Adjuvant use of melatonin for pain management in endometriosis-associated pelvic pain-A randomized double-blinded, placebo-controlled trial

Endometriosis is estimated to affect approximately 10% of women of reproductive age [ 1 ] defined by presence of ectopic endometrial-like tissue. It is a chronic systemic inflammatory disease. The most common symptoms are severe dysmenorrhea and chronic pelvic pain which are disabling to many women and, in many cases, refractory to treatment. Available treatments for endometriosis-associated pain (EAPP) consist of analgesics, hormonal therapy and sometimes surgery. Melatonin could be a treatment option due to its anti-inflammatory [ 2 ], analgesic [ 3 ], anti-estrogenic properties [ 4 ] as well as its favorable safety profile [ 5 ]. Melatonin has been shown to impair the cellular invasion, migration and proliferation of endometriosis [ 6 – 10 ]. We recently explored the analgesic effect of 10 mg melatonin daily in women suffering from severe dysmenorrhea, however we could not show this regimen to be superior to placebo [ 11 ]. To our knowledge, there is one previous study assessing the effect of melatonin showing 10 mg melatonin daily to reduce EAPP [ 12 ] in a clinically relevant manner. There is a need for clinical studies to further examine the effects in women with EAPP. We conducted our study with 20 mg melatonin daily to further assess melatonin as a possible adjunct treatment in EAPP.

We screened 120 volunteers to include 40 participants, 20 in each group ( Fig 1 ). Sixteen participants completed the study in the placebo group and 18 in the melatonin group. One participant in the melatonin group was lost to follow up and one chose to discontinue due to restless legs (a known side effect). In the placebo group, two were lost to follow up and two discontinued, one due to abdominal pain and one due to forgetting to take the study drug. Main characteristics of the study population are presented in Table 1 . The outcomes are presented as means and SD of the screening cycle. a One participant in the placebo group had both hormonal IUS and Dienogest. LNG = levonorgestrel. COCP = combined oral contraceptive pill. All variables were similar between the groups except for dyspareunia. In the placebo group endometriosis had been diagnosed in 17 (85%) participants through laparoscopy, 2 (10%) with ultrasound and one (5%) through laparotomy. In the melatonin group 13 (65%) had been diagnosed through laparoscopy, 5 (25%) with ultrasound, 1(5%) with MRI and one (5%) with symptomatology (the latter was wrongfully included in the study). In the placebo group 16 (80%) participants reported no comorbidities, one (5%) participant had migraine, one (5%) had fibromyalgia and one (5%) had a hiatal hernia. In the melatonin group 15 (75%) participants reported no comorbidities, one (5%) had hypothyroidism, one (5%) had treatment for high blood pressure, one (5%) had had a gastric sleeve done and PCOS. One (5%) in each group had had a gastric by-pass surgery. Two (20%) participants in the melatonin group took antidepressants. No significant differences in primary or secondary outcomes were seen between the treatment groups ( Table 2 ). Analyzed with linear mixed models. Adjusting for weight or for hormonal therapy did not affect the results. A sensitivity analysis stratified by hormonal treatment showed no differences between the groups ( S1 Table ). No statistically significant interaction between time and treatment group was seen, which imply similar output between the groups for all time points. The level of insomnia was similar between the treatment groups ( Table 3 ). Four (25%) participants in the placebo group had clinical insomnia (ISI >14), 3 (17%) in the melatonin group. Quality of life, assessed with EHP as well as the level of PCS was similar between the groups ( Table 3 ). No difference in the mean number of days per cycle with EAPP, which were 22.6 (SD 9.17) and 21.0 (SD 8.14) or mean number of days with analgesics, which were 10.8 (SD 13.18) and 10.4 (SD 9.20) in the placebo and melatonin group respectively. Out of the analgesics taken during the study period acetaminophen made up 48%, NSAIDs 46%, codeine 5% and 1% other opioids. In the placebo group there were 4 whole days of absenteeism and 6 half days, in the melatonin group there was one half day of absenteeism, but 68 days off compared to 41 in the placebo group. Independent t-test was used to evaluate sleep, quality of life and pain catastrophization. The participants answered to questionnaire on day 21 in the last treatment cycle. When assessing acceptability at the end of the study period 11 (61%) in the melatonin group had a good experience with the drug, 1 (6%) had a bad experience and 6 (33%) did not know. In the placebo group the reports showed 9 (56%), 3 (19%) and 4 (25%) respectively. 11 (61%) and 10 (62.5%) would recommend the study drug to a friend in the melatonin group and placebo group respectively. No statistically significant differences were seen between the treatment groups in regard to having had a good or bad experience with the drug or recommending it to a friend.

Our study could not show that 20 mg of melatonin given orally at bedtime during two menstrual cycles/ two months had better analgesic effect on endometriosis-associated pain as compared with placebo. We studied a higher dose melatonin than previously studied for EAPP, however no major adverse effects were observed. The mean levels of pain in our study might have been too low to detect any effect compared with placebo. Further studies with larger study populations may be of use to this group of patients suffering from chronic pain and limited treatment options.

We conducted this randomized, double-blinded, parallel placebo-controlled trial at Södersjukhuset, a hospital in Stockholm, Sweden. Participants were recruited between August 2019 and March 2021. Prior to enrolment, a written informed consent was obtained from the participants. The principles expressed in the Declaration of Helsinki were respected in conducting the trial. The trial was approved by The Regional Ethical Review Board at Karolinska Institutet (2017/1177-21/2). Registration number at Clincaltrials.gov NCT03782740 . The first participant was enrolled 21August 2019, last patient last visit was 27 June 2021. Call for participation was advertised on posters in the hospital, in gynecological outpatient clinics, in maternity care outpatient clinics, and on social media. We recruited women who rated their dysmenorrhea 7 or higher on a numeric rating scale (NRS) during the most painful day or a mean NRS of 3 or more for seven days with endometriosis diagnosed through laparoscopy, ultrasound or MRI. All participants were aged 18–50, in good general health and speaking and understanding Swedish. Initial recruitment included only women without any hormonal treatment with regular menstrual bleedings. The inclusion criteria was expanded to women with hormonal treatment, with or without amenorrhea, due to recruitment difficulties. Screening was made by phone followed by one observational menstrual cycle or 30-day period, during this period of time pain was recorded daily, and evaluated prior to inclusion. The first patients enrolled came for a visit with one of the doctors in charge of the trial, at the clinics Research Center for Womens’ Health for inclusion. The visit included medical history and a pregnancy test. Due to Covid-19 the study was converted entirely to remote visits. The inclusion visit was substituted with a video call for identification with id-card, after randomization the study drugs were sent by registered mail or delivered personally to participants living in Stockholm, a pregnancy test and a consent form were included in the consignment. A picture of the taken pregnancy test was returned prior to commencing with the study drug and the signed form of consent was returned by mail. Exclusion criteria were smoking, pregnancy, prior or current liver or kidney disease, ongoing use of melatonin, alteration of any medication during the last three months, regular use of opioids. After inclusion, participants were randomized to 20 mg melatonin or placebo, each dose identical and dispersed in four capsules of 5 mg melatonin or placebo (both manufactured for the trial by APL, Stockholm, Sweden). The study drug was taken at bedtime daily for two consecutive menstrual cycles with start on the first day of menstrual bleeding or 60 consecutive days for amenorrhoeic participants. If needed, the participants continued with their usual pain medication regime, for the three menstrual cycles/ months of the study. The study period was one observational cycle/month followed by two interventional cycles/months, i.e. three menstrual cycles/months in total. The primary outcome was EAPP recorded daily. The secondary outcomes were use of analgesics, dysuria, dyschezia, dyspareunia, days of pain, absenteeism, quality of life, pain catastrophization and sleep. Daily assessments of pain (using NRS), use of analgesics (with specified dosage and number of tabelts), absenteeism and bleeding using a pictorial blood loss assessment chart (PBAC) were made at bedtime using an online questionnaire sent by email every day. Recording of potential adverse events was included in the questionnaire. General experience of the study drug was evaluated at completion of the study. Quality of life was assessed with Endometriosis Health Profile-30 (EHP-30), chronic pain was assessed with the Pain Catastrophizing Scale and sleep was assessed with Insomnia sleep index (ISI) all done with online questionnaires during the first and last cycle/month respectively. The study data was collected and then managed using REDCap electronic data capture tools (9.5.9 Vanderbuilt University, Nashville TN, USA) hosted at Karolinska Institutet. Participants were randomized consecutively in order of inclusion. The manufacturer of the study drug provided consecutively numbered drug containers and made the randomization in blocks of 4. The randomization key was retrieved and opened after the last participant had completed the study asserting double blinding. To detect a clinically significant reduction of NRS of 1.3 units [ 13 ] and an assumed standard deviation (SD) of 1.2 with a power of 80%, and a 2-sided alpha value of 0.05, 15 participants in each group were needed. Like a similar study [ 12 ], we included 20 participants in each group, in total 40 participants, to compensate for those lost to follow up ( Fig 1 ). The calculation was done with sample-size.net [ 14 ] using independent t-test. We used mixed models analysis to study if there were any differences between the groups over all time points for each of the five continuous outcomes recorded daily (EAPP, amount of ingested analgesic, dysuria, dyschezia and dyspareunia) [ 15 ]. Covariance structure AR(1) with random intercept, without random slope proved to be the best fit for all outcomes. We used a model with time and treatment as fixed factors. To study if the effect between the treatments differed over time, we added an interaction between time and treatment. In addition, to study if weight and hormonal therapy would adjust the treatment effect these factors were added one at a time to the model with treatment and time and thereafter interactions between weight and treatment as well as for hormonal therapy and treatment were tested one at a time. We averaged the daily recorded values to generate one value per week, keeping the menstrual cycle as an independent variable. The first day was excluded from inference since the study drug was ingested at bedtime. Data was analysed according to the randomized treatment assignments, for patients who fulfilled the inclusion and exclusion criteria (intention-to-treat), as well as according to the per-protocol. Because the results were similar, we only present the ITT analysis. Independent t-tests were used to compare continuous data in baseline characteristics, mean days of dysmenorrhea, days of using analgesics, EHP-30, pain catastrophizing scale and insomnia sleep index. Categorical data was assessed with chi square or Fisher’s exact test. Acceptability was compared with Fisher’s exact test. A two-tailed p-value of less than 0.05 was considered to indicate statistical significance. SPSS version 26 (SPSS, Chicago, IL) was used for data analyses.

(DOC) Click here for additional data file. Stratification by hormonal treatment end weight, respectively. (XLSX) Click here for additional data file. (DOCX) Click here for additional data file. (XLSX) Click here for additional data file.