To determine the safety, clinical effectiveness and cost-effectiveness of circular stapled haemorrhoidopexy (SH) for the treatment of haemorrhoids. Main electronic databases were searched up to July 2006. Randomised controlled trials (RCTs) with 20 or more participants that compared SH with any conventional haemorrhoidectomy (CH) technique in people of any age with prolapsing haemorrhoids for whom surgery is considered a relevant option, were used to evaluate clinical effectiveness. An economic model of the surgical treatment of haemorrhoids was developed. The clinical effectiveness review included 27 RCTs (n = 2279; 1137 SH; 1142 CH). All had some methodological flaws; only two reported recruiting patients with second, third and fourth degree haemorrhoids, and 37% reported using an appropriate method of randomisation and/or allocation concealment. In the early postoperative period 95% of trials reported less pain following SH; by day 21 the pain reported following SH and CH was minimal, with little difference between the two techniques. Significantly fewer patients had unhealed wounds at 6 weeks following SH [odds ratio (OR) 0.08, 95% confidence interval (CI) 0.03 to 0.19, p < 0.001]. Residual prolapse was more common after SH (OR 3.38, 95% CI 1.00 to 11.47, p = 0.05, nine RCTs, results of a sensitivity analysis). There was no difference between SH and CH in the incidence of bleeding or postoperative complications. SH resulted in shorter operating times, hospital stay, time to first bowel movement and return to normal activity. In the short term (between 6 weeks and a year) prolapse was more common after SH (OR 4.68, 95% CI 1.11 to 19.71, p = 0.04, six RCTs). There was no difference in the number of patients complaining of pain between SH and CH. In the long term (1 year and over), there was a significantly higher rate of prolapse after SH (OR 4.34, 95% CI 1.67 to 11.28, p = 0.003, 12 RCTs). There was no difference in the number of patients experiencing pain, or the incidence of bleeding, between SH and CH. There was no difference in the total number of reinterventions, or reinterventions for pain, bleeding or complications, between SH and CH. Significantly more reinterventions were undertaken after SH for prolapse at 12 months or longer (OR 6.78, 95% CI 2.00 to 23.00, p = 0.002, six RCTs). Overall, there was no statistically significant difference in the rate of complications between SH and CH. In the economic assessment it was found that, on average, CH dominated SH. However, CH and SH had very similar costs and quality-adjusted life-years (QALYs). On average, the difference in costs between the procedures was 19 pounds and the difference in QALY was -0.001, favouring CH, over 3 years. In terms of QALYs, the superior quality of life due to lower pain levels in the early postoperative period with SH was offset by the higher rate of symptoms over the follow-up period, compared with CH. The results are very sensitive to modelling assumptions, particularly the valuation of utility in the early postoperative period. The probabilistic sensitivity analysis showed that, at a threshold incremental cost-effectiveness ratio of 20,000-30,000 pounds per QALY, SH had a 45% probability of being cost-effective. SH was associated with less pain in the immediate postoperative period, but a higher rate of residual prolapse, prolapse in the longer term and reintervention for prolapse. There was no clear difference in the rate or type of complications associated with the two techniques and the absolute and relative rates of recurrence and reintervention for both are still uncertain. CH and SH had very similar costs and QALYs, the cost of the staple gun being offset by savings in hospital stay. Should the price of the gun change, the conclusions of the economic analysis may also change. Some training may be required in the use of the staple gun; this is not expected to have major resource implications. Given the currently available clinical evidence and the results of the economic analysis, the decision as to whether SH or CH is conducted could primarily be based on the priorities and preferences of the patient and surgeon. An adequately powered, good-quality RCT is required, comparing SH with CH, recruiting patients with second, third and fourth degree haemorrhoids, and having a minimum follow-up period of 5 years to ensure an adequate evaluation of the reintervention rate. Other areas for research are the effectiveness of SH in patients with fourth degree haemorrhoids and patients with co-morbid conditions, the reintervention rates for all treatments for haemorrhoids, utilities of patients up to 6 months postoperatively, the trade-offs of patients for short-term pain versus long-term outcomes, and the ability of SH to reduce hospital stays in a real practice setting.