque cela était indiqué, en utilisant le modèle à effets fixes de Mantel-Haenszel ou la méthode de la variance inverse à effets fixes présentant une hétérogénéité. Nous avons effectué des méta-analyses sur des essais évaluant la modification du régime alimentaire par rapport à un traitement standard pour la diarrhée à la fin de la radiothérapie et pour le changement de poids de l'inclusion à la fin de la radiothérapie.

Résultats principaux

Les recherches ont identifié 7 558 titres et nous en avons exclu 7 513 au cours de des recherches de titres et de résumés. Nous avons examiné 45 articles dans leur intégralité et exclu 39. Nous avons identifié quatre études à travers des recherches manuelles dans les références bibliographiques, qui, avec les six articles éligibles suite aux recherches dans les bases de données, ont permis d’inclure 10 études. Quatre études, dont trois étaient des ECR et une étude était prospective, ont étudié l'effet du régime élémentaire sur les symptômes GI; un ECR étudiait l'effet de modification du régime alimentaire et du régime élémentaire; et cinq ECR étudiaient la modification du régime alimentaire. Les études variaient en termes de risque de biais. Les données ont été dichotomisées pour la présence ou l’absence de diarrhée en fin de radiothérapie pour quatre essais évaluant la modification du régime alimentaire comprenant des graisses modifiées, du lactose, des fibres ou une combinaison de ces changements alimentaires. Une réduction de la diarrhée a été démontrée avec l'intervention nutritionnelle (risque relatif (RR) 0,66; intervalle de confiance (IC) à 95% de 0,51 à 0,87, quatre études, 413 participants, preuves de qualité modérée) avec une hétérogénéité faible (Chi 2 =3,50, I2 =14%). Deux essais évaluant les modifications du régime alimentaire sur le changement de poids (comparant l'inclusion et la fin de la radiothérapie) n’a montré aucune différence entre les groupes d'intervention ou le groupe témoin (différence moyenne (DM) -0,57 kg; IC à 95% - de 1,22 à 0,09) avec une hétérogénéité faible (Chi 2 =1,41, I2 =29%). En général, les effets indésirables étaient mal documentés dans les études incluses. Le régime élémentaire en particulier était mal toléré. Les symptômes GI ou la toxicité, 6 mois après la radiothérapie, n'ont pas été rapportés dans les études incluses.

Conclusions des auteurs

Des effets bénéfiques ont été associés à des modifications du régime alimentaire pendant la radiothérapie au niveau de la région pelvienne pour réduire la diarrhée. Les régimes inclus des interventions simples ou des combinaisons de lipides modifiés, de restriction de lactose, de restriction de graisses et de supplémentation en fibre. Nous n'avons pas pu faire de méta-analyse du régime élémentaire, les données n’étant pas disponibles. Nous avons considéré certaines de ces études présentant un risque de biais élevé. De récents progrès ont été effectués sur des techniques de radiothérapie plus ciblées, de telle sorte que les résultats des études plus anciennes doivent être interprétés avec prudence. De plus, des problèmes se sont avérés au niveau de l'observance et de la sapidité de certaines interventions, notamment pour le régime élémentaire, ce qui limite son utilité dans la pratique clinique.

アブストラクト

根治的骨盤放射線治療を受ける成人の消化器有害性を軽減するために実施する栄養療法

背景

先進国では、年間15~30万人が骨盤放射線治療を受けると推定され、治療中は80%に消化器(GI)症状が認められる。急性のGI症状がみられると、慢性で、しばしば消耗性のGI症状のリスクが大きくなる。最大1/3の患者に、骨盤放射線治療の開始前に栄養不良がみられ、最大4/5が治療中に体重が減少する。栄養不良とGI有害性のリスク上昇には関係が認められ、放射線治療の中断および化学療法の早期中止につながることもある。このことが主要な癌治療の有効性を損なうことになりかねない。

目的

根治的骨盤放射線治療を受ける成人を対象に、GI有害性の軽減を目的として実施される栄養療法の効果を評価すること。

検索戦略

2012年5月までのCochrane Gynaecological Cancer Group’s Trials Register、Cochrane Central Register of Controlled Trials(CENTRAL)の2012年第4号、MEDLINEおよびEMBASEを検索した。関連のある試験を幅広く検索するため、採用した研究や、検索で見つかったこれまでのシステマティック・レビューの引用文献リストをハンドサーチした。

選択基準

ランダム化比較試験(RCT)、あるいは非ランダム化並行群間比較試験(準ランダム化試験、クラスターRCT、非ランダム化試験、前向きおよび後ろ向きコホート研究および30例以上の症例集積など)とした。原発性骨盤腫瘍に対する治療の一部として、根治的骨盤放射線治療を受ける18歳以上の成人を対象に、栄養介入の効果を評価した研究を採用した。ストーマを持つ患者および炎症性腸疾患の既往のある患者を除外した。栄養サポートは、放射線治療の開始前または治療中であれば、介入のタイミングは問わなかった。介入は、栄養相談;食物繊維、乳糖または脂肪による食事改善;栄養補助の食品や飲料または栄養強化食品;ポリマー系、ペプチド系またはアミノ酸系のサプリメントや、新たな基質が追加されたサプリメントなど標準的な経口栄養剤;経腸チューブ栄養;非経口栄養(部分的または完全)などとした。プロバイオティクス、プレバイオティクスおよびシンバイオティクスは除外した。

データ収集と分析

2名のレビュー著者が、独立して試験の質を評価し、データを抽出した。不足のデータは、研究著者に問い合わせた。採用した各研究のバイアスは、Review Manager5(コクラン・ソフトウェア)のバイアス評価表を用いて評価した。メタアナリシスが必要な場合は、マンテル-ヘンツェルの固定効果法あるいは異質性を示す逆分散固定効果法を用いて行った。放射線治療終了時の下痢および放射線治療の開始前と終了時の体重の変化に関して、食事改善と標準治療とを比較した試験のメタアナリシスを実施した。

主な結果

検索で7558報の論文のタイトルを集め、タイトルと抄録の検索に基づいて7513報を除外した。45報の論文の全文をレビューし、39報を除外した。参考文献のハンドサーチから4件およびデータベースの検索から6件の計10件の適格研究を採用した。4件の研究(RCT 3件および前向き研究1件)が、GI症状に対する成分栄養剤の効果を検討していた。1件のRCTが食事改善および成分栄養剤の効果を検討していた。5件のRCTが食事改善の効果を検討していた。研究によって、バイアスのリスクの程度に差があった。放射線治療終了時の下痢の有無に関して、脂肪、乳糖または食物繊維の調整や、この食事変更のうちいくつかの組み合わせによる食事改善を評価した4件の試験では、結果がふたつに分かれた。栄養療法によって下痢の減少が示された(リスク比(RR)0.66;95%信頼区間(CI)0.51~0.87、研究4件、413例、エビデンスの質は中等度)。異質性は低い(Chi2 = 3.50, I2 = 14%)。体重の変化(放射線治療の開始前と終了時の比較)に対する食事改善の効果を評価した試験2件では、介入群とコントロール群に差はみられなかった(平均差(MD)-0.57kg;95%CI -1.22~0.09)。異質性は低い(Chi2 = 1.41, I2 = 29%)。採用した研究では、有害作用の報告は一般に不完全であった。成分栄養剤の忍容性が特に不良であった。放射線治療後 6カ月以降のGI症状または有害性に関しては、採用した研究はいずれも報告していない。

著者の結論

骨盤放射線治療中の食事改善が、下痢を低減させるのに利益があることが示されてきた。この食事改善は、脂肪の調整、乳糖の制限、脂肪の制限および食物繊維の補充のいずれか、あるいはこれらを組み合わせたものであった。データがなかったため、成分栄養剤のメタアナリシスは実施できなかった。一部の研究はバイアスのリスクが高いと判定した。最近、放射線治療の新たな技術が進歩し、精度の高いものになっているため、以前の研究の結果は慎重に解釈する必要がある。加えて、特に成分栄養剤をはじめとする一部の介入にはコンプライアンス(遵守)および食味の問題があり、診療における有用性を限られたものにしている。

訳注

《実施組織》厚生労働省「「統合医療」に係る情報発信等推進事業」(eJIM:http://www.ejim.ncgg.go.jp/)[2016.8.9]
《注意》この日本語訳は、臨床医、疫学研究者などによる翻訳のチェックを受けて公開していますが、訳語の間違いなどお気づきの点がございましたら、eJIM事務局までご連絡ください。なお、2013年6月からコクラン・ライブラリーのNew review, Updated reviewとも日単位で更新されています。eJIMでは最新版の日本語訳を掲載するよう努めておりますが、タイム・ラグが生じている場合もあります。ご利用に際しては、最新版(英語版)の内容をご確認ください。

Plain language summary

Do nutritional interventions reduce gastrointestinal toxicity in adults undergoing radical pelvic radiotherapy?

Background

Research has shown that nutrition and radiotherapy have effects on each other. People with malnutrition tend to get more bowel side effects during radiotherapy. It has also been shown that many people lose weight during radiotherapy due to the treatment side effects. These side effects can lead to some people needing gaps in their radiotherapy and sometimes they even need to stop it entirely. This can reduce the chance of cancer cure. This review looked at the literature for providing extra nourishment or changes in diet to patients before or during radiotherapy, to determine whether this is of benefit in terms of reducing bowel symptoms, improving nutritional status and quality of life.

Study characteristics

Searches of all relevant sources of medical information identified 7558 articles, and, after initial screening of all these articles, we selected 45 as being suitable for this review. On reading the summaries of these 45 trials, 10 were suitable to be included in this review. We included trials that looked at the effects of a nutritional intervention in adults aged 18 years or over having radical pelvic radiotherapy as part of anticancer treatment for a primary pelvic cancer. We excluded patients with stomas and a previous history of inflammatory bowel disease.

Key results

Results showed that studies evaluating dietary changes, including changes in dietary fat, fibre and lactose, during radiotherapy reduced diarrhoea at the end of treatment. However, these dietary changes did not seem to improve patients' weights.

Quality of the evidence

While some changes to diet during radiotherapy may benefit patients by reducing diarrhoea, it is important to recognise that some of the studies reviewed were dated and there have been changes in radiotherapy technique since then that may give patients less diarrhoea anyway. In addition, many studies were of poor quality and, therefore, only a small number were able to be included in this review. Some interventions described, for example, having to take a liquid diet, were not acceptable as many patients found them difficult to tolerate and were unable to take them completely.

Résumé simplifié

Les interventions nutritionnelles réduisent-elles la toxicité gastro-intestinale chez les adultes subissant une radiothérapie pelvienne radicale?

Contexte

La recherche a montré que la nutrition et la radiothérapie ont des effets intéractifs. Les personnes souffrant de malnutrition ont tendance à ressentir plus d’effets secondaires au niveau des intestins pendant la radiothérapie. Il a également été démontré que de nombreuses personnes perdaient du poids pendant la radiothérapie en raison des effets secondaires du traitement. Ces effets secondaires peuvent conduire certains patients à recevoir une radiothérapie par temps d’intervalles et parfois même recours à arrêter totalement. Cela peut réduire la chance de guérison du cancer. Cette revue a examiné l'alimentation supplémentaire apportée ou les modifications du régime alimentaire chez les patients avant ou pendant la radiothérapie, afin de déterminer si cela est bénéfique pour réduire les symptômes de l'intestin, améliorer l'état nutritionnel et la qualité de vie.

Les caractéristiques de l'étude

Des recherches dans toutes les sources d'information médicale ont identifié 7 558 articles. Après la sélection initiale de tous ces articles, nous en avons sélectionné 45 comme étant éligibles pour cette revue. Lors de la lecture des résumés de ces 45 essais, 10 ont été adaptés pour être inclus dans cette revue. Nous avons inclus les essais ayant examiné les effets d'une intervention nutritionnelle chez des adultes âgés de 18 ans ou plus et recevant une radiothérapie pelvienne radicale dans le cadre du traitement anticancéreux pour un cancer pelvien primaire. Nous avons exclu les patients atteints de stomies et possédant des antécédents d'une maladie intestinale inflammatoire.

Résultats principaux

Les résultats ont montré que les études évaluant les changements alimentaires pendant la radiothérapie, notamment en matières grasses, en fibres et en lactose, ont réduit la diarrhée à la fin du traitement. Cependant, ces changements alimentaires n'ont pas semblé améliorer le poids des patients.

Qualité des preuves

Bien que certaines modifications alimentaires pendant la radiothérapie puissent être bénéfiques chez les patients en réduisant la diarrhée, il est important de noter que certaines des études examinées étaient anciennes et certaines techniques de radiothérapie ont changé, ce qui pourrait également entrainer une baisse de la diarrhée. De plus, de nombreuses études étaient de mauvaise qualité, seulement un petit nombre a donc pu être inclus dans cette revue. Certaines interventions décrivant, par exemple, devoir suivre un régime alimentaire liquide, n'étaient pas acceptables car de nombreux patients avaient des difficultés à le tolérer et ne parvenaient pas à le suivre entièrement.

Notes de traduction

Traduit par: French Cochrane Centre 14th January, 2014
Traduction financée par: Minist�re Fran�ais des Affaires sociales et de la Sant�, Instituts de Recherche en Sant� du Canada, Minist�re de la Sant� et des Services Sociaux du Qu�bec, Fonds de recherche du Qu�bec Sant� et Institut National d'Excellence en Sant� et en Services Sociaux

平易な要約

根治的骨盤放射線治療を受ける成人の消化器有害性に対する栄養療法の緩和効果

背景

栄養と放射線治療が互いに影響を及ぼし合うことが研究によって明らかになっている。栄養不良の人の方が、放射線治療中は腸の副作用の頻度が高い傾向にある。さらに、放射線治療の副作用によって、治療中に体重が減少する人が多いことも明らかになっている。このような副作用によって、放射線治療を一時休止する人や、中には放射線治療を完全にやめなければならない人もいる。これでは、癌の治癒の可能性を妨げることになりかねない。このレビューは、文献を調べ、放射線治療の開始前または治療中の患者に対する栄養補充や食事改善が、腸の症状の軽減および栄養状態やQOLの改善の点で利益があるかどうかを検討した。

研究の特性

関連のある医学情報を検索して集めた7558報の論文をすべて審査し、今回のレビューに適した論文45報を選別した。その45件の試験の要約を読み、適格であった10件を採用した。原発性骨盤腫瘍に対する抗癌治療の一部として、根治的骨盤放射線治療を受ける18歳以上の成人を対象に栄養療法入の効果を評価した試験を採用した。ストーマを持つ患者および炎症性腸疾患の既往のある患者を除外した。

主な結果

放射線治療中の脂肪、食物繊維または乳糖の変更をはじめとする食事改善を評価した研究では、治療終了時に下痢が低減したことが明らかになった。しかし、このような食事改善によって患者の体重は改善しないように思われた。

エビデンスの質

放射線治療中の食事改善には、下痢を減らすことによって患者に利益をもたらすものもあるが、評価した研究のなかには古いものもあり、介入とは関係なく、それ以降の放射線治療技術の進歩によって下痢の頻度が下がった可能性もある。さらに、多くの研究が質の低いもので、採用することができた研究はわずかであった。例えば流動食など、ここに記載した栄養療法のなかには、忍容性が低く、多くの患者が最後まで続けることができなかったものもあった。

訳注

《実施組織》厚生労働省「「統合医療」に係る情報発信等推進事業」(eJIM:http://www.ejim.ncgg.go.jp/)[2016.8.9]
《注意》この日本語訳は、臨床医、疫学研究者などによる翻訳のチェックを受けて公開していますが、訳語の間違いなどお気づきの点がございましたら、eJIM事務局までご連絡ください。なお、2013年6月からコクラン・ライブラリーのNew review, Updated reviewとも日単位で更新されています。eJIMでは最新版の日本語訳を掲載するよう努めておりますが、タイム・ラグが生じている場合もあります。ご利用に際しては、最新版(英語版)の内容をご確認ください。

Summary of findings(Explanation)

Summary of findings for the main comparison. Dietary modification compared with no nutritional intervention for adults undergoing pelvic radiotherapy
Dietary modification compared with no nutritional intervention for adults undergoing pelvic radiotherapy

Patient or population: adults with pelvic cancer undergoing radiotherapy

Settings: acute

Intervention: dietary modification

Comparison: no nutritional intervention/standard care

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
No nutritional intervention/standard careDietary modification

Diarrhoea

at end of treatment

Low risk population RR 0.66 (0.51 to 0.87)413
(4)
Moderate4 studies with low risk of selection bias and reporting bias but all had incomplete outcome data and, given the nature of the intervention, none were blinded. Moderate given as an overall statement
66 per 1000 43 per 1000
(34 to 57)

Change in weight

(kg)

comparing end of treatment with baseline

The mean weight change ranged across control groups from
-1.7 kg to -0.6 kg
The mean weight change in the intervention groups ranged from
-2.6 kg to -0.7 kg
-235
(2)
Moderate2 studies with low risk of selection bias and reporting bias but both had incomplete outcome data and, given the nature of the intervention, neither were blinded. Moderate given as an overall statement
*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.

Background

Description of the condition

Radiotherapy-related gastrointestinal (GI) disease is increasingly recognised as an important yet under-researched area. Seventeen thousand patients per year are treated with pelvic radiotherapy in the UK (West 2009). Across the developed world, an estimated 150,000 to 300,000 people are treated annually (Andreyev 2005; Hauer-Jensen 2003). Around 11% to 33% of patients are malnourished before pelvic radiotherapy, and up to 83% of people lose weight during treatment (McGough 2004). Eighty per cent of people will develop GI symptoms during treatment (Wedlake 2008), and 50% of patients will develop chronic GI symptoms that adversely affect quality of life (QoL) (Andreyev 2007; Gami 2003). Common GI symptoms experienced during and after pelvic radiotherapy include increased frequency of defecation, urgency to open bowels, alteration in stool consistency, change in bowel habit, pain, bleeding, bloating and nausea. In addition, chemotherapy is used increasingly alongside pelvic radiotherapy in curative treatment and acute GI symptoms are increased in this context (Coles 2003; Kirwan 2003).

Nutritional status has a complex relationship with radiotherapy. Pelvic radiotherapy can cause injury to the GI tract, which may compromise the nutritional status of patients through changes in the structure and function of the small and large bowel (Bentzen 2006; Dörr 2009). It is also increasingly recognised that acute GI symptoms can lead to debilitating chronic GI symptoms. This is referred to as consequential late effects (Bourne 1983; Jereczek-Fossa 2002; Jereczek-Fossa 2010; Wang 1998; Wedlake 2010; Weiss 1999). Nutritional intervention during pelvic radiotherapy could therefore be of benefit in improving nutritional status and reducing GI side effects in both the short and long term.

While the effect of radiation on the GI tract can lead to nutritional problems, equally nutritional status influences the tolerance and therefore potentially the completion of anticancer treatment. Malnutrition is associated with a higher risk of GI toxicity (Theis 2010), which is often responsible for breaks in radiotherapy and early cessation of chemotherapy (Coles 2003). Overall duration of treatment is important in radiotherapy (Fyles 1992), and interruptions in treatment are associated with tumour recurrences (RCR 2008). In addition, the number of cycles of chemotherapy received is associated with progression-free survival and overall survival (Peters 2000). It has been suggested that nutritional intervention during radiotherapy could be of benefit in reducing GI side effects and improving nutritional status in both the short and long term.

Description of the intervention

Nutrition support interventions may be provided at any stage before or during pelvic radiotherapy.

Nutrition support may be one or more interventions comprising dietary counselling; dietary modification of fibre, lactose or fat; supplementary foods or drinks or fortified foods including supplementary fibre; standard oral nutritional supplements (ONS) and total or partial parenteral nutrition. Standard ONS include polymeric-, peptide- or amino acid-based supplements and those where novel substrates have been added such as glutamine, fish oils, arginine or antioxidants. Only nutrient-based novel substrates will be included in this review. We excluded prebiotics, probiotics and synbiotics as they are the subject of a separate Cochrane review.

How the intervention might work

Nutritional status has a complex relationship with radiotherapy. Poor nutritional status can be caused by the effect of radiation on the gut, but also poor nutritional status is a risk factor for GI side effects of radiotherapy. Nutritional interventions therefore may directly reduce GI symptoms or toxicity in both the short and long term or may indirectly reduce GI symptoms or toxicity by improving nutritional status.

Nutritional intervention may increase the likelihood of completing the anticancer treatment as planned, which is linked to improved survival in certain pelvic cancers, by reducing GI side effects. The intervention may also improve QoL and patient-reported outcomes.

Why it is important to do this review

Although there have been some studies of nutritional interventions to reduce symptoms of radiation injury to the gut, their role and impact has not been formally evaluated. A systematic review of studies at low risk of bias on the use of nutritional interventions during pelvic radiotherapy is of benefit to determine which interventions are effective at reducing radiation-related GI symptoms in the short and long term; improving nutritional status; improving anticancer treatment completion rates and survival; and improving QoL. Furthermore, it is important to determine whether nutritional support is effective in improving clinical outcome (GI symptoms, ability to complete planned anticancer treatment, reduction in the rate of unplanned inpatient stays, rates of recurrence, survival, mortality), nutritional status or QoL measures in patients undergoing pelvic radiotherapy.

As malnutrition has been shown to be a significant risk in patients with cancer in terms of outcome and treatment toxicity and therefore requires purposeful identification and treatment, there is a need to develop practical guidance on the use of nutrition support derived from a systematic review of high-quality studies undertaken in this patient group. 

Objectives

To assess the effects of nutritional interventions for reducing GI toxicity in adults undergoing radical pelvic radiotherapy.

Primary objectives

  1. To assess the effects of nutritional intervention on GI symptoms or toxicity in adults undergoing radical pelvic radiotherapy.

  2. To assess the effects of nutritional intervention on nutritional status in adults undergoing radical pelvic radiotherapy.

Secondary objectives

  1. To assess the effects of nutritional intervention on QoL in adults undergoing radical pelvic radiotherapy.

  2. To assess the effects of nutritional intervention on completion of planned anticancer treatment (including duration of treatment) in adults undergoing radical pelvic radiotherapy.

  3. To assess the effects of nutritional intervention on unplanned hospital stay, cancer recurrence and mortality in adults undergoing radical pelvic radiotherapy.

  4. To identify adverse effects of nutritional interventions.

  5. To make comparisons between different types of nutritional interventions.

Methods

Criteria for considering studies for this review

Types of studies

  • Randomised controlled trials (RCT) including conference abstracts of RCTs where sufficient data could be obtained.

As we expected to find few RCTs of nutritional interventions, we also included non-randomised studies with concurrent comparison groups:

  • Quasi-randomised trials, cluster RCTs, non-randomised trials, prospective and retrospective cohort studies, and case series of 30 or more patients.

We excluded case-control studies and case series of fewer than 30 patients. We excluded cross-over trials as GI symptoms typically worsen throughout treatment. A cross-over designed trial is, therefore, not appropriate in this patient group.

We excluded all other types of study.

Types of participants

Adults aged 18 years or over undergoing radical pelvic radiotherapy (external beam radiotherapy, brachytherapy, or both) as part of anticancer treatment for a primary pelvic malignancy, including gynaecological (cervix or uterus), lower GI (rectal or anal) and urological (prostate or bladder) malignancies. We excluded patients with stomas and a previous history of inflammatory bowel disease.

Types of interventions

Nutritional support interventions may be provided at any stage before or during pelvic radiotherapy.

Nutritional support may be one or more of the following:

  • dietary counselling;

  • dietary modification of non-starch polysaccharides (NSP) (dietary fibre), including supplementation with psyllium, metamucil ispaghula, plantago ovata;

  • dietary lactose modification;

  • dietary fat modification;

  • supplementary foods or drinks or fortified foods;

  • standard oral nutrition supplements including polymeric-, peptide- or amino acid-based supplements and those where novel substrates have been added such as glutamine, fish oils, arginine or antioxidants. Only nutrient-based novel substrates will be included;

  • enteral tube feeds;

  • parenteral nutrition (partial or total).

The nutritional support may be supplementary to standard food and drink intake or the sole source of nourishment. Trials may examine the provision of nutritional support in comparison with not providing nutritional support or may be comparisons of alternative types of nutritional support. We will ensure that the pooling of data for the ONS interventions is appropriate and will not combine ONS that are different in composition including in terms of novel substrates. This potentially involves multiple comparisons. This review analysed those comparisons studied to date, but in future this review may be split to analyse each individual nutritional intervention compared with placebo or standard care.

Types of outcome measures

Primary outcomes
  1. GI symptoms or toxicity within six months of the start of radiotherapy as determined by validated GI toxicity scoring, patient-reported outcome measures or other symptom-based questionnaires/interviews.

  2. GI symptoms or toxicity six months or more after the start of radiotherapy determined by validated GI toxicity scoring, patient-reported outcome measures or other symptom-based questionnaires/interviews.

  3. Nutrition status as determined by weight, body mass index (BMI), anthropometric measures, subjective global assessment, biochemical measures (e.g. prealbumin, transferrin, retinol-binding protein, urinary nitrogen balance), functional measures (change in performance status, hand-grip strength and sit to stand time) or other validated nutrition assessment methods. We will assess this at the study end point or at a maximum of 12 months.

Secondary outcomes
  1. QoL measures including patient-reported outcome measures and questionnaires/interviews.

  2. Completion and duration of planned anticancer treatment.

  3. Adverse effects and compliance with nutritional intervention including those relating to nutritional product or route of delivery. Method of recording these was noted. This was assessed at the study end point or at a maximum of 12 months.

  4. Unplanned hospital stay within three months of starting radiotherapy.

  5. Recurrence of cancer. This was assessed at the study end point or at a maximum of 12 months.

  6. Mortality, specifically cancer-related and overall mortality. This was assessed at the study end point or at a maximum of 12 months.

Search methods for identification of studies

Electronic searches

We searched the following electronic databases:

  • The Cochrane Gynaecological Cancer Collaborative Review Group's Trial Register;

  • the Cochrane Central Register of Controlled Trials (CENTRAL), Issue 4, 2012;

  • MEDLINE to May 2012;

  • EMBASE to May 2012.

The MEDLINE, EMBASE and CENTRAL search strategies are presented in Appendix 1, Appendix 2 and Appendix 3. All relevant articles were identified on PubMed and we used the 'related articles' feature to carry out a further search for newly published articles. We sought reports in all languages and carried out translations if necessary.

Searching other resources

Unpublished and grey literature

We searchedm etaRegister (www.controlled-trials.com), Physicians Data query (nci.nih.gov), www.clinicaltrials.gov and www.cancer.gov/clinicaltrials for ongoing trials. If we had identified ongoing trials that had not been published through these searches, we would have approached the principal investigators to request relevant data. We searched conference proceedings and abstracts through ZETOC (zetoc.mimas.ac.uk) and WorldCat Dissertations.

 Reference lists

We handsearched the citation lists of included studies and previous systematic reviews identified to identify further relevant trials.

Data collection and analysis

Selection of studies

Two review authors independently assessed all titles and abstracts retrieved by electronic searching to determine relevance and eligibility. We discarded reports that did not meet the eligibility criteria. Where there was insufficient information to make a decision based on the abstract and title, the full article was obtained in order to make a decision. We obtained copies of the full text of potentially relevant references and two review authors independently reviewed them to ensure they meet the eligibility criteria with diversity of opinion being resolved by discussion between the two review authors and with recourse to a third review author if necessary. We documented reasons for exclusion of studies. We linked multiple reports of the same study to ensure that no data were included in the meta-analysis more than once.

Data extraction and management

We devised a data collection form for the study to facilitate the data collection from the included studies. We piloted and modified the data form as required. Two review authors undertook the process of data extraction independently with discrepancies discussed between themselves and a third review author if required. We recorded the following information for each trial:

  • year of publication, country of origin and source of funding;

  • patient details, number of participants, gender, age, inclusion and exclusion criteria;

  • details of nutritional status of patients including the proportion of malnourished patients including BMI less than 20 kg/m2, or weight loss greater than 10% in the previous three to six months, subjective global assessment or nutrition risk derived from a validated tool; functional measures (change in performance status, hand-grip strength, sit to stand time) and biochemical measures (including prealbumin, transferrin, retinol-binding protein, urinary nitrogen balance);

  • cancer diagnosis including staging if indicated, type of surgery, presence of co morbidities;

  • details of cancer treatment (including dose, volume, fractionation for external beam radiotherapy and details of brachytherapy, chemotherapy and surgery);

  • details of cancer treatment delivery (including completion of planned treatment, unplanned hospital stay, breaks to radiotherapy, total duration of radiotherapy and number of cycles of chemotherapy);

  • details of nutrition intervention (including type of food or formulation, route of intervention, length of intervention, quantity delivered and adverse effects);

  • details of GI symptoms (including toxicity scores, patient-reported outcome measure score or other symptom-based questionnaires or interviews scores);

  • details of QoL measures including patient-reported outcomes;

  • details of pharmacological confounders (angiotensin-converting enzyme (ACE) inhibitors, statins);

  • details of clinical outcome including unplanned hospital stay, recurrence of cancer, survival and mortality.

Details of primary and secondary outcomes collected included the time points at which they were collected and reported.

Assessment of risk of bias in included studies

We assessed the risk of bias in included studies using The Cochrane Collaboration's tool (Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011)). We rated the risk of bias for each study in the following areas:

  • random sequence generation;

  • allocation concealment;

  • blinding;

  • complete outcome data;

  • selective outcome reporting;

  • other sources of bias.

Measures of treatment effect

We used the following measures of the effect of treatment: for dichotomous outcomes, we used the risk ratio (RR) and for continuous outcomes, we used the mean difference (MD) between treatment arms. We used standardised mean difference (SMD) to compare studies in which the outcomes were measured on different scales, for example GI symptoms and toxicity scores.

Unit of analysis issues

Unit of analysis issues could arise due to multiple observations for the same outcome, for example repeated measurements of GI symptoms or toxicity over time. We looked at short- and long-term GI symptoms or toxicity, as defined above in the outcomes section, as these were clinically significant end points. If there were still multiple time points within these two discrete categories, this was addressed either by obtaining individual participant data to determine the individual's trend over time or by selecting a single time point within the two time categories for analysis.

We handled studies with multiple control groups by combining all relevant control groups into one group for comparison with the intervention. We combined studies with multiple intervention groups into one group for comparison with the control group if possible. If this was not possible, then we performed a pair-wise comparison with shared intervention or control groups divided out approximately evenly among the comparisons. For dichotomous outcomes, we divided up both the number of events and the total number of patients. For continuous outcomes, we only divided up the total number of participants and left the means and standard deviations (SD) unchanged.

Dealing with missing data

In studies where we identified missing data, we contacted the study authors to seek relevant information. We imputed no missing data. We performed available case analysis.

Assessment of heterogeneity

We assessed clinical heterogeneity by examining the type of participant, interventions and outcomes in each study. We only conducted meta-analyses if there were studies reporting similar comparisons for the outcome measures. We assessed statistical heterogeneity using the Chi2 test to determine whether the observed differences in results were compatible with chance alone, and using the I2 statistic to determine the impact of heterogeneity on the meta-analysis.

Assessment of reporting biases

We used funnel plots to make a visual assessment of whether small-study effects may be present. We considered publication bias as one of a number of possible explanations for small-study effects.

Data synthesis

We only conducted meta-analyses if studies reported similar comparisons for the same outcomes. We did not directly compare studies using brachytherapy alone versus studies using external beam radiotherapy, as the pattern of GI symptoms is different.

We performed meta-analyses using the Mantel-Haenszel fixed-effect method. We used the generic inverse variance method for dichotomous data if the published paper reported only odds ratio or RR and its standard error and for continuous data if the published paper reported only the difference between the means for the two groups and the standard error of this difference and raw data could not be obtained from the authors. We also used the generic inverse variance to facilitate the analysis of cluster randomised and non-randomised trials if appropriate. We only combined RCTs and non-randomised studies in a meta-analysis if the non-randomised studies were properly analysed and at low risk of bias.

Subgroup analysis and investigation of heterogeneity

We undertook subgroup analysis, if data allowed, on studies as follows:

  • comparison by baseline nutritional status (malnourished versus nourished participants);

  • comparisons by primary tumour;

  • comparisons by type of radiotherapy, including external beam conformal radiotherapy, intensity-modulated radiotherapy (IMRT), brachytherapy and combinations of these;

  • comparisons by treatment modalities including external beam radiotherapy, brachytherapy, chemotherapy, surgery and combinations of these.

Sensitivity analysis

We performed sensitivity analysis if relevant issues were identified during the review process. If sensitivity analyses identified particular decisions or missing information that greatly influenced the findings of the review, we attempted to resolve uncertainties and obtain extra information through contacting trial authors and obtaining individual patient data. If this could not be achieved, we interpreted the results with an appropriate degree of caution. We excluded studies at high risk of bias from meta-analysis as part of sensitivity analysis.

Results

Description of studies

We included nine RCTs and one prospective uncontrolled study in the meta-analysis.

See Characteristics of included studies and Characteristics of excluded studies tables.

Results of the search

The searches identified 7558 titles, of which 7513 were excluded during title and abstract searches. We reviewed 45 papers in full, and excluded 39. We identified four studies on handsearching of the references, which, along with the six eligible papers from the database search, led to 10 studies being included. Four studies, three of which were RCTs and one prospective study, investigated the effect of elemental diet on GI symptoms; one RCT investigated the effect of dietary modification and elemental diet; and five RCTs investigated dietary modification. The results of the search are shown in Figure 1 and Table 1.

Figure 1.

Flow chart of study selection

Table 1. Results from search strategies from electronic searches
 Number of studies
DatabaseInitial searchTitle reviewAbstract review
MEDLINE26942316
EMBASE42132611
CENTRAL5782816
Systematic reviews7332
Total75588045

Included studies

Four studies, three of which were RCTs (Brown 1980; Capirci 2000; McGough 2008), and one prospective uncontrolled study (Craighead 1998), investigated the effect of elemental diet on GI symptoms; one RCT investigated the effect of elemental diet and dietary modification; and five RCTs investigated dietary modification with a lactose restricted diet (Stryker 1986), a low or modified fat diet (Wedlake 2012), a low fat and low lactose diet (Bye 1992), fibre modification (Murphy 2000), and lactose and fibre modification (Pettersson 2012).

Methods of measuring nutritional status in each study are summarised in Table 2. Methods of measuring GI symptoms and QoL in each study are summarised in Table 3. The nutritional intervention used in each study is summarised in Table 4. The comparison of the timing of assessments in each study is shown in Table 5.

Table 2. Method of assessing nutritional status used in each included study
  1. BMI: body mass index.

Study ID Nutritional status measurement tool
Weight BMI Grip strength
Brown 1980X - -
Bye 1992X - -
Capirci 1993X--
Capirci 2000X - -
Craighead 1998X - -
McGough 2008XX -
Murphy 2000 - - -
Pettersson 2012 - - -
Stryker 1986 - - -
Wedlake 2012X -X
Table 3. Method of assessing gastrointestinal symptoms and quality of life used in each included study
  1. *Gastrointestinal (GI) symptom tools that give clear definition and classification of diarrhoea in terms of stool frequency and stool consistency.

    EORTC-QLQ-C30: European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC-QLQ-PR25: European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire - prostate-specific module; GISEQ: Gastrointestinal Side Effects Questionnaire; IBDQ: Inflammatory Bowel Disease Questionnaire; IBDQ-B: Inflammatory Bowel Disease Questionnaire-Bowel subset score; MDS: Murphy Diarrhoea Scale; RTOG: Radiation Therapy Oncology Group.

Study ID GI symptom/Quality of Life Measurement Tool
EORTC-QLQ-PR25 EORTC-QLQ-C30 GISEQ RTOG Diarrhoea Scale* MDS* Stool frequency Duration of diarrhoea IBDQ* IBDQ-B* Vaizey
Brown 1980 - - - - -X - - - -
Bye 1992 -X - -X - - - - - -
Capirci 1993---X-------
Capirci 2000 - - -X - - - - - - -
Craighead 1998 - - -X - - -X - - -
McGough 2008 - - -X - - - - -XX
Murphy 2000 - - - -X - - - - -
Pettersson 2012XXX - - - - - - - -
Stryker 1986 - - - - - -X - - - -
Wedlake 2012 - - - - - - - -XXX
Table 4. Nutritional interventions used in each study
Study ID Nutritional intervention
  Elemental diet Lactose modification Fat modification Fibre modification Computerised diet
Brown 1980X - - -  -
Bye 1992 -XX -  -
Capirci 1993X - - -X
Capirci 2000X - - -  -
Craighead 1998X - - -  -
McGough 2008X - - -  -
Murphy 2000 - - -X  -
Pettersson 2012 -X -X  -
Stryker 1986 -X -  -
Wedlake 2012 -X -  -
Table 5. Timing of assessment point for each study
  1. *measured from start of radiotherapy unless otherwise specified.

    EBRT: external beam radiotherapy; RT: radiotherapy.

Study ID Pre-EBRT During EBRT Post-EBRT Unclear
Baseline Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Weekly during treatment Pre and post EBRT Summary of symptoms during RT Week 8 Week 10 Week 12 2 months after end of RT 12 Months
Brown 1980 - - - - - -X - - - - - - - -
Bye 1992X - - - - -X - - - - -X - - -
Capirci 1993 - - - - - - - - - - - - - - -X
Capirci 2000 - - - - - - -XX - - - - - -
Craighead 1998 - - - - - - - - -X - - - - - -
McGough 2008X - -X -X - - - -X - - - - -
Murphy 2000 - - - - - - - - -X - - - - - -
Pettersson 2012X - - -X - - - - -X - -X - -
Stryker 1986 -XXXXX -X - - - - - - - -
Wedlake 2012X - - -X - - - - - - - - -X -
Elemental diet

McGough 2008

One single-centre parallel group unblinded RCT of 50 patients (29 women and 21 men) with histologically confirmed gynaecological, urological or lower GI malignancy was performed to determine the feasibility and efficacy of replacing one-third of the normal diet with elemental diet during the first three weeks of radiotherapy. The radiotherapy technique used was predominantly a three-field technique with a fraction size of 1.8 to 2.0 Gy per fraction. The primary outcome was GI symptoms at week five using the Inflammatory Bowel Disease Questionnaire Bowel subset (IBDQB) as a measure. Other outcomes were GI symptoms using the Vaizey Incontinence Questionnaire (VIQ), Radiation Therapy Oncology Group (RTOG) toxicity grade and nutritional status (weight and BMI). Assessments were made at baseline (immediately before the start of radiotherapy), week three, week five and week 10. Faecal calprotectin was also measured, but this was not a relevant outcome for this review.

Patients were randomised to either intervention (25 people) or no intervention (25 people). The no intervention group continued to take their normal diet throughout radiotherapy and the intervention group were asked to replace one meal per day (33% of daily calories) with elemental diet in the form of E028 Extra ready to drink cartons and E028 Extra flavoured powder sachet (SHS International, Liverpool, UK). Compliance was monitored with a diary, pharmacy dispensing records and counts of patients returned unused sachets. Patients in the intervention group were encouraged not to snack for one to two hours either side of the elemental meal.

The intervention group contained a slightly larger proportion of gynaecological cancer patients (44%) and the control arm had a higher proportion of urological cancer patients (36%). The median intake of elemental diet was 571 mL (range 71 to 1000) in week one, 500 mL (range 0 to 1000) in week two and 500 mL (range 0 to 1000) in week three. There was no significant decrease over time (P value > 0.1) but, unfortunately, this was lower than the prescribed volume required to provide 33% of caloric requirement. Six patients were non-compliant.

There were no differences in GI symptoms found at baseline using IBDQ, IBDQB, VIQ and RTOG between the two groups. GI symptoms increased between baseline and both weeks three and five (P value < 0.001) in both groups but there was no difference between groups at week three (P value = 0.214).

For the week 10 follow-up assessment, three patients were missing with no reason given. GI symptoms for both groups had improved compared with week five using IBDQ, IBDQB and RTOG (P value < 0.001). VIQ scores improved for patients in the intervention group comparing week 10 versus week five (P value < 0.001) but not in the control group (P value = 0.06). There was no difference between GI toxicity data comparing compliant and non-compliant patients on post-hoc analysis.

The authors reported that nutritional status was comparable between both groups at baseline and did not change between baseline and other time points, although actual data for this were not published.

Overall there was poor compliance with the intervention (6/25, 24%) and no improvement was seen in terms of GI symptoms or nutritional status. Total IBDQ score was used as a QoL measure. Furthermore, this was a clinically heterogeneous group of patients with different baseline characteristics in terms of primary tumour and, therefore, treatment regimen. This may have confounded the results. Elemental diet was poorly tolerated.

Brown 1980

One single-centre parallel group unblinded RCT was performed to compare reduced fibre diet (control) versus three sachets of Vivonex HN (Eaton Laboratories Ltd, Woking, UK) daily along with a reduced fibre diet (intervention). The Vivonex HN supplementation, if complied with, provided 900 kcal and 40g of mixed amino acids daily. Dietary changes were implemented from day one of radiotherapy. There was no description of who gave the reduced fibre diet instruction or what the level of fibre intake was or how this was standardised. Data for 68 patients were reported, although it was unclear if this is the total number entered into the study. Seventeen patients were in the control group, with 30 patients who adhered to the Vivonex HN diet and 21 who did not adhere to the full period of supplementation. The radiotherapy technique used was not clearly described, but patients received 4000 to 6000 rads in 20 to 24 fractions.

Biochemical and metabolic data from this study were reported in a separate publication (Foster 1980), but this was not an outcome for our review. This paper reported weight and stool frequency. Weight was measured weekly and patients recorded stool pattern during treatment. The endpoint of this study was not clear. Twenty-one patients did not adhere to the Vivonex HN diet reporting that it was time-consuming to ingest (six people), gave an unacceptable degree of bloating (six people), potentiated nausea (14 people), was unpalatable (12 people) and induced vomiting (six people). The mean stool frequency was four per day for all three groups (i.e. controls versus Vivonex HN adherents versus Vivonex HN non-adherents). It was not clear how this was determined (i.e. whether this represents the mean stool frequency over the whole of treatment or at a particular time point). No measure of variance was given as either SD or standard error. Mean weight loss was 1.6 kg, 0.5 kg and 1.7 kg for the control, Vivonex HN adherent and Vivonex HN non-adherent groups, respectively. Again no measure of variance was given. The authors reported that this difference was not significant, but the exact statistical test used and the P values were not reported.

Participants had mainly urological tumours, but this study also contained patients with many other types of primary tumours, thus creating a clinically heterogeneous group. The study was underpowered and only means were reported for stool frequency with no measurement of variance given. There was no benefit of elemental diet in terms of stool frequency or change in weight. Elemental diet was poorly tolerated with 41% (21/51) of the reported patients being non-adherent with the intervention. QoL was not assessed.

Capirci 2000

One conference abstract of a multicentre randomised study evaluating the efficacy of "natural diet" plus elemental diet compared with standard diet on acute enteric toxicity during pelvic radiotherapy was published. Although this abstract was published in 2000, there is no full report in the public domain and there was no response from the authors when contacted. Therefore, available data were limited.

The abstract reported that 677 patients were recruited into this trial with 332 patients randomised into the intervention group (natural diet plus elemental diet) and 345 patients randomised into the control group. Most patients had a primary rectal cancer (439 people), with 228 patients having primary uterine cancer and 10 patients with prostate cancer. The cases were intracentre stratified for amount of irradiated small bowel and concomitant chemotherapy as both these variables correlate with incidence of radiation-related diarrhoea. The outcomes were RTOG score and change in weight comparing pre- and post-radiotherapy weight. It was not clear which elemental diet was used and in what quantities, what the "natural diet" comprised and when the elemental diet was taken. Radiotherapy technique was not reported, neither were the number of patients treated with chemotherapy.

The authors reported that in the intervention group, 17% of patients had grade 1 toxicity, 12% had grade 2 toxicity and 1% had grade 3-4 toxicity, whereas in the control group, 25%, 27% and 4% had grade 1, grade 2 and grades 3-4 toxicity, respectively. The authors also reported that in the intervention group patients with grade 1 toxicity put on 1 kg of weight, those with grade 2 toxicity had no change in weight and those with grade 3-4 toxicity lost 5.5 kg of weight. For the control group, patients with grade 1 toxicity had no weight change, those with grade 2 toxicity lost 1.3 kg of weight and those with grade 3-4 toxicity lost 4 kg of weight. Twelve versus 44 patients required a break in radiotherapy due to GI toxicity in the intervention and control group, respectively. The authors reported that the toxicity reduction remained when data were stratified for concomitant chemotherapy or toxicity grade.

The total numbers with each grade of diarrhoea were not reported. It was not clear whether change in weight reported was mean change in weight and no measure of variance was reported. The authors reported that the differences between the intervention and control group were significant, but the statistical methods used were not described and the P values have not been published.

This was a clinically heterogeneous group. Given the large numbers of patients, it is surprising that subgroup analysis was not performed for rectal cancer and uterine cancer patients. RTOG grade is not validated and is a coarse tool that is known to underestimate toxicity. While the reported results suggest that natural diet with elemental diet reduces radiation enteritis and weight loss during treatment, the lack of reported data makes further analysis difficult.

Craighead 1998

One single-centre uncontrolled phase II prospective cohort study was performed to assess the feasibility of using elemental supplements to reduce acute enteritis in patients receiving radical pelvic radiotherapy and to assess compliance in primary cervical and endometrial cancer patients. Seventeen patients were selected to have elemental supplements (Vital HNR, Ross Products Division, Abbott Laboratories, Illinois) and a cohort of 45 patients was used to determine the baseline data for radiation enteritis. Both cohorts had a restricted diet which comprised restricted lactose, low fibre (12g daily), moderate fat intake (< 30% calories from fat) with adequate protein and carbohydrate. Intake of fruits, caffeine and other bowel stimulants was restricted. The authors did not document who delivered the information regarding the restricted diet. Patients in the intervention cohort had two or three servings of Vital HNR, each of which comprised 39g of Vital HNR powder mixed with 250 mL of liquid, which provided between 1050 and 1260 J. The treatment period was from three days before the start of radiotherapy to the last day of radiotherapy. Radiotherapy was delivered using a standard four-field arrangement using 10 mV or 15 mV photons to the whole pelvis to a dose of 4500 cGy.

Compliance was assessed using patient diaries and a sachet count at the end of radiotherapy. Bowel function and weight were recorded at baseline, weekly during radiotherapy and at 12 months. Bowel function was assessed using the RTOG grade for acute lower GI enteritis during treatment as assessed by a clinical oncologist. The presence of diarrhoea was assessed at 12 months. Duration of diarrhoea during treatment was also calculated in days. The study was powered to assess compliance not bowel function.

Most patients in the intervention cohort were having adjuvant radiotherapy following surgery for endometrial cancer (nine people). Two patients had adjuvant radiotherapy following surgery for cervical cancer. Six patients were treated with radical radiotherapy for cervical cancer and two of these had intracavity treatment. Fourteen patients reported normal stool consistency at baseline. Baseline data were not published for the control cohort. Thirteen patients complied with the enteral diet as defined by taking two or three sachets of Vital HNR daily at least 80% of the time. Four patients did not comply due to nausea.

The authors combined the non-compliant intervention group patients with the non-intervention cohort and found a greater proportion of patients in this new group had RTOG grade 2/3 diarrhoea as compared the compliant intervention cohort (55% versus 15%, Chi2 = 0.01, P value < 0.001). Mean (SD) duration of diarrhoea during treatment was 5.85 days (4.44) for the intervention compliant cohort, 18.75 days (3.48) in the intervention non-compliant cohort and 12.2 days (6.95) in the comparison cohort. At 12 months, no patients in the intervention compliant cohort had diarrhoea, all patients in the intervention non-compliant group had diarrhoea and 19/45 patients in the control group had diarrhoea. It is not reported how this was assessed and "diarrhoea" was not defined. Weight data for the control cohort were not reported, but intervention compliant group (13 people) gained a mean 1.19 kg and the intervention non-compliant group (four people) lost a mean 1.80 kg of weight. SDs or standard errors were not reported.

The authors concluded that there was acceptable compliance with the elemental diet with dietician support and that this was associated with weight gain and no diarrhoea at 12 months. The study was small and not powered for GI symptoms or weight. In addition, the study was neither randomised nor controlled. Diarrhoea was not clearly defined. QoL was not assessed. RTOG is a non-validated and coarse tool for detecting GI symptoms and has been noted to under-report these symptoms. It was not clear whether there were any differences between baseline characteristics of the cohorts.

Elemental diet with dietary modification

Capirci 1993

One conference abstract was published of a five-centre parallel group RCT of 275 patients undergoing pelvic radiotherapy to determine the effect of a computerised diet on acute enteric toxicity. Despite the fact that this abstract was published in 1993, there is no full report in the public domain and there was no response from the authors when contacted. Therefore, the data available were limited.

Patients were randomised into one of two groups. The intervention group comprised a diet, selected by a computer, which was "ipolipidic (predominantly medium chain triglycerides), ipercaloric and without lactose" with a supplement of 500 kcal of elemental diet as an adjunct. Patients started this diet four days before radiotherapy and finished four days after radiotherapy was completed. The exact nature of the diet and whether all patients in the intervention group had the same diet were not reported, neither was compliance with the diet. Patients were stratified for concomitant chemotherapy and amount of small bowel within the planned treatment volume (high dose volume). Radiotherapy technique and number of patients receiving chemotherapy were not reported. Outcomes were weight and acute enteric toxicity using European Organisation for Research and Treatment of Cancer (EORTC), although the authors reported grades that appear to be more like the RTOG grading system rather than the EORTC.

Data were analysed from 268 patients although the reasons for the drop-out or withdrawal of seven patients were not given. The authors reported that there was no difference in weight for the control group comparing before and after treatment weight but there was an increase of mean 1.1 kg for intervention group. No measure of variance was reported. Weight loss during radiotherapy was reported to be statistically significantly associated with toxicity. In terms of GI symptoms, in the intervention group, 69% had grade 0, 16% had grade 1, 14% had grade 2 and 2% had grades 3-4 toxicity. In the control group, 47% had grade 0, 25% had grade 1, 20% had grade 2 and 2% had grades 3 to 4 toxicity. The authors reported that this represents a significant difference between the two groups, although no P values were published.

The authors reported that they used EORTC to measure toxicity, but the scores that they used seem to be more in keeping with the RTOG grade. EORTC has several questionnaires that measure QoL and not toxicity and this is not reported as grades. RTOG grade is not validated and is a coarse tool that is known to underestimate toxicity. EORTC is not validated to measure toxicity. While the reported results suggest that computerised diet with elemental diet reduces acute enteric toxicity and weight loss during treatment, the lack of reported data makes further analysis difficult.

Dietary modification
Fibre modification

Murphy 2000

One single centre (two sites) parallel group unblinded randomised controlled pilot study of 84 patients (72 men and 12 women) with prostate or gynaecological cancer was performed to determine the effect of fibre modification with Metamucil on the severity and incidence of diarrhoea. Patients were randomised to take or not take Metamucil. All patients received a booklet called "Nutritional guidelines to help control diarrhoea", which advised on a low-fibre diet with limited fat, caffeine and alcohol intake. Researchers gave the dietary information. Subjects in the intervention group were instructed to take 5 mL of Metamucil powder mixed into 250 mL of water once a day in the morning and could take a second dose in the afternoon if diarrhoea persisted. Radiotherapy technique was not reported. Compliance was measured through patient diaries. Outcomes were based on patient diaries completed from the first day of radiotherapy to 28 days after completion of treatment. Patients recorded stool frequency, stool consistency, antidiarrhoeal use and daily dose of Metamucil. Data from diaries were converted into the Murphy Diarrhoea Scale (MDS), which gives a summary score for diarrhoea over the whole treatment period (Table 6), which takes account of stool frequency and texture, and use of antidiarrhoeals to determine "days-with-diarrhoea" as a numerical score. This is not a validated tool.

Table 6. Murphy Diarrhoea Scale
  1. *one 'day-with-diarrhoea' is defined as any day with one of the following:

    4-6 bowel movements more than normal

    ≥ 1 watery bowel movement

    2-3 loose or poorly formed bowel movements more than normal

    Use of antidiarrhoeal medication

Description Score
Mild diarrhoea (< 11% days-with-diarrhoea*)1
Moderate diarrhoea (11-20% days-with-diarrhoea*)2
Mild diarrhoea (> 20% days-with-diarrhoea*)3

Data from 24 subjects were excluded from analysis due withdrawal (one person), incomplete diaries (two people), inaccurate diary entries (six people), failure to return diaries (13 people) and using Metamucil while in the non-Metamucil group (two people). There were no differences between these two groups for mean age or weight at baseline. Participants in the Metamucil group (intervention) had lower mean MDS scores compared with the control group (1.8 versus 2.33, P value = 0.03) and lower incidence of diarrhoea compared with the control group (605 versus 83%, P value = 0.049). There was no difference with respect to the mean percentage days that patients took antidiarrhoeal medication (P value = 0.062). The mean time to onset of diarrhoea was not different between the two groups (P value = 0.895), neither was the mean duration of diarrhoea (P value = 0.905).

The use of Metamucil did appear to reduce the incidence and severity of diarrhoea during radiotherapy. There are no long-term follow-up data and no QoL data. This was the only study in which the intervention was applied when GI symptom occurred, rather than being used prophylactically in all patients regardless of symptoms.

Lactose and fibre modification

Pettersson 2012

One parallel group single-centre unblinded RCT of 130 patients with prostate cancer was performed to determine the effect of a lactose- and insoluble fibre-restricted diet (64 people) compared with no dietary change (66 people) on acute GI side effects and other aspects of health-related QoL. Patients were treated with external beam radiotherapy and either high-dose rate brachytherapy or proton therapy and patients were stratified according to radiation technique (brachytherapy versus proton therapy). The intervention was standardised dietary advice from a research dietician in face-to-face sessions at baseline (prior to radiotherapy), after four weeks of radiotherapy and via a telephone consultation at eight weeks following the start of radiotherapy (one week after radiotherapy completion) as well as a study-specific brochure at all time points. The advice was to avoid foods high in insoluble fibre and lactose and instead consume foods with a higher proportion of soluble fibres and low in lactose for the whole of the study period (from baseline up to 24 months after the end of radiotherapy). The control group comprised patients who received no dietary intervention and continued on their normal diet. The exact radiotherapy technique used was not described, but daily fractions of 2 Gy were delivered to 50 Gy with a treatment volume comprising the prostate only and seminal vesicles for T3 tumours.

Assessments were made at baseline, four weeks after onset of radiotherapy, eight weeks after onset of radiotherapy and two months after radiotherapy completion. Nutritional status was only assessed pre-treatment with the patient-generated subjective global assessment (PGSGA) and weight/BMI. Outcomes were GI symptoms using the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC-QLQ-C30) and European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire - prostate-specific module (EORTC-QLQ-PR25) along with the study-specific Gastrointestinal Side Effects Questionnaire (GISEQ), which assessed how much patients were "bothered" by GI symptoms. Global health status, functioning and symptoms were assessed using EORTC-QLQ-C30 and EORTC-QLQ-PR25. The study was powered for EORTC-QLQ-PR25 mean value for bowel symptoms. Adherence to dietary instructions was monitored by the study-specific Food Frequency Questionnaire (FFQ), in which patients reported how often they consumed certain foods (with unspecified portion sizes) over the preceding month. Foods were categorised on the basis of insoluble fibre and lactose content. The scale ranged from 0 (never) to 7 (three or more times per day).

There was no difference between the two groups at baseline in terms of clinical characteristics, FFQ scores, GI side effects and other aspects of QoL. There was no difference between the two groups in terms of bowel symptoms using EORTC-QLQ-C30, EORTC-QLQ-PR25 or GISEQ at eight weeks (one week after the completion of radiotherapy) except in the GISEQ item intestinal cramps (P value = 0.039). There was also no difference between the two groups in terms of QoL using EORTC-QLQ-C30 and EORTC-QLQ-PR25 although P values were not reported. There was no relationship between adherence to diet and bowel symptoms using FFQ scores. FFQ data showed that the control group did not change their diet but the intervention group had lower scores in the FFQ at all time points compared with baseline.

Overall, there was no benefit found for dietary modification of insoluble fibre and lactose in men undergoing pelvic radiotherapy for prostate cancer in terms of GI symptoms as determined by the EORTC-QLQ-C30, EORTC-QLQ-PR25 or GISEQ. However, these tools are not validated for detecting GI toxicity and, in particular, an exact definition of diarrhoea does not appear to have been given including both a change in stool frequency and consistency. This may have influenced the findings.

Lactose modification

Stryker 1986

One single-centre parallel group unblinded RCT of 64 patients (57 women and 7 men) undergoing pelvic radiotherapy for either gynaecological, prostate or sigmoid cancers was performed to determine whether modification of lactose in the diet affects the severity of GI symptoms experienced during pelvic radiotherapy. Patients were treated using parallel opposing anterior-posterior (AP) portals except in larger patients who were treated with four portals (box technique). Patients were randomised into three groups: normal diet containing at least 480 mL of milk daily (control); lactose-restricted diet with calcium replacement of 625 mg three times a day or lactose-hydrolysed diet containing 480 mL of milk daily where 90% of the lactose was hydrolysed to glucose and galactose. The details of the diets have not been specified and the professional modifying the diet was not specified. Outcomes were mean stool frequency and antidiarrhoeal use, although the latter was not an outcome for the purpose of this review.

There was discrepancy in the numbers reported. Sixty-four patients were randomised into one of the three groups. The author reported that 21 patients were randomised to the control group, 20 to the lactose-restricted group and 22 in the lactose-hydrolysed group. Several patients were excluded or dropped out: three patients were withdrawn from analysis for the control group; one patient was withdrawn from analysis for the lactose-restricted group and six patients were withdrawn from analysis for the lactose-hydrolysed group. All withdrawals were because patients did not complete four weeks of treatment on the protocol. There was no reported power calculation.

Baseline mean (SD) stool frequency was not significantly different between the groups (P value = 0.7071) with a frequency of 1.3 (0.8) for the control group, 1.0 (0.6) for the lactose-restricted group and 1.1 (1.0) for the lactose-hydrolysed group daily. The mean weekly stool frequency at week five was 22.5, 19 and 26 for the control, lactose-restricted and lactose-hydrolysed groups, respectively. A measure of variance was not reported. There was no significant difference between the three dietary regimens for mean weekly stool frequency (P value = 0.2153). Mean numbers of diphenoxylate tablets at week five were 8.5, 7 and 18 for the control, lactose-restricted and lactose-hydrolysed groups, respectively. A measure of variance was not reported. There was no significant difference between the three dietary regimens for mean number of diphenoxylate tablets used weekly (P value = 0.7526).

This study showed no benefit of lactose-restricted diet during radiotherapy, although the radiotherapy techniques used are no longer standard practice and the details regarding the lactose-restricted diet were lacking so it is difficult to comment on whether this would be comparable with a lactose-restricted diet used in the 2010s. In addition, 10 recruited patients were excluded from the analysis potentially leading to attrition bias. QoL was not measured.

Lactose and fat modification

Bye 1992

One single-centre parallel group unblinded RCT of 143 patients with either endometrial (49 women), cervical (93 women) or ovarian (one women) cancer undergoing pelvic radiotherapy was performed to determine the effect of a fat- and lactose-restricted diet on radiation-induced diarrhoea. Radiotherapy was delivered using AP fields followed by intracavity treatment for 90 of those with cervix cancer. Patients were randomised into two groups: regular hospital diet (control) or lactose- and fat-restricted diet comprising a maximum of 40g of fat and 5g of lactose daily prescribed by a dietician and individualised to patient tastes (intervention). The dietary modification was started before the start of radiotherapy and continued for six weeks after radiotherapy. Outcomes measured were weight, stool frequency/consistency, use of antidiarrhoeals, QoL using EORTC-QLQ-C36, arm muscle circumference and serum transferrin at baseline, six and 12 weeks.

There were no baseline imbalances in stool frequency, use of antidiarrhoeals, total energy intake or total fat intake. Fourteen patients were excluded from analysis: seven patients did not comply with the dietary modification, one control patient changed over to lactose- and fat-restricted diet, six patients left for personal reasons (three from each group).

At week six, there was a difference in stool frequency between control and intervention group (P value < 0.01) with patients passing a mean of 1.1 loose and watery stools per day in the control group and 1.7 loose and watery stools per day in the intervention group. When converted to the diarrhoea scale reported in the paper (Table 7), 14 of 61 patients versus 32 of 67 patients had diarrhoea in the intervention and control group, respectively (P value < 0.01). Patients in the intervention group took fewer antidiarrhoeals compared with the control group (P value < 0.01). Patients in the intervention group had a greater reduction in body weight during treatment (mean 2.6 kg) compared with controls (mean 1.7 kg) (P value = 0.006).

Table 7. Diarrhoea scale used in Bye 1992
Score Description
0No change in stool frequency
1Increase of 1-3 stools a day, normal or soft
2Increase in 4-6 stools a day, all watery
3Increase in > 6 stools a day

At week 12, there were no differences between the groups with respect to stool frequency and antidiarrhoeal use. Patients in the intervention group had gained a mean of 0.6 kg of weight whereas the control had gained a mean of 1.1 kg of weight.

There was no difference in QoL. No patients had an arm muscle circumference (AMC) less than 19 cm at baseline. No patients in the control group and one (2%) in the intervention group had an AMC less than 19 cm at week six. Two patients (2%) and one patient (1%) in the control and intervention group, respectively, had a serum transferrin less than 1.4 g/L at baseline. Four patients (6%) and eight patients (13%) in the control and intervention group, respectively, had a serum transferrin less than 1.4 g/L at week six. The significance of these differences between groups was not reported.

This study did demonstrate that a low-fat, low-lactose diet reduced the incidence of diarrhoea during radiotherapy, with the control group taking twice as many antidiarrhoeals as the intervention group. Efforts were made to define diarrhoea clearly in terms of changes in stool frequency and consistency, although the scoring system used was not a validated tool.

Fat modification

Wedlake 2012

One two-centre three-arm parallel group unblinded RCT of 117 patients (79 men and 38 women) was performed to determine the efficacy of low- or modified-fat diets on reducing GI toxicity. Patients were randomised into one of three groups: low-fat (40 people), modified-fat (38 people) and control (39 people) group. The low-fat group was prescribed a low-fat diet with long-chain triglyceride dietary fats calculated to comprise 20% of total energy intake. Patients in this group were advised to have a stable diet and maintain total energy intake from carbohydrate- and protein-based sources. The modified-fat group were prescribed a diet with fats calculated to comprise 40% of total energy intake with 50% of this derived from long-chain triglycerides and 50% from a medium-chain triglyceride-based emulsion (Liquigen). The control group were prescribed a normal fat diet with long-chain triglyceride dietary fats calculated to comprise 40% of total energy intake. Prescriptions were given in terms of "fat points" where one fat point was equal to 5 g of fat. The intervention period was the first four weeks of radiotherapy. The radiotherapy technique was not described.

At baseline, the groups were well matched for age, gender, radiotherapy, weight and long-chain triglyceride fat intake but a higher proportion of patients in the modified-fat group were treated with chemotherapy. The low-fat group contained a higher proportion of patients with urological cancer and a lower proportion of patients with lower GI cancer. Median radiotherapy dose was 54 Gy (range 36 to 74) with a dose per fraction of 1.8 to 2.0 Gy.

Outcomes were daily fat point consumption recorded in patient diaries; GI toxicity, measured by IBDQB, VIQ and RTOG; nutritional parameters, that is, weight, BMI and handgrip strength; and QoL, measured using total IBDQ score. Outcomes were assessed at baseline, after two weeks of radiotherapy, after four weeks of radiotherapy and at one year post-treatment. Percentage compliance with Liquigen was also recorded at two and four weeks. The study was powered for change in mean IBDQ score from baseline to week four (35 patients per arm were required).

Ten patients withdrew during the first four weeks of the study due to intolerance to Liquigen (six people), voluntary withdrawal (three people) and entering a mutually exclusive study (one person). These patients were not included in the analysis. A further 32 patients were lost to follow-up by one year (12 deceased, six were too ill, there was no reply from 10 patients and four patients had moved residence).

Mean IBDQB score fell in all patients between baseline and week four, indicating an increase in GI toxicity (66.2 versus 58.9, 106 people). There were no differences in the change in paired scores comparing baseline and week four between low-fat and modified-fat groups (P value = 0.914), low-fat and normal-fat groups (P value = 0.793) and modified-fat and normal-fat groups (P value = 0.890). There was no report of whether there was a difference in VIQ scores between baseline and week four or baseline and one year; IBDQB score between baseline and one year and RTOG grades.

There was a reduction in mean weight in all groups between baseline and week four and in the modified- and normal-fat groups between baseline and one year, although the level of statistical significance is not reported. By study group, mean grip strength fell by 0.2 kg in the low-fat group, rose by 2.2 kg in the modified-fat group and rose by 1.2 kg in the normal-fat group. No measure of variance of this change was reported. QoL, as determined by change in total IBDQ score during treatment, improved for all groups, as reflected in a fall in total IBDQ score. The magnitude of this fall in score was -13 for the low-fat and modified-fat groups and by -14.7 for the normal-fat groups, where an increase represents a deterioration in QoL. Change in QoL comparing baseline versus one year was a change in total IBDQ score of 2.5 for the low-fat, -9.6 for the modified-fat and -4.0 for the normal-fat groups, respectively. Compliance was variable in the different groups, ranging from 93% in the low-fat group to 76% in the modified-fat group to 9% in the normal-fat group. Compliance was poor in the normal-fat group as they reduced their fat intake too much. Compliance with Liquigen was poor with only 58% achieving 76% or greater compliance.

This study did not demonstrate that low- or modified-fat diet reduced GI symptoms. This may reflect the fact that the mean peri-radiotherapy self reported fat points (1 fat point = 5 g of fat) was 6.7 (SD 1.98) in the low-fat group, 8.2 (SD 1.40) in the modified-fat group and 12.6 (SD 3.28) in the normal-fat group, where the mean prescription was 19 points. These are quite similar and perhaps the similarity of fat intake between groups meant that differences could not be observed. In addition, the control group of "normal fat diet" is not necessarily reflective of the normal diet of most patients, rather a reflection of recommended fat intake.

Excluded studies

Seventeen studies did not meet inclusion criteria, of which two had incorrect outcomes; one was a cross-over design; two included palliative patients; one included ineligible primary tumour groups; 11 were not nutritional interventions as defined by the protocol. One study was excluded on discussion as design methodologically unjustified. The latter started as an RCT then (after four patients had been enrolled) changed to using a retrospective control group. Details of these are given in the Characteristics of excluded studies table. From the trials identified, we determined that 10 fitted our inclusion criteria (see Characteristics of included studies).

Risk of bias in included studies

Allocation

Only three of the included RCTs adequately described their methods of randomisation and allocation concealment (McGough 2008; Pettersson 2012; Wedlake 2012). We obtained information for Bye 1992 via author contact, which confirmed a low risk of selection bias. Murphy 2000 described using random tables but no information regarding allocation concealment and so we rated this as unclear risk of bias. All other RCTs did not describe the randomisation process and we were, therefore, rated them as an unclear risk of selection bias. Craighead 1998 was not an RCT and so we rated the study as a high risk of selection bias.

Blinding

All of the studies were unblinded. It would be difficult to blind a study involving dietary modification and it would be difficult to use a placebo control for elemental diet due to taste, consistency and the risk of causing the control group to have a reduction in their caloric intake. The studies did not report blinding of the researchers or the participants taking part, introducing a high risk of bias. Therefore, we rated all studies as being at high risk of bias for performance and detection bias with the understanding that it would be extremely difficult to blind the studies given the nature of the interventions.

Incomplete outcome data

Five of the studies had high risk of attrition bias on the basis that there were a number of patients who dropped out of the studies (Brown 1980; Bye 1992; Murphy 2000; Stryker 1986; Wedlake 2012). This was largely because they did not comply with the intervention often for reasons of taste, palatability or side effects of the intervention.

Selective reporting

All of the trials included in the review reported at least one of the primary outcomes listed in the Methods section.

Other potential sources of bias

There were temporal biases with some of the older studies, given that radiotherapy technique has changed over time, as have the treatment regimens. This could lead to a different pattern of GI symptoms and toxicity. This is like to affect Brown 1980; Bye 1992; Capirci 1993; and Stryker 1986 in particular, although not all of these studies clearly reported the radiotherapy technique used.

The risk of bias summary is given in Figure 2 and judgements about each risk of bias item presented as percentages across all included studies is shown in Figure 3.

Figure 2.

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

Figure 3.

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

Effects of interventions

See: Summary of findings for the main comparison Dietary modification compared with no nutritional intervention for adults undergoing pelvic radiotherapy

The 10 studies in this review all used different interventions and did not utilise the same outcome measures. Therefore, data were dichotomised for the primary outcome of GI symptoms to the presence or absence of diarrhoea at the end of treatment, that is weeks four to six from the start of radiotherapy. We chose this as it was the most consistently reported outcome in the trials. For studies using the IBDQ, data for the item of loose stool were dichotomised. We performed analysis for dietary modification versus no nutritional intervention or standard nutritional intervention. Data were not available regarding diarrhoea at the end of treatment from Brown 1980; Capirci 1993; Capirci 2000; Craighead 1998; McGough 2008; and Stryker 1986 so analysis was not possible for all nutritional interventions compared with no nutritional intervention or standard nutritional intervention, or for elemental diet versus no nutritional intervention or standard nutritional intervention. Data for change in weight comparing pre-radiotherapy and end of radiotherapy weight were available for Bye 1992 and Wedlake 2012 comparing dietary modification versus no nutritional intervention or standard nutritional intervention. There were no data on which to analyse QoL comparing all nutritional interventions to no nutritional intervention or standard nutritional intervention, dietary modification versus no nutritional intervention or standard nutritional intervention, or elemental diet versus no nutritional intervention or standard nutritional intervention.

Wedlake 2012 included two intervention groups for fat modification and one control group. For these studies, we pooled data for the two intervention groups to form one intervention group for the purposes of meta-analysis. We did this by simply adding the number of events and participants in each group for dichotomised data and using Review Manager 5 (RevMan 2012) calculator for continuous data to combine the means and SDs.

We contacted all study authors to obtain relevant information, but there was no response from the corresponding author for Brown 1980; Capirci 1993; Capirci 2000; Craighead 1998; Stryker 1986. Contact was made with the corresponding author for McGough 2008, but unfortunately, the raw data for the IBDQ data were not accessible.

Dietary modification and diarrhoea

We found five trials comparing dietary modification versus no nutritional intervention or standard nutritional intervention and reporting diarrhoea as an outcome (Bye 1992; McGough 2008; Murphy 2000; Pettersson 2012; Wedlake 2012), of which four were included in the meta-analysis involving 413 participants (Bye 1992; Murphy 2000; Pettersson 2012; Wedlake 2012) (Figure 4). We analysed dichotomous data using RRs with Mantel-Haenszel in a fixed-effect method. Absolute risk of diarrhoea was 44% (85/191) in the control group and 30% (66/222) in the intervention group and heterogeneity between the studies was low to moderate (Chi2 = 3.50, P value = 0.32, I2 = 14%). The relative effect was 0.66 (95% CI 0.51 to 0.87) for dietary modification (Analysis 1.1).

Figure 4.

Forest plot of comparison: 1 Nutritional intervention versus no nutritional intervention, outcome: 1.1 Diarrhoea.

Dietary modification and weight

We found five trials comparing dietary modification versus no nutritional intervention or standard nutritional intervention and reporting change in weight as an outcome (Bye 1992; McGough 2008; Murphy 2000; Pettersson 2012; Wedlake 2012), of which two were included in the meta-analysis involving 235 participants (Bye 1992; Wedlake 2012) (Figure 5). Continuous data were analysed using MD with inverse variance in a fixed-effect method. The heterogeneity was low to moderate (Chi2 = 1.41, P value = 0.24, I2 = 29%). The MD was -0.57 (95% CI -1.22 to 0.09) (Analysis 1.2).

Figure 5.

Forest plot of comparison: 1 Nutritional intervention versus no nutritional intervention, outcome: 1.2 Change in weight.

Other outcomes

The other outcomes, that is GI symptoms or toxicity six months or more after the start of radiotherapy, QoL measures, completion and duration of planned anticancer treatment, unplanned hospital stay within three months of starting radiotherapy, recurrence of cancer and mortality, were not reported in any of the trials. We have reported adverse effects and compliance with nutritional intervention descriptively above.

We did not undertake funnel plots on any of the comparisons in the review, as the number of trials in each of the analyses was too small to determine risk of publication bias.

Discussion

Summary of main results

Ten studies met our inclusion criteria and were of varying quality. This quality assessment is essential to place the results within the context of current clinical practice. The first point of note is that the included studies used a variety of outcome measures for GI symptoms ranging from physician reported (RTOG) to patient reported (IBDQ) and from static assessments made at one-off time points to those attempting to reflect overall patterns and symptom burden throughout radiotherapy treatment (MDS). Some studies used a composite score covering the breadth of GI symptoms (IBDQ) and some used specific aspects of diarrhoea as an outcome (mean stool frequency). None of these methods were validated. Some tools, for example, QoL measures, were used outside the area for which they were validated (EORTC). These questionnaires were developed and validated to measure QoL not GI toxicity, and such use is methodologically invalid.

The other notable heterogeneity between studies was the range of hypotheses. Some studies aimed to assess whether the intervention prevented GI symptoms, some to assess whether it controlled GI symptoms, some to assess whether it decreased GI symptoms and some to assess whether it maintained or improved nutritional status. This results in a confused picture for analysis. Studies were powered to several different end points, not all of which were GI symptoms. This may lead to underpowering of studies for the main outcome of this review and may lead to an underestimation of the intervention effect in the individual studies.

All of the studies except Murphy 2000 used interventions in all patients from the start of treatment. This approach is based on the assumption that there may be a 'one size fits all' solution or a 'magic bullet' to prevent or treat toxicity in individual patients. This approach is unlikely to be the case, particularly with nutritional interventions or if baseline nutritional status is not taken into account, or both. It is known that certain GI diagnoses are common during radiotherapy in patients who develop GI symptoms, for example lactose intolerance (15% to 44%) (Fernández-Bañares 1991; Wedlake 2008), small bowel bacterial overgrowth (26%) (Wedlake 2008), and bile acid malabsorption (44% to 57%) (Fernández-Bañares 1991; Yeoh 1984). Clearly, a low-fat diet is unlikely to treat lactose intolerance or weight loss effectively and a low-lactose diet will not lead to a resolution of symptoms in small bowel bacterial overgrowth or bile acid malabsorption. However, nutritional interventions tailored to symptoms and diagnoses are most likely to be efficacious. Only Murphy 2000 used a dietary intervention if and when diarrhoea began, and this intervention was Metamucil, a soluble fibre-based bulking agent. The results of this study suggested that increasing soluble fibre intake improves GI symptoms, which runs counter to the current initial advice given to radiotherapy patients, which is to reduce dietary fibre intake.

There was a large amount of clinical heterogeneity in terms of tumour groups included within the studies. Some studies included mixed pelvic tumours, some purely gynaecological tumours and some purely prostate tumours. This will alter radiotherapy parameters including technique, dose, fractionation and treatment volume, all of which will impact on outcomes including likelihood of developing GI symptoms and the nature and pattern of GI symptoms. Furthermore, the nutritional status at baseline of the different tumours may differ.

Overall completeness and applicability of evidence

The evidence presented is applicable for the current management of diarrhoea during radical pelvic radiotherapy with regard to dietary modification.

There is a potential issue with temporal bias in that patients in older studies underwent radiotherapy using techniques that are no longer standard practice. Radiotherapy techniques are becoming more refined and, as such, the data from these older studies may not reflect what would be found if the study was repeated again now. Older techniques, including parallel opposing pairs, delivered the total radiation dose to a larger area of the pelvis. Techniques such as 3D conformal radiotherapy and intensity modulated radiotherapy allows the radiation beams to be shaped to match the tumour and, in the latter case, the doses can be altered; indeed the radiotherapy technique used changes the volume and dose delivered to normal tissue, therefore potentially changing the pattern of GI toxicity. However, due to missing data, many of these older studies were excluded from meta-analysis, which reduced this risk in our findings.

Quality of the evidence

The quality of the evidence is variable with some high quality trials included and others with a high risk of bias. We downgraded the quality of evidence to moderate due to attrition bias in the studies.

Potential biases in the review process

We chose to dichotomise the data for diarrhoea into "any severity of diarrhoea" vs "no diarrhoea" as this was the most frequently reported GI symptom outcome either as a stand-alone symptom or as an item on a questionnaire (IBDQ, EORTC). Even with this, there were discrepancies in that a clear definition of diarrhoea was not always given. Some authors used stool frequency, although this is not the complete definition of diarrhoea, which also encompasses a change in stool consistency. It may be that with no definition provided, there were differences in the interpretation and reporting of diarrhoea between individuals. Murphy 2000 and Bye 1992 attempted to take into consideration the full definition of diarrhoea and Murphy 2000 used a measure that was a composite score of change in bowel habit across the whole of the course of radiotherapy. McGough 2008 and Wedlake 2012 used IBDQ data which includes items for loose stool and frequency.

Agreements and disagreements with other studies or reviews

A previous review looked at the role of nutritional interventions in patients with pelvic cancers undergoing radiotherapy (McGough 2004). McGough 2004 had a broader remit including interventions during radiotherapy but also for chronic GI symptoms. It also included studies of children, palliative treatment, retrospective methodology and a wider range of tumour groups.

McGough 2004 reported a benefit in terms of GI symptoms during pelvic radiotherapy with low-fat diets, but no change with modified-lactose diets. It also reported statistically significant reductions in GI symptoms with elemental diet, however of the five studies reviewed, two included the same patient group (Brown 1980; Foster 1980), one was a prospective uncontrolled cohort study (Craighead 1998), one was methodologically unsound (McArdle 1986), and one was only published in abstract form in a non-peer reviewed booklet and was lacking in detail (Capirci 2000).

In summary, McGough 2004 stated that low-fat diets and elemental diets may be of benefit in preventing acute GI symptoms during pelvic radiotherapy. This differs from our systematic review and meta-analysis. We report a clear benefit from dietary modification in terms of reduction in diarrhoea during radiotherapy. Data were not available to meta-analyse the effect that elemental diet had on GI symptoms during pelvic radiotherapy, but there were significant issues with compliance and tolerability of the intervention and the studies reporting benefit were methodologically unsound, used poor measurement tools, were only published in abstract form with limited details, or a combination of these. McGough 2004 also highlighted the inadequate tools used for measuring GI toxicity, the inadequate methodology in many of the studies and the failure to use clinically important end points.

Authors' conclusions

Implications for practice

The results for dietary modification favoured the intervention compared with control for reduction in any severity of diarrhoea during pelvic radiotherapy. There were not enough data in the appropriate form to analyse elemental diet as an intervention further, but there were real issues regarding tolerability of this intervention and compliance, which make it of questionable relevance in day-to-day clinical practice. There were no data on the effects on nutritional interventions on chronic gastrointestinal (GI) symptoms or toxicity, unplanned hospital stay within three months of starting radiotherapy, recurrence of cancer and mortality. There were not enough data to analyse the effect of nutritional intervention on nutritional status or QoL fully, therefore, those data were inconclusive.

Implications for research

This review highlights the paucity of high-quality, robust research using nutritional interventions during radiotherapy to improve GI symptoms. With the understanding of the strong association between acute GI symptoms and chronic GI symptoms (consequential late effects) and the link between GI symptoms, length of radiotherapy and survival, there is a need for high-quality multicentre randomised controlled trials (RCTs) using validated outcome measures that focus on ameliorating GI symptoms during pelvic radiotherapy.

In future trials there must be a consistent, validated way of measuring GI symptoms in these patients to facilitate the comparison of interventions and be able to extract meaningful data regarding the efficacy of a given intervention. The CTCAE pelvic symptom questionnaire, a patient-reported outcome measure comprising rectum-bowel, bladder-urethra and sexual function subsets, is the only validated patient-reported outcome measure in this patient group, but was not used in any of the studies.

Nutritional interventions require further evaluation with regard to acute and chronic GI symptoms, quality of life and ability to complete planned cancer treatment, particularly when used selectively in those patients who develop GI symptoms during radiotherapy. Studies need to be planned to reduce heterogeneity in terms of tumour groups included and also interventions tailored to baseline nutritional status. This will facilitate informed clinical decisions to be made with increased confidence in the use of nutritional interventions during pelvic radiotherapy.

Acknowledgements

We thank Jo Morrison for her clinical expertise.

The National Institute for Health Research (NIHR) is the largest single funder of the Cochrane Gynaecological Cancer Group. 

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR, National Health Service (NHS) or the Department of Health.

Data and analyses

Download statistical data

Comparison 1. Nutritional intervention versus no nutritional intervention
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Diarrhoea4413Risk Ratio (M-H, Fixed, 95% CI)0.66 [0.51, 0.87]
2 Change in weight2235Mean Difference (IV, Fixed, 95% CI)-0.57 [-1.22, 0.09]
Analysis 1.1.

Comparison 1 Nutritional intervention versus no nutritional intervention, Outcome 1 Diarrhoea.

Analysis 1.2.

Comparison 1 Nutritional intervention versus no nutritional intervention, Outcome 2 Change in weight.

Appendices

Appendix 1. MEDLINE search strategy

MEDLINE Ovid

1 Pelvic Neoplasms/
2 exp Uterine Neoplasms/
3 Urinary Bladder Neoplasms/
4 Prostatic Neoplasms/
5 exp Rectal Neoplasms/
6 ((pelvi* or cervi* or uter* or endometr* or bladder* or prostat* or rect* or anal or anus) adj5 (cancer* or tumor* or tumour* or neoplas* or malignan* or carcinoma* or adenocarcinoma*)).mp.
7 1 or 2 or 3 or 4 or 5 or 6
8 exp radiotherapy/
9 radiotherapy.fs.
10 (radiotherap* or brachytherap* or radiat* or irradiat* or chemoradiotherap* or chemoradiat*).mp.
11 8 or 9 or 10
12 exp Diet/
13 Food Habits/
14 exp Food/
15 exp Nutrition Therapy/
16 exp Nutrition Disorders/
17 exp Feeding Methods/
18 exp Polysaccharides/
19 exp Amino Acids/
20 exp Peptides/
21 exp Antioxidants/
22 exp Fish Oils/
23 (nutrition* or nutrient* or malnutrition or diet* or food* or feed* or drink* or supplement* or formula* or vitamin* or mineral* or fat* or carbohydrate* or protein*).mp.
24 (lactose or fibre or fiber or bran or psyllium or plantago ovata or ispaghula or polysaccharide* or polymeric* or peptide* or amino-acid* or glutamine* or fish oil* or arginine* or antioxidant* or anti-oxidant*).mp.
25 (enteral* or nasogastric* or gastrostomy of jejunostomy or oral* or tube* or sip*).mp.
26 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25
27 randomized controlled trial.pt.
28 controlled clinical trial.pt.
29 randomized.ab.
30 placebo.ab.
31 clinical trials as topic.sh.
32 randomly.ab.
33 trial.ti.
34 exp Cohort Studies/
35 (cohort* or prospective* or retrospective*).mp.
36 Comparative Study/
37 (observational or comparative).mp.
38 (case* adj5 series).mp.
39 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38
40 7 and 11 and 26 and 39
41 exp animals/ not humans.sh.
42 40 not 41

key:

mp=title, abstract, original title, name of substance word, subject heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier, sh=subject heading, ti=title, ab=abstract

Appendix 2. EMBASE search strategy

EMBASE Ovid

1 pelvis tumor/
2 exp uterus cancer/
3 bladder tumor/
4 exp prostate tumor/
5 rectum tumor/
6 ((pelvi* or cervi* or uter* or endometr* or bladder* or prostat* or rect* or anal or anus) adj5 (cancer* or tumor* or tumour* or neoplas* or malignan* or carcinoma* or adenocarcinoma*)).mp.
7 1 or 2 or 3 or 4 or 5 or 6
8 exp radiotherapy/
9 rt.fs.
10 (radiotherap* or brachytherap* or radiat* or irradiat* or chemoradiotherap* or chemoradiat*).mp.
11 8 or 9 or 10
12 exp nutrition/
13 exp nutritional disorder/
14 exp polysaccharide/
15 exp amino acid/
16 peptide/
17 exp antioxidant/
18 fish oil/
19 (nutrition* or nutrient* or malnutrition or diet* or food* or feed* or drink* or supplement* or formula* or vitamin* or mineral* or fat* or carbohydrate* or protein*).mp.
20 (lactose or fibre or fiber or bran or psyllium or plantago ovata or ispaghula or polysaccharide* or polymeric* or peptide* or amino-acid* or glutamine* or fish oil* or arginine* or antioxidant* or anti-oxidant*).mp.
21 (enteral* or nasogastric* or gastrostomy of jejunostomy or oral* or tube* or sip*).mp.
22 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21
23 7 and 11 and 22
24 crossover procedure/
25 double-blind procedure/
26 randomized controlled trial/
27 single-blind procedure/
28 random*.mp.
29 factorial*.mp.
30 (crossover* or cross over* or cross-over*).mp.
31 placebo*.mp.
32 (double* adj blind*).mp.
33 (singl* adj blind*).mp.
34 assign*.mp.
35 allocat*.mp.
36 volunteer*.mp.
37 controlled clinical trial/
38 cohort analysis/
39 prospective study/
40 retrospective study/
41 exp comparative study/
42 observational study/
43 case study/
44 (cohort* or prospective* or retrospective* or observational or comparative or (case* adj5 series)).mp.
45 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44
46 23 and 45
47 (exp Animal/ or Nonhuman/ or exp animal Experiment/) not Human/
48 46 not 47

key:

mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer, device trade name, keyword

Appendix 3. CENTRAL search strategy

CENTRAL

#1 MeSH descriptor Pelvic Neoplasms, this term only
#2 MeSH descriptor Uterine Neoplasms explode all trees
#3 MeSH descriptor Urinary Bladder Neoplasms, this term only
#4 MeSH descriptor Prostatic Neoplasms, this term only
#5 MeSH descriptor Rectal Neoplasms explode all trees
#6 ((pelvi* or cervi* or uter* or endometr* or bladder* or prostat* or rect* or anal or anus) near/5 (cancer* or tumor* or tumour* or neoplas* or malignan* or carcinoma* or adenocarcinoma*))
#7 (#1 OR #2 OR #3 OR #4 OR #5 OR #6)
#8 MeSH descriptor Radiotherapy explode all trees
#9 Any MeSH descriptor with qualifier: RT
#10 (radiotherap* or brachytherap* or radiat* or irradiat* or chemoradiotherap* or chemoradiat*)
#11 (#8 OR #9 OR #10)
#12 MeSH descriptor Diet explode all trees
#13 MeSH descriptor Food Habits, this term only
#14 MeSH descriptor Food explode all trees
#15 MeSH descriptor Nutrition Therapy explode all trees
#16 MeSH descriptor Nutrition Disorders explode all trees
#17 MeSH descriptor Feeding Methods explode all trees
#18 MeSH descriptor Polysaccharides explode all trees
#19 MeSH descriptor Amino Acids explode all trees
#20 MeSH descriptor Peptides explode all trees
#21 MeSH descriptor Antioxidants explode all trees
#22 MeSH descriptor Fish Oils explode all trees
#23 (nutrition* or nutrient* or malnutrition or diet* or food* or feed* or drink* or supplement* or formula* or vitamin* or mineral* or fat* or carbohydrate* or protein*)
#24 (lactose or fibre or fiber or bran or psyllium or plantago ovata or ispaghula or polysaccharide* or polymeric* or peptide* or amino-acid* or glutamine* or fish oil* or arginine* or antioxidant* or anti-oxidant*)
#25 (enteral* or nasogastric* or gastrostomy of jejunostomy or oral* or tube* or sip*)
#26 (#12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25)
#27 (#7 AND #11 AND #26)

 

Contributions of authors

All the named review authors have been involved in the development of this review. CH wrote proposal, CH/SB/SED/SL commented on proposal.

Declarations of interest

None known.

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • Dr S Burden is supported by a Macmillan Cancer Care Post Doctoral Fellowship, UK.

Differences between protocol and review

None.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Brown 1980

MethodsParallel group RCT
Participants

68 participants undergoing pelvic radiotherapy with mainly urological malignancy

Male:female 53:15

21 did not find Vivonex tolerable, 30 intervention and 17 controls

Interventions

Group 1: low-fibre diet with Vivonex HN (Eaton Laboratories Ltd, Woking, UK) 3 sachets daily (n = 30) from day 1 of radiotherapy

Group 2: low-fibre diet (standard care) (n = 17)

21 did not find Vivonex tolerable

The Vivonex HN supplementation provided 900 kcal and 40 g of mixed amino acids daily

OutcomesStool frequency, weight and reason for non-compliance
Notes

Dose/regimen of Vivonex defined in Foster 1980

Need to consider relevance of the study given the change in radiotherapy techniques over time and the age of this study

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot described
Allocation concealment (selection bias)Unclear riskNot described
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
High risk21 patients in the intervention group found the elemental diet unacceptable
Selective reporting (reporting bias)Low risk 
Other biasUnclear riskPossible temporal bias related to radiotherapy technique

Bye 1992

MethodsParallel group RCT
Participants

143 patients with gynaecological cancer undergoing pelvic radiotherapy

Mean age (intervention): 51 years

Mean age (control): 56 years

All female

Interventions

Group 1: lactose- and fat-restricted diet to 5 g/day and 40 g/day, respectively

Group 2: no nutritional intervention

Duration of intervention was from the start of radiotherapy to 6 weeks after completion of radiotherapy

OutcomesWeight, stool frequency/consistency (Table 7) number of patients with arm muscle circumference < 19 cm, number of people with transferrin < 1.4 g/L, stool frequency, QoL (EORTC-QLQ-C36)
NotesNeed to consider relevance of the study given the change in radiotherapy techniques over time and the age of this study
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskInformation from author
Allocation concealment (selection bias)Low riskInformation from author
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
High risk14 patients excluded from analysis
Selective reporting (reporting bias)Low riskNot all of EORTC reported - data obtained from author
Other biasUnclear riskPossible temporal bias related to radiotherapy technique

Capirci 1993

MethodsMulticentre parallel group RCT
Participants275 patients with mixed pelvic tumours undergoing pelvic radiotherapy
Interventions

Group 1: computerised diet (ipolipidic, ipercaloric and lactose free) with 500 kcal of elemental diet starting 4 days before radiotherapy and finishing 4 days after radiotherapy was completed

Group 2: no nutritional intervention

Duration of intervention was from 4 days before the start of radiotherapy to 4 days after the completion of radiotherapy

OutcomesEORTC (more likely to mean RTOG), weight
NotesNeed to consider relevance of the study given the change in radiotherapy techniques over time and the age of this study
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot enough information to assess bias
Allocation concealment (selection bias)Unclear riskNot enough information to assess bias
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot enough information to assess bias
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot enough information to assess bias
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot enough information to assess bias
Selective reporting (reporting bias)Unclear riskNot enough information to assess bias

Capirci 2000

MethodsMulticentre parallel RCT
Participants

677 patients undergoing pelvic radiotherapy for mixed pelvic tumours

Most patients had a primary rectal cancer (n = 439), with 228 having primary uterine cancer and 10 patients with prostate cancer

Interventions

Group 1: natural diet with elemental diet (n = 332)

Group 2: standard diet (n = 345)

Duration of intervention unclear

OutcomesRTOG grade and change in weight comparing before and after radiotherapy
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot enough information to assess bias
Allocation concealment (selection bias)Unclear riskNot enough information to assess bias
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot enough information to assess bias
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot enough information to assess bias
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot enough information to assess bias
Selective reporting (reporting bias)Unclear riskNot enough information to assess bias
Other biasUnclear riskNot enough information to assess bias

Craighead 1998

MethodsUncontrolled prospective cohort study
Participants

62 uterus or cervical cancer patients undergoing pelvic radiotherapy

Mean age for the intervention group: 48 years

All female

Interventions

Group 1: 2-3 servings of 39 g of Vital HNR powder (Ross Products Division, Abbott Laboratories, Illinois) mixed with 250 mL of liquid which provided 1050 to 1260 J per day (n = 17)

Group 2: no nutritional intervention (n = 45)

Duration of intervention was from 3 days pre-radiotherapy to last day of radiotherapy

OutcomesCompliance, RTOG grade for acute lower GI enteritis as assessed by a clinical oncologist, diarrhoea at 12 months, duration of diarrhoea (days)
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskNot randomised
Allocation concealment (selection bias)High riskNot randomised
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
Low risk 
Selective reporting (reporting bias)Low risk 
Other biasUnclear riskPossible temporal bias related to radiotherapy technique

McGough 2008

MethodsParallel group RCT
Participants

50 patients with histologically confirmed gynaecological, urological or lower GI malignancy undergoing pelvic radiotherapy

Median age: 61.5 years

Male:female 21:29

Interventions

Group 1: E028 Extra ready to drink cartons (SHS International, Liverpool, UK) or E028 Extra flavoured powder sachets (SHS International, Liverpool, UK) to replace one-third of daily calories

Group 2: no nutritional intervention

Duration of intervention was during the first 3 weeks of radiotherapy

OutcomesIBDQ, IBDQB, VIQ, RTOG, weight and BMI
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskPerformed by independent Institute of Cancer Research Office
Allocation concealment (selection bias)Low riskPermuted blocks
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
Low riskAll analyses reported conducted on an intention-to-treat basis
Selective reporting (reporting bias)Low risk 
Other biasUnclear riskSmall single-centre study

Murphy 2000

MethodsParallel group RCT
Participants

84 patients with prostate or gynaecological cancer undergoing pelvic radiotherapy

Male:female 72:12

Mean age (group 1): 63.7 years

Mean age (group 2): 66.7

Interventions

Group 1: Metamucil

Group 2: standard treatment

For the intervention group, when diarrhoea started during treatment 5 mL of Metamucil powder was mixed into 250 mL water once a day in the morning. A second dose could be taken in the afternoon if diarrhoea persisted

OutcomesMurphy Diarrhoea Score (Table 6)
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskTable of random numbers
Allocation concealment (selection bias)Unclear riskNot adequately described
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
High risk24 patients excluded from analysis
Selective reporting (reporting bias)Low risk 
Other biasUnclear riskPossible temporal bias related to radiotherapy technique

Pettersson 2012

MethodsParallel group single-centre unblinded RCT
Participants

130 patients with prostate cancer undergoing pelvic radiotherapy

Median age: 66 years

All male

Interventions

Group 1: lactose- and insoluble fibre-restricted diet (n = 64)

Group 2: no dietary change (n = 66)

OutcomesGI symptoms from the EORTC-QLQ-C30 and EORTC-QLQ-PR25, study-specific Gastrointestinal Side Effects Questionnaire (GISEQ), QoL was assessed using EORTC-QLQ-C30
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskEfron's biased coin design
Allocation concealment (selection bias)Low riskPerformed by administrative personnel outside the research project
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
Low risk 
Selective reporting (reporting bias)Low risk 
Other biasLow risk 

Stryker 1986

MethodsParallel group RCT
Participants

64 patients undergoing pelvic radiotherapy for either gynaecological, prostate or sigmoid cancers

Mean age (group 1): 61 years

Mean age (group 2): 58 years

Mean age (group 3): 57 years

Male:female 7:57

Interventions

Group 1: lactose-restricted diet with calcium replacement of 625 mg 3 times daily

Group 2: lactose-hydrolysed diet containing 480 mL of milk daily wherein 90% of the lactose hydrolysed to glucose and galactose

Group 3: normal diet containing at least 480 mL of milk daily (control)

OutcomesMean weekly stool frequency
NotesNeed to consider relevance of the study given the change in radiotherapy techniques over time and the age of this study
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot adequately described
Allocation concealment (selection bias)Unclear riskNot adequately described
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
High risk 
Selective reporting (reporting bias)Low risk 
Other biasUnclear riskTemporal bias may be relevant here as radiotherapy technique has progressed since this study was performed

Wedlake 2012

  1. a

    BMI: body mass index; EORTC-QLQ-C30: European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC-QLQ-PR25: European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire - prostate-specific module; GI: gastrointestinal; IBDQ: Inflammatory Bowel Disease Questionnaire; IBDQB: Inflammatory Bowel Disease Questionnaire Bowel subset; QoL: quality of life; RCT: randomised controlled trial; RTOG: Radiation Therapy Oncology Group; VIQ: Vaizey Incontinence Questionnaire.

MethodsParallel group RCT
Participants

117 patients

Mean age: 65 years

Male:female 79:38

Interventions

Group 1: low-fat diet

Group 2: modified-fat diet

Group 3: control

OutcomesDaily fat point consumption, IBDQB, VIQ and RTOG, weight, BMI, handgrip strength and QoL (total IBDQ score)
NotesNeed to consider generalisability of the study given the acceptability of the intervention to patients - 418 patients refused consent
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomisation by Cancer Research
Allocation concealment (selection bias)Low riskPermuted blocks
Blinding of participants and personnel (performance bias)
All outcomes
High riskNot blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskNot blinded
Incomplete outcome data (attrition bias)
All outcomes
High risk117 consented, 107 followed up at weeks 2 and 4 and only 75 followed up to 1 year
Selective reporting (reporting bias)Low riskAll outcomes reported
Other biasLow risk 

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Ahmad 2010Soy isoflavines is not a nutritional supplement as defined by the protocol
Bartalone 2009Incorrect outcomes as defined by the protocol
Bounous 1975Includes patients with pancreatic tumours
De Maria 1992Glutathione is not a nutritional supplement as defined by the protocol
Foster 1980Incorrect outcomes as defined by the protocol
Kinsella 1981Palliative patients included
Kozelsky 2003Glutamine is not a nutritional supplement as defined by the protocol
Lodge 1995Cross-over trial
Mateos Dominguez 2010Glutamine is not a nutritional supplement as defined by the protocol
McArdle 1986Methodologically unsound - started as randomised controlled trial then used retrospective control group
Membrive Conejo 2011Glutamine is not a nutritional supplement as defined by the protocol
Mucke 2004Selenium is not a nutritional supplement as defined by the protocol
Mucke 2006Selenium is not a nutritional supplement as defined by the protocol, contains a subset of participants included in Mucke 2004
Muecke 2010Selenium is not a nutritional supplement as defined by the protocol, contains a subset of participants included in Mucke 2004 and Mucke 2006
Ohshima 1977L-cystein is not a nutritional supplement as defined by the protocol
Rotovnik Kozjek 2011Glutamine is not a nutritional supplement as defined by the protocol
Shah 2011Multivitamins and minerals are not nutritional supplements as defined by the protocol
Valerio 1978Palliative patients included

Ancillary