This review concluded that flexible sigmoidoscopy was a better investigative tool in colorectal clinics than rigid sigmoidoscopy. This conclusion reflects the data presented but should be interpreted cautiously due to the relatively poor methodological quality of included studies and high levels of clinical and statistical heterogeneity.

To compare the effectiveness of flexible sigmoidoscopy with rigid sigmoidoscopy for the investigation of colorectal symptoms.

MEDLINE, EMBASE and The Cochrane Library were searched without language or date restrictions. Search terms but not dates were reported. Bibliographies of included studies were screened for additional articles.

Prospective, randomised and non-randomised trials that compared flexible and rigid sigmoidoscopy in symptomatic or asymptomatic patients at colorectal clinics were eligible for inclusion. Studies were required to report at least one of the outcome measures: detection of colorectal cancer (primary outcome); detection of polyps; detection of inflammatory bowel disease; detection of diverticulosis; depth of examination; procedure-associated discomfort; and duration of procedure.

Trials of patients with a known diagnosis of cancer or other anal or perianal pathologies were excluded.

All studies were conducted in Europe or USA. Use of a barium enema and examination position varied between studies (reported in the article).

Two reviewers independently assessed studies for inclusion and any disagreements were resolved by discussion.

The methodological quality of included studies was assessed using the Downs and Black checklist (maximum score 33).

The authors did not report how many reviewers performed the quality assessment.

Dichotomous data were extracted as the number of events for each sigmoidoscopy method and estimates of risk difference (RD) and 95% confidence intervals (CIs) were calculated. Continuous data were extracted as the mean value for each sigmoidoscopy method.

Data were extracted by one reviewer and checked by a second.

Authors were contacted for missing data as necessary.

Pooled estimates of risk difference with 95% confidence intervals were calculated for dichotomous outcomes using a random-effects model. Weighted mean difference (WMD) or standardised mean difference (SMD) where endpoints were measured on different scales, with 95% confidence intervals, were calculated for continuous outcome measures using a random-effects model.

Randomised and non-randomised studies were pooled separately and sensitivity analyses were conducted to assess the effects of excluding individual studies.

Publication bias was assessed using a funnel plot and the fail safe N method.

Ten studies (3,915 participants, range 71 to 1,012) were included in the review. Four studies were randomised and six were non-randomised. The mean and standard deviation Downs and Black score was 16±2. The main areas of poor scoring were insufficient explanation of and adjustment for main confounders, missing data and absence of description of source population.

Compared with rigid sigmoidoscopy, flexible sigmoidoscopy was associated with significantly higher rates of detection of colorectal cancer (RD 0.02, 95% CI 0.006 to 0.034; Ι²=69.6%) and any abnormality (RD 0.138, 95% CI 0.077 to 0.200; Ι²=94.7%). Both pooled estimates were based on nine studies.

Rigid sigmoidoscopy caused significantly more patient discomfort than flexible sigmoidoscopy (WMD 0.981 (95% CI 0.693 to 1.269; six studies; Ι²=92.9%).

Flexible sigmoidoscopy was associated with significantly greater depth of examination (SMD 3.175, 95% CI 2.397 to 3.954; nine studies; Ι²=99%). Compared with flexible sigmoidoscopy, rigid sigmoidoscopy was associated with significantly shorter examination times (SMD -1.601, 95% CI -2.728 to -0.474; six studies; Ι²=99%).

Separate pooled estimates for randomised and non-randomised studies were also reported.

The funnel plot suggested a limited possibility of publication bias.

Flexible sigmoidoscopy was a better investigative tool in colorectal clinics than rigid sigmoidoscopy.

This article reported a clear research objective and defined appropriate inclusion criteria. Various sources were searched for relevant studies without language and date restrictions, which reduced the likelihood of missing relevant studies. The study selection and data extraction processes included measures to minimise error and bias; it was not clear whether these measures were also applied to quality assessment. The methodological quality of included studies was assessed appropriately and results were reported.

Meta-analytic pooling of included studies may not have been appropriate given the high levels of between-study heterogeneity and apparent clinical heterogeneity (for example, the review included both diagnostic and screening studies).

The authors' conclusions reflect the data presented but should be interpreted cautiously due to the relatively poor methodological quality of included studies and high levels of between-study heterogeneity.

Practice: The authors stated that the limitations of rigid sigmoidoscopy justified a switch to flexible sigmoidoscopy.

Research: The authors stated that economic aspects of the switch to flexible sigmoidoscopy required further investigation.

Not stated.