Twenty studies in total were included.
Ten studies assessed presacral neurectomy (905 patients) including two RCTs (97 patients), three comparative but non-randomised retrospective studies (196 patients), and five non-comparative studies prospective or retrospective studies (612 patients).
Ten studies (including one RCT, 6 prospective and 3 retrospective studies) were used to assess uterosacral ligament resection (607 patients).
A. Presacral neurectomy.
1. Dysmenorrhoea.
Results were inconsistent.
In the five non-comparative studies, the frequency of dysmenorrhoea recurrence, or persistence after treatment ranged from 4% to 40%. The pooled frequency of non-responders at the end of follow-up was 23% (95% CI: 19%, 27%). Breslow-Day test showed significant heterogeneity (chi-squared = 12.36, df = 4, p = 0.014).
Only two of the three non-randomised trials demonstrated a significant treatment benefit of pre-sacral neurectomy, and the results of the two identified RCTs were discordant. Significant quantitative heterogeneity among studies prevented pooling of data on dysmenorrhoea.
2. Deep dyspareunia (3 studies with 113 patients, including one RCT, on part RCT and one retrospective comparative): there were no statistically significant differences between the intervention and the control group. Pooled OR = 0.69 (95% CI: 0.31, 1.54). Breslow-Day test showed no significant heterogeneity (chi-squared = 0.72, df = 2, p = 0.69).
3. Complications included: urinary tract infection; fever; parotid gland infection; temporary urge incontinence; retroperitoneal bleeding; constipation; urinary urgency; painless first stage of labour; vaginal dryness; and intraoperative bleeding.
B. Uterosacral ligament resection (10 non-comparative arms of studies).
1. Dysmenorrhoea: the frequency varied from 0% to 50% after treatment and, at the end of follow-up was 23% (95% CI: 20%, 27%). The forest plot shows heterogeneity.
2. Deep dyspareunia (4 non-comparative studies: frequency varied from 6% to 42% after treatment, and at the end of follow-up was 13% (95% CI: 8%, 18%). The forest plot suggests heterogeneity.
3. Complications included: haematomas at trochar insertion site; intraoperative bleeding; retroperitoneal fluid extravasation; urinary tract infection; bladder injury; uterine injury; bleeding from operating site; vaginal cuff perforation; vaginal cuff wound diastasis; pelvic pain; urinary retention; nerve injury; and rectovaginal fistula.
Quality of studies.
Methodological problems in the primary studies included: lack of classification of disease stage; lack of evaluation of pain using a valid and reliable instrument; drop-outs and withdrawals were either not clearly identified or not included in the analysis; lack of randomised control group; and lack of description of methods used to assess side-effects.
Sources of clinical heterogeneity included: differences in patient selection; modalities of patient evaluation; and variable follow-up.
Hysterectomy for pelvic pain of different aetiology.
Five studies were described (684 patients) including three prospective and two retrospective observational studies. Percentage obtaining relief or symptomatic improvement ranged from 83% to 97%.
Adhesions and pelvic pain.
Four studies were described including one RCT. The RCT (48 women with pelvic adhesions) reported no significant difference between patients allocated to either adhesiolysis or non-surgical management.