A double-blind randomised controlled trial of laparoscopic uterine nerve ablation for women with chronic pelvic pain

Transcription

1 BJOG: an International Journal of Obstetrics and Gynaecology September 2004, Vol. 111, pp DOI: /j x A double-blind randomised controlled trial of laparoscopic uterine nerve ablation for women with chronic pelvic pain N.P. Johnson, C.M. Farquhar, S. Crossley, Y. Yu, A.M. Van Peperstraten, M. Sprecher, J. Suckling Objective To determine the effectiveness of laparoscopic uterine nerve ablation (LUNA) for chronic pelvic pain in women with endometriosis and women with no laparoscopic evidence of endometriosis. Design A prospective double-blind randomised controlled trial (RCT). Setting Single-centre, secondary-level gynaecology outpatient service and tertiary-level pelvic pain and endometriosis outpatient service in Auckland, New Zealand. Population One hundred and twenty-three women undergoing laparoscopy for investigation and management of chronic pelvic pain, 56 with no laparoscopic evidence of endometriosis and 67 with endometriosis. Methods Women were randomised from the two populations, firstly those with no evidence of endometriosis and secondly those undergoing laparoscopic surgical treatment for endometriosis, to receive LUNA or no LUNA. Participant and assessor blinding was employed. Follow up for pain outcomes was undertaken at 24 hours, 3 months and 12 months. Main outcome measures Changes in non-menstrual pelvic pain, dysmenorrhoea, deep dyspareunia and dyschezia were assessed primarily by whether there was a decrease in visual analogue score for these types of pain of 50% or more from baseline and additionally whether there was a significantly different change in median visual analogue score. The numbers requiring further surgery or starting a new medical treatment for pelvic pain and complications were also measured. Results There was a significant reduction in dysmenorrhoea at 12 month follow up in women with chronic pelvic pain in the absence of endometriosis who underwent LUNA (median change in visual analogue scale (VAS) from baseline 4.8 versus 0.8 (P ¼ 0.039), 42.1% versus 14.3% experiencing a successful treatment defined as a 50% or greater reduction in visual analogue pain scale for dysmenorrhoea (P ¼ 0.045). There was no significant difference in non-menstrual pelvic pain, deep dyspareunia or dyschezia in women with no endometriosis undergoing LUNA versus no LUNA. The addition of LUNA to laparoscopic surgical treatment of endometriosis was not associated with a significant difference in any pain outcomes. Conclusions LUNA is effective for dysmenorrhoea in the absence of endometriosis, although there is no evidence of effectiveness of LUNA for non-dysmenorrhoeic chronic pelvic pain or for any type of chronic pelvic pain related to endometriosis. INTRODUCTION Chronic pelvic pain is a common condition that often has a profound impact on a woman s personal health and quality of life, but also an economic impact through loss of working hours. 1 Such pain may present as dysmenorrhoea (primary or secondary), non-menstrual pain, deep dyspareunia (pain with sexual intercourse) or dyschezia (defaecatory pain). Treatment of the condition is sometimes unrewarding owing to a lack of effective interventions and Department of Obstetrics and Gynaecology, National Women s Hospital, University of Auckland, Auckland, New Zealand Correspondence: Dr N. P. Johnson, Department of Obstetrics and Gynaecology, National Women s Hospital, University of Auckland, Auckland, New Zealand. D RCOG 2004 BJOG: an International Journal of Obstetrics and Gynaecology more radical surgery, such as hysterectomy, often becomes the final option. 1 If conservative surgery can be shown to be effective, this would represent a major improvement in the management of chronic pelvic pain. Endometriosis is the most common identifiable pathological condition associated with chronic pelvic pain. Although laparoscopic surgical treatment of endometriosis has been shown to be effective for pain relief, 2 not all women experience a sustained benefit from surgery. Medical treatment for endometriosis-related pain carries a high incidence of side effects and the benefit is not sustained once treatment ceases. 3 If a simple procedure to ablate the important sensory afferent nerves carrying pain stimuli from the pelvis could be carried out, this could theoretically be beneficial for all women with chronic pelvic pain. Uterine nerve ablation involves the transection of the uterosacral ligaments close to their insertion into the cervix. The procedure interrupts pelvic afferent sensory nerve fibres of the Lee Frankenhauser nerve plexus (Fig. 1).

2 LUNA EFFECTIVE FOR CHRONIC PELVIC PAIN ONLY IN WOMEN WITH DYSMENORRHOEA IN THE ABSENCE OF ENDOMETRIOSIS 951 Fig. 1. Sensory afferent nerve supply of the female pelvic organs. C ¼ afferent nerve supply of cervix (illustrated on right side of diagram); O ¼ afferent nerve supply of ovary (illustrated on left side of diagram); U ¼ afferent nerve supply of uterus (illustrated on right side of diagram). P Parasympathetic nerve; S sympathetic nerve. Copyright approval for reproduction of this figure has been kindly granted by Blackwell Science, publisher of Gynaecological Endoscopy. 12 In 1955 Doyle described a technique of vaginal transection of the uterosacral nerves apparently effective for dysmenorrhoea. 4 Uncontrolled studies claimed support for the use of laparoscopic uterine nerve ablation (LUNA) for both primary and secondary dysmenorrhoea with either complete relief or substantial reduction in menstrual pain claimed in the majority of subjects. 5 7 Pain outcomes are susceptible to placebo effect, thus the importance of randomisation and double-blinding in the elimination of bias in the assessment of pain outcomes following any intervention. A systematic review of surgical pelvic neuroablation, including four randomised trials of LUNA, 8 11 found limited evidence that LUNA may provide effective relief from primary dysmenorrhoea but no evidence of effectiveness when LUNA was added to conservative surgery for endometriosis. 12 It was concluded that there was insufficient evidence to recommend the routine use of LUNA in the management of chronic pelvic pain, but suggested there may be a difference in response between women with endometriosis and those with chronic pelvic pain in the absence of endometriosis. There have been isolated case reports of uterine prolapse 13,14 and bladder dysfunction 15 following LUNA, although the complication rate attributable directly to LUNA appears to be extremely low. 12 Other adverse effects of neuroablation, such as haematoma formation, constipation and subsequent painless labour, appear to be more associated with presacral neurectomy (PSN) than LUNA. Indeed PSN, a more invasive procedure involving the transection of a greater number of nerve fibres than LUNA is associated with a much higher complication rate than LUNA, 12 although a recent randomised trial has supported the effectiveness

3 952 N.P. JOHNSON ET AL. of laparoscopic PSN for severe dysmenorrhoea related to endometriosis. 16 Our randomised controlled trial (RCT) was undertaken to determine the effectiveness of LUNA in two different populations: women with no evidence of endometriosis and women undergoing laparoscopic surgical treatment for endometriosis. METHODS We undertook a single-centre double-blind parallel randomised trial of LUNA in women with chronic pelvic pain. Approval for the trial had been granted by the Auckland Ethics Committee. The inclusion criteria for the study population were as follows: women aged years inclusive; a history of chronic pelvic pain (either dysmenorrhoea, non-menstrual pelvic pain, defaecatory pain or deep dyspareunia for more than six months); no change in medication for the three months prior to trial recruitment. Exclusion criteria were previous hysterectomy or pelvic malignancy, previous LUNA, known ovarian cysts, plan for a pregnancy within 12 months, intention to change other medical treatment that could influence pelvic pain scores within 12 months, laparoscopic findings rendering LUNA impossible (for example frozen pelvis with no access to uterosacral ligaments), if complete transection of one or both uterosacral ligaments was required to remove endometriosis, the finding of pelvic adhesions that did not appear to be due to endometriosis. Eligible women who had given consent were divided into one of two separate populations based on laparoscopic findings those with no laparoscopic evidence of endometriosis and those with laparoscopic appearances of endometriosis. In women without endometriosis, participants were randomised to receive LUNA versus no LUNA. In women undergoing laparoscopic surgical treatment for endometriosis, participants were randomised to receive laparoscopic excision and/or ablation of endometriosis and LUNA versus laparoscopic excision and/or ablation of endometriosis alone. Routine diagnostic laparoscopy was employed and the laparoscopic procedures for endometriosis involved a combination of excision and/or ablation. Excision of endometriosis involved the use of laparoscopic scissors; ablation was performed with bipolar electrocautery on a power setting of 30 W. The majority of the surgical procedures (104 out of 123) were performed, respectively, by one of two clinicians (N.J. and C.F.). The LUNA procedure was performed as follows. Bilateral uterosacral ligament electrocautery was applied (bipolar current set at 30 W) 0.5 cm from the ligamentous insertion to the cervix until the tissue was blanched. Complete transection of the ligaments was then undertaken with laparoscopic scissors and bipolar electrocautery reapplied to the base of the incision to secure haemostasis. The primary outcomes studied were the four pain domains non-menstrual pelvic pain, dysmenorrhoea, deep dyspareunia and dyschezia at the 3 and 12 month follow up from the surgery. The visual analogue scale (VAS) for pain was assessed by the patient from 0 to 10 in a clinic interview with the research nurse (0 ¼ no pain; 10 ¼ worst imaginable pain). Participants were asked to give a VAS for each pain domain as follows: (a) nonmenstrual pelvic pain was pain unrelated to menses over the most recent four weeks; (b) dysmenorrhoea related to the previous menstrual period; (c) deep dyspareunia related to deep pain with intercourse over the most recent four weeks (women who were not sexually active were not included in this statistic); (d) dyschezia related to pain during defaecation over the most recent four weeks. Secondary outcome measures were pain scores on the first post-operative day (to explore whether LUNA might have an influence on immediate relief of chronic pelvic pain or post-operative pain), satisfaction, whether further surgery for pelvic pain had been performed by 12 months, whether new medical treatment was being used at 12 months, whether prolapse or other surgery-related complications had occurred by 12 months. Numbers are presented as counts (percentages) or median {interquartile range} unless otherwise specified. Differences between LUNA and no LUNA within each population were tested using m 2 test or Fisher s exact test if more appropriate for categorical data and Mann Whitney U test for continuous data using SPSS software. Changes in each of the four pain domains over the study period were quantified (a) dichotomously, with an improvement defined as at least a 50% reduction from the baseline VAS pain score, 17 determined a priori as the primary outcome; (b) a change in median VAS pain scores from baseline. An intention-to-treat (ITT) analysis using imputation was employed for the primary outcome to include women lost to follow up or for whom there were missing data, in addition to those who had undergone further surgery for chronic pelvic pain (all of whom were considered to have had unsuccessful primary treatment). A power calculation has been undertaken to determine an appropriate sample size to demonstrate what was deemed to be a clinically important treatment effect of LUNA. 12 For women with chronic pelvic pain in the absence of endometriosis, to have 80% power at the 95% confidence level to detect benefit in 50% of women, assuming benefit in 10% controls, at least 48 participants would be required for analysis following randomisation. For women with endometriosis, to have 80% power at 95% confidence level to detect benefit in 90%, assuming benefit in 60% controls undergoing conventional endometriosis surgery, 18 at least 76 participants would be required for analysis following randomisation. Allowing for losses to follow up, it was planned to recruit 50 women with chronic pelvic pain in the absence of endometriosis and 80 women with endometriosis. It was

4 LUNA EFFECTIVE FOR CHRONIC PELVIC PAIN ONLY IN WOMEN WITH DYSMENORRHOEA IN THE ABSENCE OF ENDOMETRIOSIS 953 Fig. 2. Flow of participants through the trial. initially hoped to achieve recruitment within three years. However, when this did not occur after three years, it was determined by consensus among investigators to allow a further 20 months until the end of 2001 then to close recruitment (a total recruitment period of four years and nine months). Women were recruited from publicly funded and private secondary-level gynaecology clinics and a publicly funded

5 954 N.P. JOHNSON ET AL. Table 1. Baseline characteristics of randomised participants. Values are given as n (%), mean [SD] or median {interquartile range}. LUNA (n ¼ 22) No LUNA (n ¼ 34) LUNA (n ¼ 32) No LUNA (n ¼ 35) Age (years) 29 [5.83] 29 [6.49] 30 [6.71] 29 [5.31] Ethnicity European 16 (73) 28 (82) 24 (75) 25 (71) Maori 3 (14) 1 (3) 2 (6) 1 (3) Pacific Island 1 (5) 1 (3) 2 (6) 5 (14) Indian 0 2 (6) 0 1 (3) Other 2 (9) 2 (6) 4 (13) 3 (9) Parity 0 8 (36) 15 (44) 22 (69) 26 (74) 1 14 (64) 19 (56) 10 (31) 9 (26) Previous laparoscopy 12 (55) 20 (59) 21 (55) 25 (71) Previous laparotomy 2 (9) 8 (24) 4 (13) 8 (23) Ever Used Medication Paracetamol 12 (55) 15 (44) 13 (41) 15 (43) NSAID 9 (41) 21 (62) 19 (59) 13 (37) Opiate analgesics 1 (5) 2 (6) 4 (13) 2 (6) Combined OCP 1 (5) 0 4 (13) 2 (6) Progestogen 2 (10) 6 (18) 6 (19) 7 (20) Danazol, Gestrinone or GnRHa (9) 8 (23) Other 5 (23) 5 (15) 4 (13) 6 (17) Duration pelvic pain (months) 33 {17, 87} 42 {20, 78} 60 {24, 108} 37 {13, 108} Days in pain in last 28 days 15 {5, 24} 15 {7, 26} 12 {4, 28} 20 {8, 28} No. women with Non-menstrual pelvic pain 21 (95) 32 (94) 28 (88) 35 (100) Dysmenorrhoea 19 (86) 28 (82) 30 (94) 31 (89) Deep dyspareunia 14 (64) 20 (59) 19 (59) 21 (60) Dyschezia 11 (50) 20 (59) 19 (59) 26 (74) VAS pain at baseline Non-menstrual pelvic pain 6 {4, 8} 6 {4, 8} 6 {3, 7} 6 {5, 9} Dysmenorrhoea 8 {6, 8} 6 {3, 9} 8 {7, 10} 9 {7, 10} Deep dyspareunia 4 {1, 7} 8 {7, 10} 6 {1, 7} 8 {4, 9} Dyschezia 1 {0, 6} 4 {0, 6} 3 {0, 6} 6 {2, 7} NSAID ¼ non-steroidal anti-inflammatory drug; OCP ¼ oral contraceptive pill; GnRHa ¼ gonadotrophin releasing hormone analogue. tertiary-level endometriosis and pelvic pain clinic in Auckland, New Zealand. Having given informed consent to participate prior to the laparoscopic procedure, women were randomised at the time of the laparoscopy once the laparoscopic diagnosis of endometriosis had been made or were excluded if a LUNA procedure was not technically possible. Randomisation was performed using computer-generated random number sequences (unknown to the research nurse and the surgeons who were the executors of the assignment). Concealment of randomisation was maintained securely by storage in sealed, sequentially numbered opaque envelopes for both the non-endometriosis and the endometriosis groups until the interventions were assigned during the laparoscopic procedure. The numbers were dissimilar in the population with no endometriosis (22 LUNA, 34 no LUNA) owing to unblocked randomisation, although this did not occur in the endometriosis population (32 LUNA, 35 no LUNA). Double-blinding of participants and assessor (a research nurse who was not involved in treatment assignment or the surgery itself) was employed. The assignment envelope was opened intraoperatively. Neither the participant nor the assessor at follow up was aware of the assignment. No record was made in the patient chart whether or not LUNA was performed (this record was maintained in a study file to maintain participant and assessor blinding) and surgical, anaesthetic and nursing staff were aware that participant blinding was necessary. To our knowledge, there were no breaches of blinding. No sham incisions were deemed necessary since all surgeons felt that, in most cases with normal pelvic anatomy, LUNA would be technically feasible using only two incisions (one for the laparoscope and

6 LUNA EFFECTIVE FOR CHRONIC PELVIC PAIN ONLY IN WOMEN WITH DYSMENORRHOEA IN THE ABSENCE OF ENDOMETRIOSIS 955 Table 2. Operative details. Values are given as n (%). LUNA (n ¼ 22) No LUNA (n ¼ 34) LUNA (n ¼ 32) No LUNA (n ¼ 35) Number of laparoscopic incisions 2 12 (55) 23 (68) 4 (13) 4 (11) 3 9 (41) 9 (26) 24 (75) 24 (69) (3) 4 (13) 6 (17) Not Specified 1 (5) 3 (9) 2 (6) 1 (3) Revised ASRM endometriosis score Stage 1 (1 5) (31) 12 (34) Stage 2 (6 15) (41) 13 (37) Stage 3 (16 40) (19) 5 (14) Stage 4 (>40) (9) 3 (9) Pelvic congestion present 11 (50) 17 (50) 7 (22) 3 (9) Surgeon Surgeon 1 11 (50) 18 (53) 17 (53) 21 (60) Surgeon 2 7 (32) 11 (32) 10 (31) 9 (26) Other 4 (18) 5 (15) 5 (16) 5 (14) The revised American Society for Reproductive Medicine score (ASRM) 19 was used to stage endometriosis, although no women had a completely obliterated pouch of Douglas. one for surgical instruments). The assessor and participants were unblinded once the 12 month follow up had been completed. The major financial support for the trial came from the Princess of Wales Memorial Trust with seed funding contributed by Johnson & Johnson (New Zealand) (the suppliers of the bipolar electrocautery laparoscopic surgical instruments). The funding contributed towards the salary of a research nurse. No other external funding was sought. RESULTS The flow of participants through the trial is shown in Fig. 2. During the recruitment phase of the trial, from April 1997 to December 2001 inclusive, 200 women were offered the trial. Of the women offered the trial, 63 declined, the majority citing concerns over future fertility or a concern that nerve ablation could lead to a masking of symptoms of significant pathology in the future. One hundred and thirty-seven women gave informed consent. Of these, 14 were excluded based on laparoscopic findings (in five LUNA was deemed either technically impossible or complete transaction of uterosacral ligaments would have been required to remove the endometriosis; in nine adhesions were present in the absence of typical findings of endometriosis and were felt to be related to previous pelvic inflammatory disease). One hundred and twenty-three women were randomised; 56 in the group with chronic pelvic pain without evidence of endometriosis and 67 in the group with endometriosis. Recruitment was terminated at the end of 2001 with the number of recruited participants required Table 3. Twenty-four hours after surgery. Values are given as n (%) or median {interquartile range}. LUNA (n ¼ 22) No LUNA (n ¼ 34) LUNA (n ¼ 32) No LUNA (n ¼ 35) Pelvic pain VAS 24 hours post-op 0 {0, 6.0} 0 {0, 4.0} 0.5 {0, 4.0} 1.0 {0, 5.0} Pelvic pain Completely resolved 1 (4.6) 5 (14.7) 13 (40.6) 6 (17.1) Partially resolved 3 (13.6) 7 (20.7) 4 (12.5) 6 (17.1) No change 3 (13.6) 3 (8.8) 2 (6.3) 1 (2.9) Increased 0 1 (2.9) 0 2 (5.8) Not able to decide 14 (63.6) 17 (50.0) 13 (40.6) 20 (57.1) Not recorded 1 (4.6) 1 (2.9) 0 0 Surgical pain VAS 24 hours post-op 7.0 {4.0, 8.0} 7.0 {4.0, 8.0} 5.0 {3.5, 6.5} 8.0 {5.3, 8.8}

7 956 N.P. JOHNSON ET AL. Table 4. Pain outcomes: available case analysis. Values are given as n (%) or median {interquartile range}. LUNA (n ¼ 22 randomised) No LUNA (n ¼ 34 randomised) P LUNA (n ¼ 32 randomised) No LUNA (n ¼ 35 randomised) P Non-menstrual pelvic pain Pre-operative 6 {4, 8} 6 {4, 8} 6 {3, 7} 6 {5, 9} 3 months 50% reduction VAS 10/19 (53) 15/32 (47) 9/28 (32) 18/34 (53) VAS change from baseline 2.5 { 5, 0.6} 1 { 5.4, 0} 1.3 { 3.1, 1.0} 3.5 { 5, 2} 12 months 50% reduction VAS 8/17 (47) 13/30 (44) /22 (50) 15/30 (50) VAS change from baseline 4 { 6, 0} 1 { 5, 1.3} { 6, 0} 3.5 { 5.8, 1} Dysmenorrhoea Pre-operative 8 {6, 8} 6 {3, 9} 8 {7, 10} 9 {7, 10} 3 months 50% reduction VAS 3/16 (19) 10/27 (37) 6/26 (23) 11/28 (39) VAS change from baseline 2 { 3.3, 1} 1.5 { 5, 0.4} 0 { 3.5, 0} 2 { 5, 0} 12 months 50% reduction VAS 8/12 (67) 4/21 (19) f 7/21 (33) 11/24 (46) VAS change from baseline 4.8 { 6, 2.9} 0.8 { 3.8, 0} { 7, 1} 3 { 5.5, 0} Deep dyspareunia Pre-operative 4 {1, 7} 8 {7, 10} 6 {1, 7} 8 {4, 9} 3 months 50% reduction VAS 6/12 (50) 10/18 (56) 5/17 (29) 10/18 (56) VAS change from baseline 2 { 5, 0} 1.5 { 5, 0} 0 { 3, 0} 3.5 { 7, 1} 12 months 50% reduction VAS 7/9 (78) 8/14 (57) f 6/10 (60) 8/16 (50) f VAS change from baseline 3 { 4, 0} 2 { 5.5, 0} { 5, 0} 2 { 6, 0.5} Dyschezia Pre-operative 1 {0, 6} 4 {0, 6} 3 {0, 6} 6 {2, 7} 3 months 50% reduction VAS 7/11 (64) 9/20 (45) 9/19 (47) 18/25 (72) VAS change from baseline 0.8 { 3.3, 0} 0 { 3, 0} 0 { 4.5, 0.8} 3 { 5.5, 0} 12 months 50% reduction VAS 8/11 (73) 7/19 (37) f 7/14 (50) 10/23 (43) f VAS change from baseline 1.3 { 4.8, 0} 0 { 2, 0} { 3, 0.25} 1 { 5, 0} % reduction VAS was defined as a 50% or greater reduction in visual analogue pain score in women who had not undergone further surgery for pelvic pain. VAS change from baseline was VAS minus baseline VAS only in women who had not undergone further surgery for pelvic pain. f Fisher s exact test probability value. by the power calculation having been surpassed in the group with no evidence of endometriosis (56 recruited; 50 required) but not attained in the group with endometriosis (67 recruited; 80 required). Data were available for all except one woman at 24 hours. Owing to losses from follow up, data were missing for three women at three months (2 with LUNA and 0 with no LUNA in the population with no endometriosis; 0 with LUNA and 1 with no LUNA in the endometriosis population). At 12 months, 17 women had been lost to follow up (4 with LUNA and 2 with no LUNA in the population with no endometriosis; six with LUNA and five with no LUNA in the endometriosis population), there were also other missing data in women who were followed up and six women had undergone further surgery to treat chronic pelvic pain (four hysterectomies and two further laparoscopic surgical procedures for endometriosis). Thus, at 12 months, the majority of data were available for 50 women with no endometriosis and 56 women with endometriosis (although visual analogue pain scale data were not collected for those who had undergone further surgery). The baseline characteristics of the women at entry to the trial are presented in Table 1. Randomisation led to equitable distribution of baseline characteristics between the randomised groups. Operative details are presented in Table 2. There was no difference in the number of ports used whether LUNA was performed in the endometriosis population (m 2 ¼ 0.34, P ¼ 0.844) or in the population with no endometriosis (m 2 ¼ 0.89, P ¼ 0.346). None of the women with endometriosis had complete obliteration of the pouch of Douglas. No study participants required a return to theatre post-operatively. No important intra-operative or postoperative complications occurred (specifically, there were no cases of ureteric injury, intra-operative bleeding or post-operative haematoma formation), other than two women who had urinary retention requiring catheterisation within 24 hours of the surgery (both in the endometriosis population not undergoing LUNA). At 24 hours after surgery (Table 3), there was no apparent difference between those undergoing LUNA or not in terms of VAS for pelvic pain overall, whether or not endometriosis was present. Table 4 shows an available case analysis of pain outcomes in the various groups. Table 4 includes women

8 LUNA EFFECTIVE FOR CHRONIC PELVIC PAIN ONLY IN WOMEN WITH DYSMENORRHOEA IN THE ABSENCE OF ENDOMETRIOSIS 957 Table 5. Successful treatment for primary pain outcomes at 12 months: ITT analysis. Values are given as n (%). LUNA No LUNA P LUNA No LUNA P Non-menstrual pain 8/21 (38.1) 13/32 (40.6) /28 (39.3) 15/35 (42.9) Dysmenorrhoea 8/19 (42.1) 4/28 (14.3) f 7/30 (23.3) 11/31 (35.5) Dyspareunia 7/14 (50.0) 8/20 (40.0) f 6/19 (31.6) 8/21 (38.1) Dyschezia 8/11 (72.7) 7/20 (35.0) f 7/19 (36.8) 10/26 (38.5) Note that success was defined as a 50% or greater reduction in VAS pain score in women who had not had further surgery for pelvic pain, been lost to follow up or had missing data for the outcomes concerned. An assumption was made that women lost to follow up or with missing data did not have a successful treatment. f Fisher s exact test probability value. undergoing further surgery to treat chronic pelvic pain as an unsuccessful treatment in terms of 50% reduction of VAS (but not for the outcome VAS change from baseline, from which these women were excluded). It does not include those lost to follow up or for whom there were missing data. The number of women with a successful treatment (achieving a 50% or greater reduction in the VAS score at 12 months for dysmenorrhoea compared with that at baseline, in the absence of further surgery for chronic pelvic pain) was significantly higher in the population with no endometriosis undergoing LUNA versus no LUNA. This effect was not seen at three months. No such statistically significant effects were seen for deep dyspareunia, dyschezia and non-menstrual pelvic pain. There was also a significantly greater reduction in median VAS from baseline in the non-endometriosis population undergoing LUNA for dysmenorrhoea. In the endometriosis population, there were no results showing significant improvement in the LUNA group. Table 5 shows an ITT analysis at 12 months, where women lost to follow up or with missing data have been included as an unsuccessful treatment. A successful outcome is defined as a 50% or greater reduction in VAS pain score in the absence of further surgery for pelvic pain or loss to follow up. A similar effect is seen for dysmenorrhoea in the non-endometriosis group undergoing LUNA. Table 6 shows the secondary outcomes at 12 months. There were no differences in satisfaction rates and the numbers requiring further surgery for pelvic pain or starting a new medical treatment were too few to show any differences, between those undergoing LUNA or not in both the endometriosis and non-endometriosis populations. Nine women described symptoms possibly attributable to utero-vaginal prolapse (only three of whom had actually undergone LUNA), but no significant utero-vaginal prolapse could be identified on examination of any of these women. DISCUSSION We have found that a significant improvement in dysmenorrhoea at 12 month follow up may be gained from LUNA in women with chronic pelvic pain with no laparoscopic evidence of endometriosis. We found no evidence that LUNA produces a significant improvement Table 6. Secondary outcomes. Values are given as n (%). LUNA (n ¼ 22) No LUNA (n ¼ 34) LUNA (n ¼ 32) No LUNA (n ¼ 35) Satisfaction as 12 months 15/18 (83) 22/32 (69) 18/26 (69) 24/30 (80) Further surgery for pelvic pain by 12 months Laparoscopic surgery Hysterectomy Using new medical treatment at 12 months Combined oral contraceptive Progestogens Danazol, Gestrinone or GnRHa Mirena Prolapse by 12 months Suggestive symptoms Confirmed by examination

9 958 N.P. JOHNSON ET AL. in non-menstrual pelvic pain, deep dyspareunia or dyschezia in these women. We found no evidence to support an additional LUNA procedure in women undergoing laparoscopic surgical treatment for endometriosis. There was no evidence to suggest an influence of LUNA on immediate relief of chronic pelvic pain or post-operative pain. Other than the randomised design, the main strengths of this study were the individual assessment of two separate populations (women with or without endometriosis), maintenance of double blinding until after the 12 month follow up assessment and the verification of pain outcomes by two assessment methods: 50% reduction in VAS and median VAS change from baseline. It was noteworthy that dysmenorrhoea showed a significant improvement at 12 months in the non-endometriosis population, whichever pain assessment method was used in both the available case and ITT analyses. Unfortunately, the power of the study was reduced by failure to achieve the required sample size in the endometriosis population (despite an extension to the predefined recruitment period, from three to four years and nine months, to compensate for an unexpectedly high proportion of women declining randomisation) and the missing data owing to a higher than expected rate of loss to follow up by 12 months (16.4% in the endometriosis population and 10.7% in the non-endometriosis population). Generalisability might be restricted in a single-centre trial. Our data support the findings of a meta-analysis of previous RCTs, 12 showing evidence of effectiveness of LUNA for chronic pelvic pain in the absence of endometriosis, but insufficient evidence of the effectiveness of adding LUNA to laparoscopic surgical removal of endometriosis. One further recent RCT of LUNA in women undergoing laparoscopic endometriosis surgery has also shown no evidence that addition of LUNA is effective, 20 supporting our findings in the endometriosis population. The suggestion that the effectiveness of LUNA may decline over time towards 12 months 8,9 was not supported by our data. In common with an RCT assessing the effectiveness of laparoscopic surgical removal of endometriosis, 18 the effectiveness of LUNA in women with no endometriosis became apparent only beyond the 3 month follow up stage, when the placebo effect of a laparoscopy may still be apparent. A long term study suggested that success rates declined rapidly over time from 72% in the first year to 39% in the fourth for LUNA, 21 although non-randomised data, especially where patients are aware of whether they had LUNA performed, must be interpreted cautiously when assessing a subjective outcome such as pain, as such data could be easily biased by a placebo effect. The longer term effect of LUNA could not be assessed in our trial, as it was not considered reasonable to maintain blinding beyond 12 months. The absence of LUNA-related complications in our trial (in particular, there were no ureteric injuries and no confirmed cases of uterine prolapse during follow up) supports the finding from other trials that LUNA is a very low risk procedure 12 that may be performed easily by adequately trained gynaecological laparoscopic surgeons, adding only a few minutes to the time of a laparoscopic procedure. The ideal neuroablative surgical procedure for pelvic pain would transect all afferent sensory nerves from all the pelvic organs and leave all other nerves unaffected. While pelvic neuroanatomy is complicated and still not completely understood, what is known makes it clear that no such ideal neuroablative surgical procedure exists (see Fig. 1). The body of the uterus is widely considered to be innervated only by sympathetic nerves. 22 The cervix has predominantly parasympathetic (but also sympathetic) innervation. The afferent sensory nerves from both the uterus and cervix traverse the cervical division of the Lee Frankenhauser plexus, which lies within and around the site of attachment of the uterosacral ligaments to the posterior aspect of the cervix. 23,24 From the uterosacral ligament, the parasympathetic afferent nerves reach the dorsal root ganglia of the first to fourth sacral spinal nerves (S1 S4) via the pelvic splanchnic nerves (nervi erigentes) and inferior hypogastric nerve plexus (also known as the pelvic plexus) then superior hypogastric nerve plexus (also known as the presacral nerve or hypogastric plexus). 25 The sympathetic afferent nerves emerging from the Lee Frankenhauser plexus accompany the uterine, iliac and inferior mesenteric arteries to the sacral sympathetic trunk via the sacral splanchnic nerves, some of which by-pass the superior hypogastric nerve plexus. Afferent nerves accompany both parasympathetic and sympathetic nerves from the ovary: pain fibres by-pass the uterosacral ligament course through corresponding plexuses to their cells of origin in the dorsal root ganglia of the 10th and 11th thoracic spinal nerves (T10 T11) some of the upper ovarian plexus afferent nerves course directly via renal and aortic plexuses and by-pass the presacral nerve. It is no surprise that LUNA has not been shown to be an effective adjunct to laparoscopic surgical removal of endometriosis. The operation interrupts only some of the afferent sensory nerve fibres from the pelvis, thus LUNA may be less effective for pelvic pain associated with more extensive pelvic pathology. Careful study of Fig. 1 clarifies why transection of the Lee Frankenhauser nerve plexus in LUNA could be particularly ineffective for pain arising from the ovary or adjacent tissues, as could be the case in ovarian or paraovarian endometriosis, since all ovarian afferent nerves by-pass the Lee Frankenhauser plexus and many of the pain fibres also by-pass the presacral nerve. Without performing a periarterial sympathectomy of the iliac, inferior mesenteric and ovarian vessels, a number of afferent fibres will always be left intact. It has been argued persuasively that the effectiveness of a LUNA procedure for women with endometriosis-related pain could be due more to a debulking of endometriotic lesions whose most common site on the uterosacral ligament is the very site where a uterine nerve ablation is performed. Indeed the so-called arcus taurinus procedure 26 would be an effective method of cytoreduction in endometriosis this, rather

10 LUNA EFFECTIVE FOR CHRONIC PELVIC PAIN ONLY IN WOMEN WITH DYSMENORRHOEA IN THE ABSENCE OF ENDOMETRIOSIS 959 than the uterine nerve ablation aspect of the procedure, is likely to be responsible for its effectiveness. 26 Although there is evidence of effectiveness of LUNA for dysmenorrhoea in women with chronic pelvic pain in the absence of endometriosis, the pros and cons of a LUNA procedure should be considered carefully in women for whom LUNA may be indicated. In particular, there is no indication for LUNA in women undergoing endometriosis surgery and no evidence that LUNA is beneficial for nonmenstrual pelvic pain, dyspareunia or dyschezia. Furthermore, although the complication rate from LUNA appears to be low, we have observed adhesions at the LUNA site on uterosacral ligaments at second-look laparoscopy, in one case, tubal fimbriae were adherent to a previous LUNA site in an otherwise normal pelvis. A number-needed-to-treat calculation suggests that four (95% confidence interval 2 27) women with dysmenorrhoea unrelated to endometriosis would need to undergo LUNA to gain one additional benefit of a 50% or greater reduction in dysmenorrhoea VAS at 12 months (assuming benefit in 14.3% of women undergoing only diagnostic laparoscopy). Further evidence will be available on completion of an ongoing multicentre RCT assessing the efficacy of LUNA (K. Khan, personal communication). Acknowledgements The authors would like to thank Cassie Mosko, research nurse in the early phase of the study, and Teena West, biostatistician, who assisted with the statistical analyses. The clinic staff at Fertility Plus and National Women s Hospital were highly co-operative in facilitating recruitment. Above all, we are grateful to our patients for participating in the trial. References 1. Dawood MY. Nonsteroidal anti-inflammatory drugs and changing attitudes towards dysmenorrhea. Am J Med 1988;84: Jacobson TZ, Barlow DH, Koninckx PR, Olive D, Farquhar C. Laparoscopic surgery for subfertility associated with endometriosis [Cochrane review]. The Cochrane Library, Issue 1. Chichester: Wiley, Prentice A, Deary AJ, Goldbeck-Wood S, Farquhar C, Smith SK. Gonadotrophin-releasing hormone analogues for pain associated with endometriosis [Cochrane review]. The Cochrane Library, Issue 1. Chichester: Wiley, Doyle JB. Paracervical uterine denervation by transection of the cervical plexus for the relief of dysmenorrhoea. Am J Obstet Gynecol 1955;70: Ewen SP, Sutton CJG. A combined approach to painful heavy periods: laparoscopic laser uterine nerve ablation and endometrial resection. Gynaecol Endosc 1994;3: Perez JJ. Laparoscopic presacral neurectomy. Results of the first 25 cases. J Reprod Med 1990;35: Wiborny R, Pichler B. Endoscopic dissection of the uterosacral ligaments for the treatment of chronic pelvic pain. Gynaecol Endosc 1998;7: Lichten EM, Bombard J. Surgical treatment of primary dysmenorrhea with laparoscopic uterine nerve ablation. J Reprod Med 1987;32: Chen FP, Chang SD, Chu KK, Soong YK. Comparison of laparoscopic presacral neurectomy and laparoscopic uterine nerve ablation for primary dysmenorrhea. J Reprod Med 1996;41: Vercellini P, Aimi G, Busacca M, Uglietti A, Viganali M, Crosignani PG. Laparoscopic uterosacral ligament resection for dysmenorrhea associated with endometriosis: results of a randomized controlled trial [abstract]. Fertil Steril 1997;(Suppl):S3 (October). 11. 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: Johnson N, Wilson M, Farquhar C. Surgical pelvic neuroablation for chronic pelvic pain. Gynaecol Endosc 2000;9: Davis GD. Uterine prolapse after laparoscopic uterosacral transection in nulliparous airborne trainees: a report of three cases. J Reprod Med 1996;41: Good MC, Copas Jr PR, Doody MC. Uterine prolapse after laparoscopic uterosacral transection: a case report. J Reprod Med 1992;37: Fitzpatrick CC, Flood H, Punch M, Hilgers TW, Elkins TE, McGuire EJ. Bladder dysfunction after repeat laparoscopic uterine nerve ablation (LUNA). Int Urogynecol J 1995;6: Zullo F, Palomba S, Zupi E, et al. Effectiveness of presacral neurectomy in women with severe dysmenorrhoea caused by endometriosis who were treated with laparoscopic conservative surgery: a 1-year prospective double-blind controlled trial. Am J Obstet Gynecol 2003;189: Farquhar CM, Rogers V, Franks S, Pearce S, Wadsworth J, Beard RW. A randomized controlled trial of medroxyprogesterone acetate and psychotherapy for the treatment of pelvic congestion. Br J Obstet Gynaecol 1989;96: Sutton CJG, Ewen SP, Whitelaw N, Haines P. Prospective randomized double-blind controlled trial of laser laparoscopy in the treatment of pelvic pain associated with minimal, mild and moderate endometriosis. Fertil Steril 1994;62: American Society for Reproductive Medicine. Revised American Society for Reproductive Medicine classification of endometriosis: Fertil Steril 1997;67: Vercellini P, Aimi G, Busacca M, Apolone G, Uglietti A, Crosignani PG. Laparoscopic uterosacral ligament resection for dysmenorrhea associated with endometriosis: results of a randomized, controlled trial. Fertil Steril 2003;80: Papasakelariou C. Long-term results of laparoscopic uterosacral nerve ablation. Gynaecol Endosc 1996;5: Owman C, Rosenbren E, Sjoberg NO. Adrenergic innervation of the human female reproductive organs: a histochemical and chemical investigation. Obstet Gynecol 1967;30: Frankenhauser G. Die Bewegungenerven der Gebarmutter. Z Med Nat Wiss 1864;1: Cleland JP. Paravertebral anaesthesia in obstetrics. Surg Gynecol Obstet 1933;57: Williams PL, Warwick R. Reproductive organs of the female. In: Williams PL, Warwick R, editors. Gray s Anatomy, 36th edition. Edinburgh: Churchill Livingstone, 1980: Jones KD, Sutton C. Arcus taurinus: the mother and father of all LUNAs. Gynaecol Endosc 2001;10: Accepted 21 March 2004