Description of studies

After evaluation of the abstracts from the searches, we excluded articles that clearly were not randomized or controlled trials or did not focus on interventions or outcomes included in this review. We retrieved copies of 55 potentially relevant trials, of which two met the inclusion criteria (Sokal 1999; Christensen 2002). We excluded six reports (Black 1989; Nirapathpongporn 1990; Black 2003; Chen 2004; Chen 2005; Song 2006). No new studies have been found since the review was first published.

The two eligible studies compared similar interventions but differed in their timing and nature of post‐operative evaluations, including the evaluation of sterility. They also differed in operator experience with the no‐scalpel technique. No surgeon in the Christensen 2002 trial had substantial experience with the technique while some (3/8) in the Sokal 1999 trial were experienced with the method.

Christensen 2002 compared vasectomy with bilateral incision versus vasectomy with no‐scalpel approach to the vas. Outcome measures included vasectomy success (defined as azoospermia at three months post‐operation), duration of procedure, pain and discomfort immediately following the procedure using visual analog scales, adverse events, need for second doctor's assistance, and conversion to a different type of procedure. Participants were mailed a survey regarding pain and wound problems after one month, and they were to return after three months for semen analysis for establishing azoospermia. Semen analysis methods were not described.

Sokal 1999 compared vasectomy (small segment of the vas excised and ligated both ends) using a no‐scalpel incision versus vasectomy with a standard incision (i.e., a double vertical incision used in Guatemala and Semarang, Indonesia and a single vertical incisions in the remaining sites). Safety was the primary outcome measure. Other measures were ease of use, duration of procedure and sterility (defined as absence of live spermatozoa). A second measure of vasectomy effectiveness (i.e., failure) was determined at the surgeons' discretion without standardized criteria between centers. Participants were to return between 3 and 15 days for post‐operative evaluation and at 10 weeks for semen analysis. Semen analysis methods were not described. The vas occlusion technique in most cases was ligation (99.8% for the no‐scalpel and 99.6% for the standard incision group). Excision of the vas segment was completed for most participants (99.7% for the no‐scalpel and 99.9% for the incision group) and ligation of both ends of the vas was performed in 99.7% of all vasectomies. Sutures were used for wound closure in 2.2% of no‐scalpel and 28.9% of standard incision procedures.

Risk of bias in included studies

Christensen 2002 was a small (N=100) randomized controlled trial using sealed, sequentially‐numbered envelopes. Blinding was not reported for the outcome assessors or participants. Two randomization envelopes were opened by mistake (both for no‐scalpel arm) and were not replaced. One post‐randomization exclusion was reported (no‐scalpel arm) for a case with high testis due to inability to perform vasectomy under local anesthetic. Two men assigned to the no‐scalpel group were converted to bilateral incision during the procedure but, consistent with intent‐to‐treat principle, they were analyzed according to their randomized group (i.e., not treatment group). The study had a high loss to follow up at the one‐month survey (13%) and at the three‐month assessment for azoospermia (74%), which could have biased the study results greatly.

Sokal 1999 conducted a large randomized controlled trial (N=1429) using randomly‐generated numbers, and opaque, sealed envelopes to conceal the allocation process. Although the outcome assessor was blinded as to the group assignments, the participant blinding was unclear. After randomization, 35 men failed to receive their assigned type of incision due to allocation errors (N=17) or intraoperative obstacles (N=8). The primary analysis included participants with protocol violations, random allocation errors or technical failures (except for the exclusion of one technical failure that used different approaches on each vas). However, the analysis groups were based on the treatment received instead of the randomized assignment (N=705 for the no‐scalpel group and N=723 for the standard incision group). The follow‐up rates were about 77% for short‐term (at 15 days post‐vasectomy) and 91% for long‐term (after 15 days post‐vasectomy).

Both trials were conducted by non‐profit organizations.

Effects of interventions

Although the smaller study did not find any significant difference in perioperative bleeding (Christensen 2002), the no‐scalpel group in the larger trial (Sokal 1999) was half as likely to bleed during surgery as the scalpel group (OR 0.49; 95% CI 0.27 to 0.89). Furthermore, both studies found less hematoma during follow up for the men who received the no‐scalpel technique than those who had the standard incision (OR 0.23; 95% CI 0.15 to 0.36).

Christensen 2002 reported that the median operating times for the no‐scalpel and incisional groups (20 and 24 minutes, respectively) were not significantly different, but provided only the P value without any measure of the precision of the estimates. In contrast, the larger trial of Sokal 1999 showed that the no‐scalpel technique required less surgery time. The no‐scalpel group was more likely to have an operation time of 6 minutes or less (OR 2.37; 95% CI 1.92 to 2.91) and less likely to be 11 minutes or longer in duration (OR 0.56; 95% CI 0.43 to 0.73) than the incisional group (Sokal 1999). The 6 and 11 minute cut points, though, were selected post hoc.Christensen 2002 did not find significant differences between groups in perioperative difficulties (i.e., tight ductus, difficulty in identifying the ductus, hydrocele testis, need for assistance from a second operator). However, Sokal 1999 reported more difficulty isolating the vas (OR 1.80; 95% CI 1.18 to 2.76), short scrotum or thin deferens, and adhesions for this group compared to the scalpel group despite the shorter operation time with the no‐scalpel approach. The no‐scalpel group, though, was less likely to have (unspecified) equipment difficulties than the scalpel group (OR 0.28; 95% CI 0.10 to 0.77) (Sokal 1999). No statistically significant differences were observed for difficulty entering the scrotum, closing the incision, occluding the vas, difficulties due to fatty, adipose or fibrous tissue, chronic infection, pain or patient restlessness (only P value was provided) (Sokal 1999).

The no‐scalpel group was less likely to report perioperative pain than the incisional group (OR 0.75; 95% CI 0.61 to 0.93) in the larger trial (Sokal 1999). In contrast, the smaller study found no differences in visual analog scale scores for perioperative pain or discomfort between the two groups (Christensen 2002). Similarly, no differences in the reporting during follow up of general pain, scrotal pain or pain at ejaculation was found for the smaller study (Christensen 2002). In the larger trial (Sokal 1999), the no‐scalpel group was less likely to report scrotal pain (OR 0.63; 95% CI 0.50 to 0.80) than the standard incision group. The no‐scalpel group in the larger trial (Sokal 1999) also had less incisional infection during follow up than the scalpel group (OR 0.21; 95% CI 0.06 to 0.78). The smaller trial found no statistically significant difference in infection or wound problems (Christensen 2002). Pain or tenderness was the most common long‐term (i.e., at least 15 days post‐operation) side effect reported in Sokal 1999, but the outcome did not differ significantly between groups. Three men in each group in this trial required hospitalization during the follow‐up period, of which three complications appeared to be vasectomy‐related. Sokal 1999 also found a quicker resumption of sexual intercourse among the no‐scalpel than the incision group (P < 0.05 from Kaplan‐Meier survival estimate in article). In addition, nearly 90% of the participants in both groups reported that they were 'satisfied' or 'very satisfied' with the procedure.

In the smaller trial (Christensen 2002), microscopic analysis of resected tissue showed sufficient bilateral vasectomy from 98/99 surgeries. The exception was one case of vasectomy with bilateral incision, in which the operators had difficulty identifying the right‐sided ductus deferens during the surgery, and the microscopic analysis afterwards showed missing material. Only 26 of 99 patients (13 in each group) adhered to the study protocol of returning to provide semen samples at three months post‐operation. Two cases in the scalpel group and three cases in the no‐scalpel group were determined to be insufficient vasectomy. The larger trial (Sokal 1999) found similar rates of vasectomy success for the two groups whether measured as sterility (i.e., azoospermia) (OR 0.94; 95% CI 0.50 to 1.76) or the center‐specific evaluation of vasectomy failure (OR 0.87; 95% CI 0.37 to 2.07). The authors did not find evidence of interaction by clinical sites, and they also found almost identical results when they repeated the analysis using a per‐protocol population (i.e., excluding the 108 men with protocol violations, random allocation errors or technical failures).