Laparoscopic uterosacral ligament resection for dysmenorrhea associated with endometriosis: results of a randomized, controlled trial

Transcription

1 FERTILITY AND STERILITY VOL. 80, NO. 2, AUGUST 2003 Copyright 2003 American Society for Reproductive Medicine Published by Elsevier Inc. Printed on acid-free paper in U.S.A. Laparoscopic uterosacral ligament resection for dysmenorrhea associated with endometriosis: results of a randomized, controlled trial Paolo Vercellini, M.D., a Giorgio Aimi, M.D., a Mauro Busacca, M.D., b Giovanni Apolone, M.D., c Anna Uglietti, M.D., a and Pier Giorgio Crosignani, M.D. a Istituto Ostetrico e Ginecologico Luigi Mangiagalli, University of Milan, and Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy Received September 26, 2002; revised and accepted January 7, Reprint requests: Paolo Vercellini, M.D., Istituto Luigi Mangiagalli, Clinica Ostetrica e Ginecologica I, Università di Milano, Via Commenda 12, 20122, Milan, Italy (FAX: ; paolo.vercellini@unimi.it). a Reproductive Surgery Unit, First Department of Obstetrics and Gynecology, University of Milan. b Second Department of Obstetrics and Gynecology, University of Milan. c Laboratorio di Epidemiologia Clinica, Istituto di Ricerche Farmacologiche Mario Negri /03/$30.00 doi: /s (03) Objective: To evaluate the efficacy of laparoscopic resection of the uterosacral ligaments in women with endometriosis and predominantly midline dysmenorrhea. Design: Randomized controlled trial. Setting: Two academic departments. Patient(s): One hundred eighty patients undergoing operative laparoscopy as first-line therapy for stage I to IV symptomatic endometriosis. Intervention(s): Operative laparoscopy including uterosacral ligament resection or conservative surgery alone. Main Outcome Measure(s): Proportion of women with recurrence of moderate or severe dysmenorrhea 1 year after surgery. Result(s): No complications occurred. Among the patients who were evaluable 1 year after operative laparoscopy, 23 of 78 (29%) women who had uterosacral ligament resection and 21 of 78 (27%) women who had conservative surgery only reported recurrent dysmenorrhea. The corresponding numbers of patients at 3 years were 21 of 59 (36%) women and 18 of 57 (32%) women, respectively. Time to recurrence was similar in the two groups. Pain was substantially reduced, and patients in both groups experienced similar and significant improvements in health-related quality of life, psychiatric profile, and sexual satisfaction. Overall, 68 of 90 (75%) patients in the uterosacral ligament resection group and 67 of 90 (74%) patients in the conservative surgery group were satisfied at 1 year. Conclusion(s): Addition of uterosacral ligament resection to conservative laparoscopic surgery for endometriosis did not reduce the medium- or long-term frequency and severity of recurrence of dysmenorrhea. (Fertil Steril 2003;80: by American Society for Reproductive Medicine.) Key Words: Endometriosis, dysmenorrhea, deep dyspareunia, pelvic pain, laparoscopy, uterosacral ligaments, conservative pelvic surgery, quality of life Endometriosis is frequently associated with disabling dysmenorrhea. Conservative surgical treatment at laparoscopy or laparotomy is generally satisfactory, but patients often report only partial resolution of pain or recurrence (1, 2). Two uterine denervation techniques have been proposed to improve long-term results (2, 3). In a previous study from our institution, conservative surgery alone was not significantly less effective than presacral neurectomy in limiting the rate of recurrence of dysmenorrhea (4). With the latter technique, only the sympathetic nerves of the superior hypogastric plexus are sectioned; however, the uterus is innervated also by parasympathetic fibers, the nervi erigentes, which originate from S2 to S4 (5, 6). Both sympathetic and parasympathetic fibers run in the uterosacral folds and reach the parametrium, posterolateral to the uterine cervix, where they unite to form the plexuses of Frankenhauser (5, 6). Resection of the uterosacral ligaments, which in theory should achieve more complete 310

2 uterine denervation than presacral neurectomy, is routinely performed at laparoscopy in women with endometriosis and midline dysmenorrhea (7, 8). However, the intervention carries the risk for such complications as bleeding, ureteral lesions, and pelvic support disorders (2). Moreover, the efficacy of uterosacral ligament resection in relieving pain associated with endometriosis was not confirmed in a small, randomized, controlled study or in a recent meta-analysis of literature data (9, 10). We sought to compare the efficacy and safety of conservative surgery at laparoscopy with or without resection of the uterosacral ligaments in a large series of women with endometriosis. The primary outcome of interest was the rate of recurrence of dysmenorrhea 1 year after surgery. We also assessed differences between the groups in relief of deep dyspareunia and nonmenstrual pain, health-related quality of life, psychoemotional profile, sexual functioning, and overall satisfaction with treatment. MATERIALS AND METHODS In this open-label, parallel-group, randomized clinical trial, we compared conservative surgery at laparoscopy including uterosacral ligament resection with operative laparoscopy alone for treatment of midline dysmenorrhea associated with endometriosis. The investigation was performed in two academic departments specializing in the study and management of endometriosis. Institutional review board approval was obtained, and the patients gave informed consent to participate in the study. Eligible patients were women 18 to 40 years of age who were undergoing first-line operative laparoscopy for symptomatic minimal to severe endometriosis, as diagnosed according to the revised American Fertility Society classification (11), and who reported pelvic pain of more than 6 months duration. Patients with a previous clinical or endoscopic diagnosis of endometriosis or with other diseases that might cause pelvic pain (such as chronic pelvic inflammatory disease, pelvic varices, or genital malformations) were excluded. Other exclusion criteria were treatment for endometriosis other than nonsteroid anti-inflammatory drugs up to 6 months before study entry; presence of vaginal endometriotic lesions; previous diagnosis of gastrointestinal, urologic, and orthopedic diseases in which pain may radiate to the pelvic area; and known psychiatric disturbances. Each patient was asked to complete a questionnaire on the presence and severity of dysmenorrhea, deep dyspareunia, and nonmenstrual pelvic pain graded according to a 0- to 3-point multidimensional categorical rating modified from that devised by Biberoglu and Behrman (12). The original defines dysmenorrhea according to loss of work efficiency and need for bed rest (absence of pain, 0; some loss of work efficiency, mild, 1; in bed part of 1 day, occasional loss of work, moderate, 2; in bed for 1 or more days, incapacitation, severe, 3), nonmenstrual pain according to various degrees of discomfort and use of analgesics (absence of pain, 0; occasional pelvic discomfort, mild, 1; noticeable discomfort for most of the cycle, moderate, 2; pain persisting during the cycle or requiring strong analgesics, severe, 3), and deep dyspareunia according to limitation of sexual activity (no discomfort, 0; tolerated discomfort, mild, 1; intercourse painful to the point of interruption, moderate, 2; intercourse avoided because of pain, severe, 3). The women were also requested to grade the severity of dysmenorrhea, nonmenstrual pelvic pain, and deep dyspareunia by using a 100-mm visual analogue, on which the extreme left indicated the absence of pain and the extreme right indicated pain as bad as possible. A score of 1 to 50 was considered mild pain, 51 to 80 moderate pain, and 81 to 100 severe pain. Cut-off points defining different categories of pain were chosen on the basis of a previous correlation analysis with the Biberoglu and Behrman multidimensional categorical rating (13). Participants had predominantly midline and moderate or severe dysmenorrhea on both s. At baseline, the patients completed the Short Form-36 general health survey questionnaire (14 16), the Hospital Anxiety and Depression Scale (17), and the revised Sabbatsberg Sexual Rating Scale (which was submitted anonymously) (18). The Short Form-36 questionnaire is self-administered, contains 36 items, and measures health in eight multi-item dimensions covering functional status, well-being, and overall health. For each dimension, item scores are coded, summed, and transformed onto a from 0 (worst health) to 100 (best health). This instrument has been developed to measure health perception in a general population, and its clinical validity and internal consistency have been demonstrated in large samples in many countries, including Italy (14, 15). For this study, we used the International Quality of Life Assessment Short Form-36, Italian version, release 1.6 (16). The Hospital Anxiety and Depression Scale is a selfassessment mood specifically designed for use in nonpsychiatric hospital outpatients to determine states of anxiety and depression. It comprises 14 items: 7 for the anxiety sub and 7 for the depression sub. Five mutually exclusive answers, rated from 0 to 4 according to increasing psychiatric severity, are provided for each of the 14 questions. The points are then summed to give anxiety and depression subtotal scores and a total score. The Hospital Anxiety and Depression Scale has been shown to be a reliable instrument for screening for clinically significant anxiety and depression and a valid measure of the severity of these mood disorders (17). The revised Sabbatsberg Sexual Rating Scale is a simple and acceptable 12-item questionnaire suitable for self-as- FERTILITY & STERILITY 311

3 sessment of sexual functioning. It evaluates various aspects separately, including libido, arousal or pleasure, orgasm capacity, and satisfaction. For each of the 12 questions, there are five possible answers scored from 0 to 4 points (from the lowest to the highest sexual satisfaction rating). The 12 scores are then summed and transformed onto a of 0 to 100. Scores between these values represent the percentage of the total possible score. The validity and reliability of this instrument have been demonstrated (18). Women not engaging in intercourse were instructed to not complete the Sabbatsberg Sexual Rating Scale form. Conservative surgery at laparoscopy was performed according to the method of Cook and Rock (19). After excision or coagulation of all endometriotic lesions, eligible patients were randomized in a proportion of 1:1 to conservative surgery including resection of the uterosacral ligaments or to operative laparoscopy alone. Treatment allocation was performed with a computer-generated randomization sequence by using serially numbered, opaque, sealed envelopes. The envelopes and the original randomization list were kept by the chief nurse and were not available to attending physicians. The envelope was opened in the operating room. If the patient was allocated to denervation, the ureters were identified and the peritoneum was incised longitudinally along the lateral margins of the uterosacral ligaments, which were then grasped and pulled toward the midline. The ureters were pushed laterally, and the full thickness of the ligaments was coagulated with bipolar forceps. A segment 1 cm or greater in depth and length was excised with scissors close to the uterine junction. Only mechanical instruments and electrosurgery were used for the endoscopic procedures. Histopathologic identification of nerve fibers in resected specimens was not attempted systematically. After operative laparoscopy, patients underwent clinical examination every 3 months and completed the pain questionnaire monthly. Patients were withdrawn from the study if after surgery they used any type of hormonal therapy that could affect pain symptoms associated with endometriosis. However, they were allowed to take nonsteroidal anti-inflammatory drugs for dysmenorrhea (naproxen sodium, one 550-mg tablet b.i.d. unless contraindicated). Alcohol intake was not monitored during the study, because abusers of alcohol were identified and excluded at pretrial screening. At the 12-month evaluation, the women were requested to rate the overall degree of satisfaction with their treatment on a five-category Likert (very satisfied, satisfied, uncertain, dissatisfied, very dissatisfied) and to complete again the Short Form-36 questionnaire, the Hospital Anxiety and Depression Scale, and the Sabbatsberg Sexual Rating Scale. The preplanned duration of follow-up was 1 year. However, in view of the suggested progressive decrease in relief of pain with increasing time after surgery, we decided to prolong the follow-up to 3 years. During the second and third postoperative year, the patients underwent clinical and ultrasonographic evaluation every 6 months. On these occasions, only pain diaries were collected and the degree of satisfaction with treatment was recorded. In calculating the required sample size, the primary outcome of interest was the rate of dysmenorrhea recurrence 1 year after surgery. On the basis of our previous experience, we estimated the medium-term rate of recurrence of dysmenorrhea after conservative surgery for symptomatic endometriosis to be around 30%. With the addition of paracervical denervation, we expected the percentage of dysmenorrhea recurrence to be lower, about 10%. A difference of 20% between study groups was considered clinically relevant. To have an 80% chance of detecting such a difference at an overall significance level of 5%, 72 patients per group were required. Allowing for dropouts, the aim was to recruit a total of 180 participants. We computed the rates of recurrence of moderate to severe pain symptoms and satisfaction with treatment (very satisfied satisfied vs. uncertain dissatisfied very dissatisfied) for both study arms and tested the differences by using 2 statistics. The different categories of patient satisfaction were not assigned individual scores. To allow evaluation of participants with different lengths of follow-up, the Kaplan Meier method was used to calculate the 36-month cumulative proportion of women who reported moderate or severe dysmenorrhea recurrence, and the curves obtained were compared by using the log-rank test. The event dates used were the date of surgery and the date of recurrence of pain, or last follow-up visit, or last menstrual period in case of conception. Women who became pregnant during follow-up or were withdrawn for reasons other than symptoms recurrence were not excluded from this analysis. Variations in pelvic symptoms were evaluated by computing pain scores on the multidimensional categorical rating and visual analogue s; differences between baseline and final median scores in the two groups were compared by using the Mann Whitney U test. The same method was chosen to compare variations between the two study groups in scores on the Short Form-36, the Hospital Anxiety and Depression Scale, and the Sabbatsberg Sexual Rating Scale. Within-group variations between baseline and follow-up values were evaluated by using the Wilcoxon matched pairs test. We used nonparametric tests to avoid any distributional assumption. When appropriate, the 95% confidence interval was calculated for the observed differences. All statistical tests were two-sided. P.05 was considered significant. RESULTS Two hundred seventy-three women evaluated in two endometriosis outpatient clinics were considered for the study. 312 Vercellini et al. Uterosacral ligament resection for dysmenorrhea Vol. 80, No. 2, August 2003

4 FIGURE 1 Trial profile. TABLE 1 Distribution of patients according to age, parity, endometriosis stage, and treatment allocation. Characteristic No. in the uterosacral ligament resection group (%) (n 90) No. in the laparoscopic surgery only group (%) (n 90) Age (y) (51) 48 (53) (49) 42 (47) Parity 0 77 (86) 78 (87) 1 13 (14) 12 (13) Endometriosis stage I 35 (39) 29 (32) II 20 (22) 17 (19) III 18 (20) 23 (26) IV 17 (19) 21 (23) Ninety-three women were excluded for various reasons (Fig. 1). The remaining 180 patients, who had a mean ( SD) age of 30 5 years, were enrolled; 90 were allocated to conservative surgery including resection of the uterosacral ligaments and 90 to conservative surgery alone. Table 1 shows the distribution of randomized women according to age, parity, endometriosis stage, and treatment allocation. The distribution of these variables was similar in the two groups. No patient experienced complications during or after surgery, and no procedure was converted to laparotomy. The mean time required to resect the uterosacral ligaments was 10 4 minutes. The groups did not differ in postoperative morbidity rates (data not shown), and 161 women left the hospital within 24 hours of the intervention. Moderate abdominal or thoracic pain was reported by 6 patients in the experimental group and 7 in the control group, and 3 patients in each group experienced short-term fever. These women were discharged within 48 hours. Twenty-four patients (13%) (12 in each group) were withdrawn within 1 year of randomization (Fig. 1). Patients who were lost to follow-up were included as failures in the evaluation of satisfaction with treatment, whereas in the case of pregnant women and those excluded for reasons other than symptom recurrence, the last assessment before conception or withdrawal was considered. These 24 patients were excluded from analyses of pain, quality of life, psychological profile, and sexual satisfaction scores because complete diaries were not available. Table 2 shows pain intensity at 12 months. A statistically significant decrease was observed in both groups in all symptom scores on the visual analogue and multidimensional categorical rating. Among patients who were evaluable 1 year after operative laparoscopy, 23 of 78 (29%) women in the uterosacral ligament resection group and 21 of 78 (27%) women in the conservative surgery group reported recurrent dysmenorrhea, defined as moderate or severe pain on both pain s ( ; P.859) (odds ratio, 1.13 [95% CI, 0.56 to 2.28]). With the exception of 4 patients in the uterosacral ligament resection group and 3 in the conservative surgery group, recurrent dysmenorrhea was predominantly of the hypogastric, midline type. Sixty-four (36%) women (31 in the uterosacral ligament resection group and 33 in the laparoscopic surgery only group) were withdrawn within 3 years of treatment (Fig. 1). Table 3 shows pain symptoms at 3 years of follow-up. The overall benefit of operative laparoscopy was confirmed in both groups, as follow-up scores were significantly lower than baseline scores. However, scores for women who underwent paracervical denervation were not significantly better than those for women treated with conservative surgery only. FERTILITY & STERILITY 313

5 TABLE 2 Pain symptom scores in patients with endometriosis before and 1 year after laparoscopic surgical treatment, according to treatment allocation and pain. Uterosacral ligament resection group Laparoscopic surgery only group Symptom Visual analogue Multi-dimensional Visual analogue Multi-dimensional Dysmenorrhea 78 a 78 a Baseline 78 (69 85) 2 (2 2) 76 (65 86) 2 (2 2) 1 year of follow-up 25 (0 49) 1 (0 1) 12 (0 35) 0 (0 1) reduction b,c,d 52 (24 70) 1 (1 2) 58 (40 74) 2 (1 2) Deep dyspareunia e 44 a 39 a Baseline 60 (47 75) 1 (1 2) 54 (30 75) 1 (1 2) 1 year of follow-up 0 (0 25) 0 (0 1) 0 (0 21) 0 (0 1) reduction b,c,d 43 (30 61) 1 (1 2) 33 (20 55) 1 (1 1) Nonmenstrual pain 58 a 49 a Baseline 55 (30 70) 1 (1 1) 35 (25 50) 1 (1 1) 1 year of follow-up 18 (0 30) 1 (0 1) 0 (0 20) 0 (0 1) reduction b,c,d 32 (14 58) 0 (0 1) 31 (22 42) 1 (1 1) Note: Values are medians (interquartile ranges). a Number of patients reporting the symptom. b Baseline values minus values at the end of follow-up. c The within-group difference is statistically significant (P.01). d The between-group difference is not statistically significant. e Seven patients in the uterosacral ligament resection group and six in the laparoscopic surgery only group reported no sexual activity. Among patients who were evaluable 3 years after surgery, 21 of 59 (36%) women in the uterosacral ligament resection group and 18 of 57 (32%) women in the conservative surgery group reported recurrent dysmenorrhea ( ; P.794) (odds ratio, 1.19 [95% CI, 0.55 to 2.59]). According to the Kaplan Meier method, no significant difference was observed in time to symptom recurrence. The cumulative probability of recurrence of moderate to severe dysmenorrhea was 27% at 1 year and 43% at 3 years in the experimental group and 25% at 1 year and 36% at 3 years in the control group (Fig. 2). Subgroup analysis by endometriosis stage after stratification based on the revised American Fertility Society classification (11) did not reveal significant differences between the two study groups in the effect of treatments on dysmenorrhea. Similarly, no significant between-group difference was detected in the frequency and duration of use of nonsteroidal anti-inflammatory drugs during follow-up (data not shown). Table 4 shows scores on the Short Form-36 questionnaire obtained 12 months after surgery in the study groups. Compared with baseline values, mean scores increased significantly in all dimensions except physical functioning and emotional role limitation in the uterosacral ligament resection group and general health perception, vitality, and mental health in the conservative surgery group. No between-group difference was statistically significant. Figure 3 compares the scores observed in the two study groups with normative data for a general Italian female population of corresponding age (13, 15). The comparison demonstrates the direction and relevance of the effect of the two treatment approaches on patients health-related quality of life. The distribution of the 1-year Short Form-36 scores for the eight dimensions was similar in the two study groups and the general female population, whereas baseline scores were generally lower. Significant mean improvements in the Hospital Anxiety and Depression Scale as well as in the Sexual Rating Scale scores were seen in both groups. However, no variable considered at 1 year of follow-up differed significantly between groups (Table 3). We also analyzed data according to the intention-to-treat principle. At 1 year, 20 women in the uterosacral ligament resection group were very satisfied with their treatment, 48 were satisfied, 11 uncertain, 10 dissatisfied, and 1 very dissatisfied. Corresponding figures for the laparoscopic surgery only group were 14, 53, 10, 13, and 0 women, respectively. Overall, 68 of 90 (75%) patients in the uterosacral ligament resection group and 67 of 90 (74%) in the laparoscopic surgery only group were very satisfied or satisfied. At 3 years, 12 women in the uterosacral ligament resection 314 Vercellini et al. Uterosacral ligament resection for dysmenorrhea Vol. 80, No. 2, August 2003

6 TABLE 3 Pain symptom scores in patients with endometriosis before and 3 years after laparoscopic surgical treatment, according to treatment allocation and pain. Uterosacral ligament resection group Laparoscopic surgery only group Symptom Visual analogue Multi-dimensional Visual analogue Multi-dimensional Dysmenorrhea 59 a 57 a Baseline 75 (69 84) 2 (2 2) 80 (70 86) 2 (2 2) 3-year follow-up 38 (8 68) 1 (0 1) 40 (10 60) 1 (1 1) reduction b,c,d 37 (20 56) 1 (1 2) 43 (26 64) 1 (1 2) Deep dyspareunia e 37 a 28 a Baseline 58 (45 72) 1 (1 2) 54 (26 67) 1 (1 2) 3-year follow-up 22 (0 35) 0 (1 1) 18 (0 30) 0 (0 1) reduction b,c,d 24 (16 36) 1 (1 1) 20 (17 38) 1 (0 1) Nonmenstrual pain 42 a 35 a Baseline 50 (29 74) 1 (1 1) 39 (30 52) 1 (1 1) 3-year follow-up 25 (0 35) 1 (1 1) 20 (0 50) 0 (0 1) reduction b,c,d 28 (14 40) 0 (0 1) 22 (0 37) 1 (0 1) Note: Values are medians (interquartile ranges). a Number of patients reporting the symptom. b Baseline values minus values at the end of follow-up. c The within-group difference is statistically significant (P.01). d The between-group difference is not statistically significant. e Five patients in the uterosacral ligament resection group and four in the laparoscopic surgery only group reported no sexual activity. FIGURE 2 Cumulative 36-month probability of recurrence of moderate or severe dysmenorrhea, as assessed by a linear analogue in 180 symptomatic women with endometriosis who had laparoscopic surgery with (solid line) or without (dashed line) uterosacral ligament resection (log-rank test, ; P.59). FERTILITY & STERILITY 315

7 TABLE 4 Comparison of Short Form-36, Hospital Anxiety and Depression Scale, and revised Sabbatsberg Sexual Rating Scale scores 1 year after laparoscopic surgery with or without uterosacral ligament resection. Uterosacral ligament resection group Laparoscopic surgery only group Scale Baseline Follow-up Baseline Follow-up Short Form-36 a Physical functioning Role limitation (physical) Pain General health perceptions Vitality Social functioning Role limitation (emotional) Mental health Hospital Anxiety and Depression Scale a Anxiety Depression Total Revised Sabbatsberg Sexual Rating Scale a Note: Values are means ( SD). a None of the between-group differences is statistically significant. group were very satisfied, 43 satisfied, 10 uncertain, 23 dissatisfied, and 2 very dissatisfied. Corresponding figures for the laparoscopic surgery only group were 10, 49, 6, 24, and 1, respectively. Overall, 55 of 90 (61%) paients in the uterosacral ligament resection group and 59 of 90 (65%) patients in the laparoscopic surgery only group were very satisfied or satisfied ( ; P.763) (odds ratio, 0.87 [95% CI, 0.48 to 1.57]). This analysis included all randomized patients, with those lost to follow-up considered as dissatisfied. Categorization of these latter patients as very satisfied or as satisfied or partially satisfied in the same proportion as patients who completed the trial did not significantly change the results. No side effects attributable to pelvic denervation, specifically changes in bladder and intestinal function, were observed. Twelve patients in the experimental group and 17 in the control group achieved a pregnancy within 3 years of surgery. The first stage of labor in women who underwent uterosacral ligament resection and delivered vaginally was not painless. DISCUSSION Both sympathetic and parasympathetic fibers are found in the anterior two thirds of the uterosacral ligaments, around the area of attachment to the cervix (5, 6). Doyle attributed the reported failures of presacral neurectomy in ameliorating dysmenorrhea to incomplete denervation namely, resection of the superior hypogastric plexus that spared the nervi erigentes and hypothesized that surgical interruption of both the sympathetic and parasympathetic pathways was fundamental to achieve satisfactory pain relief (20). He proposed paracervical uterine denervation, a simple procedure that consists of amputation of the uterosacral ligaments as usually performed during the first steps of a vaginal hysterectomy. Complete or partial relief of dysmenorrhea was reported in 69 of the 73 (94%) patients who had this surgery; however, most were unaffected by endometriosis (20). Uterosacral ligament resection in addition to presacral neurectomy has been performed at laparotomy in women with endometriosis with inconsistent results (21, 22). With the widespread use of operative laparoscopy, many surgeons began to resect the uterosacral ligaments systematically during operations for endometriosis in patients with predominantly midline dysmenorrhea (2, 7 10). This accepted surgical custom is based on a convincing neuroanatomic rationale but not on robust scientific data. The quality of the evidence from the available studies is suboptimal (2, 10), and we are aware of only one small randomized controlled trial conducted on this topic, the results of which did not demonstrate an additional antalgic effect of paracervical denervation over and above that of ablation of endometriotic lesions (9). Endometriosis was an exclusion criterion in another published randomized study showing that laparoscopic uterosacral ligament resection has an effect on dysmenorrhea (23). 316 Vercellini et al. Uterosacral ligament resection for dysmenorrhea Vol. 80, No. 2, August 2003

8 FIGURE 3 The Short Form-36 general health survey profile of dimension-specific scores observed in patients with endometriosisassociated dysmenorrhea before and 12 months after laparoscopic surgery with (A) (n 78) or without (B) (n 78) uterosacral ligament resection compared with Italian normative data obtained from a sample of 1,032 women 25 to 34 years of age. The main limitation of our study was its open-label design; for organizational reasons, we could not perform a double-blind trial. However, we believe that the placebo effect potentially associated with knowing that denervation was performed would have enhanced a between-group difference in pain variation, which was not observed. Furthermore, the placebo effect of treatments for dysmenorrhea vanishes within a few months (24). We took great care to ensure that the randomization procedure was performed properly and to conceal treatment allocation sequence to surgeons on the operating list. FERTILITY & STERILITY 317

9 In addition, two s were administered for pain assessment: a visual analogue that measured mainly the subjective aspect of symptoms, and a multidimensional categorical rating that measured the functional impairment caused by pelvic pain. The recurrence rate of moderate to severe dysmenorrhea 1 year after operative laparoscopy was 29% after uterosacral ligament resection and 27% after conservative surgery alone. The sample size was calculated on the basis of the expected rate of pain recurrence, which proved correct for the standard treatment group but too optimistic for the experimental group. Absolute figures at 3 years of follow-up should be interpreted with caution because a substantial proportion of patients in both groups were lost to follow-up or excluded for major protocol violations or miscellaneous reasons. Survival analysis confirmed that the antalgic effect of surgery gradually vanishes, and the groups did not differ in time to recurrence of pain. According to the Kaplan Meier method, the cumulative probability of experiencing troublesome or intense pain at menstruation increased progressively, reaching almost 40% 3 years after operative laparoscopy. This finding underscores the importance of a sufficiently long follow-up in this type of research setting, to prevent falsely optimistic conclusions. The absence of an effect of paracervical denervation is hard to explain, particularly because almost all women in this series had recurrent dysmenorrhea of the midline hypogastric type, as was also the case in our previous trial on presacral neurectomy. We hypothesize that endometriosis causes pelvic inflammation, which may stimulate pain sensory fibers that innervate the pelvic peritoneum. If pain at menstruation in women with endometriosis is mainly due to intraperitoneal bleeding (25), interruption of uterine innervation may not be more effective than simple ablation of lesions. Alternatively, sympathetic and parasympathetic fibers that innervate the uterus may lie in, under, and around the attachments of the uterosacral ligaments to the cervix (5, 6). If this is the case, simply severing the ligaments may create only partial denervation. This idea is supported by the absence of bladder and intestinal dysfunction among our patients; this complication has been reported as occurring often after presacral neurectomy (4). Moreover, no antalgic effect during the first stage of labor was observed in patients who delivered after resection of the uterosacral ligaments. Finally, we cannot exclude an incorrect or inconsistent surgical technique. However, laparoscopies were always performed or supervised by four expert endoscopists with extensive training in treatment of endometriosis. We did not detect significant between-group differences in reduction of deep dyspareunia. This is perplexing, because pain at intercourse should be related to pain stimuli originating in the retrocervical area, which is most likely to be struck during deep penetration. However, dyspareunia was not a selection criterion, and our study power may not have been sufficient to exclude the possibility that an appreciable beneficial effect of uterosacral ligament resection on this variable went unrecognized. Similarities in variation of nonmenstrual pain were expected, because this symptom is mainly related to ovulation and stimuli arising in the gonads and peritoneum, which are not affected by paracervical denervation (3, 5, 6). In our opinion, evaluation of therapies only in terms of biomedical results (e.g., variations in pain scores) can be misleading, especially in chronic diseases such as endometriosis, which has symptoms that may disrupt working ability, social relationships, and sexual functioning. Accordingly, evaluation of therapeutic interventions in such clinical conditions should include the use of instruments to assess the overall effects on general well-being, not only pain-related benefits. The s for evaluating health-related quality of life, anxiety and depressive symptoms, and sexual functioning were chosen for their demonstrated validity and reliability and their simplicity and acceptability (14 18). No patient refused to provide information on these s. Differences between the groups in the eight health-related quality-of-life scores were limited and probably of marginal clinical significance. Overall, laparoscopic surgery appreciably improved general health, psychoemotional status, and sexual satisfaction, regardless of whether paracervical denervation was performed. Satisfaction with treatment was also substantial, with only one fourth of the participants being uncertain or dissatisfied at 1 year of follow-up. The figure increased to 37% (66 of 180 women) at the end of 3 years of follow-up. However, this estimate considers all patients lost to follow-up as treatment failures. This conservative approach was adopted to prevent inappropriately optimistic conclusions. Moreover, the number of patients lost to follow-up was similar in the two groups; this should have avoided biases in the interpretation of the results. In conclusion, we found that the addition of uterosacral ligament resection to conservative laparoscopic surgery in women with endometriosis did not reduce the medium- or long-term frequency and severity of recurrence of dysmenorrhea. Complications and side effects of paracervical denervation were not observed. However, the procedure entails some risks, and further information on its neuroanatomical rationale and efficacy should be obtained before accepting its routine performance in clinical practice. References 1. Crosignani PG, Vercellini P, Biffignandi F, Costantini W, Cortesi I, Imparato E. Laparoscopy versus laparotomy in conservative surgical treatment for severe endometriosis. Fertil Steril 1996;66: Vercellini P, De Giorgi O, Pisacreta A, Pesole AP, Vicentini S, Crosignani PG. Surgical management of endometriosis. Baillieres Clin Obstet Gynaecol 2000;14: Vercellini et al. Uterosacral ligament resection for dysmenorrhea Vol. 80, No. 2, August 2003

10 3. Vercellini P, Fedele L, Bianchi S, Candiani GB. Pelvic denervation for chronic pain associated with endometriosis: fact or fancy? Am J Obstet Gynecol 1991;165: Candiani GB, Fedele L, Vercellini P, Bianchi S, Di Nola G. Presacral neurectomy for the treatment of pelvic pain associated with endometriosis: a controlled study. Am J Obstet Gynecol 1992;167: Rapkin AJ. Neuroanatomy, neurophysiology, and neuropharmacology of pelvic pain. Clin Obstet Gynecol 1990;33: Bonica JJ. General considerations of pain in the pelvis and perineum. In: Bonica JJ, (ed). The management of pain, 2nd ed. Malvern (PA): Lea & Febiger, 1990: Khan KS, Khan SF, Nwosu CR, Dwarakanath LS, Chien W. Laparoscopic uterosacral nerve ablation in chronic pelvic pain: an overview. Gynaecol Endosc 1999;8: Daniels J, Gray R, Khan KS, Gupta JK. Laparoscopic uterine nerve ablation: a survey of gynaecological practice in UK. Gynaecol Endosc 2000;9: Sutton C, Pooley AS, Jones KD, Dover RW, Haines P. A prospective, randomized, double-blind controlled trial of laparoscopic uterine nerve ablation in the treatment of pelvic pain associated with endometriosis. Gynaecol Endosc 2001;10: Proctor ML, Farquhar CM, Sinclair OJ, Johnson NP. Surgical interruption of pelvic nerve pathways for primary and secondary dysmenorrhoea (Cochrane Review). In: The Cochrane Library, Issue 4, Oxford: Update Software. 11. The American Fertility Society. Revised American Fertility Society classification of endometriosis: Fertil Steril 1985;43: Biberoglu KO, Behrman SJ. Dosage aspects of danazol therapy in endometriosis: short-term and long-term effectiveness. Am J Obstet Gynecol 1981;139: Vercellini P, Trespidi L, De Giorgi O, Cortesi I, Parazzini F, Crosignani PG. Endometriosis and pelvic pain: relation to disease stage and localization. Fertil Steril 1996;65: Ware JE, Sherbourne CD. The MOS 36-Item Short Form Health Survey (SF-36), I: conceptual framework and item selection. Med Care 1992; 30: McHorney CA, Ware JE, Raczek AE. The MOS 36-Item Short Form Health Survey (SF-36), II: psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 1993;31: Apolone G, Mosconi P. The Italian SF-36 health survey: translation, validation and norming. J Clin Epidemiol 1998;51: Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983;67: Garrat AM, Torgerson DJ, Wyness J, Hall MH, Reid DM. Measuring sexual function in premenopausal women. Br J Obstet Gynaecol 1995; 102: Cook AS, Rock JA. The role of laparoscopy in the treatment of endometriosis. Fertil Steril 1991;55: Doyle JB. Paracervical uterine denervation by transection of the cervical plexus for the relief of dysmenorrhea. Am J Obstet Gynecol 1955; 70: Damario MA, Horowitz IR, Rock JA. The role of uterosacral ligament resection in conservative operation for recurrent endometriosis. J Gynecol Surg 1994;10: Lee RB, Stone K, Magelssen D, Belts RP, Benson WL. Presacral neurectomy for chronic pelvic pain. Obstet Gynecol 1986;68: Lichten EM, Bombard J. Surgical treatment of primary dysmenorrhea with laparoscopic uterine nerve ablation. J Reprod Med 1987;32: Fedele L, Marchini M, Acaia A, Garagiola V, Tiengo M. Dynamics and significance of placebo response in primary dysmenorrhea. Pain 1989; 36: Brosens IA. Endometriosis a disease because it is characterized by bleeding. Am J Obstet Gynecol 1997;176: FERTILITY & STERILITY 319