Twenty-nine RCTs (n=2,056) were included in the review. Sample sizes ranged from 20 to 200 participants. Follow-up periods ranged from six weeks to a median of 62 months. Two funnel plots (based on post-operative bleeding and recurrent prolapse) indicated the possibility of publication bias.
Study quality
Most studies used sealed envelopes to conceal allocation and a per-protocol data analysis. The randomisation method was not stated in five trials. The remaining trials mostly used either computer-generated random numbers or drawing of sealed envelopes. Eleven trials stated that they used independent blinded observers to record data (this was unclear in most other trials).
Efficacy
Stapled haemorrhoidopexy resulted in a reduced requirement for analgesics compared to conventional haemorrhoidectomy both during hospital stay and after discharge (seven trials, SMD -2.98, 95%CI: -4.76 to -1.20, p=0.001). But, significant heterogeneity was seen between the studies (I2=98.5%). Stapled haemorrhoidopexy treatment was also associated with a shorter hospital stay (eight trials, WMD -0.95 days, 95%CI: -1.32, -0.59, p<0.001, also with considerable heterogeneity, I2=93%) and operating time (nine trials, WMD -11.42, 95%CI: -18.26, -4.59, p<0.001, heterogeneity I2=99%). Stapled haemorrhoidopexy was associated with a faster return to normal activities (10 trials, WMD -11.75 days, 95%CI: -21.42, -2.08, p=0.017, but with great heterogeneity between studies I2=99.8%). However, stapled haemorrhoidopexy was also associated with more episodes of recurrent prolapse at a minimum of one year follow up (14 trials, RR 2.07, 95% CI: 1.33, 3.20, p=0.001) with no heterogeneity found in the analysis, I2=0%.
Safety
No significant difference in total complications was found between the two treatment groups, RR 1.08 (95% CI: 0.80, 1.45, p=0.63, I2=31.6% for heterogeneity). Stapled haemorrhoidopexy treatment was associated with a higher incidence of bleeding requiring re-intervention (25 trials, RR 1.57, 95% CI: 1.06, 2.33, p=0.023, with no heterogeneity, I2=0%).
For the sensitivity analyses, it was found that analysing only high-quality trials (at least 3 on the Jadad scale) or only studies with at least 40 patients, resulted in no significant differences in post-operative haemorrhage (unlike when all studies were included, where a significant difference was found).
Further results were reported.