| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Woman’s Health Research Institute (K.E.-H.), Woman’s Hospital, and Metabolic Center of Louisiana Research Foundation (O.M., M.B., D.V., R.B.), Baton Rouge, Louisiana 70815
Address all correspondence and requests for reprints to: Karen Elkind-Hirsch, M.Sc., Ph.D, Woman’s Health Research Institute, 9050 Airline Highway, Baton Rouge, Louisiana 70815. E-mail: Karen.Elkind-Hirsch{at}womans.org.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: We evaluated exenatide (EX) and metformin (MET), alone and in combination (COM), on menstrual cyclicity, hormonal parameters, metabolic profiles, and inflammatory markers in overweight, insulin-resistant women with PCOS.
Design, Setting, and Participants: Sixty overweight oligoovulatory women with PCOS (body mass index > 27; 18–40 yr) were randomized to one of three treatment groups: MET [1000 mg twice daily (BID)], EX (10 µg BID), or COM (MET 1000 mg BID, EX 10 µg BID) for 24 wk. The primary outcome was menstrual frequency; secondary outcome measures included changes in ovulation rate, insulin action, anthropometric measures, androgen levels, and inflammatory markers.
Results: Forty-two (70%) patients completed the study. COM therapy was superior to EX or MET monotherapy in improving menstrual cyclicity, ovulation rate, free androgen index, and insulin sensitivity measures and reducing weight and abdominal fat. Both EX arms were more effective in promoting weight loss than MET (P = 0.003).
Conclusions: COM appears better than either EX or MET alone on menstrual cycle frequency and hormonal and metabolic derangements. A marked decrease in central adiposity could partly explain the improvements in reproductive function, insulin-glucose parameters, and adiponectin observed in these overweight women with PCOS treated with COM therapy. Larger trials of longer duration are warranted to assess the long-term efficacy and safety of combined EX-MET therapy in overweight women with PCOS.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
PCOS is a prediabetic state for which early recognition and intervention, such as weight control, diet modification, and/or pharmacological intervention may prevent or delay the development of diabetes (9). Both the hyperinsulinemia and the hyperandrogenism can be reduced with weight loss, resulting in better regularity of menses and fertility potential (10, 11). In women with PCOS, both thiazolidinediones and metformin (MET) decrease IR and androgen levels and increase ovulation in a dose-responsive fashion (12, 13, 14, 15). However, although progressive IR plays a key role in the predisposition to diabetes in PCOS, subtle alterations in I secretion also appear to contribute to the susceptibility toward DM2 (2, 3, 4). A novel antidiabetic medication, exenatide (EX), is an incretin mimetic that shares similar glucoregulatory properties of the hormone glucagon-like peptide-1 (GLP-1) including Glu-dependent enhancement of I secretion (16, 17). Among the primary actions of EX in subjects with DM2 is the ability to restore first- and second-phase I secretion, which is attenuated in this population (18). Therapy with EX often results in weight loss, which further assists in decreasing IR (19, 20, 21). Optimal treatment of PCOS would not only correct specific clinical consequences of anovulation but also reduce the comorbidities such as obesity and DM2 linked to this syndrome. We report the results of a 24-wk trial designed to directly compare the therapeutic effects of EX and MET alone or in combination (COM) on menstrual cyclicity and metabolic and endocrinological parameters in overweight/obese women with PCOS.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This was an open-label prospective, randomized, outpatient clinical trial with three treatment groups over 24 wk; the primary outcome was change in menstrual frequency. Prespecified secondary outcomes included changes in anthropometric measurements [body mass index (BMI), abdominal girth (AG), and absolute body weight], rate of ovulation, I sensitivity and secretion, reproductive hormone levels, lipid profiles [total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglycerides (TRG)], inflammatory markers [high-sensitivity C-reactive protein (hsCRP), IL-6, and TNF-
]), and adiponectin levels.
Women with PCOS were eligible for enrollment if they met the following criteria: 18–40 yr and overweight/obese (BMI > 27). The diagnosis of PCOS was determined using a modification of the criteria from Rotterdam 2003 (1). To be eligible for the study, subjects had to have menstrual disorders (fewer than six menstruations in 12 months) and one of the following two criteria: either clinical and/or biochemical hyperandrogenism (excluding secondary causes) and/or polycystic ovaries. Diabetic subjects, smokers, those who used injectable hormonal contraceptive within 6 months, or those taking sex hormones, drugs that affect gastrointestinal motility or carbohydrate metabolism, or lipid-lowering and/or antiobesity drugs within 3 months of the study were excluded.
The Institutional Review Board of the Woman’s Hospital Foundation approved the study. After obtaining written consent, all subjects were screened with a serum TSH, β-human chorionic gonadotropin, and lipid profile to exclude thyroid disorder, severe hypertriglyceridemia, and/or pregnancy. Eighty-one women with PCOS were consented and 21 participants were excluded from the study. Eligible patients were allocated one of three treatment groups by a simple randomization process according to a computer-generated code to allow equal patient distribution of 20 patients per group. The randomization code for the study was generated electronically using a block-of-three randomization technique to create a treatment assignment spreadsheet. Subjects and investigators were not blinded as to the treatment intervention.
Treatment protocol
Subjects underwent the following clinical, metabolic, and laboratory evaluations before, after 12 wk, and after 24 wk of study treatment. Vital signs as well as the anthropometric measurements were obtained at all visits. AG was measured at the greatest protuberance. The BMI was calculated using the standard formula (kilograms per square meter).
All the patients were on an unrestricted diet. At baseline and final evaluation (24 wk), a 75-g oral Glu tolerance test (OGTT) was performed in the morning (0700–1000 h) after a 12-h overnight fast with Glu and I levels measured at 0 (baseline before the Glu load), 30, 60, and 120 min. Fasting baseline blood samples were also used for measures of estradiol, testosterone (T), SHBG, dehydroepiandrosterone sulfate (DHEAS), inflammatory markers, and adiponectin.
Patients received either extended-release MET [1000 mg twice daily (BID)], EX (10 µg BID), or COM (MET plus EX) for 24 wk. For all patients on MET, the initial dose was 500 mg before dinner for at least 2 wk and gradually increased to a final dose of 1000 mg BID (breakfast and dinner). All patients receiving EX were started on 5 µg administered by sc injection BID by means of a prefilled pen and increased to 10 µg sc BID after 1 month (Fig. 1
). Subjects recorded the presence/absence of menstrual bleeding daily in a menstrual cycle diary distributed at the time medication was dispensed.
|
All laboratory testing and anthropometric measures described for the pretreatment evaluation was repeated at the final 24-wk visit. All subjects’ complete menstrual diaries were reviewed and occurrence of ovulation (luteal serum progesterone level > 3 ng/dl) documented. Adverse events were recorded throughout the study by direct questioning, by self-reporting by patients, and from the results of physical examinations and clinical laboratory tests. Counting the tablets and/or measuring excess EX in the syringe returned by the participants after each treatment period assessed adherence to therapy.
Laboratory measures and assays
Serum T, DHEAS, progesterone, estradiol, SHBG, I, Glu, total cholesterol, HDL-C, TRG, calculated LDL-C, and TSH were determined as previously described (22). The proinflammatory cytokines and adiponectin were determined by ELISA (Quest Diagnostic/Nichols, San Juan Capistrano, CA). The free androgen index (FAI) was calculated as the quotient 100 x T/SHBG (23). After the Glu and I levels had been measured, the following mathematical models were used to assess I action. Using fasting baseline Glu and I concentrations, IR was calculated by the homeostasis model assessment for IR (HOMA-IR) (24). From the 2-h OGTT, the Matsuda index (ISOGTT) was derived from both fasting and stimulated values of Glu and I (25). The insulinogenic index (IGI; 
I/Glu30–0) and corrected I response at Glu peak (CIRgp), indexes of β-cell activity, were computed from Glu-I values obtained during the OGTT (26, 27). An I secretion-sensitivity index (ISSI) was derived by applying the concept of the disposition index to measurements obtained during the 2-h OGTT (28).
Statistical analyses
The intent-to-treat population (ITT) was defined as all 60 randomized subjects who received at least one injection of medication and/or took one oral dose of medication starting from the evening of d 1. The evaluable population was defined as all 42 randomized subjects who completed treatment through wk 24 and received at least 90% of the study medication injections and/or oral drug.
The primary outcome measure was menstrual frequency, and secondary outcome measures included changes in anthropometric parameters, I sensitivity and pancreatic β-cell function, hormonal levels, and inflammatory cytokines. The normality of all variables was checked using the Kolmogorov-Smirnov test. The results are presented as means ± SEM unless otherwise indicated. Menstrual frequency (expressed as a ratio variable) before and after different drug treatment (evaluable population) was compared with the McNemar test. The menstrual frequency ratio (MFR) was calculated using the ratio of expected menses to observation weeks. Menstrual events were normalized such that an individual with normal menstrual rhythm would have 12 cycles in 52 wk (MFR of 1.00) and have five cycles in 24 drug intervention weeks (3–4 treatment weeks to impact the cycle), yielding a ratio of 100% (treatment MFR = 1.0). Based on eligibility criteria, all participants had a pretreatment MFR in the preceding year of less than 0.5. Treatment impact on the occurrence of ovulation was analyzed using the Kruskal-Wallis and median test. For all other analyses, in which the measures are continuous, data from evaluable subjects were submitted to a repeated-measures general linear model (subjects/treatments x repeated-measures ANOVA) including the arm of drug treatment as the between-subjects effect, and the visit (baseline and 24 wk) as the within-subjects effect. To evaluate the differences in the response to each treatment over visits, the interaction effect was calculated. Only where a statistically significant interaction effect was found (P 
0.05) was the contrast test applied to locate the differences between the three medication groups. Adjustments for multiple testing were performed using Bonferroni correction. Baseline comparisons between groups (ITT) were made by one-way ANOVA and post hoc comparisons performed with the Bonferroni-Dunn test to analyze the variation among the three groups if the ANOVA showed overall baseline differences were significant (P < 0.05). Regression analysis between weight loss and AG reduction was performed using Pearson product-moment correlation test. We further explored the association of changes in adiponectin with alterations in weight, AG, I measures, FAI, T, and SHBG levels using Pearson’ correlation procedures; the resulting estimates are reported as r values.
A priori sample size analysis was performed using the online calculator provided by the Massachusetts General Hospital Mallinckrodt General Clinical Research Center (http://hedwig.mgh.harvard.edu/sample_size/size.html). The primary endpoint of this study, for which a power calculation was used to determine the sample size, was resumed regular menstrual rhythm. Given there were no previous studies using EX or combination treatment for PCOS, we based our calculations on a previous study with comparator drug intervention (29). It was determined that 20 subjects per treatment group would be needed to give a power of 80% to detect a statistically significant difference ( = 0.05) of approximately 60% regular menstrual frequency with treatment by comparison with a pretreatment menses prevalence projected as 20% based on the inclusion criterion of oligomenorrhea. The calculation was adapted to allow for significant dropout rate (25%) for completion of the full trial of 24 wk. Other than sample size calculations, all analyses were performed using SPSS 15.1 for Windows (SPSS, Inc., Chicago, IL). All P values are two-tailed; P < 0.05 was considered statistically significant.
Safety endpoints included adverse events, laboratory tests, physical exams, and vital signs. All safety analyses were performed using the ITT and performed with the use of summary statistics and frequency tables.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The study was conducted from August 2006 to June 2007. Forty Caucasian and 20 African-American women were randomized and received treatment, and 42 patients (70%) completed the study per protocol (Fig. 1
). Race was equally distributed across treatment arms. Basal and Glu-stimulated anthropometric, hormonal, and metabolic characteristics of the three groups of patients with PCOS (ITT) are shown in Table 1. At baseline, all studied parameters for the three groups were similar.
|
Cycle changes
Clinical, metabolic, anthropometric, and laboratory study outcomes at baseline and after 24 wk on completed treatment patients are summarized in Tables 2 and 3. Menses occurrence was significantly increased in all groups after treatment (P < 0.001; Table 2). Changes in the menstrual index with drug treatment are further illustrated in Fig. 2. There was a significant interaction effect on bleeding periods (P = 0.047) with menses more regular with COM therapy vs. single-agent therapy with MET (P = 0.018). Compared with baseline, ovulatory rates improved on both single-agent and COM treatments with the rate of ovulation with COM therapy significantly higher than either agent alone (P < 0.01). In the COM group, the ovulation rate was 86% (12 of 14) compared with 50% (seven of 14) in the EX group and 29% (four of 14) in the MET group.
|
|
|
). Both EX arms were more effective in enabling weight loss than MET (P = 0.003). After 24 wk, the COM group had a mean weight loss of 6 ± 0.5 kg, the EX group had a mean weight loss of 3.2 ± 0.1 kg, and the MET group had a mean loss of 1.6 ± 0.2 kg (Fig. 3). Comparable results were shown for BMI with a significant reduction in BMI after 24 wk in all groups (P < 0.001; Table 2) with COM significantly better than MET in lowering BMI. Reduction in body weight was significantly correlated with an increase in menstrual frequency (r = 0.42; P < 0.006).
|
). As seen in Fig. 3, AG was steadily reduced at 12 and 24 wk with COM therapy, whereas MET treatment resulted in a decrease at 12 wk with a slight increase by 24 wk (P < 0.017). Reductions in weight and AG were significantly correlated across all treatment regimens (r = 0.7; P < 0.0001).
Endocrine changes
Total testosterone (P < 0.02) and FAI (P < 0.001) were significantly decreased, whereas SHBG levels were increased, but not significantly, with all treatments (Table 2). FAI was significantly more reduced with COM treatment compared with MET (P < 0.35) but not EX alone (Table 2). Levels of DHEAS and TSH were not significantly altered by treatment (Table 2).
Metabolic changes Eighteen women with PCOS (30% ITT) had Glu intolerance; 11 completed the study. Seven (64%) of these women had normal Glu tolerance at the completion of therapy (one of three on EX; three of five on MET, and three of three on COM).
The HOMA-IR values significantly decreased with all treatments (P = 0.043; Table 3). Likewise, I sensitivity determined by the ISOGTT was significantly improved with treatments (P < 0.002). Subjects’ ISOGTT was significantly higher on COM therapy compared with EX therapy (P < 0.02) but not compared with MET therapy (P < 0.085; Table 2). Although there were no statistically significant changes in the IGI with treatment, I secretion as measured by the CIRgp was significantly reduced by EX and COM therapy (P < 0.016; Table 3). ISSI, derived from the product of the IGI and the ISOGTT, showed a progressive increasing mean score from MET (232 ± 116) to EX (395 ± 112) to COM (516 ± 117) (P < 0.005) indicative of improved β-cell function with enhanced I sensitivity (Table 3). The mean ISSI score in the COM treatment group was significantly higher than the mean score on MET (P < 0.049).
Total cholesterol and TRG levels decreased significantly with COM therapy compared with MET monotherapy that did not consistently improve (COM vs. MET, P < 0.035 and P < 0.02, respectively; Table 3). HDL-C and LDL-C levels did not change significantly with treatment (Table 3).
Inflammatory markers
Adiponectin levels were significantly increased with all treatments (P = 0.04; Table 3). Although there was no association between weight loss or AG reduction and serum adiponectin (r = –0.1; P = 0.587), increased adiponectin levels were significantly correlated with reduction of I secretion (CIRgp) after treatment (r = 0.42; P < 0.01). An increase in adiponectin was also related to a decrease in FAI (r = 0.33; P < 0.039) but not T or SHBG levels (r = 0.18; P = 0.28). No consistent change in TNF-, IL-6, or hsCRP levels were observed despite improvement in I action and weight loss with treatment (Table 3).
Adverse events
The most frequent adverse events were mild or moderate and were gastrointestinal in nature (Table 4). Nausea was the most frequent adverse event (overall 27%), and it was higher in COM-treated subjects than in those on monotherapy (Table 4). Nausea was generally mild or moderate in intensity and was reported at a higher incidence during the initial weeks of EX therapy (weeks 0–8) and declined thereafter. Other reported adverse events associated with EX were vomiting (7%) or headache (2%). Diarrhea (overall 13%) was more likely to occur with combination or MET treatment. Gradual dose titration reduced the gastrointestinal side effects associated with EX and MET. None of the participants treated with EX, MET, or COM discontinued the study because of gastrointestinal adverse effects. Four patients were excluded because they became pregnant during the study. All the pregnancies were detected and medications discontinued within a maximum of 7 d of expected menses.
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Clinical improvement in menstrual cycle regularity was evident with all treatment regimens with significantly more women experiencing menstrual periods on COM therapy than MET. Although a progesterone level was documented in only one cycle, the rate of ovulation was over 85% with COM therapy compared with less than 30% with MET, indicating the superiority of COM therapy vs. either agent alone. In all groups, improvement in bleeding frequency was related to weight loss and preceded the normalization of carbohydrate metabolism and other metabolic disturbances.
Both MET and EX were associated with weight loss with COM having an additive effect. Studies have reported modest weight reduction in patients taking MET with more consistent effects with higher doses of MET (30). Although not tested in women with PCOS, our findings are consistent with King et al. (31) who reported after 12 wk of EX in patients with DM2 a mean weight loss of 3.89 kg in the entire group and 4.23 kg in those on thiazolidinediones. The improvement in central adiposity was most dramatic with COM therapy where mean AG measurements steadily declined over 24 wk, whereas mean AG was not improved at 24 wk on MET. This finding is in agreement with Lord et al. (32), who showed that MET had no clinically significant effect on reducing visceral fat in women with PCOS.
Total T levels, as well as free androgen values, were significantly decreased in all treated PCOS groups. These changes are most likely caused by a decrease in circulating I in all patients responding to treatment. The mean FAI in the combined group was significantly lower compared with MET alone but not EX at study completion. This trend could be the result of increased levels of SHBG observed with EX therapy but not MET, further reducing the bioavailability of circulating androgens. Metformin, however, did improve the FAI, secondary to a significant fall in total T without a change in SHBG. The data on improvements in SHBG with MET in women with PCOS vary and are less convincing when considered together with placebo data (32). Our finding that MET failed to significantly influence SHBG is an observation that has been recorded previously in obese PCOS women (33). The failure to significantly change these values with MET treatment may reflect the confounding effect of obesity in the patient cohort and the complex nature of SHBG control mechanisms. The fact that subjects were still relatively hyperinsulinemic and obese in the MET-treated PCOS group may further explain no increase of the SHBG.
All of the subjects were IR at baseline. Although the improvement in HOMA-IR was greater with COM therapy, the lack of statistically significant differences between treatments in the present study may be due to the large variability in IR calculated using the HOMA method in women with PCOS (34). An apparently unique form of IR characterizes PCOS. Even in the absence of changes in fasting Glu and I levels, the results of dynamic studies indicate that Glu utilization is variable in overweight women with PCOS. These abnormalities are not detected using the HOMA indices that are calculated from fasting values. The ISOGTT was significantly improved with all treatments, with COM treatment more effective than EX alone. Subtle alterations in I secretion were also demonstrable, even in subjects whose Glu tolerance was normal. The CIRgp was significantly improved with EX and COM therapy, suggesting more efficient I secretion, which was not observed with MET treatment. Given the inverse relationship between I secretion and I sensitivity, it requires estimation of both variables for correct assessment in any individual. We observed COM treatment significantly reduced fasting I levels as well as improved first-phase I responses to oral Glu administration. The mean ISSI, an index of I secretion in relation to I sensitivity, improved to the greatest extent with COM therapy followed by EX and then MET after 24 wk. The minimal effect of MET in our study may result from the research participants having moderate to extreme obesity. Other findings of MET alone in comparably obese women noted a similar lack of benefit with MET (33, 35).
We found that the increase in circulating adiponectin with all treatments was correlated with reduced I secretion but not to changes in body weight or AG. This is in accordance with previous findings that circulating adiponectin levels in women with PCOS are not directly affected by BMI or fat mass but with the degree of Glu intolerance and IR (36). Interestingly, in this study, no significant changes in levels of hsCRP, TNF-, and IL-6 were observed with any treatment despite decreases in body weight, abdominal adiposity, and improved I action. Similarly, others have reported that modest weight loss or MET treatment had no effect on CRP, IL-6, and TNF- in overweight women with PCOS (37, 38, 39). Given the small number of patients in this study, the absence of association between some of the serum proinflammatory cytokines and treatment intervention could potentially be a result of the power limitation of the study.
In summary, this study provides preliminary evidence that the combination of EX and MET resulted in greater improvements in menstrual cyclicity, endogenous ovulatory function, body weight, AG, IR, and hyperandrogenism than single-agent treatment. The small sample size, high dropout rate (30%), and known allocation of treatment by the patient, practitioner, and researcher limit the generalizability of results. Large, blinded, randomized clinical trials of longer duration are warranted to assess the long-term efficacy and safety of combined EX-MET therapy in overweight women with PCOS.
|
Footnotes |
|---|
The ClinicalTrials.gov number for this work is NCT00344851.
Disclosure Statement: K.E.H. is on advisory boards for Organon USA and Schrafts/Walgreen’s Specialty Pharmacy and received lecture fees from Organon USA and Amylin Pharmaceuticals/Eli Lilly & Co., Inc. R.B. is on an advisory board for for Sanofi-Aventis Pharmaceuticals and received lecture fees from Merck & Co., Inc.; Amylin Pharmaceuticals/Eli Lilly & Co., Inc.; GlaxoSmithKline Pharmaceuticals; Sanofi-Aventis Pharmaceuticals; Novartis Pharamaceuticals Corp.; and Medtronics. O.M., M.B., and D.V. have nothing to declare.
First Published Online May 6, 2008
Abbreviations: AG, Abdominal girth; BID, twice daily; BMI, body mass index; CIRgp, corrected insulin response at Glu peak; COM, combination; DHEAS, dehydroepiandrosterone sulfate; DM2, type 2 diabetes mellitus; EX, exenatide; FAI, free androgen index; Glu, glucose; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, homeostasis model assessment for insulin resistance; hsCRP, high-sensitivity C-reactive protein; I, insulin; IGI, insulinogenic index; IR, insulin resistance; ISOGTT, Matsuda index; ISSI, insulin secretion-sensitivity index; ITT, intent-to-treat population; LDL-C, low-density lipoprotein cholesterol; MET, metformin; MFR, menstrual frequency ratio; OGTT, oral glucose tolerance test; PCOS, polycystic ovary syndrome; T, testosterone; TRG, triglyceride.
Received January 16, 2008.
Accepted April 24, 2008.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K.A. Toulis, D.G. Goulis, D. Farmakiotis, N.A. Georgopoulos, I. Katsikis, B.C. Tarlatzis, I. Papadimas, and D. Panidis Adiponectin levels in women with polycystic ovary syndrome: a systematic review and a meta-analysis Hum. Reprod. Update, May 1, 2009; 15(3): 297 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Palomba, A. Falbo, F. Zullo, and F. Orio Jr. Evidence-Based and Potential Benefits of Metformin in the Polycystic Ovary Syndrome: A Comprehensive Review Endocr. Rev., February 1, 2009; 30(1): 1 - 50. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, P = 0.018, between-treatment effect (COM vs. MET).