Menopause: The Journal of The North American Menopause Society Vol. 21, No. 6, pp. 567/573 DOI: 10.1097/gme.0b013e3182a7c073 * 2013 by The North American Menopause Society
Phase 3 randomized controlled study of gastroretentive gabapentin for the treatment of moderate-to-severe hot flashes in menopause JoAnn V. Pinkerton, MD,1 Risa Kagan, MD,2 David Portman, MD,3 Rekha Sathyanarayana, BS,4 Michael Sweeney, MD,4 for the Breeze 3 Investigators Abstract Objective: The goal of this study was to evaluate the efficacy and safety of gastroretentive gabapentin (G-GR) for the treatment of moderate-to-severe menopausal hot flashes. Methods: The primary endpoints of this randomized, placebo-controlled study of G-GR (600 mg AM/1,200 mg PM) were the mean daily frequency and severity of hot flashes at weeks 4 and 12. Secondary endpoints included Patients’ Global Impression of Change, Clinicians’ Global Impression of Change, and daily sleep interference at week 24. Results: Six hundred women with 7 or more moderate-to-severe hot flashes/day enrolled; 66.2% completed 24 weeks of treatment. At weeks 4 and 12, G-GRYtreated women experienced significantly greater reductions in mean hot flash frequency and severity than placebo-treated women (frequency: week 4, j1.7, P G 0.0001; week 12, j1.14, P = 0.0007; severity: week 4, j0.21, P G 0.0001; week 12, j0.19, P = 0.012). Similar reductions were maintained up to week 24. On the Patient Global Impression of Change, more women receiving G-GR than placebo were Bmuch[ or [very much[ improved (week 12: 58% vs 44%, P = 0.0008; week 24: 76% vs 55%, P G 0.0001). G-GR significantly reduced sleep interference compared with placebo at week 12 (P = 0.0056) and week 24 (P = 0.0084). Approximately 5% more women taking G-GR withdrew because of adverse events (G-GR/placebo, 16.7%/11.5%). The most common adverse events were dizziness (12.7%/3.4%), headache (9.3%/8.1%), and somnolence (6.0%/2.7%); incidences dropped to sustained low levels after a few weeks. Conclusions: G-GR is a modestly effective nonhormone therapy option for the treatment of moderate-to-severe hot flashes due to menopause and is well tolerated with titration. Key Words: Menopause Y Vasomotor symptoms Y Hot flashes Y Gastroretentive Y Gabapentin Y Clinical meaningfulness. pproximately 50% to 80% of perimenopausal and postmenopausal women experience vasomotor symptoms (VMS), which include hot flashes and night sweats.1,2 Together, hot flashes and their effects on sleep may leave women tired, irritable, and unable to concentrate, and may diminish their overall sense of well-being.1,3 Although hormone therapy (HT) is very effective for the treatment of VMS, nonhormone treatment alternatives are needed for women who are unwilling or unable to use HT.
A
Gabapentin has been shown to reduce hot flash frequency and severity in women who had had natural menopause and chemically or surgically induced menopause.4,5 In addition, randomized controlled clinical studies of immediate-release gabapentin for the treatment of hot flashes reported that gabapentin, compared with placebo, is efficacious in reducing the frequency and severity of hot flashes (9006 and 2,4007,8 mg/d). Immediate-release formulations of gabapentin have been shown to have relatively high rates of the adverse events (AEs)
Received May 31, 2013; revised and accepted August 1, 2013. From the 1Division of Midlife, Department of Obstetrics and Gynecology, University of Virginia Health Sciences Center, Charlottesville, VA; 2East Bay Physicians Medical Group, Sutter East Bay Medical Foundation, Berkeley, CA; 3Columbus Center for Women’s Health Research, Columbus, OH; and 4 Depomed Inc, Newark, CA. Clinical trial registration: NCT01080300 (www.clinicaltrials.gov). Funding/support: This study was funded by Depomed Inc. Financial disclosure/conflicts of interest: J.V.P. has served as a consultant (fees to the University of Virginia) for Pfizer, Noven Pharmaceuticals, Depomed, and Shionogi; received grants/research support (fees to the University of Virginia) from Depomed, Bionova, and Endoceutics; and received travel funds from Pfizer, Noven Pharmaceuticals, Shionogi, and Depomed. R.K. has served as a consultant and advisory board member for Merck, Pfizer, Amgen, Noven, Novo Nordisk, Depomed, and Shionogi; has received research support and travel support from Depomed; has served on the speakers
bureaus of Novo Nordisk, Noven, Novogyne, and Shionogi; is a member of the scientific advisory boards of the American Bone Health, The North American Menopause Society (through October 2012), and the American Orthopedics AssociationVOwn the Bone; and has provided legal consulting for Merck and Warner Chilcott. D.P. has served as a consultant to Noven, Meda, and Teva; has received research grants from Noven, Depomed, Teva, Pfizer, and Bayer (paid to Columbus Center for Women’s Health Research); and has served on the speakers bureaus of Noven, Teva, and Warner Chilcott. M.S. and R.S. are employees of Depomed Inc and are shareholders in the company. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal_s Web site (www.menopause.org). Address correspondence to: JoAnn V. Pinkerton, MD, Box 801104, University of Virginia Health System, Charlottesville, VA 22908. E-mail: [email protected] Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
567
PINKERTON ET AL
dizziness and somnolence (28% and 21%, respectively),9 and its thrice-daily dosing is less convenient for an active patient population. To address the high incidence of AEs, the pharmacokinetic issues of complex titration, and the relatively short half-life of the immediate-release formulation, a gastroretentive formulation of gabapentin (gastroretentive gabapentin [G-GR]) was developed. With its improved pharmacokinetics, G-GR can be administered once or twice daily.10 When taken with meals, G-GR tablets swell to three to four times their original size, causing them to be retained in the stomach for approximately 8 to 9 hours. While they are in the stomach, G-GR tablets provide a steady and continuous delivery of gabapentin to its optimal site of absorption in the upper small intestine.11,12 Consequently, the plasma concentration of gabapentin following G-GR administration twice per day with a meal dosed twice per day asymmetrically (600 mg AM/1,200 mg PM), results in exposure comparable to that of 1,800 mg/day immediate-release gabapentin administered as 600 mg three times a day.10,13 The slower release afforded by the gastroretentive formulation seems to enhance tolerability and permits a more rapid titration to an efficacious dose.14 A phase 2 dose-finding clinical study of G-GR in women with VMS15 demonstrated that 1,800 mg/day was efficacious and had fewer adverse effects than 2,400 mg/day. Earlier phase 3 placebo-controlled studies comparing 1,200 mg once daily and 1,800 mg dosed asymmetrically twice per day (600 mg AM/ 1,200 mg PM) found that the 1,800-mg dose was more efficacious than the 1,200-mg dose and was well tolerated.16,17 This study evaluated the efficacy and safety of G-GR in a large, randomized, multicenter, placebo-controlled phase 3 clinical study (Breeze 3) of 1,800 mg/day G-GR (600 mg with breakfast/ 1,200 mg with dinner) in women experiencing 7 or more moderate-to-severe hot flashes/day (or Q50 moderate-to-severe hot flashes/wk). METHODS Study design Women from 67 investigative sites in the United States participated in this prospective, randomized, double-blind, placebo-controlled, multicenter trial conducted from August 2010 to September 2011 (Appendix 1, Supplemental Digital Content 1, http://links.lww.com/MENO/A67). The study protocol and informed consent form were approved by the institutional review boards, and the study was conducted in accordance with Good Clinical Practice/International Conference on Harmonization guidelines. An informed consent form was obtained from each participant before screening. Eligible participants included healthy postmenopausal women who experienced 7 or more moderate-to-severe hot flashes/ day during a 14-day baseline, had no history of malignancy within the previous 2 years (except for basal cell carcinoma), had amenorrhea for 12 months or more or had amenorrhea for 6 to 12 months and a follicle-stimulating hormone level higher than 40 mIU/mL, or were 6 weeks or more post-bilateral oophorectomy with or without hysterectomy. Severity was defined as follows: mild, sensation of heat without sweating
568
Menopause, Vol. 21, No. 6, 2014
(scored 1); moderate, sensation of heat with sweating, able to continue activity (scored 2); severe, sensation of heat with sweating, causing cessation of activity (scored 3). Exclusion criteria included current treatment with HT, hormone agonists, antiestrogens, or aromatase inhibitors; creatinine gonadotropinreleasing clearance less than 60 mL/minute; history of gastric reduction; history of substance abuse within the past year; or any serious medical condition that, in the opinion of the investigators, would jeopardize the safety of the participant or affect the validity of study results. Women using HT to treat VMS had to discontinue such treatments before baseline as follows: vaginal HT, 1 week; transdermal HT, 4 weeks; oral HT, 8 weeks. No concomitant treatment of hot flashes was permitted. Women taking antidepressants could stay on their medications at the same dosage throughout the study as long as the dosage had not changed in the month before enrollment. At the end of the baseline period, women who continued to meet the study eligibility requirement of an average of 7 or more moderate-to-severe hot flashes/day were randomized to either G-GR or placebo with a parallel 1:1 randomization, stratified by site in blocks of four, using computerized randomization. Placebo was matched to look exactly like G-GR. The sponsor, investigators, participants, and study center personnel were blinded to treatment group assignment. No unblinding occurred in this study. The study included a 1-week titration to 1,800 mg of G-GR (dosed as 1,200 mg PM/600 mg AM) or matched placebo, followed by a 23-week stable treatment period and a 4-week poststudy observation period. G-GR titration (AM/PM) occurred on days 1 to 3 (0 mg/600 mg), days 4 to 6 (0 mg/1,200 mg), and day 7 (600 mg/1,200 mg). All study drugs were taken with meals. There was no tapering of study medication at the end of the study. The study diary was completed daily by participants from baseline to the end of week 12 and then during weeks 16, 20, 23, and 24. Women visited the clinic at the end of treatment weeks 4, 12, 20, and 24. Investigator sites made telephone contact with participants at weeks 8 and 16. Hot flash recording All qualified women were given a cellphone-based electronic diary (mPRO; t-Clinical, Oxford UK) and instructed to record the occurrence and severity of each hot flash in real time. The 2-week period was used to assess qualification for the study and to establish baseline hot flash measurements. Throughout the study, women could add to the diary unrecorded events that occurred during the night on the following morning. The device was also used to record the participant’s Sleep Interference Scale (SIS) score, reflecting the extent to which hot flashes interfered with sleep on the previous night. The SIS score was self-rated (1, did not interfere; 10, did not sleep because of hot flashes). Efficacy endpoints The primary efficacy endpoints were the mean changes in the frequency and severity of hot flashes with G-GR compared with placebo at weeks 4 and 12. The key secondary endpoints were these parameters at week 24. Other secondary endpoints * 2013 The North American Menopause Society
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
G-GR FOR MENOPAUSAL HOT FLASHES
included the Patients’ Global Impression of Change (PGIC) and Clinical Global Impression of Change (CGIC) instruments completed at visits on weeks 12 and 24; changes from baseline in sleep quality, as measured by the SIS; the Columbia Suicide Severity Rating Scale, which was administered to detect suicidal ideation at screening; and visits on weeks 4, 12, and 24 (or early termination) and 4 weeks after the end of treatment (week 28). Blood samples for routine hematology and chemistry tests were taken at screening and at study end. Analyses of primary and secondary efficacy parameters used a modified intent-to-treat population that included all women who were randomized, took one or more doses of study therapy, and had one or more postrandomization efficacy assessments. The week 24 endpoints analyzed only the completer population (defined as women who completed 24 wk of therapy and had one or more diary entries from days 158 to 168 postrandomization). A sample size of approximately 600 women (300 per group) was planned for this study; assuming a 30% noncompletion rate, 420 or more women (210 per group) would have available efficacy data for analysis through week 24. This sample size ensured that the observed endpoints at week 24 would have 90% power to detect a difference of 1.3 or more episodes in the median change in the mean daily number of moderate-to-severe hot flashes from baseline, compared with placebo, and an approximately 80% power to detect a difference of 0.2 or more in the median change in the mean daily severity score from baseline between treatment groups. Safety assessments Safety variables consisted of AEs, clinical laboratory test results, vital signs, new physical examination abnormalities, and premature terminations. The safety population included all randomized participants who received study medications. AEs were coded using Medical Dictionary for Regulatory Activities (MedDRA) version 9.0 to give a preferred term and a primary body system for each event. A treatment-emergent AE was defined as any AE occurring after randomization through 3 days after the end of treatment. Statistical analysis The van Elteren (ie, stratified Wilcoxon) test procedure was originally proposed for the analysis of primary and key secondary efficacy measurements, with study site as stratification factor. However, review of available data showed that 19 of 67 (28.4%) participating sites enrolled less than 5 women, with 38 of 67 (56.7%) sites enrolling less than 10 women. The sensitivity of the van Elteren test becomes limited when the sample sizes in so many strata (ie, centers in this analysis) are small.18 Therefore, a parametric analysis of covariance, which was prespecified, but is not sensitive to small center sizes, was considered the most appropriate statistical test for the analysis of these data. Only results from analyses of covariance (or its analysis of variance analog for the continuous secondary efficacy endpoints) are presented in this article. All treatment comparisons were made using a two-sided 0.05 level of significance.
For the analysis of ordinal categorical data (CGIC/PGIC), a Cochran-Mantel-Haenszel test of general association with modified ridit scores was used for the overall comparison among all treatment groups. For the analysis of dichotomous outcome data, a two-sided Fisher’s exact test was used for the overall comparison among all treatment groups. Differences between two proportions (G-GR treatment j placebo) and the associated 95% CIs were obtained using a normal approximation to the binomial. For week 12, data were obtained from the modified intent-to-treat population, and last-observationcarried-forward imputation was used to account for missing data. For week 24, data were obtained from the completer population as observed. RESULTS Six hundred women were enrolled and randomized in this study: 298 to the placebo arm and 302 to the G-GR arm. Five women did not receive study therapy, and two women had no postrandomization efficacy measures recorded, resulting in a modified intent-to-treat population of 593 women (placebo, 294; G-GR, 299) (Fig. 1). One hundred ninety-eight (33.8%) women discontinued from the study: 104 (35.9%) in the placebo arm and 94 (31.8%) in the G-GR arm. Baseline demographics and disease characteristics were similar between groups (Table 1). Participants were primarily white and had a mean age of 54 years, 38% had a body mass index of 30 kg/m2 or higher, and the majority (75%) experienced natural menopause. The baseline mean (SD) daily frequencies of hot flashes were similar between arms: 12.0 (5.5)
FIG. 1. Consort diagram. The intent-to-treat population included women who were randomized, took at least one dose of study therapy, and had at least one postrandomization efficacy assessment. The safety population included all randomized participants who received study medication. The completer population included all women who completed 24 weeks of therapy and had at least one diary entry from days 158 to 168 postrandomization. HF, hot flash; G-GR, gastroretentive gabapentin. Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
569
PINKERTON ET AL TABLE 1. Baseline participant demographics and characteristics: modified intent-to-treat population Variable
Placebo (n = 294)
G-GR (n = 299)
54.0 (6.0) 34-69 282 (95.9) 12 (4.1)
54.0 (6.1) 38-70 280 (93.6) 19 (6.4)
197 (67.0) 82 (27.9) 9 (3.1) 6 (2.0)
215 (71.9) 74 (24.7) 9 (3.0) 1 (0.3)
182 (61.9) 112 (38.3) 17.7-59.3
184 (61.5) 115 (38.5) 16.3-52.4
120.2 (90.2) 107
114.3 (36.7) 104
220 (74.8) 74 (25.2)
223 (74.6) 76 (25.4)
2.2 (1.2)
2.0 (1.3)
12.0 (5.5) 2.54 (0.28) 7.4 (2.2)
11.8 (4.7) 2.55 (0.29) 7.3 (2.2)
Age, y Mean (SD) Range G65, n (%) Q65, n (%) Race, n (%) White Black Hispanic Other Body mass index, n (%), kg/m2 G30 Q30 Range Creatinine clearance, mL/min Mean (SD) Median Menopause history, n (%) Natural Surgical Children delivered Mean (SD) Hot flashes Frequency, mean no./day (SD) Severity, mean (SD) SIS score, mean (SD) G-GR, gastroretentive gabapentin.
for placebo and 11.8 (4.7) for G-GR. The mean (SD) daily hot flash severity scores were also similar: 2.54 (0.28) for placebo and 2.55 (0.29) for G-GR. These scores indicate that, at study entry, half of the hot flashes experienced were moderate and half were severe. The mean baseline sleep interference scores were also similar between arms: 7.4 for placebo and 7.3 for G-GR (out of 10), indicating significant sleep disturbance at baseline (Table 1). Hot flash frequency and severity Women in the G-GR arm had significantly greater reductions in hot flash frequency and severity from baseline compared with women in the placebo arm (Fig. 2). At week 4, the baseline-adjusted mean (SEM) reduction in hot flash frequency was j5.01 (0.27) for the placebo group and j6.72 (0.27) for the G-GR group (difference, j1.70; 95% CI, j2.31 to j1.09; P G 0.0001). At week 12, the mean (SEM) reductions were j6.50 (0.29) for the placebo arm and j7.64 (0.29) for the G-GR arm (difference, j1.14; 95% CI, 1.8 to j0.48; P = 0.0007; Fig. 2). Reductions from baseline in mean hot flash frequency were apparent after 1 week and continued throughout 12 weeks (Fig. 3). At week 12, more women in the G-GR arm had achieved a 50% reduction in hot flash frequency (73% vs 60%). At week 4, the corresponding baseline-adjusted mean (SEM) reductions in hot flash severity were j0.22 (0.04) for the placebo arm and j0.42 (0.04) for the G-GR arm (difference, j0.2 [0.052]; 95% CI, j0.31 to j0.1; P G 0.0001); at week 12, the reductions in severity were j0.46 (0.06) in the placebo arm and j0.65 (0.06) in the G-GR arm (difference, j0.19 [0.07]; 95% CI, j0.33 to j0.04; P = 0.012; Fig. 2).
570
Menopause, Vol. 21, No. 6, 2014
Statistically significant reductions in both hot flash frequency and hot flash severity were maintained for women who completed the 24-week study (Fig. 2). The mean (SEM) reduction in hot flash frequency at week 24 was j7.91 (0.36) for the placebo arm and j8.99 (0.37) for the G-GR arm (difference, j1.08 [0.45]; 95% CI, j1.98 to j0.19; P = 0.0174; Fig. 2). The reductions in hot flash severity at week 24 were j0.86 (0.09) and j0.64 (0.09; difference, j0.22 [0.11]; 95% CI, j0.44 to j0.0; P = 0.0457; Fig. 2). Secondary endpoints: PGIC, CGIC, and SIS For the secondary endpoints addressing clinical meaningfulness, as rated by the PGIC and CGIC scales, significantly more women in the G-GR arm were Bmuch[ or Bvery much[ improved. Based on the PGIC scale, the number of women who reported feeling Bmuch[ or Bvery much[ improved was 44% for the placebo arm versus 58% for the G-GR arm (P = 0.0008) at week 12, and 55% for the placebo arm versus 76% for the G-GR arm (P G 0.0001) at week 24 (Table 2). Using the CGIC, physicians reported that the percentage of women reporting Bmuch[ or Bvery much[ improvement was 42% for the placebo arm versus 58% for the G-GR arm (P G 0.0001) at week 12, and was 60% for the placebo arm (P = 0.0002) versus 78% for the G-GR arm at week 24 (Table 2). Sleep, as measured by the mean SIS, was markedly improved for women in the G-GR arm compared with those in the placebo arm at week 12 (j2.2 [placebo] vs j3.1 [G-GR]; P G 0.0001) and at week 24 (j2.2 [placebo] vs j3.2 [G-GR]; P G 0.0001; Table 2).
FIG. 2. Baseline-adjusted mean (SEM) reductions in (A) hot flash frequency and (B) hot flash severity at weeks 4, 12, and 24. G-GR, gastroretentive gabapentin. * 2013 The North American Menopause Society
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
G-GR FOR MENOPAUSAL HOT FLASHES
the G-GR arm (anemia, infective arthritis, periorbital cellulitis, and chronic lymphocytic leukemia). The only death in the study was that of a participant in the placebo arm who had hemorrhagic stroke. No serious AE was deemed to be related to the study drug. There was one case of positive suicidal ideation in each arm during the study (placebo, at week 24; G-GR, at week 28); both were judged by investigators to be unrelated to therapy. DISCUSSION In this phase 3 study of highly symptomatic postmenopausal women with more than 7 moderate-to-severe hot flashes/day, G-GR produced modest, statistically significant reductions in hot flash frequency and severity compared with placebo. The absolute decrease in hot flash frequency relative to placebo did not meet the Food and Drug Administration guideline of 2 hot flashes/day.19 However, this guideline was designed for standard-dose HT and does not include reductions in hot flash severity or improvements in participantreported outcomes. In this study, the baseline hot flash severity for both arms was 2.5, indicating that every other hot flash was severe and thus led to cessation of personal activity. After 12 weeks on placebo, women had a mean severity score of 2.04, indicating that the participants still had some severe hot flashes. In contrast, after 12 weeks of G-GR treatment, the mean severity score decreased to less than 2, indicating that the women had no moderate or severe hot flashes at all on some days. Thus, the mean reduction of 1.1 hot flashes/ day observed in the G-GR arm corresponded to elimination of a severe hot flash (one that would force a woman to stop whatever she was doing).
FIG. 3. Change from mean baseline hot flash frequency score during the first 12 weeks of study. *Statistically significant difference from placebo at the 0.05 level. G-GR, gastroretentive gabapentin.
Safety The safety population included 595 women: 300 in the G-GR group and 295 in the placebo group. G-GR was well tolerated after titration, with only 5.2% more women in the G-GR arm discontinuing because of AEs compared with those in the placebo arm (16.7% vs 11.5%; Table 3). Discontinuations because of AEs during titration were low (2.4% in the placebo arm and 4.7% in the G-GR arm). The most common AEs leading to discontinuation throughout the study were dizziness (placebo, 0.7%; G-GR, 3.3%) and somnolence (placebo, 1.0%; G-GR, 2.0%). The most common AEs (95% incidence) reported during the study were dizziness (placebo, 3.4%; G-GR, 12.7%), headache (placebo, 8.1%; G-GR, 9.3%), somnolence (placebo, 2.7%; G-GR, 6.0%), and upper respiratory tract infection (placebo, 3.7%; G-GR, 6.0%; Table 3). The incidences of dizziness and somnolence, the AEs most commonly associated with gabapentinoids, decreased rapidly after the titration week (Fig. 4). For dizziness, prevalence decreased from 10.7% at the end of week 1 to 6.2% at the end of week 2, and to a steady rate of approximately 3% by the end of week 6. For somnolence, prevalence was 4.3% at the end of week 1 and stabilized at approximately 3% by the end of week 4. Placebo rates for dizziness and somnolence were less than 2% at the end of each study week. There were seven (2.4%) serious AEs in the placebo arm (palpitations, hemorrhagic stroke, acute sinusitis, musculoskeletal chest pain, asthma, primary atypical pneumonia, and acute myocardial infarction) and four (1.3%) serious AEs in
Clinical meaningfulness of reductions in hot flash frequency and severity Perhaps more important than the modest improvements in hot flash frequency and severity, the clinical meaningfulness of G-GR treatment was evidenced by the finding that significantly more women in the G-GR arm reported feeling Bmuch[ or Bvery much[ improved on the PGIC (58% for G-GR vs 44% for placebo at week 12). This suggests that reduction in the total burden of hot flashes, which factors in reductions in both frequency and severity, contributes to improvement in women’s overall sense of well-being. Among women who completed the study, this greater sense of improvement was
TABLE 2. Secondary efficacy endpoints on weeks 12 and 24 Week 12 Instrument a
PGIC , % CGICa, %
Week 24
Placebo (n = 295)
G-GR (n = 300)
P
Placebo (n = 177)
G-GR (n = 185)
P
44 42
58 58
0.0008 G0.0001
55 60
76 78
G0.0001 0.0002
Placebo (n = 294)
G-GR (n = 299)
P
Placebo (n = 191)
G-GR (n = 206)
P
b
j2.2 j3.1 G0.0001 j2.2 j3.2 G0.0001 SIS G-GR, gastroretentive gabapentin; PGIC, Patient Global Impression of Change; CGIC, Clinician Global Impression of Change; SIS, Sleep Interference Scale. a Patients Bmuch[ or Bvery much[ improved. b Change in score from baseline. Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
571
PINKERTON ET AL TABLE 3. Treatment-emergent adverse events Placebo (n = 295)
Category
G-GR (n = 300)
Any TEAEa 165 (52.9) 197 (65.7) Q1 TEAE leading to discontinuation 34 (11.5) 50 (16.7) Q1 TEAE during titration week 42 (14.2) 81 (27.0) Q1 TEAE leading to discontinuation 7 (2.4) 14 (4.7) during titration week Q1 serious TEAE 7 (2.4) 4 (1.3) Most common TEAEsb Dizziness 10 (3.4) 38 (12.7) Headache 24 (8.1) 28 (9.3) Somnolence 8 (2.7) 18 (6.0) Upper respiratory tract infection 11(3.7) 18 (6.0) G-GR, gastroretentive gabapentin; TEAE, treatment-emergent adverse event. a Among TEAEs, 93.4% were mild or moderate in severity. b Incidence of Q5% or Q2% higher incidence in G-GR versus placebo.
maintained: 76% of women taking G-GR versus 55% of women taking placebo reported feeling Bmuch[ or Bvery much[ improved at week 24. Another important component of clinical meaningfulness is reduction in sleep disturbance, a problem that leads many postmenopausal women to consult their physicians for the treatment of hot flashes. Sleep disturbance due to hot flashes is associated with chronic insomnia,20 leading to fatigue, mood disorders, and reduced quality of life.1,3,21 Women in this study were experiencing significant sleep disturbance at baseline, with a mean SIS score for the total population of 7.3 of 10. Women in the G-GR arm, however, experienced a significant improvement in sleep compared with those in the placebo arm, as evidenced by reduced mean SIS scores at weeks 12 and 24. The fact that the evening dose of G-GR provides peak steady-state levels of gabapentin 6 to 12 hours after dosing10,22 could explain, in part, the improvements in sleep observed with G-GR.10,22 Placebo response rate The placebo response rate observed in this study (a 54% reduction from baseline in hot flash frequency at week 12) was higher by 10% or more compared with other published studies of nonhormone VMS treatments. When treatment effects are modest, higher placebo rates make it more difficult to demonstrate that the active treatment is significantly better. Although not a direct comparison, a cross-study comparison of findings from this G-GR study with those for the low-dose HT Angeliq (0.25 mg of drospirenone and 0.5 mg of estradiol) demonstrates this. In comparable female populations, the mean reductions in hot flash frequency observed at week
12 for G-GR and Angeliq were similar (j7.71 hot flashes/d for Angeliq vs j7.64 hot flashes/d for G-GR).23 The reductions in hot flash frequency observed in the placebo arms, however, were considerably different (j4.54 hot flashes/d for Angeliq vs j6.50 hot flashes/d for G-GR). Thus, based solely on the difference in placebo response rate (~2 hot flashes/d), it appears that Angeliq reduces hot flash frequency by 3.17 hot flashes/day over placebo, whereas G-GR reduces hot flash frequency by only 1.14 hot flashes/day. Factors contributing to the placebo response should be similar for HT and non-HT. Because regression toward the mean tends to occur in women experiencing severe cases of any condition,24 one would expect that the number of hot flashes in placebo participants would decrease (ie, move toward the mean) across time on their own, thus raising the apparent rate of improvement in the placebo arm. Other potential factors include the clinical trial environment, coupled with biofeedback from the requirement to capture each hot flash daily using an electronic device, and increased interactions with healthcare professionals and fellow hot flash sufferers.25 One must also consider that women in the study who were seeking alternatives to HT and may have tried other alternative approaches, might have been especially hopeful that the study medication would provide relief. Safety and tolerability G-GR was well tolerated during the 24 weeks of this study. Given the relatively small number of discontinuations attributed to AEs during the first week of treatment (2.4% in the placebo arm vs 4.7% in G-GR arm), it would seem that AEs were manageable for most women and that the more rapid 1-week titration used in this study (as opposed to the 2-wk titration used in studies of G-GR in postherpetic neuralgia)14 was suitable for this population of highly symptomatic women, who are younger and generally healthier. The low propensity for drug interactions is also an important component of the safety profile of gabapentin. Because gabapentin is renally excreted and not metabolized, G-GR has no interaction with cytochrome P45026 and would therefore not be expected to reduce the effectiveness of tamoxifen in women with breast cancer. Aspects of study design Strengths of this study include its large, diverse study population, which is representative of the overall US population
FIG. 4. Prevalence of the common gabapentinoid-related adverse events, dizziness and somnolence, at the end of weeks 1 through 12.
572
Menopause, Vol. 21, No. 6, 2014
* 2013 The North American Menopause Society
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
G-GR FOR MENOPAUSAL HOT FLASHES
(one third was nonwhite and nearly 40% had a body mass index Q30 kg/m2). A limitation of the study was the lack of a direct comparison with immediate-release gabapentin. However, immediate-release gabapentin is not approved for the treatment of VMS, and this study was designed to have sufficient power to demonstrate the efficacy and safety of G-GR compared with placebo, not to directly compare G-GR with immediate-release gabapentin. Given the high placebo response rate in this study (54%), future studies of G-GR or other nonhormone agents should be designed to minimize the placebo rate. One approach that yielded a placebo rate of 45% to 48% was the use of a prolonged single-blind placebo run-in that eliminated women who responded well to placebo before randomization.27 In that study, which compared low-dose mesylate salt of paroxetine with placebo, hot flash frequency was reduced by 60% in the active arm versus 48% in the placebo arm at week 4 (P = 0.0072) and by 61% in the active arm versus 45% in the placebo arm at week 12 (P = 0.0001).27 Another change in study design for nonhormone agents, which tend to have smaller absolute changes in hot flash frequency and severity, would be to use a participant-reported outcome (eg, the PGIC) as a primary endpoint. Because participants often report feeling significantly improved after treatment with non-HTs despite more modest reductions in hot flash frequency and severity, such an endpoint may better reflect the overall clinical effectiveness and clinical meaningfulness of the agents for the treatment of VMS. CONCLUSIONS Even with the relatively high placebo response rate in this study, G-GR has been found to be modestly effective at reducing the frequency and severity of hot flashes for up to 24 weeks, but it does not meet the Food and Drug Administration criterion of a reduction of 2 or more hot flashes/day. Despite this, as reflected by improvements in women’s overall sense of well-being and in their ability to sleep, the reductions in hot flash frequency and severity afforded by G-GR seem to be clinically meaningful to women. G-GR is well tolerated, especially after the titration period. Acknowledgments: We gratefully thank Stephanie Kareht, PhD, of Depomed for medical writing support and Howard Hait, PhD, for assistance with biostatistics (as a consultant to Depomed).
REFERENCES 1. Bachmann GA. Vasomotor flushes in menopausal women. Am J Obstet Gynecol 1999;180:S312-S316. 2. Morrissey DR, Kirchner JT. Management of the climacteric. Options abound to relieve women’s midlife symptoms. Postgrad Med 2000;108: 85-88, 93-94, 97-100. 3. Oldenhave A, Jaszmann LJ, Haspels AA, Everaerd WT. Impact of climacteric on well-being. A survey based on 5213 women 39 to 60 years old. Am J Obstet Gynecol 1993;168:772-780. 4. Guttuso T Jr, Kurlan R, McDermott MP, Kieburtz K. Gabapentin’s effects on hot flashes in postmenopausal women: a randomized controlled trial. Obstet Gynecol 2003;101:337-345.
5. Pandya KJ, Morrow GR, Roscoe JA, et al. Gabapentin for hot flashes in 420 women with breast cancer: a randomised double-blind placebocontrolled trial. Lancet 2005;366:818-824. 6. Butt DA, Lock M, Lewis JE, Ross S, Moineddin R. Gabapentin for the treatment of menopausal hot flashes: a randomized controlled trial. Menopause 2008;15:1-9. 7. Guttuso T Jr. Hot flashes refractory to HRT and SSRI therapy but responsive to gabapentin therapy. J Pain Symptom Manage 2004;27: 274-276. 8. Reddy SY, Warner H, Guttuso T Jr, et al. Gabapentin, estrogen, and placebo for treating hot flushes: a randomized controlled trial. Obstet Gynecol 2006;108:41-48. 9. NeurontinA (gabapentin) Capsules, NeurontinA (gabapentin) Tablets, NeurontinA (gabapentin) Oral Solution [US prescribing information]. New York, NY: Pfizer Inc; 2012. Available at: http://labeling.pfizer.com/ ShowLabeling.aspx?id=630. Accessed July 18, 2012. 10. Argoff CE, Chen C, Cowles VE. Clinical development of a once-daily gastroretentive formulation of gabapentin for treatment of postherpetic neuralgia: an overview. Expert Opin Drug Deliv 2012;9:1147-1160. 11. Hou SY, Cowles VE, Berner B. Gastric retentive dosage forms: a review. Crit Rev Ther Drug Carrier Syst 2003;20:459-497. 12. Berner B, Cowles VE. Case studies in swelling polymeric gastric retentive tablets. Expert Opin Drug Deliv 2006;3:541-548. 13. Gordi T, Hou E, Kasichayanula S, Berner B. Pharmacokinetics of gabapentin after a single day and at steady state following the administration of gastric-retentive, extended-release, and immediate-release tablets: a randomized, open-label, multiple-dose, three-way crossover, exploratory study in healthy subjects. Clin Ther 2008;30:909-916. 14. Sang CN, Sathyanarayana R, Sweeney M. Gastroretentive gabapentin (G-GR) formulation reduces intensity of pain associated with postherpetic neuralgia (PHN). Clin J Pain 2013;29:281-288. 15. Sweeney M, Sathyanarayana R, Gordi T, Cowles VE, Heritier M. Positive efficacy data from a phase 2 trial of gabapentin extended-release in the treatment of menopausal hot flashes [Abstract]. Menopause 2008;15: 1225. Abstract P-61 16. Kagan R, Sathyanarayana R, Sweeney M, Pratt J, Zheng Y. The effect of outliers on hot flash frequency in the Breeze 1 Study of gabapentin-ER for the non-hormonal treatment of vasomotor symptoms [Abstract]. Menopause 2010;17:1219. Abstract S-14. 17. Baron M, Cowles VE, Utian W, Koltun W, Sathyanarayana R, Sweeney M. Efficacy of gabapentin extended-release (G-ER) on vasomotor symptoms in post-menopausal women in Breeze 2 Study [Abstract]. Menopause. 2011;18:1369. Abstract P-76 18. Stokes ME, Davis CS, Koch GG. Categorical Data Analysis Using the SAS System, 2nd ed. Cary, NC: SAS Institute Inc, 2000. 19. Center for Drug Evaluation and Research. Guidance for Industry: Estrogen and Estrogen/Progestin Drug Products to Treat Vasomotor Symptoms and Vulvar and Vaginal Atrophy SymptomsVRecommendations for Clinical Evaluation. Rockville, MD: Food and Drug Administration, 2003. http:// www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm071643.pdf. Accessed May 31, 2013. 20. Ohayon MM. Severe hot flashes are associated with chronic insomnia. Arch Intern Med 2006;166:1262-1268. 21. Regestein QR. Hot flashes, sleep, and mood. Menopause 2010;17:16-18. 22. Cowles VE, Gordi T, Hou SY. Steady-state pharmacokinetics of gabapentin after administration of a novel gastroretentive extendedrelease formulation in postmenopausal women with vasomotor symptoms. Clin Drug Investig 2012;32:593-601. 23. Angeliq [US prescribing information]. Wayne, NJ: Bayer HealthCare Pharmaceuticals; 2012. Available at: http://berlex.bayerhealthcare.com/ html/products/pi/fhc/Angeliq_PI.pdf. Accessed May 31, 2013. 24. Bland JM, Altman DG. Regression towards the mean. BMJ Open 1994;308:1499. 25. Mahon SM, Kaplan M. Placebo effect in hot flush research. Lancet Oncol 2012;13:e188. 26. Patsalos PN, Perucca E. Clinically important drug interactions in epilepsy: interactions between antiepileptic drugs and other drugs. Lancet Neurol 2003;2:473-481. 27. Simon JA, Sanacora G, Bhaskar S, Lippman J. Safety and efficacy of low-dose mesylate salt of paroxetine (LDMP) for the treatment of vasomotor symptoms (VMS) associated with menopause: a 24-week, randomized, placebo controlled phase 3 study [Abstract]. Menopause. 2012;19:1371. Abstract S-2.
Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
573
Phase 3 randomized controlled study of gastroretentive gabapentin for the treatment of moderate-to-severe hot flashes in menopause JoAnn V. Pinkerton, MD,1 Risa Kagan, MD,2 David Portman, MD,3 Rekha Sathyanarayana, BS,4 Michael Sweeney, MD,4 for the Breeze 3 Investigators Abstract Objective: The goal of this study was to evaluate the efficacy and safety of gastroretentive gabapentin (G-GR) for the treatment of moderate-to-severe menopausal hot flashes. Methods: The primary endpoints of this randomized, placebo-controlled study of G-GR (600 mg AM/1,200 mg PM) were the mean daily frequency and severity of hot flashes at weeks 4 and 12. Secondary endpoints included Patients’ Global Impression of Change, Clinicians’ Global Impression of Change, and daily sleep interference at week 24. Results: Six hundred women with 7 or more moderate-to-severe hot flashes/day enrolled; 66.2% completed 24 weeks of treatment. At weeks 4 and 12, G-GRYtreated women experienced significantly greater reductions in mean hot flash frequency and severity than placebo-treated women (frequency: week 4, j1.7, P G 0.0001; week 12, j1.14, P = 0.0007; severity: week 4, j0.21, P G 0.0001; week 12, j0.19, P = 0.012). Similar reductions were maintained up to week 24. On the Patient Global Impression of Change, more women receiving G-GR than placebo were Bmuch[ or [very much[ improved (week 12: 58% vs 44%, P = 0.0008; week 24: 76% vs 55%, P G 0.0001). G-GR significantly reduced sleep interference compared with placebo at week 12 (P = 0.0056) and week 24 (P = 0.0084). Approximately 5% more women taking G-GR withdrew because of adverse events (G-GR/placebo, 16.7%/11.5%). The most common adverse events were dizziness (12.7%/3.4%), headache (9.3%/8.1%), and somnolence (6.0%/2.7%); incidences dropped to sustained low levels after a few weeks. Conclusions: G-GR is a modestly effective nonhormone therapy option for the treatment of moderate-to-severe hot flashes due to menopause and is well tolerated with titration. Key Words: Menopause Y Vasomotor symptoms Y Hot flashes Y Gastroretentive Y Gabapentin Y Clinical meaningfulness. pproximately 50% to 80% of perimenopausal and postmenopausal women experience vasomotor symptoms (VMS), which include hot flashes and night sweats.1,2 Together, hot flashes and their effects on sleep may leave women tired, irritable, and unable to concentrate, and may diminish their overall sense of well-being.1,3 Although hormone therapy (HT) is very effective for the treatment of VMS, nonhormone treatment alternatives are needed for women who are unwilling or unable to use HT.
A
Gabapentin has been shown to reduce hot flash frequency and severity in women who had had natural menopause and chemically or surgically induced menopause.4,5 In addition, randomized controlled clinical studies of immediate-release gabapentin for the treatment of hot flashes reported that gabapentin, compared with placebo, is efficacious in reducing the frequency and severity of hot flashes (9006 and 2,4007,8 mg/d). Immediate-release formulations of gabapentin have been shown to have relatively high rates of the adverse events (AEs)
Received May 31, 2013; revised and accepted August 1, 2013. From the 1Division of Midlife, Department of Obstetrics and Gynecology, University of Virginia Health Sciences Center, Charlottesville, VA; 2East Bay Physicians Medical Group, Sutter East Bay Medical Foundation, Berkeley, CA; 3Columbus Center for Women’s Health Research, Columbus, OH; and 4 Depomed Inc, Newark, CA. Clinical trial registration: NCT01080300 (www.clinicaltrials.gov). Funding/support: This study was funded by Depomed Inc. Financial disclosure/conflicts of interest: J.V.P. has served as a consultant (fees to the University of Virginia) for Pfizer, Noven Pharmaceuticals, Depomed, and Shionogi; received grants/research support (fees to the University of Virginia) from Depomed, Bionova, and Endoceutics; and received travel funds from Pfizer, Noven Pharmaceuticals, Shionogi, and Depomed. R.K. has served as a consultant and advisory board member for Merck, Pfizer, Amgen, Noven, Novo Nordisk, Depomed, and Shionogi; has received research support and travel support from Depomed; has served on the speakers
bureaus of Novo Nordisk, Noven, Novogyne, and Shionogi; is a member of the scientific advisory boards of the American Bone Health, The North American Menopause Society (through October 2012), and the American Orthopedics AssociationVOwn the Bone; and has provided legal consulting for Merck and Warner Chilcott. D.P. has served as a consultant to Noven, Meda, and Teva; has received research grants from Noven, Depomed, Teva, Pfizer, and Bayer (paid to Columbus Center for Women’s Health Research); and has served on the speakers bureaus of Noven, Teva, and Warner Chilcott. M.S. and R.S. are employees of Depomed Inc and are shareholders in the company. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal_s Web site (www.menopause.org). Address correspondence to: JoAnn V. Pinkerton, MD, Box 801104, University of Virginia Health System, Charlottesville, VA 22908. E-mail: [email protected] Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
567
PINKERTON ET AL
dizziness and somnolence (28% and 21%, respectively),9 and its thrice-daily dosing is less convenient for an active patient population. To address the high incidence of AEs, the pharmacokinetic issues of complex titration, and the relatively short half-life of the immediate-release formulation, a gastroretentive formulation of gabapentin (gastroretentive gabapentin [G-GR]) was developed. With its improved pharmacokinetics, G-GR can be administered once or twice daily.10 When taken with meals, G-GR tablets swell to three to four times their original size, causing them to be retained in the stomach for approximately 8 to 9 hours. While they are in the stomach, G-GR tablets provide a steady and continuous delivery of gabapentin to its optimal site of absorption in the upper small intestine.11,12 Consequently, the plasma concentration of gabapentin following G-GR administration twice per day with a meal dosed twice per day asymmetrically (600 mg AM/1,200 mg PM), results in exposure comparable to that of 1,800 mg/day immediate-release gabapentin administered as 600 mg three times a day.10,13 The slower release afforded by the gastroretentive formulation seems to enhance tolerability and permits a more rapid titration to an efficacious dose.14 A phase 2 dose-finding clinical study of G-GR in women with VMS15 demonstrated that 1,800 mg/day was efficacious and had fewer adverse effects than 2,400 mg/day. Earlier phase 3 placebo-controlled studies comparing 1,200 mg once daily and 1,800 mg dosed asymmetrically twice per day (600 mg AM/ 1,200 mg PM) found that the 1,800-mg dose was more efficacious than the 1,200-mg dose and was well tolerated.16,17 This study evaluated the efficacy and safety of G-GR in a large, randomized, multicenter, placebo-controlled phase 3 clinical study (Breeze 3) of 1,800 mg/day G-GR (600 mg with breakfast/ 1,200 mg with dinner) in women experiencing 7 or more moderate-to-severe hot flashes/day (or Q50 moderate-to-severe hot flashes/wk). METHODS Study design Women from 67 investigative sites in the United States participated in this prospective, randomized, double-blind, placebo-controlled, multicenter trial conducted from August 2010 to September 2011 (Appendix 1, Supplemental Digital Content 1, http://links.lww.com/MENO/A67). The study protocol and informed consent form were approved by the institutional review boards, and the study was conducted in accordance with Good Clinical Practice/International Conference on Harmonization guidelines. An informed consent form was obtained from each participant before screening. Eligible participants included healthy postmenopausal women who experienced 7 or more moderate-to-severe hot flashes/ day during a 14-day baseline, had no history of malignancy within the previous 2 years (except for basal cell carcinoma), had amenorrhea for 12 months or more or had amenorrhea for 6 to 12 months and a follicle-stimulating hormone level higher than 40 mIU/mL, or were 6 weeks or more post-bilateral oophorectomy with or without hysterectomy. Severity was defined as follows: mild, sensation of heat without sweating
568
Menopause, Vol. 21, No. 6, 2014
(scored 1); moderate, sensation of heat with sweating, able to continue activity (scored 2); severe, sensation of heat with sweating, causing cessation of activity (scored 3). Exclusion criteria included current treatment with HT, hormone agonists, antiestrogens, or aromatase inhibitors; creatinine gonadotropinreleasing clearance less than 60 mL/minute; history of gastric reduction; history of substance abuse within the past year; or any serious medical condition that, in the opinion of the investigators, would jeopardize the safety of the participant or affect the validity of study results. Women using HT to treat VMS had to discontinue such treatments before baseline as follows: vaginal HT, 1 week; transdermal HT, 4 weeks; oral HT, 8 weeks. No concomitant treatment of hot flashes was permitted. Women taking antidepressants could stay on their medications at the same dosage throughout the study as long as the dosage had not changed in the month before enrollment. At the end of the baseline period, women who continued to meet the study eligibility requirement of an average of 7 or more moderate-to-severe hot flashes/day were randomized to either G-GR or placebo with a parallel 1:1 randomization, stratified by site in blocks of four, using computerized randomization. Placebo was matched to look exactly like G-GR. The sponsor, investigators, participants, and study center personnel were blinded to treatment group assignment. No unblinding occurred in this study. The study included a 1-week titration to 1,800 mg of G-GR (dosed as 1,200 mg PM/600 mg AM) or matched placebo, followed by a 23-week stable treatment period and a 4-week poststudy observation period. G-GR titration (AM/PM) occurred on days 1 to 3 (0 mg/600 mg), days 4 to 6 (0 mg/1,200 mg), and day 7 (600 mg/1,200 mg). All study drugs were taken with meals. There was no tapering of study medication at the end of the study. The study diary was completed daily by participants from baseline to the end of week 12 and then during weeks 16, 20, 23, and 24. Women visited the clinic at the end of treatment weeks 4, 12, 20, and 24. Investigator sites made telephone contact with participants at weeks 8 and 16. Hot flash recording All qualified women were given a cellphone-based electronic diary (mPRO; t-Clinical, Oxford UK) and instructed to record the occurrence and severity of each hot flash in real time. The 2-week period was used to assess qualification for the study and to establish baseline hot flash measurements. Throughout the study, women could add to the diary unrecorded events that occurred during the night on the following morning. The device was also used to record the participant’s Sleep Interference Scale (SIS) score, reflecting the extent to which hot flashes interfered with sleep on the previous night. The SIS score was self-rated (1, did not interfere; 10, did not sleep because of hot flashes). Efficacy endpoints The primary efficacy endpoints were the mean changes in the frequency and severity of hot flashes with G-GR compared with placebo at weeks 4 and 12. The key secondary endpoints were these parameters at week 24. Other secondary endpoints * 2013 The North American Menopause Society
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
G-GR FOR MENOPAUSAL HOT FLASHES
included the Patients’ Global Impression of Change (PGIC) and Clinical Global Impression of Change (CGIC) instruments completed at visits on weeks 12 and 24; changes from baseline in sleep quality, as measured by the SIS; the Columbia Suicide Severity Rating Scale, which was administered to detect suicidal ideation at screening; and visits on weeks 4, 12, and 24 (or early termination) and 4 weeks after the end of treatment (week 28). Blood samples for routine hematology and chemistry tests were taken at screening and at study end. Analyses of primary and secondary efficacy parameters used a modified intent-to-treat population that included all women who were randomized, took one or more doses of study therapy, and had one or more postrandomization efficacy assessments. The week 24 endpoints analyzed only the completer population (defined as women who completed 24 wk of therapy and had one or more diary entries from days 158 to 168 postrandomization). A sample size of approximately 600 women (300 per group) was planned for this study; assuming a 30% noncompletion rate, 420 or more women (210 per group) would have available efficacy data for analysis through week 24. This sample size ensured that the observed endpoints at week 24 would have 90% power to detect a difference of 1.3 or more episodes in the median change in the mean daily number of moderate-to-severe hot flashes from baseline, compared with placebo, and an approximately 80% power to detect a difference of 0.2 or more in the median change in the mean daily severity score from baseline between treatment groups. Safety assessments Safety variables consisted of AEs, clinical laboratory test results, vital signs, new physical examination abnormalities, and premature terminations. The safety population included all randomized participants who received study medications. AEs were coded using Medical Dictionary for Regulatory Activities (MedDRA) version 9.0 to give a preferred term and a primary body system for each event. A treatment-emergent AE was defined as any AE occurring after randomization through 3 days after the end of treatment. Statistical analysis The van Elteren (ie, stratified Wilcoxon) test procedure was originally proposed for the analysis of primary and key secondary efficacy measurements, with study site as stratification factor. However, review of available data showed that 19 of 67 (28.4%) participating sites enrolled less than 5 women, with 38 of 67 (56.7%) sites enrolling less than 10 women. The sensitivity of the van Elteren test becomes limited when the sample sizes in so many strata (ie, centers in this analysis) are small.18 Therefore, a parametric analysis of covariance, which was prespecified, but is not sensitive to small center sizes, was considered the most appropriate statistical test for the analysis of these data. Only results from analyses of covariance (or its analysis of variance analog for the continuous secondary efficacy endpoints) are presented in this article. All treatment comparisons were made using a two-sided 0.05 level of significance.
For the analysis of ordinal categorical data (CGIC/PGIC), a Cochran-Mantel-Haenszel test of general association with modified ridit scores was used for the overall comparison among all treatment groups. For the analysis of dichotomous outcome data, a two-sided Fisher’s exact test was used for the overall comparison among all treatment groups. Differences between two proportions (G-GR treatment j placebo) and the associated 95% CIs were obtained using a normal approximation to the binomial. For week 12, data were obtained from the modified intent-to-treat population, and last-observationcarried-forward imputation was used to account for missing data. For week 24, data were obtained from the completer population as observed. RESULTS Six hundred women were enrolled and randomized in this study: 298 to the placebo arm and 302 to the G-GR arm. Five women did not receive study therapy, and two women had no postrandomization efficacy measures recorded, resulting in a modified intent-to-treat population of 593 women (placebo, 294; G-GR, 299) (Fig. 1). One hundred ninety-eight (33.8%) women discontinued from the study: 104 (35.9%) in the placebo arm and 94 (31.8%) in the G-GR arm. Baseline demographics and disease characteristics were similar between groups (Table 1). Participants were primarily white and had a mean age of 54 years, 38% had a body mass index of 30 kg/m2 or higher, and the majority (75%) experienced natural menopause. The baseline mean (SD) daily frequencies of hot flashes were similar between arms: 12.0 (5.5)
FIG. 1. Consort diagram. The intent-to-treat population included women who were randomized, took at least one dose of study therapy, and had at least one postrandomization efficacy assessment. The safety population included all randomized participants who received study medication. The completer population included all women who completed 24 weeks of therapy and had at least one diary entry from days 158 to 168 postrandomization. HF, hot flash; G-GR, gastroretentive gabapentin. Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
569
PINKERTON ET AL TABLE 1. Baseline participant demographics and characteristics: modified intent-to-treat population Variable
Placebo (n = 294)
G-GR (n = 299)
54.0 (6.0) 34-69 282 (95.9) 12 (4.1)
54.0 (6.1) 38-70 280 (93.6) 19 (6.4)
197 (67.0) 82 (27.9) 9 (3.1) 6 (2.0)
215 (71.9) 74 (24.7) 9 (3.0) 1 (0.3)
182 (61.9) 112 (38.3) 17.7-59.3
184 (61.5) 115 (38.5) 16.3-52.4
120.2 (90.2) 107
114.3 (36.7) 104
220 (74.8) 74 (25.2)
223 (74.6) 76 (25.4)
2.2 (1.2)
2.0 (1.3)
12.0 (5.5) 2.54 (0.28) 7.4 (2.2)
11.8 (4.7) 2.55 (0.29) 7.3 (2.2)
Age, y Mean (SD) Range G65, n (%) Q65, n (%) Race, n (%) White Black Hispanic Other Body mass index, n (%), kg/m2 G30 Q30 Range Creatinine clearance, mL/min Mean (SD) Median Menopause history, n (%) Natural Surgical Children delivered Mean (SD) Hot flashes Frequency, mean no./day (SD) Severity, mean (SD) SIS score, mean (SD) G-GR, gastroretentive gabapentin.
for placebo and 11.8 (4.7) for G-GR. The mean (SD) daily hot flash severity scores were also similar: 2.54 (0.28) for placebo and 2.55 (0.29) for G-GR. These scores indicate that, at study entry, half of the hot flashes experienced were moderate and half were severe. The mean baseline sleep interference scores were also similar between arms: 7.4 for placebo and 7.3 for G-GR (out of 10), indicating significant sleep disturbance at baseline (Table 1). Hot flash frequency and severity Women in the G-GR arm had significantly greater reductions in hot flash frequency and severity from baseline compared with women in the placebo arm (Fig. 2). At week 4, the baseline-adjusted mean (SEM) reduction in hot flash frequency was j5.01 (0.27) for the placebo group and j6.72 (0.27) for the G-GR group (difference, j1.70; 95% CI, j2.31 to j1.09; P G 0.0001). At week 12, the mean (SEM) reductions were j6.50 (0.29) for the placebo arm and j7.64 (0.29) for the G-GR arm (difference, j1.14; 95% CI, 1.8 to j0.48; P = 0.0007; Fig. 2). Reductions from baseline in mean hot flash frequency were apparent after 1 week and continued throughout 12 weeks (Fig. 3). At week 12, more women in the G-GR arm had achieved a 50% reduction in hot flash frequency (73% vs 60%). At week 4, the corresponding baseline-adjusted mean (SEM) reductions in hot flash severity were j0.22 (0.04) for the placebo arm and j0.42 (0.04) for the G-GR arm (difference, j0.2 [0.052]; 95% CI, j0.31 to j0.1; P G 0.0001); at week 12, the reductions in severity were j0.46 (0.06) in the placebo arm and j0.65 (0.06) in the G-GR arm (difference, j0.19 [0.07]; 95% CI, j0.33 to j0.04; P = 0.012; Fig. 2).
570
Menopause, Vol. 21, No. 6, 2014
Statistically significant reductions in both hot flash frequency and hot flash severity were maintained for women who completed the 24-week study (Fig. 2). The mean (SEM) reduction in hot flash frequency at week 24 was j7.91 (0.36) for the placebo arm and j8.99 (0.37) for the G-GR arm (difference, j1.08 [0.45]; 95% CI, j1.98 to j0.19; P = 0.0174; Fig. 2). The reductions in hot flash severity at week 24 were j0.86 (0.09) and j0.64 (0.09; difference, j0.22 [0.11]; 95% CI, j0.44 to j0.0; P = 0.0457; Fig. 2). Secondary endpoints: PGIC, CGIC, and SIS For the secondary endpoints addressing clinical meaningfulness, as rated by the PGIC and CGIC scales, significantly more women in the G-GR arm were Bmuch[ or Bvery much[ improved. Based on the PGIC scale, the number of women who reported feeling Bmuch[ or Bvery much[ improved was 44% for the placebo arm versus 58% for the G-GR arm (P = 0.0008) at week 12, and 55% for the placebo arm versus 76% for the G-GR arm (P G 0.0001) at week 24 (Table 2). Using the CGIC, physicians reported that the percentage of women reporting Bmuch[ or Bvery much[ improvement was 42% for the placebo arm versus 58% for the G-GR arm (P G 0.0001) at week 12, and was 60% for the placebo arm (P = 0.0002) versus 78% for the G-GR arm at week 24 (Table 2). Sleep, as measured by the mean SIS, was markedly improved for women in the G-GR arm compared with those in the placebo arm at week 12 (j2.2 [placebo] vs j3.1 [G-GR]; P G 0.0001) and at week 24 (j2.2 [placebo] vs j3.2 [G-GR]; P G 0.0001; Table 2).
FIG. 2. Baseline-adjusted mean (SEM) reductions in (A) hot flash frequency and (B) hot flash severity at weeks 4, 12, and 24. G-GR, gastroretentive gabapentin. * 2013 The North American Menopause Society
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
G-GR FOR MENOPAUSAL HOT FLASHES
the G-GR arm (anemia, infective arthritis, periorbital cellulitis, and chronic lymphocytic leukemia). The only death in the study was that of a participant in the placebo arm who had hemorrhagic stroke. No serious AE was deemed to be related to the study drug. There was one case of positive suicidal ideation in each arm during the study (placebo, at week 24; G-GR, at week 28); both were judged by investigators to be unrelated to therapy. DISCUSSION In this phase 3 study of highly symptomatic postmenopausal women with more than 7 moderate-to-severe hot flashes/day, G-GR produced modest, statistically significant reductions in hot flash frequency and severity compared with placebo. The absolute decrease in hot flash frequency relative to placebo did not meet the Food and Drug Administration guideline of 2 hot flashes/day.19 However, this guideline was designed for standard-dose HT and does not include reductions in hot flash severity or improvements in participantreported outcomes. In this study, the baseline hot flash severity for both arms was 2.5, indicating that every other hot flash was severe and thus led to cessation of personal activity. After 12 weeks on placebo, women had a mean severity score of 2.04, indicating that the participants still had some severe hot flashes. In contrast, after 12 weeks of G-GR treatment, the mean severity score decreased to less than 2, indicating that the women had no moderate or severe hot flashes at all on some days. Thus, the mean reduction of 1.1 hot flashes/ day observed in the G-GR arm corresponded to elimination of a severe hot flash (one that would force a woman to stop whatever she was doing).
FIG. 3. Change from mean baseline hot flash frequency score during the first 12 weeks of study. *Statistically significant difference from placebo at the 0.05 level. G-GR, gastroretentive gabapentin.
Safety The safety population included 595 women: 300 in the G-GR group and 295 in the placebo group. G-GR was well tolerated after titration, with only 5.2% more women in the G-GR arm discontinuing because of AEs compared with those in the placebo arm (16.7% vs 11.5%; Table 3). Discontinuations because of AEs during titration were low (2.4% in the placebo arm and 4.7% in the G-GR arm). The most common AEs leading to discontinuation throughout the study were dizziness (placebo, 0.7%; G-GR, 3.3%) and somnolence (placebo, 1.0%; G-GR, 2.0%). The most common AEs (95% incidence) reported during the study were dizziness (placebo, 3.4%; G-GR, 12.7%), headache (placebo, 8.1%; G-GR, 9.3%), somnolence (placebo, 2.7%; G-GR, 6.0%), and upper respiratory tract infection (placebo, 3.7%; G-GR, 6.0%; Table 3). The incidences of dizziness and somnolence, the AEs most commonly associated with gabapentinoids, decreased rapidly after the titration week (Fig. 4). For dizziness, prevalence decreased from 10.7% at the end of week 1 to 6.2% at the end of week 2, and to a steady rate of approximately 3% by the end of week 6. For somnolence, prevalence was 4.3% at the end of week 1 and stabilized at approximately 3% by the end of week 4. Placebo rates for dizziness and somnolence were less than 2% at the end of each study week. There were seven (2.4%) serious AEs in the placebo arm (palpitations, hemorrhagic stroke, acute sinusitis, musculoskeletal chest pain, asthma, primary atypical pneumonia, and acute myocardial infarction) and four (1.3%) serious AEs in
Clinical meaningfulness of reductions in hot flash frequency and severity Perhaps more important than the modest improvements in hot flash frequency and severity, the clinical meaningfulness of G-GR treatment was evidenced by the finding that significantly more women in the G-GR arm reported feeling Bmuch[ or Bvery much[ improved on the PGIC (58% for G-GR vs 44% for placebo at week 12). This suggests that reduction in the total burden of hot flashes, which factors in reductions in both frequency and severity, contributes to improvement in women’s overall sense of well-being. Among women who completed the study, this greater sense of improvement was
TABLE 2. Secondary efficacy endpoints on weeks 12 and 24 Week 12 Instrument a
PGIC , % CGICa, %
Week 24
Placebo (n = 295)
G-GR (n = 300)
P
Placebo (n = 177)
G-GR (n = 185)
P
44 42
58 58
0.0008 G0.0001
55 60
76 78
G0.0001 0.0002
Placebo (n = 294)
G-GR (n = 299)
P
Placebo (n = 191)
G-GR (n = 206)
P
b
j2.2 j3.1 G0.0001 j2.2 j3.2 G0.0001 SIS G-GR, gastroretentive gabapentin; PGIC, Patient Global Impression of Change; CGIC, Clinician Global Impression of Change; SIS, Sleep Interference Scale. a Patients Bmuch[ or Bvery much[ improved. b Change in score from baseline. Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
571
PINKERTON ET AL TABLE 3. Treatment-emergent adverse events Placebo (n = 295)
Category
G-GR (n = 300)
Any TEAEa 165 (52.9) 197 (65.7) Q1 TEAE leading to discontinuation 34 (11.5) 50 (16.7) Q1 TEAE during titration week 42 (14.2) 81 (27.0) Q1 TEAE leading to discontinuation 7 (2.4) 14 (4.7) during titration week Q1 serious TEAE 7 (2.4) 4 (1.3) Most common TEAEsb Dizziness 10 (3.4) 38 (12.7) Headache 24 (8.1) 28 (9.3) Somnolence 8 (2.7) 18 (6.0) Upper respiratory tract infection 11(3.7) 18 (6.0) G-GR, gastroretentive gabapentin; TEAE, treatment-emergent adverse event. a Among TEAEs, 93.4% were mild or moderate in severity. b Incidence of Q5% or Q2% higher incidence in G-GR versus placebo.
maintained: 76% of women taking G-GR versus 55% of women taking placebo reported feeling Bmuch[ or Bvery much[ improved at week 24. Another important component of clinical meaningfulness is reduction in sleep disturbance, a problem that leads many postmenopausal women to consult their physicians for the treatment of hot flashes. Sleep disturbance due to hot flashes is associated with chronic insomnia,20 leading to fatigue, mood disorders, and reduced quality of life.1,3,21 Women in this study were experiencing significant sleep disturbance at baseline, with a mean SIS score for the total population of 7.3 of 10. Women in the G-GR arm, however, experienced a significant improvement in sleep compared with those in the placebo arm, as evidenced by reduced mean SIS scores at weeks 12 and 24. The fact that the evening dose of G-GR provides peak steady-state levels of gabapentin 6 to 12 hours after dosing10,22 could explain, in part, the improvements in sleep observed with G-GR.10,22 Placebo response rate The placebo response rate observed in this study (a 54% reduction from baseline in hot flash frequency at week 12) was higher by 10% or more compared with other published studies of nonhormone VMS treatments. When treatment effects are modest, higher placebo rates make it more difficult to demonstrate that the active treatment is significantly better. Although not a direct comparison, a cross-study comparison of findings from this G-GR study with those for the low-dose HT Angeliq (0.25 mg of drospirenone and 0.5 mg of estradiol) demonstrates this. In comparable female populations, the mean reductions in hot flash frequency observed at week
12 for G-GR and Angeliq were similar (j7.71 hot flashes/d for Angeliq vs j7.64 hot flashes/d for G-GR).23 The reductions in hot flash frequency observed in the placebo arms, however, were considerably different (j4.54 hot flashes/d for Angeliq vs j6.50 hot flashes/d for G-GR). Thus, based solely on the difference in placebo response rate (~2 hot flashes/d), it appears that Angeliq reduces hot flash frequency by 3.17 hot flashes/day over placebo, whereas G-GR reduces hot flash frequency by only 1.14 hot flashes/day. Factors contributing to the placebo response should be similar for HT and non-HT. Because regression toward the mean tends to occur in women experiencing severe cases of any condition,24 one would expect that the number of hot flashes in placebo participants would decrease (ie, move toward the mean) across time on their own, thus raising the apparent rate of improvement in the placebo arm. Other potential factors include the clinical trial environment, coupled with biofeedback from the requirement to capture each hot flash daily using an electronic device, and increased interactions with healthcare professionals and fellow hot flash sufferers.25 One must also consider that women in the study who were seeking alternatives to HT and may have tried other alternative approaches, might have been especially hopeful that the study medication would provide relief. Safety and tolerability G-GR was well tolerated during the 24 weeks of this study. Given the relatively small number of discontinuations attributed to AEs during the first week of treatment (2.4% in the placebo arm vs 4.7% in G-GR arm), it would seem that AEs were manageable for most women and that the more rapid 1-week titration used in this study (as opposed to the 2-wk titration used in studies of G-GR in postherpetic neuralgia)14 was suitable for this population of highly symptomatic women, who are younger and generally healthier. The low propensity for drug interactions is also an important component of the safety profile of gabapentin. Because gabapentin is renally excreted and not metabolized, G-GR has no interaction with cytochrome P45026 and would therefore not be expected to reduce the effectiveness of tamoxifen in women with breast cancer. Aspects of study design Strengths of this study include its large, diverse study population, which is representative of the overall US population
FIG. 4. Prevalence of the common gabapentinoid-related adverse events, dizziness and somnolence, at the end of weeks 1 through 12.
572
Menopause, Vol. 21, No. 6, 2014
* 2013 The North American Menopause Society
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
G-GR FOR MENOPAUSAL HOT FLASHES
(one third was nonwhite and nearly 40% had a body mass index Q30 kg/m2). A limitation of the study was the lack of a direct comparison with immediate-release gabapentin. However, immediate-release gabapentin is not approved for the treatment of VMS, and this study was designed to have sufficient power to demonstrate the efficacy and safety of G-GR compared with placebo, not to directly compare G-GR with immediate-release gabapentin. Given the high placebo response rate in this study (54%), future studies of G-GR or other nonhormone agents should be designed to minimize the placebo rate. One approach that yielded a placebo rate of 45% to 48% was the use of a prolonged single-blind placebo run-in that eliminated women who responded well to placebo before randomization.27 In that study, which compared low-dose mesylate salt of paroxetine with placebo, hot flash frequency was reduced by 60% in the active arm versus 48% in the placebo arm at week 4 (P = 0.0072) and by 61% in the active arm versus 45% in the placebo arm at week 12 (P = 0.0001).27 Another change in study design for nonhormone agents, which tend to have smaller absolute changes in hot flash frequency and severity, would be to use a participant-reported outcome (eg, the PGIC) as a primary endpoint. Because participants often report feeling significantly improved after treatment with non-HTs despite more modest reductions in hot flash frequency and severity, such an endpoint may better reflect the overall clinical effectiveness and clinical meaningfulness of the agents for the treatment of VMS. CONCLUSIONS Even with the relatively high placebo response rate in this study, G-GR has been found to be modestly effective at reducing the frequency and severity of hot flashes for up to 24 weeks, but it does not meet the Food and Drug Administration criterion of a reduction of 2 or more hot flashes/day. Despite this, as reflected by improvements in women’s overall sense of well-being and in their ability to sleep, the reductions in hot flash frequency and severity afforded by G-GR seem to be clinically meaningful to women. G-GR is well tolerated, especially after the titration period. Acknowledgments: We gratefully thank Stephanie Kareht, PhD, of Depomed for medical writing support and Howard Hait, PhD, for assistance with biostatistics (as a consultant to Depomed).
REFERENCES 1. Bachmann GA. Vasomotor flushes in menopausal women. Am J Obstet Gynecol 1999;180:S312-S316. 2. Morrissey DR, Kirchner JT. Management of the climacteric. Options abound to relieve women’s midlife symptoms. Postgrad Med 2000;108: 85-88, 93-94, 97-100. 3. Oldenhave A, Jaszmann LJ, Haspels AA, Everaerd WT. Impact of climacteric on well-being. A survey based on 5213 women 39 to 60 years old. Am J Obstet Gynecol 1993;168:772-780. 4. Guttuso T Jr, Kurlan R, McDermott MP, Kieburtz K. Gabapentin’s effects on hot flashes in postmenopausal women: a randomized controlled trial. Obstet Gynecol 2003;101:337-345.
5. Pandya KJ, Morrow GR, Roscoe JA, et al. Gabapentin for hot flashes in 420 women with breast cancer: a randomised double-blind placebocontrolled trial. Lancet 2005;366:818-824. 6. Butt DA, Lock M, Lewis JE, Ross S, Moineddin R. Gabapentin for the treatment of menopausal hot flashes: a randomized controlled trial. Menopause 2008;15:1-9. 7. Guttuso T Jr. Hot flashes refractory to HRT and SSRI therapy but responsive to gabapentin therapy. J Pain Symptom Manage 2004;27: 274-276. 8. Reddy SY, Warner H, Guttuso T Jr, et al. Gabapentin, estrogen, and placebo for treating hot flushes: a randomized controlled trial. Obstet Gynecol 2006;108:41-48. 9. NeurontinA (gabapentin) Capsules, NeurontinA (gabapentin) Tablets, NeurontinA (gabapentin) Oral Solution [US prescribing information]. New York, NY: Pfizer Inc; 2012. Available at: http://labeling.pfizer.com/ ShowLabeling.aspx?id=630. Accessed July 18, 2012. 10. Argoff CE, Chen C, Cowles VE. Clinical development of a once-daily gastroretentive formulation of gabapentin for treatment of postherpetic neuralgia: an overview. Expert Opin Drug Deliv 2012;9:1147-1160. 11. Hou SY, Cowles VE, Berner B. Gastric retentive dosage forms: a review. Crit Rev Ther Drug Carrier Syst 2003;20:459-497. 12. Berner B, Cowles VE. Case studies in swelling polymeric gastric retentive tablets. Expert Opin Drug Deliv 2006;3:541-548. 13. Gordi T, Hou E, Kasichayanula S, Berner B. Pharmacokinetics of gabapentin after a single day and at steady state following the administration of gastric-retentive, extended-release, and immediate-release tablets: a randomized, open-label, multiple-dose, three-way crossover, exploratory study in healthy subjects. Clin Ther 2008;30:909-916. 14. Sang CN, Sathyanarayana R, Sweeney M. Gastroretentive gabapentin (G-GR) formulation reduces intensity of pain associated with postherpetic neuralgia (PHN). Clin J Pain 2013;29:281-288. 15. Sweeney M, Sathyanarayana R, Gordi T, Cowles VE, Heritier M. Positive efficacy data from a phase 2 trial of gabapentin extended-release in the treatment of menopausal hot flashes [Abstract]. Menopause 2008;15: 1225. Abstract P-61 16. Kagan R, Sathyanarayana R, Sweeney M, Pratt J, Zheng Y. The effect of outliers on hot flash frequency in the Breeze 1 Study of gabapentin-ER for the non-hormonal treatment of vasomotor symptoms [Abstract]. Menopause 2010;17:1219. Abstract S-14. 17. Baron M, Cowles VE, Utian W, Koltun W, Sathyanarayana R, Sweeney M. Efficacy of gabapentin extended-release (G-ER) on vasomotor symptoms in post-menopausal women in Breeze 2 Study [Abstract]. Menopause. 2011;18:1369. Abstract P-76 18. Stokes ME, Davis CS, Koch GG. Categorical Data Analysis Using the SAS System, 2nd ed. Cary, NC: SAS Institute Inc, 2000. 19. Center for Drug Evaluation and Research. Guidance for Industry: Estrogen and Estrogen/Progestin Drug Products to Treat Vasomotor Symptoms and Vulvar and Vaginal Atrophy SymptomsVRecommendations for Clinical Evaluation. Rockville, MD: Food and Drug Administration, 2003. http:// www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm071643.pdf. Accessed May 31, 2013. 20. Ohayon MM. Severe hot flashes are associated with chronic insomnia. Arch Intern Med 2006;166:1262-1268. 21. Regestein QR. Hot flashes, sleep, and mood. Menopause 2010;17:16-18. 22. Cowles VE, Gordi T, Hou SY. Steady-state pharmacokinetics of gabapentin after administration of a novel gastroretentive extendedrelease formulation in postmenopausal women with vasomotor symptoms. Clin Drug Investig 2012;32:593-601. 23. Angeliq [US prescribing information]. Wayne, NJ: Bayer HealthCare Pharmaceuticals; 2012. Available at: http://berlex.bayerhealthcare.com/ html/products/pi/fhc/Angeliq_PI.pdf. Accessed May 31, 2013. 24. Bland JM, Altman DG. Regression towards the mean. BMJ Open 1994;308:1499. 25. Mahon SM, Kaplan M. Placebo effect in hot flush research. Lancet Oncol 2012;13:e188. 26. Patsalos PN, Perucca E. Clinically important drug interactions in epilepsy: interactions between antiepileptic drugs and other drugs. Lancet Neurol 2003;2:473-481. 27. Simon JA, Sanacora G, Bhaskar S, Lippman J. Safety and efficacy of low-dose mesylate salt of paroxetine (LDMP) for the treatment of vasomotor symptoms (VMS) associated with menopause: a 24-week, randomized, placebo controlled phase 3 study [Abstract]. Menopause. 2012;19:1371. Abstract S-2.
Menopause, Vol. 21, No. 6, 2014
Copyright © 2014 The North American Menopause Society. Unauthorized reproduction of this article is prohibited.
573
