Scolaris Content Display Scolaris Content Display

Reconstrucción en Y de Roux versus Billroth‐I tras gastrectomía distal por cáncer gástrico

Collapse all Expand all

Antecedentes

El cáncer gástrico es el quinto más diagnosticado en el mundo. Gracias a la mejoría en las tasas de detección precoz del cáncer gástrico y a los avances tecnológicos en los tratamientos, se ha conseguido una mejoría significativa en las tasas de supervivencia de las personas con cáncer sometidas a gastrectomía. Posteriormente, se ha puesto cada vez más énfasis en el síndrome posgastrectomía (p.ej., plenitud, retraso en el vaciado y sudoración fría, entre otros) y en la calidad de vida posoperatoria. Sin embargo, no se sabe con certeza qué tipo de reconstrucción produce mejores desenlaces después de la cirugía.

Objetivos

Evaluar la evidencia sobre la calidad de vida relacionada con la salud y los desenlaces de seguridad de las reconstrucciones en Y de Roux y Billroth‐I después de la gastrectomía distal para las personas con cáncer gástrico.

Métodos de búsqueda

Se realizaron búsquedas en la Biblioteca Cochrane y en el Registro Cochrane central de ensayos controlados (Cochrane Central Register of Controlled Trials; CENTRAL), MEDLINE y Embase el 4 de mayo de 2021. Se comprobaron las listas de referencias de los estudios incluidos y se estableció contacto con fabricantes y profesionales del sector. No hubo restricciones de idioma.

Criterios de selección

Ensayos controlados aleatorizados (ECA) que asignaron a los participantes a la reconstrucción en Y de Roux o a la reconstrucción Billroth‐I después de la gastrectomía distal por cáncer gástrico.

Obtención y análisis de los datos

Dos autores de la revisión examinaron de forma independiente los estudios identificados mediante la búsqueda para determinar su elegibilidad y extrajeron los datos. Los desenlaces principales fueron la calidad de vida relacionada con la salud tras la cirugía y la incidencia de fugas anastomóticas. Los desenlaces secundarios incluían la pérdida de peso corporal, la incidencia de reflujo biliar, la duración de la estancia hospitalaria y la morbilidad general. Se utilizó un modelo de efectos aleatorios para realizar los metanálisis. El riesgo de sesgo de los estudios incluidos se evaluó de acuerdo con el Manual Cochrane para revisiones sistemáticas de intervenciones y la certeza de la evidencia mediante el método GRADE.

Resultados principales

En la revisión se incluyeron ocho ECA (942 participantes). Un estudio incluyó tanto a pacientes con cáncer como a pacientes con enfermedades benignas, como las úlceras de estómago. Dos estudios compararon las reconstrucciones en Y de Roux, Billroth‐I y Billroth‐II, mientras que los otros estudios compararon directamente la Y de Roux y la Billroth‐I.

Para los desenlaces principales, la evidencia indica que podría haber poca o ninguna diferencia en la calidad de vida relacionada con la salud entre la reconstrucción en Y de Roux y la de Billroth‐I (diferencia de medias estandarizada 0,04; intervalo de confianza [IC] del 95%: ‐0,11 a 0,18; I² = 0%; seis estudios; 695 participantes; evidencia de certeza baja debido a las limitaciones de los estudios y a la imprecisión). La evidencia del efecto de la reconstrucción en Y de Roux versus la de Billroth‐I sobre la incidencia de fugas anastomóticas es muy incierta (razón de riesgos [RR] 0,63; IC del 95%: 0,16 a 2,53; I² = 0%; cinco estudios; 711 participantes; evidencia de certeza muy baja). La incidencia de fugas anastomóticas fue del 0,6% y el 1,4% en los grupos de Y de Roux y Billroth‐I, respectivamente.

Para los desenlaces secundarios, la evidencia indica que la reconstrucción con Billroth‐I podría dar lugar a poca o ninguna diferencia en la pérdida de peso corporal en comparación con la reconstrucción en Y de Roux (diferencia de medias [DM] 0,41; IC del 95%: ‐0,77 a 1,59; I² = 0%; cuatro estudios; 541 participantes; evidencia de certeza baja). La reconstrucción en Y de Roux probablemente reduce la incidencia de reflujo biliar en comparación con la reconstrucción Billroth‐I (RR 0,40; IC del 95%: 0,25 a 0,63; I² = 22%; cuatro estudios; 399 participantes; evidencia de certeza moderada). La reconstrucción Billroth‐I podría acortar la estancia hospitalaria posoperatoria, pero la evidencia de este desenlace es muy incierta (DM 0,96; IC del 95%: 0,16 a 1,76; I² = 56%; siete estudios; 894 participantes; evidencia de certeza muy baja). La reconstrucción con Billroth‐I podría reducir la morbilidad posoperatoria general en comparación con la reconstrucción en Y de Roux (RR 1,47; IC del 95%: 1,02 a 2,11; I² = 0%; siete estudios; 891 participantes; evidencia de certeza baja).

Conclusiones de los autores

Los datos indican que hay poca o ninguna diferencia entre la reconstrucción en Y de Roux y la de Billroth‐I en cuanto a la calidad de vida relacionada con la salud. La evidencia del efecto de la reconstrucción en Y de Roux versus la de Billroth‐I sobre la incidencia de fugas anastomóticas es muy incierta, ya que la incidencia de este desenlace fue baja. Aunque la certeza de la evidencia fue baja, se encontraron algunas diferencias posiblemente importantes desde el punto de vista clínico entre la reconstrucción en Y de Roux y la de Billroth‐I en los desenlaces a corto plazo. La reconstrucción en Y de Roux probablemente reduce la incidencia de reflujo biliar en el estómago remanente en comparación con la reconstrucción Billroth‐I. La reconstrucción Billroth‐I podría acortar la estancia hospitalaria posoperatoria en comparación con la reconstrucción en Y de Roux, pero la evidencia es muy incierta. La reconstrucción Billroth‐I podría reducir la morbilidad posoperatoria global en comparación con la reconstrucción en Y de Roux. Los ensayos futuros deberían incluir un seguimiento a largo plazo de la calidad de vida relacionada con la salud y la pérdida de peso corporal.

PICOs

Population
Intervention
Comparison
Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Resumen en términos sencillos

Reconstrucción en Y de Roux frente a Billroth‐I tras la cirugía por cáncer gástrico

Pregunta de la revisión

Se revisó la evidencia del efecto de la reconstrucción en Y de Roux en comparación con la reconstrucción Billroth‐I después de la cirugía para el cáncer gástrico. Se encontraron ocho estudios.

Antecedentes

El cáncer gástrico (de estómago) es uno de los más frecuentes en todo el mundo. Recientemente, los índices de detección precoz del cáncer gástrico y la tecnología de tratamiento han mejorado. Como resultado, las personas pueden sobrevivir más tiempo después de la cirugía y se ha debatido la importancia de la calidad de vida después de esta. Tanto Billroth‐I como Y de Roux son opciones para la reconstrucción de la continuidad del sistema gastrointestinal tras una gastrectomía distal (extirpación de la parte inferior del estómago).

Sin embargo, no existen normas sobre qué procedimiento reconstructivo seleccionar ya que las guías no describen qué procedimiento debe ser prioritario. Por lo tanto, era importante realizar una revisión de la evidencia disponible para ayudar en la toma de decisiones de las personas que se someten a la cirugía, los cirujanos, los médicos, el personal sanitario y los responsables de políticas sanitarias.

Características de los estudios

La evidencia está actualizada hasta el 4 de mayo de 2021.

Se identificaron ocho ensayos que incluían 942 participantes con cáncer gástrico que se sometieron a gastrectomía distal. Los estudios se realizaron en cuatro países. Un estudio incluyó a pacientes con cáncer y a pacientes con otras enfermedades (como úlceras de estómago). Dos estudios compararon la reconstrucción en Y de Roux, Billroth‐I y Billroth‐II, mientras que los demás estudios compararon directamente la Y de Roux y Billroth‐I. En cuanto a los enfoques quirúrgicos, en todos los estudios se utilizó cirugía abierta o laparoscópica (mínimo acceso), o ambas; no se utilizó la cirugía robótica. Se utilizaron seis escalas diferentes para medir la calidad de vida.

Resultados clave

La evidencia indica que la reconstrucción en Y de Roux podría dar lugar a una diferencia mínima o nula en la calidad de vida a los 12 meses después de la cirugía. Sin embargo, estos resultados se deben interpretar con cautela porque los investigadores del estudio midieron la calidad de vida de diferentes maneras. La evidencia es muy incierta en cuanto al efecto de las intervenciones sobre las fugas a través de la conexión reconstruida (fugas anastomóticas).

La reconstrucción Billroth‐I podría dar lugar a una diferencia escasa o nula en la pérdida de peso corporal; probablemente aumenta el reflujo biliar hacia el estómago remanente; y podría reducir las complicaciones generales tras la cirugía en comparación con la reconstrucción en Y de Roux. La evidencia es muy incierta en cuanto al efecto de las intervenciones sobre la duración de la estancia hospitalaria.

Certeza de la evidencia

La certeza de la evidencia para la calidad de vida fue baja debido a las limitaciones en la forma en que los estudios se diseñaron y realizaron, y porque no hay suficientes estudios para estar seguros de los resultados de este desenlace. La certeza de la evidencia para los otros desenlaces varió de muy baja a moderada. Es necesario investigar más a fondo los efectos de los métodos de reconstrucción durante un periodo de tiempo más prolongado.

Authors' conclusions

Implications for practice

The evidence suggests that there is little to no difference between Roux‐en‐Y and Billroth‐I reconstruction for the outcome health‐related quality of life. However, the certainty of the evidence is low. Roux‐en‐Y probably prevents bile reflux into the remnant stomach. Billroth‐I may reduce overall morbidity. Billroth‐I may also shorten length of hospital stay, although the evidence for this outcome is very uncertain. The certainty of the evidence was low or very low for all of the outcomes in this review except incidence of bile reflux, for which the certainty was moderate. The results of this review should therefore be interpreted in consideration of this.

Implications for research

In this review, health‐related quality of life was assessed only by disease‐specific scales, which are suitable for assessing health‐related quality of life with disease‐specific symptoms, but difficult for policymakers to use. The assessment with a preference‐based scale should also be discussed in future research. Moreover, a long‐term assessment of health‐related quality of life is needed.

We did not include an assessment of cost in the protocol because it was out of the scope of this Cochrane Review. None of the included studies reported the cost‐related outcome. Our review showed Billroth‐I reconstruction requires a shorter hospital stay than Roux‐en‐Y reconstruction. In the future, outcomes related to cost‐effectiveness should be assessed.

Summary of findings

Open in table viewer
Summary of findings 1. Roux‐en‐Y compared to Billroth‐I after distal gastrectomy for gastric cancer

Roux‐en‐Y compared to Billroth‐I after distal gastrectomy for gastric cancer

Patient or population: people undergoing distal gastrectomy for gastric cancer
Setting: operating room in a hospital
Intervention: Roux‐en‐Y
Comparison: Billroth‐I

Outcomes

Anticipated absolute effects* (95% CI)

Relative effect
(95% CI)

№ of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with Billroth‐I

Risk with Roux‐en‐Y

Health‐related quality of life

SMD 0.04 higher
(0.11 lower to 0.18 higher)

695
(6 RCTs)

⊕⊕⊝⊝
LOW1,2

Regarding the effect size of the SMD, 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect (Cohen 1988). Converting an SMD of 0.04 into a global health status score of EORTC QLQ‐C30, Roux‐en‐Y may increase it by 0.56 points (95% CI −1.53 to 2.50) (Murad 2019).

Incidence of anastomotic leakage

14 per 1000

9 per 1000
(2 to 36)

RR 0.63
(0.16 to 2.53)

711
(5 RCTs)

⊕⊝⊝⊝
VERY LOW3,4

Loss of body weight

The mean loss of body weight ranged from 8 to 9 percent.

The mean loss of body weight was 0.41% greater
(0.77 smaller to 1.59 greater).

541
(4 RCTs)

⊕⊕⊝⊝
LOW3,5

Loss of body weight is expressed as a percentage (e.g. if a person of 50 kg becomes 40 kg, the loss of body weight is 20%). Weight loss can theoretically range from negative infinity to 100%, but weight gain is uncommon after gastrectomy, and weight loss of more than 50% is also uncommon. The observed values are therefore usually expected to fall within the range of 0 to 50%.

Incidence of bile reflux

397 per 1000

159 per 1000
(99 to 250)

RR 0.40
(0.25 to 0.63)

399
(4 RCTs)

⊕⊕⊕⊝
MODERATE6

Length of hospital stay

The mean length of hospital stay ranged from 7 to 23 days.

The mean length of hospital stay was 0.96 days longer
(0.16 to 1.76 days longer).

894
(7 RCTs)

⊕⊝⊝⊝
VERY LOW3,7,8

Length of hospital stay is expressed in days, and ranges from 1 to infinite.

Postoperative morbidity

99 per 1000

146 per 1000
(101 to 209)

RR 1.47
(1.02 to 2.11)

891
(7 RCTs)

⊕⊕⊝⊝
LOW 3,9

*The risk in the intervention group (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; EORTC QLQ‐C30: EORTC Core Quality of Life Questionnaire ‐ Core Questionnaire; RCT: randomised controlled trial; RR: risk ratio; SMD: standardised mean difference

GRADE Working Group grades of evidence
High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

1Downgraded one level due to imprecision; 95% CIs are compatible with both benefit and harm.
2Downgraded one level due to risk of bias; participants were not blinded despite this being a patient‐reported outcome.
3Downgraded one level due to risk of bias; surgeons were not blinded
4Downgraded by two levels due to imprecision; 95% CIs are wide due to rare events, with possibility of substantial harm and substantial benefit, suggesting very serious imprecision.
5Downgraded one level due to imprecision; small sample size — outcome included only four studies.
6Downgraded one level due to risk of bias; outcome assessors may not have been blinded.
7Downgraded one level due to risk of bias; incomplete reporting on the results indicating data are skewed.
8Downgraded one level due to inconsistency; substantial heterogeneity was observed.
9Downgraded one level due to inconsistency; possible inconsistencies in trial results due to different diagnostic criteria used to report complications.

Background

Description of the condition

In 2012, an estimated one million cases of gastric cancer occurred worldwide, making it the fifth most common malignancy and the third leading cause of cancer death in the world (IARC 2015). More than 70% of gastric cancer occurred in low‐ and middle‐income countries; geographically, half the total gastric cancer worldwide was seen in Eastern Asia, and 17% in Europe (IARC 2015).

Gastrectomy is a surgical procedure to treat gastric cancer. Standard gastrectomy for gastric cancer is defined as a principal procedure to be performed with curative intent involving resection of at least two‐thirds of the stomach with a D2 lymphadenectomy (JGCA 2017a), which can improve disease‐specific survival (Mocellin 2015). A sufficient resection margin should be ensured to determine the resection line. Generally, a distal gastrectomy can be performed if a resection margin of at least 3 cm can be obtained in advanced cancer. If not, a total gastrectomy should be considered (JGCA 2017a).

Due to improved early detection rates of gastric cancer and technological advances in treatments (ASGE 2002), a significant improvement in survival rates has been achieved amongst people with cancer undergoing gastrectomy (Roukos 1999). Early gastric cancer limited to the mucosa can be resected endoscopically, whilst preserving the stomach. For other types of early cancers (e.g. submucosal invasion) and advanced cancers, however, surgical resection is still the central treatment for gastric cancer (JGCA 2017a). Surgeons have been aware of the importance of postgastrectomy syndrome including fullness, delayed emptying, and cold sweat, etc., and several studies have investigated reconstructive procedures following distal gastrectomy from the viewpoint of achieving a good quality of life postsurgery (Lee 2012Nakamura 2016Nunobe 2007Takiguchi 2012Terashima 2014).

See Appendix 1 for a glossary of terms.

Description of the intervention

The four reconstructive procedures commonly employed after a distal gastrectomy are gastroduodenostomy (Billroth‐I (B‐I)); gastrojejunostomy (with or without Braun anastomosis (Billroth‐II (B‐II))); Roux‐en‐Y gastrojejunostomy; and jejunal interposition (JGCA 2017a). Billroth‐I reconstruction has been historically accepted as a common procedure after distal gastrectomy in Japan (Yoshino 2000). Roux‐en‐Y reconstruction has been more common in Western countries (Kumagai 2012; Schwarz 2015), although recently Billroth‐II reconstruction has also been performed (Lee 2012; Tran 2016). Billroth‐I is considered a simple procedure, as it requires only one anastomosis (Kanaya 2011), thus it remains the mainstay of reconstruction in Japan (Hoya 2009; Kumagai 2012). Nevertheless, Roux‐en‐Y reconstruction is gradually becoming more popular in Japan due to adverse events associated with Billroth‐I reconstruction, including bile reflux and ulcer in the remnant gastric segment (Kojima 2008).

How the intervention might work

Billroth‐I reconstruction may have a physiologic advantage of letting food pass through the duodenum (Kalmar 2006), and could potentially improve postoperative weight loss (Terashima 2014). However, based on the findings of an observational study, Roux‐en‐Y reconstruction has proven superiority over Billroth‐I and Billroth‐II (Fukuhara 2002), as it may be beneficial in preventing bile reflux into the gastric remnant if an appropriate length of the Roux limb is utilised. However, the possibility of a postoperative internal hernia occurring due to the nature of the Roux‐en‐Y reconstruction procedure, which causes mesenteric defects, cannot be neglected (Kelly 2013). Additionally, the possibility of Roux‐Y stasis, which shows nausea or vomiting worsened by eating either solids or liquids without mechanical obstruction, can dissuade surgeons from opting for a Roux‐en‐Y reconstruction (Kojima 2008). There has been no definitive and straightforward rationale for selecting a particular reconstructive procedure. The surgeon's experience, training received, and institutional policies of hospitals or facilities at which they work may dictate the choice of procedure. Furthermore, medical insurance and equipment, which differ amongst nations, may also affect preferences.

Why it is important to do this review

To date, there are no standard criteria regarding which reconstructive procedure to select. Existing guidelines do not describe which procedure should be prioritised as it is unclear if any one procedure is superior to others in terms of short‐ and long‐term efficacy and safety (JGCA 2017a; Smyth 2016). It was therefore important to perform a review of the available evidence and to establish a body of evidence for people undergoing surgery, surgeons, physicians, medical staff, and policymakers.

Objectives

To assess the evidence on health‐related quality of life and safety outcomes of Roux‐en‐Y and Billroth‐I reconstructions after distal gastrectomy for people with gastric cancer.

Methods

Criteria for considering studies for this review

Types of studies

As opposed to previously published reviews which included observational studies (Xiong 2013; Zong 2011), we included only randomised controlled trials (RCTs), with no restrictions on language. We included studies reported as full text, those published as abstract only, and unpublished data. We did not apply any restriction with respect to blinding status because blinding for surgeons and participants, which prevents detection and performance biases, is typically difficult in the case of surgical RCTs, although the degree of such biases is unknown (Probst 2016); however, the blinding status of assessors should be clarified (Speich 2017). We planned to include cluster‐RCTs when available. Clinical studies with a cross‐over design are impossible for the interventions in question.

Types of participants

We included adults aged 18 years and older with a diagnosis of gastric cancer who were undergoing robotic, laparoscopic, laparoscopy‐assisted, or open distal gastrectomy.

Types of interventions

We included studies comparing Roux‐en‐Y with Billroth‐I reconstruction. When a primary anastomosis was constructed, it could either be performed as an intracorporeal procedure (stapled or hand‐sewn) or an extracorporeal procedure (stapled or hand‐sewn). We did not specify types of stapling devices: we included both linear and circular stapling.

Types of outcome measures

We assessed the following outcomes. Reporting of the outcomes listed here was not considered as inclusion criteria for the review.

Primary outcomes

  1. Health‐related quality of life after surgery (at time point closest to 12 months postoperatively), assessed by any generic and disease‐specific questionnaires on domains such as global health status and physical and mental well‐being.

  2. Incidence of anastomotic leakage (defined as Clavien‐Dindo grade II or greater) (Clavien 2009; Dindo 2004), within 30 days after surgery.

Secondary outcomes

  1. Loss of body weight, measured one year after surgery (range: 6 months to 24 months).

  2. Incidence of bile reflux, including remnant gastritis or reflux oesophagitis (at time point closest to 12 months postoperatively).

  3. Length of hospital stay (in days).

  4. Postoperative morbidity (Clavien‐Dindo grade II or greater), within 30 days after surgery.

Search methods for identification of studies

Electronic searches

We conducted a literature search to identify all published and unpublished RCTs. We placed no restrictions on the language of publication when searching the electronic databases. We translated non‐English language papers and assessed them thoroughly for potential inclusion in the review, as necessary.

We searched the following electronic databases up to 4 May 2021.

  1. The Cochrane Library databases, and the Cochrane Central Register of Controlled Trials (CENTRAL; to 2020 Issue 4; Appendix 2) (via Ovid).

  2. MEDLINE (1946 to 4 May 2021; Appendix 3) (via Ovid).

  3. Embase (1974 to 4 May 2021; Appendix 4) (via Ovid).

We also searched PubMed (www.ncbi.nlm.nih.gov/pubmed) before completing the review to obtain records that were not yet indexed in Ovid MEDLINE.

Searching other resources

We checked the reference lists of all primary studies for additional references that our original electronic searches may have missed. We searched clinical trials registers for unpublished data and planned to send a comprehensive list of relevant articles to the first author of reports of the included studies to enquire as to any further relevant studies (Higgins 2021). We also contacted manufacturers and experts in the field to ask if they knew of any ongoing and unpublished trials. We searched for errata or retractions from eligible studies on PubMed (www.ncbi.nlm.nih.gov/pubmed) and reported the date this was done in the review. We performed a citation search in the Web of Science to obtain articles citing the included studies.

We searched the following registers and grey literature sources up to 4 May 2021.

Grey literature databases

  1. OpenGrey (www.opengrey.eu).

Clinical trial registers and trial result registers

We thoroughly searched various clinical trial registers/trial result registers.

  1. US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov).

  2. Current Controlled Trials metaRegister of Controlled Trials (mRCT; www.isrctn.com ):

    1. active registers;

    2. archived registers.

  3. EU Clinical Trials Register (www.clinicaltrialsregister.eu).

  4. World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch).

  5. University Hospital Medical Information Network Clinical Trials Registry (UMIN‐CTR; www.umin.ac.jp/ctr).

We searched conference minutes of the following congress and annual meeting of societies for relevant abstracts.

  1. American Society of Clinical Oncology (ASCO); searched up to 2019.

  2. European Society of Medical Oncology (ESMO); searched up to 2018.

  3. Japan Society of Clinical Oncology (JSCO); searched up to 2019.

Data collection and analysis

Selection of studies

Two review authors (DN and RG) independently screened the titles and abstracts of all studies identified by our search, coding them as 'retrieve' (eligible, potentially eligible, or unclear) or 'do not retrieve'. We retrieved the full text of study reports or publications, and two review authors (DN and RG) independently screened the full text, identified studies for inclusion, and recorded reasons for exclusion of ineligible studies. Any disagreements were resolved through discussion or by consulting a third review author (NW) when necessary. We identified and excluded duplicates and collated multiple reports of the same study so that each study, rather than each report, was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and tabulated characteristics of excluded studies (Liberati 2009).

Data extraction and management

To determine study characteristics and outcome data, we used a standard data collection form that had been piloted on at least one study in the review. Two review authors (DN and RG) independently extracted the following study characteristics from included studies.

  1. Methods: study design, number of study centres and location, study setting, withdrawals, date of study.

  2. Participants: number, mean age, age range, gender, diagnostic criteria (for anastomotic leakage, bile reflux, and health‐related quality of life), inclusion and exclusion criteria.

  3. Interventions: intervention, comparison.

  4. Outcomes: primary and secondary outcomes specified and collected, time points reported.

  5. Notes: funding for study, notable conflicts of authors participating in the study.

Two review authors (DN and RG) independently extracted outcome data from the included studies. We presented the characteristics of included studies in the ‘ Characteristics of included studies ’ tables, as well as whether data were reported in an unusable way. Any disagreements were resolved through discussion or by involving a third review author (NW) when necessary. One review author (DN) copied the data from the data collection form into the Review Manager 5 file (Review Manager 2020). We double‐checked that data had been entered correctly by comparing the study reports with the presentation of data in the review. A second review author (RG) spot‐checked study characteristics for accuracy against the study report.

Assessment of risk of bias in included studies

Two review authors (DN and RG) independently assessed the risk of bias of each study, according to the criteria in Table 8.5.d of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Any disagreements were resolved by discussion or by involving a third review author (NW) when necessary. We assessed the risk of bias based on the following domains.

  1. Random sequence generation.

  2. Allocation concealment.

  3. Blinding of participants and personnel.

  4. Blinding of outcome assessment.

  5. Incomplete outcome data.

  6. Selective outcome reporting.

  7. Other potential bias (i.e. specific study design, deceptive study, co‐intervention, blinding of data assessors, baseline imbalance in participant characteristics, perioperative timing of randomisation, and the influence of funders).

We judged each potential source of bias as low, high, or unclear, and provided a quote from the study report and justification for our judgement in the risk of bias table.

We summarised the risk of bias judgements across studies for each of the domains listed. We considered blinding separately for different key outcomes where necessary (e.g. for unblinded outcome assessment, risk of bias for all‐cause mortality may be very different than for a participant‐reported health‐related quality of life scale). Where information on risk of bias related to unpublished data or correspondence with a trialist, we noted this in the risk of bias table.

We took into account the risk of bias for studies that contributed to a given outcome as part of the GRADE methodology when considering treatment effects.

Assessment of bias in conducting the review

We conducted the review based on the published protocol and reported any deviations from it in the Differences between protocol and review section of this review.

Measures of treatment effect

We analysed dichotomous data (incidence of anastomotic leakage, incidence of bile reflux, and postoperative morbidity) as a risk ratio (RR), and continuous data (health‐related quality of life, loss of body weight, and length of hospital stay) as a mean difference (MD) or standardised mean difference (SMD), using a 95% confidence interval (CI) in both cases. We ensured that higher scores for continuous outcomes had the same meaning for a particular outcome; explained the direction to the reader; and reported where the directions were reversed if this was necessary.

We deemed health‐related quality of life scores to be continuous (not arbitrarily categorised) and articulated them as SMDs with 95% CIs following the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). We delineated a positive SMD as showing better effects for quality of life.

We undertook meta‐analyses only where this was meaningful, that is if the treatments, participants, and the underlying clinical question were similar enough for pooling to make sense. Considering substantial heterogeneity of surgical outcomes, we used the inverse‐variance random‐effects model unless the events were sparse.

Length of hospital stay is known to be skewed data in most cases. Although some studies ignored the skewness of data and reported mean and standard deviation (SD) for length of hospital stay, trials reporting medians and interquartile ranges (IQRs) only indicated the data were most likely to be skewed (Altman 1996). For studies that did not report mean and SD but did report median and IQRs, we assumed the median was similar to the mean, and IQRs were approximately 1.35 SDs (Higgins 2021). When median and range were reported, we calculated the mean and SD from the median, range, and sample size (Hozo 2005). We excluded these studies in a sensitivity analysis.

Unit of analysis issues

We planned to include cluster‐RCTs in the analysis along with individually randomised trials. We planned to adjust the sample sizes of cluster‐RCTs using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021), with an estimate of the intracluster correlation coefficient (ICC) derived from the trial if possible, or from a similar trial or a study of a similar population. If we used ICCs from other sources, we would report this and conduct a sensitivity analysis to investigate the effect of variation in the ICC. If we identified both cluster‐RCTs and individually randomised trials, we would synthesise the relevant information. We considered it reasonable to combine the results from both if there was little heterogeneity amongst the study designs, and interaction between the effect of intervention and the choice of randomisation unit was considered to be unlikely.

If we included multiple‐arm studies (e.g. comparison between Roux‐en‐Y, Billroth‐I and ‐II), we formed a pair‐wise comparison between Roux‐en‐Y and Billroth‐I reconstruction in the meta‐analysis, and excluded the other arms.

Dealing with missing data

We contacted investigators or study sponsors to verify key study characteristics and to obtain missing numerical data pertaining to outcomes as indicated (e.g. when a study was identified as abstract only). If we were unable to obtain information from investigators or study sponsors, we imputed the mean from the median (i.e. consider median as the mean) and the SD from the standard error, interquartile range, or P values, based on the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). When only the median and range were reported, we imputed the mean and the SD calculated from the median, range, and sample size (Hozo 2005). We assessed the impact of including such studies in a sensitivity analysis and indicated our results. When we were unable to calculate the SD from any numerical data including standard error, IQR, range, or P values, we imputed SD as the highest SD in the remaining studies included in the outcome, as it appears safe to borrow SDs from other studies (Furukawa 2006; Higgins 2021), being mindful of the fact that this method of imputation decreases the weight of the studies in the meta‐analysis of MD and shifts the effect towards no effect for SMD.

Assessment of heterogeneity

We used the I² statistic to measure heterogeneity amongst studies in each analysis (Higgins 2003); we considered that an I² of 30% to 60% may represent moderate heterogeneity, and that an I² greater than 60% may represent substantial heterogeneity (Higgins 2021). Furthermore, we explored heterogeneity by prespecified subgroup analyses (Higgins 2021), regardless of the measurement of I² statistics. We additionally assessed heterogeneity by evaluating whether there was good overlap of CIs by visual inspection of the forest plots.

Assessment of reporting biases

If we are able to pool more than 10 studies, we would create and examine a funnel plot to explore possible publication biases. We planned to use Egger's test to determine the statistical significance of reporting bias (Egger 1997). We considered P < 0.05 to be a statistically significant reporting bias.

Data synthesis

We performed meta‐analyses based on the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions using Review Manager 5 (Higgins 2021; Review Manager 2020). We used a random‐effects model for analysis as we considered that the different studies estimated different, yet related intervention effects (DerSimonian 1986), especially pertaining to surgical outcomes.

Subgroup analysis and investigation of heterogeneity

We planned to carry out the following subgroup analyses.

  1. Open resection versus laparoscopic resection versus others.

  2. Early clinical‐stage gastric cancer (stage I) versus others according to the 15th edition of Japanese Classification of Gastric Carcinoma (JGCA 2017b).

We chose two characteristics for subgroup analyses: surgical approach and clinical stage. The rationale for the choice of characteristics was as follows: i) difference of surgical approach may affect the outcomes (Best 2016Vinuela 2012); and ii) the difference of clinical stage alters the extent of lymph node dissection (JGCA 2017a), which may affect the outcomes.

If the study included laparoscopic and laparoscopy‐assisted cases between 0% and 30%, 31% and 70%, 71% and 100%, and an unknown proportion, we classified them as 'open (laparotomy) study', 'mixed study', 'laparoscopic study', and 'unknown approach study', respectively. If the study included stage I cases between 0% and 30%, 31% and 70%, 71% and 100%, and an unknown proportion, we classified them as 'advanced stage study', 'mixed study', 'early‐stage study', and 'unknown stage study', respectively.

We planned to use the following outcomes in subgroup analyses. To avoid multiplicity issues, we focused on a total of three outcomes that included the two primary outcomes and one of the secondary outcomes.

  1. Health‐related quality of life after surgery (12 months postoperatively).

  2. Loss of body weight (6 to 24 months postoperatively).

  3. Incidence of anastomotic leakage (within 30 days postoperatively).

We used the formal statistical test for subgroup differences to test for subgroup interactions as described in Section 9.6.6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021).

Sensitivity analysis

We performed an a priori sensitivity analysis to assess the robustness of our conclusions that involved exclusion of studies with high risk of bias in random sequence generation, studies with high risk of bias in allocation concealment, and studies with high risk of attrition bias. We performed additional sensitivity analyses by excluding studies that contained stage IV participants for the outcomes of health‐related quality of life and loss of body weight. We also performed additional sensitivity analyses by excluding studies that reported only median and range in length of hospital stay; excluding studies with unclear usage of the Clavien‐Dindo classification in overall morbidity; and using a fixed‐effect model for the outcome with sparse events (anastomotic leakage).

Reaching conclusions

Our conclusions were only based on findings obtained from the quantitative or narrative synthesis of studies included in this review. We avoided making recommendations for practice. Our implications for research intended to give the reader a clear sense of the focus needed for future research and clarifications required for the uncertainties that remain in the field.

Summary of findings and assessment of the certainty of the evidence

We created a summary of findings table for the comparison Roux‐en‐Y versus Billroth‐I reconstruction with the following outcomes.

  1. Health‐related quality of life (at the time point closest to 12 months postoperatively).

  2. Incidence of anastomotic leakage (within 30 days postoperatively).

  3. Loss of body weight (6 to 24 months after surgery).

  4. Incidence of bile reflux (at time point closest to 12 months postoperatively).

  5. Length of hospital stay (in days).

  6. Postoperative morbidity (within 30 days after surgery).

We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the quality of the body of evidence based on studies that contributed data to the meta‐analyses for each outcome, classifying it as high, moderate, low, or very low. We used the methods and recommendations described in Section 8.5 and Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions, employing GRADEpro GDT software (GRADEpro GDT; Higgins 2011). Two review authors (DN, NH) independently assessed the certainty of the evidence. We justified all decisions to downgrade or upgrade the certainty of evidence in the footnotes, providing comments to aid the reader's understanding of the review where necessary. We considered whether there was additional outcome information that was not incorporated into the meta‐analyses, noted this in the comments, and stated if it supported or contradicted the information obtained from the meta‐analyses.

Results

Description of studies

Results of the search

Our initial search strategy conducted on 4 May 2021 yielded a total of 1559 articles, with one article identified from other sources (ASCO). Based on title and abstract screening, 17 articles appeared to meet the inclusion criteria and were obtained in full text. One article was written in Chinese, one in Italian, and the others in English. Two of the 17 full‐text articles were conference abstracts. Our check of the references of these 17 articles identified one article relevant to the current review. After full‐text screening, we excluded two articles and assessed one article as awaiting classification. We therefore included 15 articles on eight studies in the review (Figure 1).


Study flow diagram.

Study flow diagram.

We also contacted manufacturers of surgical equipment (Ethicon and Medtronic) to enquire as to whether they had data or literature, which revealed that they did not have this information. We found no grey literature in our search of OpenGrey. Searching other sources including ClinicalTrials.gov also revealed no relevant ongoing trials.

Included studies

See: Characteristics of included studies.

Design

All of the included trials employed randomisation at the level of the individual. We identified no cluster‐randomised studies.

Setting

Three studies were conducted in Japan (Ishikawa 2005; Nakamura 2016; Takiguchi 2012), three in South Korea (Choi 2017; Hur 2017; Lee 2012), one in China (Yang 2017), and one in Italy (D'Amato 1999). Six studies were single‐centre (Choi 2017; D'Amato 1999; Hur 2017; Ishikawa 2005; Lee 2012; Yang 2017), and two were multicentre (Nakamura 2016; Takiguchi 2012).

Participants

All of the included studies enrolled gastric cancer patients who underwent distal gastrectomy. Benign disease patients were also enrolled in one study (D'Amato 1999), and stage IV patients were enrolled in three studies (Ishikawa 2005; Takiguchi 2012; Yang 2017). The included studies enrolled a total of 942 participants.

Interventions

Two studies compared Roux‐en‐Y, Billroth‐I, and Billroth‐II reconstructions (D'Amato 1999; Lee 2012), whilst the remaining studies compared Roux‐en‐Y and Billroth‐I directly.

Surgical approaches

Open surgery (laparotomy), laparoscopic surgery, and robotic surgery are generally used in gastric cancer surgery. Open or laparoscopic approach was employed in all of the studies included in this review. Robotic approach was not employed.

Outcome measures
Health‐related quality of life

Six studies reported health‐related quality of life (HRQoL) (D'Amato 1999; Hur 2017; Lee 2012; Nakamura 2016; Takiguchi 2012; Yang 2017). One study did not report HRQoL, although HRQoL was prespecified in the study protocol (Choi 2017).

HRQoL was assessed using six scales, as follows.

  1. EORTC Core Quality of Life Questionnaire ‐ Core Questionnaire (EORTC QLQ‐C30) (0 to 100 points, higher scores mean better quality of life (QoL)).

  2. EORTC Quality of Life Questionnaire ‐ Gastric Cancer Module (EORTC QLQ‐STO22) (0 to 100 points, higher scores mean better QoL).

  3. Functional Assessment of Cancer Therapy‐General (FACT‐G) (summation of raw scores; higher scores indicate better QoL).

  4. Functional Assessment of Cancer Therapy‐Gastric (FACT‐Ga) (summation of raw scores; higher scores indicate better QoL).

  5. Dysfunction After Upper Gastrointestinal Surgery 20 (DAUGS20) (0 to 100 points, lower scores mean better QoL).

  6. Gastrointestinal Quality of Life Index (GIQLI) (0 to 144 points, higher scores mean better QoL) (Eypasch 1995).

EORTC QLQ‐C30 and FACT‐G are questionnaires in general for cancer patients, whilst EORTC QLQ‐STO22, FACT‐Ga, DAUGS20, and GIQLI are the specific modules for gastric cancer patients. EORTC QLQ‐STO22 is used to complement the EORTC QLQ‐30, and FACT‐Ga is used to complement the FACT‐G.

The included studies collected HRQoL at different time points, as follows:

  • every 3 months from preoperative baseline to 12 months after surgery (Yang 2017);

  • at baseline and six months after surgery, and reported the ratio of follow‐up QoL score to the baseline score (Hur 2017);

  • at baseline and 5 days, 6 months, and 12 months after surgery (Lee 2012);

  • at baseline, one year after surgery and three years after surgery, and reported the difference from baseline to follow‐up in the intervention and control groups (Nakamura 2016);

  • at three months after the last case had been registered (Takiguchi 2012).

One study did not report the time point of data collection (D'Amato 1999).

We contacted the trial authors of two studies because the original data were necessary for meta‐analysis (Hur 2017; Nakamura 2016), and were provided this information.

Incidence of anastomotic leakage

Seven studies reported incidence of anastomotic leakage (Choi 2017; Hur 2017; Ishikawa 2005; Lee 2012; Nakamura 2016; Takiguchi 2012; Yang 2017).

Loss of body weight

Patient's body weight was reported as body weight itself (Ishikawa 2005; Nakamura 2016; Takiguchi 2012), or as body mass index (Choi 2017). One study reported body weight change at six months after surgery (Ishikawa 2005); two studies reported body weight change one year after surgery (Choi 2017; Takiguchi 2012); whilst the last study reported body weight change three years after surgery but also measured body weight one year after surgery (Nakamura 2016). We contacted the author of this trial (Nakamura 2016), who provided us the data for body weight change one year after surgery.

Incidence and severity of bile reflux, including remnant gastritis and reflux oesophagitis

Seven studies performed endoscopic investigation after gastric surgery: two studies six months after surgery (Hur 2017; Ishikawa 2005), two studies 12 months after surgery (Lee 2012; Yang 2017), one study 36 months after surgery (Nakamura 2016), and one study at both 12 months and five years after surgery (Takiguchi 2012); the remaining study did not report the time of investigation (D'Amato 1999).

Bile reflux was assessed endoscopically by RGB (Residue, Gastritis, Bile) classification, Kubo 2002, in four studies (Hur 2017; Lee 2012; Nakamura 2016; Yang 2017).

Reflux oesophagitis was assessed endoscopically by Los Angeles classification in four studies (Ishikawa 2005; Lee 2012; Nakamura 2016; Takiguchi 2012).

Length of hospital stay

Seven studies reported length of hospital stay. Definition of 'hospital stay' differed amongst studies. Postoperative hospital stay was used in four studies (Ishikawa 2005; Nakamura 2016; Takiguchi 2012; Yang 2017), and hospital stay was used in three studies (Choi 2017; Hur 2017; Lee 2012).

Postoperative morbidity (Clavien‐Dindo grade II or greater), within 30 days after surgery

Seven studies reported postoperative morbidity. Only four studies used the Clavien‐Dindo classification (Choi 2017; Hur 2017; Nakamura 2016; Yang 2017). Three studies did not mention using the Clavien‐Dindo classification (Ishikawa 2005; Lee 2012; Takiguchi 2012); all the reported surgical complications in one of these studies could be deemed as Clavien‐Dindo grade II or greater (Ishikawa 2005); however, the other two studies might comprise complications equivalent to Clavien‐Dindo grade I (Lee 2012; Takiguchi 2012).

Excluded studies

We excluded two studies for which the randomisation process was conditional (Osugi 2004; Wang 2011). These studies compared Billroth‐I, Billroth‐II, and Roux‐en‐Y. However, randomisation was performed between Roux‐en‐Y and Billroth‐II only when Billroth‐I was impossible (Characteristics of excluded studies).

We assessed one study as awaiting classification as information regarding inclusion criteria and whether the study was appropriately conducted was lacking (Ren 2019). This study compared Billroth‐I, Billroth‐II, and Roux‐en‐Y. We contacted the trial author by email for further information, but received no response.

Risk of bias in included studies

Graphical presentations of risk of bias in the included studies are presented in Figure 2 and Figure 3. Any disagreements between the two review authors (DN and RG) on risk of bias assessment were resolved by discussion.


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

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


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

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

See: Characteristics of included studies.

Allocation

Random sequence generation

All of the included studies were described as randomised studies. However, three studies did not mention the method of random sequence generation (D'Amato 1999; Hur 2017; Ishikawa 2005). The other studies employed appropriate methods for sequence generation.

Allocation concealment

Three studies adequately concealed allocation by sealed opaque envelopes or central allocations (Nakamura 2016; Takiguchi 2012; Yang 2017). The other five studies provided insufficient information about the method of concealment to make a judgement of low or high risk of bias.

Blinding

Blinding of participants and personnel

All of the included trials were at high risk of performance bias due to the nature of the intervention because surgeons cannot perform reconstruction without knowing the allocation. There was also no mention of efforts to keep the other clinical staff blinded to the treatment allocation in any of the included studies. As for blinding of participants, two studies clearly stated that participants were blinded (Choi 2017; Lee 2012). Three studies stated that no blinding for participants was done, although they focused on patient‐reported outcomes (Hur 2017; Nakamura 2016; Yang 2017). The other three studies did not report blinding status for participants (D'Amato 1999; Ishikawa 2005; Takiguchi 2012).

Blinding of outcome assessment

Since one of the primary outcomes in this review was patient‐reported, blinding for participants was directly linked to detection bias. We assessed four studies as at high risk of bias because participants were not blinded, although their outcomes were patient‐reported (Choi 2017; Hur 2017; Nakamura 2016; Yang 2017). One study that reported blinding for participants described in their study that "all medical documents were recorded without specifying the actual surgical approach" (Lee 2012). However, it would seem to be impossible to make medical records whilst concealing the procedures performed because surgeons generally make operative records. We therefore judged this study as at high risk of bias (Lee 2012). We assessed the other three studies as at unclear risk of bias because they did not report blinding status for outcome assessors or participants making a judgement (D'Amato 1999; Ishikawa 2005; Takiguchi 2012).

Incomplete outcome data

We judged two studies as at low risk of bias because they had a small number of dropouts and these were balanced in numbers and reasons (Choi 2017; Ishikawa 2005). We assessed one study as at unclear risk of bias because the reasons for attrition were not clearly described, although the reasons for exclusion were clearly stated, and the numbers of dropouts were small and well balanced in the intervention and control arms (Takiguchi 2012). We also judged the remaining five studies as at unclear risk of bias because they did not mention missing numbers or reasons for attrition (D'Amato 1999; Hur 2017; Lee 2012; Nakamura 2016; Yang 2017).

Selective reporting

Five studies had been registered in clinical trial databases before the commencement of trials (Choi 2017; Hur 2017; Nakamura 2016; Takiguchi 2012; Yang 2017). Of these, four studies reported outcomes adequately (Hur 2017; Nakamura 2016; Takiguchi 2012; Yang 2017), whilst one study did not report some prespecified outcomes and was therefore rated as at high risk of bias (Choi 2017).

Other potential sources of bias

The timing of randomisation is an important issue in surgical trials because preoperative randomisation allows surgeons to know allocation before resection, which may introduce an element of bias. Preoperative randomisation was used in three trials, which we deemed as at high risk of bias (Choi 2017; Hur 2017; Ishikawa 2005); intraoperative randomisation was used in three trials (Nakamura 2016; Takiguchi 2012; Yang 2017); and the other two trials did not report the timing of randomisation (D'Amato 1999; Lee 2012).

We suspected baseline imbalances in one study because stage IV patients were only included in the Roux‐en‐Y group, which may introduce an element of bias (Takiguchi 2012). Moreover, the acquisition method of HRQoL data in this study may have resulted in large variation in duration from surgery to the questionnaire survey, which could distort the outcome when attrition bias existed. We concluded that this variation was unclear, but considering baseline imbalances, we judged this study to be at high risk of bias (Takiguchi 2012).

In one study the proportion of participants who underwent postoperative chemotherapy differed between the Roux‐en‐Y group and the Billroth‐I group (Yang 2017). This may introduce co‐intervention bias, therefore we judged this study as at high risk of bias (Yang 2017).

There seemed to be little influence of funders in the included studies.

Effects of interventions

See: Summary of findings 1 Roux‐en‐Y compared to Billroth‐I after distal gastrectomy for gastric cancer

Primary analysis

1.1 Primary outcomes
1.1.1 Health‐related quality of life

Six studies contributed to meta‐analysis (D'Amato 1999Hur 2017Lee 2012Nakamura 2016Takiguchi 2012Yang 2017). Different scales were used to investigate HRQoL. We prioritised the use of scales to measure HRQoL including gastric cancer‐specific symptoms. If such a scale was not used, the scores for global health status were adopted. Lower scores indicate better quality of life in DAUGS20, whilst higher scores indicate better quality of life in FACT‐Ga, GIQLI, and global health status of QLQ‐C30. The timing of QoL acquisition in contributing studies was at six months after surgery in one study (Hur 2017), one year after surgery in three studies (Lee 2012Nakamura 2016Yang 2017), unknown in one study (D'Amato 1999), and at various time points after surgery in one study (Takiguchi 2012).

The evidence suggests that there is little to no difference in health‐related quality of life between Roux‐en‐Y and Billroth‐I reconstructions (standardised mean difference (SMD) 0.04, 95% confidence interval (CI) −0.11 to 0.18, P = 0.64, I² = 0%; 6 studies; 695 participants; low‐certainty evidence; Analysis 1.1) (Figure 4). We downgraded the certainty of the evidence by two levels due to study limitations (lack of blinding of participants and outcome assessment) and imprecision (95% CI overlapped no effect and was compatible both with benefit and harm).


Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.1 Health‐related quality of life.

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.1 Health‐related quality of life.

1.1.2 Incidence of anastomotic leakage

Five studies contributed to meta‐analysis (Hur 2017Ishikawa 2005Lee 2012Nakamura 2016Takiguchi 2012). Only five cases of anastomotic leakage were reported. Incidence was 0.6% and 1.4% in the Roux‐en‐Y and Billroth‐I groups, respectively. Two studies reported anastomotic leakage as zero (Choi 2017Yang 2017). The evidence is very uncertain for the effect of Roux‐en‐Y and Billroth‐I reconstruction on incidence of anastomotic leakage (risk ratio (RR) 0.63, 95% CI 0.16 to 2.53, P = 0.51, I² = 0%; 5 studies; 711 participants; very low‐certainty evidence; Analysis 1.2) (Figure 5). We downgraded the certainty of the evidence by three levels due to risk of bias (lack of blinding) and imprecision (wide confidence interval because of rare events; a possibility of substantial harm and substantial benefit).


Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.2 Incidence of anastomotic leakage.

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.2 Incidence of anastomotic leakage.

1.2 Secondary outcomes
1.2.1 Loss of body weight

Four studies contributed to meta‐analysis (Choi 2017Ishikawa 2005Nakamura 2016Takiguchi 2012). We chose the ratio of the postoperative body weight to the preoperative body weight as the outcome measure. Since one study did not report the precise value of SD (Ishikawa 2005), which was drawn as an error bar in the figure, we imputed the highest SD amongst the included studies in the analysis (Nakamura 2016), according to the method prespecified in the protocol. The evidence suggests that Billroth‐I reconstruction reduces loss of body weight by 0.41% compared to Roux‐en‐Y reconstruction, but it may increase loss of body weight by 0.77%, or reduce it by 1.59% (mean difference (MD) 0.41, 95% CI −0.77 to 1.59, P = 0. 50, I² = 0%; 4 studies; 541 participants; low‐certainty evidence; Analysis 1.3) (Figure 6). We downgraded the certainty of the evidence by two levels due to risk of bias (lack of blinding) and imprecision (small sample size).


Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.3 Loss of body weight.

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.3 Loss of body weight.

1.2.2 Incidence of bile reflux, including remnant gastritis or reflux oesophagitis

Incidence of bile reflux into remnant stomach

Four studies contributed to meta‐analysis (Hur 2017Lee 2012Nakamura 2016Yang 2017). Roux‐en‐Y reconstruction probably reduces the incidence of bile reflux compared to Billroth‐I reconstruction (RR 0.40, 95% CI 0.25 to 0.63, P < 0.001, I² = 22%; 4 studies; 399 participants; moderate‐certainty evidence; Analysis 1.4). No substantial heterogeneity was observed. We downgraded the certainty of the evidence by one level because of study limitations (lack of blinding of outcome assessment and incomplete outcome data).

Incidence of reflux oesophagitis

Four studies reported reflux oesophagitis using the Los Angeles classification (Ishikawa 2005Lee 2012Nakamura 2016Takiguchi 2012). One study reported this outcome both one year and five years after surgery (Takiguchi 2012); we employed the result of one year after surgery. Heterogeneity was substantial (I² = 65%), and the CIs did not overlap well upon visual inspection. There was inconsistency in the direction of effect, and we judged that quoting an average value for the intervention effect could be misleading.

1.2.3 Length of hospital stay (days)

One study reported hospital stay in median and range as 11 days (range: 7 to 88 days) in the Roux‐en‐Y group and 11 days (range: 7 to 63 days) in the Billroth‐I group (Nakamura 2016). We calculated the mean and SD from the median, range, and sample size (Hozo 2005), and imputed them. Subsequently, seven studies contributed to the meta‐analysis (Choi 2017Hur 2017Ishikawa 2005Lee 2012Nakamura 2016Takiguchi 2012Yang 2017). Billroth‐I reconstruction may shorten hospital stay compared to Roux‐en‐Y reconstruction, but the evidence is very uncertain (MD 0.96, 95% CI 0.16 to 1.76, P = 0.02; 7 studies; 894 participants; very low‐certainty evidence; Analysis 1.5). There was moderate heterogeneity (I² = 56%). We downgraded the certainty of the evidence by three levels because of inconsistency and study limitations (lack of blinding and incomplete reporting on results which indicates the data are skewed).

1.2.4 Postoperative morbidity (Clavien‐Dindo grade II or greater), within 30 days after surgery

Seven studies contributed to meta‐analysis (Choi 2017Hur 2017Ishikawa 2005Lee 2012Nakamura 2016Takiguchi 2012Yang 2017). Billroth‐I reconstruction may reduce postoperative morbidity within 30 days after surgery compared to Roux‐en‐Y reconstruction (RR 1.47, 95% CI 1.02 to 2.11, P = 0.04, I² = 0%; 7 studies; 891 participants; low‐certainty evidence; Analysis 1.6). We downgraded the certainty of the evidence by two levels because of the trial design (lack of blinding) and inconsistency. Details of the reported complications were summarised in Table 1.

Open in table viewer
Table 1. Details of complications reported in included studies

Study

Roux‐en‐Y

Affected participants in Roux‐en‐Y

Billroth‐I

Affected participants in Billroth‐I

Choi 2017

1 Atelectasis

1 Bleeding

2/20

1 Atelectasis

1 Complicated fluid

2/20

Hur 2017

3 Postoperative ileus

1 Pneumonia

1 Leakage

1 Wound seroma

1 Voiding difficulty

6/58

1 Postoperative ileus

1 Pneumonia

1 Intra‐abdominal fluid

1 Cholecystitis

1 Voiding difficulty

2 Unknown fever

7/56

Yang 2017

10 Pulmonary complications

1 Acute cholecystitis

2 Superficial surgical site infection

2 Intra‐abdominal infection

1 Adhesive ileus

1 Acute urinary retention

1 Gastroplegia

14/70

10 Pulmonary complications

1 Acute cholecystitis

1 Acute urinary retention

1 Gastroplegia

11/70

Nakamura 2016

4 Delayed gastric emptying

1 Pancreatic fistula

2 Anastomotic stricture

1 Pancreatic fistula

1 Postoperative bleeding

9/59

1 Surgical site infection

2 Anastomotic leakage

1 Anastomotic stricture

4/60

Lee 2012

1 Anastomotic leakage

2 Gastorojejunostomy outlet obstruction

2 Intra‐abdominal abscess

1 Deep vein thrombosis

6/47

2 Bleeding

1 Wound problem

1 Chyle leakage

4/49

Takiguchi 2012

3 Pancreatic fistula

3 Abdominal abscess

2 Bowel obstruction

2 Postoperative pancreatitis

2 Surgical site infection

2 Anastomotic stricture

23/169

2 Pancreatic fistula

2 Anastomotic leakage

3 Abdominal abscess

1 Bowel obstruction

2 Postoperative pancreatitis

3 Surgical site infection

3 Anastomotic stricture

14/163

Ishikawa 2005

3 Gastric stasis

1 Intestinal obstruction

1 Pneumonia

2 Anastomotic stricture

6/24

1 Leakage

1 Intestinal obstruction

2/26

See summary of findings Table 1 for the results of the primary analysis.

Secondary analysis

We categorised the included trials considering disease stages and surgical approaches for subgroup and sensitivity analysis, according to the prespecified methods. Regarding proportion of stage I patients included, we categorised four studies as 'early‐stage study' (Choi 2017Hur 2017Nakamura 2016Takiguchi 2012); two studies as 'mixed‐stage study' (Ishikawa 2005Yang 2017); and two studies as 'unknown‐stage study' (D'Amato 1999Lee 2012).

Regarding surgical approach, we judged two studies as 'laparoscopic study' (Choi 2017Hur 2017); two studies as 'mixed study' (Lee 2012Nakamura 2016); two studies as 'open (laparotomy) study' (Takiguchi 2012Yang 2017); and two studies as 'unknown study' (D'Amato 1999Ishikawa 2005).

1.3 Subgroup analysis and investigation of heterogeneity
1.3.1 Health‐related quality of life

We judged two studies as open (laparotomy) studies (Takiguchi 2012Yang 2017). If limited to open (laparotomy) studies, the estimated SMD was 0.15 (95% CI −0.05 to 0.34, P = 0.14), greater than the primary analysis. In contrast, the estimated SMD of the laparoscopic studies was −0.21 (95% CI −0.59 to 0.17, P = 0.28), and that of mixed and unknown studies was −0.07 (95% CI −0.35 to 0.22, P = 0.66). The SMD of laparoscopic studies had the opposite direction to that of the open studies (Analysis 2.1). The test for subgroup analysis did not show significant heterogeneity (I² = 39.9%).

We judged three studies as early‐stage studies (Hur 2017Nakamura 2016Takiguchi 2012). If limited to these three studies, the estimated SMD was −0.02 (95% CI −0.20 to 0.17, P = 0.87) (Analysis 2.3), which had the opposite direction to the primary analysis. Test for subgroup analysis did not show heterogeneity (I² = 0%).

1.3.2 Loss of body weight

We judged one study as an open (laparotomy) study (Takiguchi 2012); its estimated MD was 0.60 (95% CI −0.87 to 2.07, P = 0.42) (Analysis 2.2). Test for subgroup analysis did not show heterogeneity (I² = 0%). Regarding disease stage, we categorised three studies as early‐stage studies (Choi 2017Nakamura 2016Takiguchi 2012). If limited to these studies, the estimated MD was 0.37 (95% CI −0.86 to 1.59, P = 0.56) (Analysis 2.4).

1.3.3 Incidence of anastomotic leakage

Since the incidence of anastomotic leakage was sparse, we did not conduct the prespecified subgroup analysis.

1.4 Sensitivity analysis
1.4.1 Studies without selection and attrition bias

We judged no studies as at high risk of bias for random sequence generation, allocation concealment, or attrition bias.

1.4.2 Studies without stage IV patients

Three studies included stage IV cancer patients (Ishikawa 2005Takiguchi 2012Yang 2017), whose percentages were 2.0%, 1.5%, and 5.0%, respectively. Due to the small proportion of stage IV patients, we did not conduct the prespecified sensitivity analysis.

1.4.3 Studies without skewed data

In primary analyses, length of hospital stay (days) had a moderate amount of heterogeneity (I² = 56%). After excluding the study reporting only median value and range, Nakamura 2016, and the study under suspicion of being skewed with large SD (Ishikawa 2005), the I² statistic was 13% (Analysis 3.1).

1.4.4 Studies without unclear usage of Clavien‐Dindo classification

In the primary analysis, three studies were unclear regarding the use of the Clavien‐Dindo classification. One study reported complications in detail (Ishikawa 2005), and all the events reported can be deemed Clavien‐Dindo grade II or greater. Although the other two studies did not report their events in detail (Lee 2012Takiguchi 2012), we deemed their reported events as equivalent to the Clavien‐Dindo classification grade II or greater. As a sensitivity analysis, we excluded two studies (Lee 2012Takiguchi 2012). As a result, the difference shown in the primary analysis did not remain statistically significant (Analysis 3.2). This sensitivity analysis was not prespecified in the protocol.

1.4.5 Anastomotic leakage with a fixed‐effect model

In the primary analysis, we used a random‐effects model to assess anastomotic leakage, although events were sparse. However, use of the random‐effects model for rare events should be avoided (Higgins 2021), therefore we performed a post hoc sensitivity analysis using a fixed‐effect model (Analysis 3.3). When the fixed‐effect model was used, the magnitude and the direction of the result did not change, whilst the confidence interval narrowed.

1.4.6 Studies without benign disease patients

One study included participants with benign disease, whose percentage was 55% (D'Amato 1999). The population in which participants with benign diseases were predominantly included may differ from that of only gastric cancer; hence, we performed a post hoc sensitivity analysis on health‐related quality of life. After excluding this study, the magnitude and the direction of the result did not change, whilst the confidence interval widened (Analysis 3.4).

1.4.7 Studies without patients with diabetes mellitus

All the participants in one study had type 2 diabetes mellitus (Choi 2017). Since the characteristics of participants may be different from those of the other included studies, we performed a post hoc sensitivity analysis on body weight loss. After excluding this study, the direction of the result did not change. Still, the magnitude of the result became larger in favour of Billroth‐I (Analysis 3.5).

1.4.8 Studies without co‐intervention bias

In one study (Yang 2017), 62.9% of participants in the Roux‐en‐Y group underwent adjuvant chemotherapy, whilst 47.1% of the Billroth‐I group did. This imbalance may bring about co‐intervention bias, thus we performed a post hoc sensitivity analysis on health‐related quality of life. After excluding this study, the point estimate changed slightly toward favouring Billroth‐I, with a wider confidence interval (Analysis 3.6).

Discussion

Summary of main results

This systematic review and meta‐analysis assessed the evidence on health‐related quality of life and safety outcomes of Roux‐en‐Y reconstruction for a short‐term phase compared with Billroth‐I reconstruction after distal gastrectomy for people with gastric cancer.

With regard to our primary outcomes, the evidence suggests that there is little to no difference between reconstruction methods for the outcome health‐related quality of life. The evidence is very uncertain regarding the effect of Roux‐en‐Y versus Billroth‐I reconstruction on incidence of anastomotic leakage. Subgroup analysis implied that the difference in surgical approaches might affect the direction of the effect for health‐related quality of life, although this was not statistically significant.

For our secondary outcomes, Billroth‐I reconstruction may reduce loss of body weight by 0.41% compared to Roux‐en‐Y reconstruction one year after surgery, but it may increase loss of body weight by 0.77%, or reduce it by 1.59%.

Roux‐en‐Y reconstruction probably reduces the incidence of bile reflux into the remnant stomach; however, the incidence of bile reflux into the remnant stomach was assessed only by endoscopic investigation, thus the difference might not be clinically relevant.

In contrast, Billroth‐I reconstruction may shorten length of hospital stay, but the evidence is very uncertain. Furthermore, Billroth‐I reconstruction may reduce overall morbidity. The sensitivity analyses suggested that the difference in length of hospital stay was robust, but the risk ratio in overall morbidity was not. Although not robust, the risk ratio was 1.47 in the primary analysis, which may be clinically meaningful. The average difference in length of hospital stay was 0.96 days, which may not be relevant to an individual patient. The value for such difference may differ amongst people and depend on the medical system of the individual country.

The certainty of the evidence was low for health‐related quality of life, moderate for incidence of bile reflux, very low for length of hospital stay and incidence of anastomotic leakage, and low for the other outcomes.

Overall completeness and applicability of evidence

Participants

Seven out of the eight included studies were conducted in East‐Asian countries. Hence, most of the participants in the current review could be Asian. This might weaken the applicability of the evidence.

Interventions

The current review includes eight studies conducted from 1995 to 2014. Surgical equipment and technique have changed over these 20 years. Recent studies tended to employ laparoscopic surgery. Moreover, methods for reconstruction varied amongst studies, as some trials employed hand‐sewn sutures, whilst others employed staplers. These differences might affect the applicability of the evidence.

Outcomes

Given that the Clavien‐Dindo classification was introduced in 2004 (Dindo 2004), some trials that started before 2004 did not employ this classification. The Clavien‐Dindo classification is currently the standard criteria for evaluating surgical complications (Clavien 2009). If appropriately used, Clavien‐Dindo grade II or greater precisely shows complications requiring additional treatment. When surgical complications are reported without use of this classification, complications may be overestimated because those that do not require additional treatment cannot be discriminated. In the current review, we included all studies reporting morbidity in the primary analysis. In the sensitivity analysis, we excluded two studies in which use of the Clavien‐Dindo classification was uncertain; the results were not statistically significant. Although the result of the primary analysis and the sensitivity analysis had the same direction, this led to downgrading of the certainty of the evidence.

Quality of the evidence

We summarised the certainty of the evidence of each outcome in the Effects of interventions section according to the GRADE approach.

Limitations in study design and implementation

Firstly, we assessed all of the included studies as at high risk of bias for blinding of personnel. This is unavoidable because surgeons cannot perform surgery without knowing the allocated procedure. Secondly, six studies measured and reported patient‐reported outcomes, but all six studies were not judged as at low risk of bias for outcome assessment. Patient‐reported outcomes should be assessed with adequate participant blinding. Thirdly, two studies reported their length of hospital stay using potentially skewed data.

Indirectness of the evidence

Firstly, bile reflux was assessed by endoscopic investigation. Roux‐en‐Y was significantly more effective in reducing the incidence of bile reflux into the remnant stomach. This evidence was based on endoscopic findings that are objective but could be trivial for patients without symptoms caused by bile reflux. Secondly, the study conducted in Europe comprised patients both with benign disease and with gastric cancer; almost 50% of participants in this study had gastric cancer, whilst the other studies included only patients with gastric cancer. This might affect the outcomes with regard to cancer‐specific symptoms.

Inconsistency of the results

We reviewed the incidence of reflux oesophagitis narratively since there was inconsistency in the results from each trial. This outcome could be affected by the missing of participants.

Imprecision of the results

Incidence of anastomotic leakage was sparse. When events are rare, the results of meta‐analysis can be biased (Higgins 2021). As described above, this led to downgrading of the certainty of the evidence.

Publication bias

We did not draw funnel plots, as we included fewer than 10 studies. We were unable to determine whether or not there was a publication bias. However, one study did not report QoL in the paper (Choi 2017), although the acquisition of QoL was stated in the protocol, which raises the suspicion of publication bias.

Potential biases in the review process

We followed Cochrane searching methodology. As a result of our initial search, we obtained non‐English reports as well as English‐language reports.

The strength of this review is the acquisition of unpublished data from trial authors, as recommended in the Cochrane Handbook for Systematic Reviews of Interventions. We obtained unpublished data for HRQoL and loss of body weight and performed data synthesis with this information.

The limitation of this review is that we included only eight studies.

Agreements and disagreements with other studies or reviews

Several reviews have been published recently investigating the efficacy of reconstruction methods. Xiong 2013 included RCTs and observational studies but assessed them separately. Their pooled analysis of RCTs had the same concept as ours but did not include newly published RCTs. Zong 2011 included observational studies and RCTs, which was different from our methodology. A newly published network meta‐analysis, Cai 2018, assessing Roux‐en‐Y, Billroth‐I, and Billroth‐II included only RCTs, but the results of the risk of bias assessment differed from our results. The newest review, Kim 2019, also assessed three reconstruction methods and included mainly observational studies. None of these reviews focused on health‐related quality of life. Since treatments should be evaluated not only by surgical outcomes or endoscopic findings but also by general functioning and quality of life, we focused our review on HRQoL as one of the primary outcomes, making our review different from the others.

Study flow diagram.

Figures and Tables -
Figure 1

Study flow diagram.

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

Figures and Tables -
Figure 2

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

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

Figures and Tables -
Figure 3

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

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.1 Health‐related quality of life.

Figures and Tables -
Figure 4

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.1 Health‐related quality of life.

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.2 Incidence of anastomotic leakage.

Figures and Tables -
Figure 5

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.2 Incidence of anastomotic leakage.

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.3 Loss of body weight.

Figures and Tables -
Figure 6

Forest plot of comparison: 1 Roux‐en‐Y versus Billroth‐I reconstruction, outcome: 1.3 Loss of body weight.

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 1: Health‐related quality of life

Figures and Tables -
Analysis 1.1

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 1: Health‐related quality of life

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 2: Incidence of anastomotic leakage

Figures and Tables -
Analysis 1.2

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 2: Incidence of anastomotic leakage

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 3: Loss of body weight

Figures and Tables -
Analysis 1.3

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 3: Loss of body weight

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 4: Incidence of bile reflux

Figures and Tables -
Analysis 1.4

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 4: Incidence of bile reflux

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 5: Length of hospital stay

Figures and Tables -
Analysis 1.5

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 5: Length of hospital stay

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 6: Postoperative morbidity

Figures and Tables -
Analysis 1.6

Comparison 1: Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 6: Postoperative morbidity

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 1: Health‐related quality of life based on surgical approach

Figures and Tables -
Analysis 2.1

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 1: Health‐related quality of life based on surgical approach

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 2: Loss of body weight based on surgical approach

Figures and Tables -
Analysis 2.2

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 2: Loss of body weight based on surgical approach

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 3: Health‐related quality of life based on cancer stage

Figures and Tables -
Analysis 2.3

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 3: Health‐related quality of life based on cancer stage

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 4: Loss of body weight based on cancer stage

Figures and Tables -
Analysis 2.4

Comparison 2: Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 4: Loss of body weight based on cancer stage

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 1: Length of hospital stay in studies without skewed data

Figures and Tables -
Analysis 3.1

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 1: Length of hospital stay in studies without skewed data

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 2: Postoperative morbidity in studies in which use of Clavien‐Dindo classification was not unclear

Figures and Tables -
Analysis 3.2

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 2: Postoperative morbidity in studies in which use of Clavien‐Dindo classification was not unclear

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 3: Incidence of anastomotic leakage with a fixed‐effect model

Figures and Tables -
Analysis 3.3

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 3: Incidence of anastomotic leakage with a fixed‐effect model

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 4: Health‐related quality of life in studies without benign disease patients

Figures and Tables -
Analysis 3.4

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 4: Health‐related quality of life in studies without benign disease patients

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 5: Loss of body weight in studies not limited to diabetic patients

Figures and Tables -
Analysis 3.5

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 5: Loss of body weight in studies not limited to diabetic patients

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 6: Health‐related quality of life in studies without co‐intervention bias

Figures and Tables -
Analysis 3.6

Comparison 3: Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction, Outcome 6: Health‐related quality of life in studies without co‐intervention bias

Summary of findings 1. Roux‐en‐Y compared to Billroth‐I after distal gastrectomy for gastric cancer

Roux‐en‐Y compared to Billroth‐I after distal gastrectomy for gastric cancer

Patient or population: people undergoing distal gastrectomy for gastric cancer
Setting: operating room in a hospital
Intervention: Roux‐en‐Y
Comparison: Billroth‐I

Outcomes

Anticipated absolute effects* (95% CI)

Relative effect
(95% CI)

№ of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with Billroth‐I

Risk with Roux‐en‐Y

Health‐related quality of life

SMD 0.04 higher
(0.11 lower to 0.18 higher)

695
(6 RCTs)

⊕⊕⊝⊝
LOW1,2

Regarding the effect size of the SMD, 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect (Cohen 1988). Converting an SMD of 0.04 into a global health status score of EORTC QLQ‐C30, Roux‐en‐Y may increase it by 0.56 points (95% CI −1.53 to 2.50) (Murad 2019).

Incidence of anastomotic leakage

14 per 1000

9 per 1000
(2 to 36)

RR 0.63
(0.16 to 2.53)

711
(5 RCTs)

⊕⊝⊝⊝
VERY LOW3,4

Loss of body weight

The mean loss of body weight ranged from 8 to 9 percent.

The mean loss of body weight was 0.41% greater
(0.77 smaller to 1.59 greater).

541
(4 RCTs)

⊕⊕⊝⊝
LOW3,5

Loss of body weight is expressed as a percentage (e.g. if a person of 50 kg becomes 40 kg, the loss of body weight is 20%). Weight loss can theoretically range from negative infinity to 100%, but weight gain is uncommon after gastrectomy, and weight loss of more than 50% is also uncommon. The observed values are therefore usually expected to fall within the range of 0 to 50%.

Incidence of bile reflux

397 per 1000

159 per 1000
(99 to 250)

RR 0.40
(0.25 to 0.63)

399
(4 RCTs)

⊕⊕⊕⊝
MODERATE6

Length of hospital stay

The mean length of hospital stay ranged from 7 to 23 days.

The mean length of hospital stay was 0.96 days longer
(0.16 to 1.76 days longer).

894
(7 RCTs)

⊕⊝⊝⊝
VERY LOW3,7,8

Length of hospital stay is expressed in days, and ranges from 1 to infinite.

Postoperative morbidity

99 per 1000

146 per 1000
(101 to 209)

RR 1.47
(1.02 to 2.11)

891
(7 RCTs)

⊕⊕⊝⊝
LOW 3,9

*The risk in the intervention group (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; EORTC QLQ‐C30: EORTC Core Quality of Life Questionnaire ‐ Core Questionnaire; RCT: randomised controlled trial; RR: risk ratio; SMD: standardised mean difference

GRADE Working Group grades of evidence
High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

1Downgraded one level due to imprecision; 95% CIs are compatible with both benefit and harm.
2Downgraded one level due to risk of bias; participants were not blinded despite this being a patient‐reported outcome.
3Downgraded one level due to risk of bias; surgeons were not blinded
4Downgraded by two levels due to imprecision; 95% CIs are wide due to rare events, with possibility of substantial harm and substantial benefit, suggesting very serious imprecision.
5Downgraded one level due to imprecision; small sample size — outcome included only four studies.
6Downgraded one level due to risk of bias; outcome assessors may not have been blinded.
7Downgraded one level due to risk of bias; incomplete reporting on the results indicating data are skewed.
8Downgraded one level due to inconsistency; substantial heterogeneity was observed.
9Downgraded one level due to inconsistency; possible inconsistencies in trial results due to different diagnostic criteria used to report complications.

Figures and Tables -
Summary of findings 1. Roux‐en‐Y compared to Billroth‐I after distal gastrectomy for gastric cancer
Table 1. Details of complications reported in included studies

Study

Roux‐en‐Y

Affected participants in Roux‐en‐Y

Billroth‐I

Affected participants in Billroth‐I

Choi 2017

1 Atelectasis

1 Bleeding

2/20

1 Atelectasis

1 Complicated fluid

2/20

Hur 2017

3 Postoperative ileus

1 Pneumonia

1 Leakage

1 Wound seroma

1 Voiding difficulty

6/58

1 Postoperative ileus

1 Pneumonia

1 Intra‐abdominal fluid

1 Cholecystitis

1 Voiding difficulty

2 Unknown fever

7/56

Yang 2017

10 Pulmonary complications

1 Acute cholecystitis

2 Superficial surgical site infection

2 Intra‐abdominal infection

1 Adhesive ileus

1 Acute urinary retention

1 Gastroplegia

14/70

10 Pulmonary complications

1 Acute cholecystitis

1 Acute urinary retention

1 Gastroplegia

11/70

Nakamura 2016

4 Delayed gastric emptying

1 Pancreatic fistula

2 Anastomotic stricture

1 Pancreatic fistula

1 Postoperative bleeding

9/59

1 Surgical site infection

2 Anastomotic leakage

1 Anastomotic stricture

4/60

Lee 2012

1 Anastomotic leakage

2 Gastorojejunostomy outlet obstruction

2 Intra‐abdominal abscess

1 Deep vein thrombosis

6/47

2 Bleeding

1 Wound problem

1 Chyle leakage

4/49

Takiguchi 2012

3 Pancreatic fistula

3 Abdominal abscess

2 Bowel obstruction

2 Postoperative pancreatitis

2 Surgical site infection

2 Anastomotic stricture

23/169

2 Pancreatic fistula

2 Anastomotic leakage

3 Abdominal abscess

1 Bowel obstruction

2 Postoperative pancreatitis

3 Surgical site infection

3 Anastomotic stricture

14/163

Ishikawa 2005

3 Gastric stasis

1 Intestinal obstruction

1 Pneumonia

2 Anastomotic stricture

6/24

1 Leakage

1 Intestinal obstruction

2/26

Figures and Tables -
Table 1. Details of complications reported in included studies
Comparison 1. Roux‐en‐Y versus Billroth‐I reconstruction

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1.1 Health‐related quality of life Show forest plot

6

695

Std. Mean Difference (IV, Random, 95% CI)

0.04 [‐0.11, 0.18]

1.2 Incidence of anastomotic leakage Show forest plot

5

711

Risk Ratio (M‐H, Random, 95% CI)

0.63 [0.16, 2.53]

1.3 Loss of body weight Show forest plot

4

541

Mean Difference (IV, Random, 95% CI)

0.41 [‐0.77, 1.59]

1.4 Incidence of bile reflux Show forest plot

4

399

Risk Ratio (M‐H, Random, 95% CI)

0.40 [0.25, 0.63]

1.5 Length of hospital stay Show forest plot

7

894

Mean Difference (IV, Random, 95% CI)

0.96 [0.16, 1.76]

1.6 Postoperative morbidity Show forest plot

7

891

Risk Ratio (M‐H, Random, 95% CI)

1.47 [1.02, 2.11]

Figures and Tables -
Comparison 1. Roux‐en‐Y versus Billroth‐I reconstruction
Comparison 2. Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

2.1 Health‐related quality of life based on surgical approach Show forest plot

6

695

Std. Mean Difference (IV, Random, 95% CI)

0.04 [‐0.11, 0.18]

2.1.1 Open studies

2

404

Std. Mean Difference (IV, Random, 95% CI)

0.15 [‐0.05, 0.34]

2.1.2 Laparoscopic studies

1

106

Std. Mean Difference (IV, Random, 95% CI)

‐0.21 [‐0.59, 0.17]

2.1.3 Mixed and unknown studies

3

185

Std. Mean Difference (IV, Random, 95% CI)

‐0.07 [‐0.35, 0.22]

2.2 Loss of body weight based on surgical approach Show forest plot

4

541

Mean Difference (IV, Random, 95% CI)

0.41 [‐0.77, 1.59]

2.2.1 Open studies

1

332

Mean Difference (IV, Random, 95% CI)

0.60 [‐0.87, 2.07]

2.2.2 Laparoscopic studies

1

40

Mean Difference (IV, Random, 95% CI)

‐0.70 [‐4.48, 3.08]

2.2.3 Mixed and unknown studies

2

169

Mean Difference (IV, Random, 95% CI)

0.34 [‐2.01, 2.68]

2.3 Health‐related quality of life based on cancer stage Show forest plot

6

695

Std. Mean Difference (IV, Random, 95% CI)

0.04 [‐0.11, 0.18]

2.3.1 Early stage studies

3

490

Std. Mean Difference (IV, Random, 95% CI)

‐0.02 [‐0.20, 0.17]

2.3.2 Mixed and unknown studies

3

205

Std. Mean Difference (IV, Random, 95% CI)

0.15 [‐0.13, 0.42]

2.4 Loss of body weight based on cancer stage Show forest plot

4

541

Mean Difference (IV, Random, 95% CI)

0.41 [‐0.77, 1.59]

2.4.1 Early stage studies

3

491

Mean Difference (IV, Random, 95% CI)

0.37 [‐0.86, 1.59]

2.4.2 Mixed and unknown studies

1

50

Mean Difference (IV, Random, 95% CI)

0.90 [‐3.43, 5.23]

Figures and Tables -
Comparison 2. Subgroup analysis in Roux‐en‐Y versus Billroth‐I reconstruction
Comparison 3. Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

3.1 Length of hospital stay in studies without skewed data Show forest plot

5

722

Mean Difference (IV, Random, 95% CI)

0.62 [0.16, 1.09]

3.2 Postoperative morbidity in studies in which use of Clavien‐Dindo classification was not unclear Show forest plot

5

463

Risk Ratio (M‐H, Random, 95% CI)

1.39 [0.87, 2.24]

3.3 Incidence of anastomotic leakage with a fixed‐effect model Show forest plot

5

711

Risk Ratio (M‐H, Fixed, 95% CI)

0.61 [0.19, 1.93]

3.4 Health‐related quality of life in studies without benign disease patients Show forest plot

5

665

Std. Mean Difference (IV, Random, 95% CI)

0.04 [‐0.12, 0.19]

3.5 Loss of body weight in studies not limited to diabetic patients Show forest plot

3

501

Mean Difference (IV, Random, 95% CI)

0.53 [‐0.72, 1.77]

3.6 Health‐related quality of life in studies without co‐intervention bias Show forest plot

5

559

Std. Mean Difference (IV, Random, 95% CI)

‐0.01 [‐0.18, 0.15]

Figures and Tables -
Comparison 3. Sensitivity analysis in Roux‐en‐Y versus Billroth‐I reconstruction