Exenatide, Dapagliflozin, or Phentermine/Topiramate Differentially Affect Metabolic Profiles in Polycystic Ovary Syndrome

Karen E. Elkind-Hirsch; N. Chappell; Ericka Seidemann; John Storment; Drake Bellanger


J Clin Endocrinol Metab. 2021;106(10):3019-3033. 

In This Article


Participant Disposition

A total of 130 participants provided consented and were screened and 119 met criteria and were randomly assigned. Eighty-five non-Hispanic White (71.4%) and 34 non-Hispanic Black women (28.6%) were started on medical treatment. Ethnicity was participant defined and race equally distributed across treatment arms. Ninety-two of 119 (77%) participants completed the study per protocol [69 (83%) White and 23 (68%) Black participants]. Intolerable side effects from the medications (n = 10) and pregnancy (n = 4) were the main reason for participant loss from trial completion. The highest loss rate was from the PHEN/TPM arm, from which 9 participants dropped out before study completion. The study was conducted from April 2016 to July 2020. A summary of participant flow and disposition over time is illustrated in Figure 1.

Figure 1.

Participant flow and patient disposition over 24 weeks of trial.

Baseline Characteristics

Baseline anthropometric and hormonal characteristics as well as basal and glucose-stimulated metabolic measures of the 5 treatment groups that completed study trial are shown in Table 1. When baseline comparisons were analyzed, there were no consistent differences between the treatment groups for intent-to-treat participants as well as participants completing the trial (see Table 1). Additionally, no significant differences in menstrual cycle frequency, prolactin, TSH, and SHBG between treatment groups at baseline were determined for intent-to-treat participants and completers.

Effects of Treatment at 24 Weeks

Pretreatment baseline characteristics of participants completing the trial (see Table 1) were compared with glycemic, anthropometric, and cardiometabolic parameters measured after 24 weeks of completed treatment and are summarized in Table 2 and Table 3.

Fasting and Mean Glucose Levels. FBG and MBG levels during OGTT were significantly decreased in all treatment groups (P ≤ .0001). Participants on all therapies showed significantly improved glucose excursion during an OGTT after 24 weeks of treatment, with patients on EQW/DAPA showing a significantly greater drop in MBG levels compared with DAPA, DAPA/MET, and PHEN/TPM (P < .03; see Table 2). At the 24-week assessment, no participants treated with any of the 5 treatment drugs had progressed to diabetes based on the OGTT.

Insulin Sensitivity and Secretion. Fasting insulin sensitivity as determined by the HOMA-IR was improved by all treatments (P < .035). Similarly, the OGTT-derived insulin sensitivity index (SIOGTT) was significantly increased after 24 weeks with all drug therapies (P = .0001; see Table 3). The corrected early insulin response to a glucose challenge (IGI/HOMA-IR) was not significantly changed with all treatments; a significant increase was seen in the EQW and EQW/DAPA groups as compared with DAPA, DAPA-MET, and PHEN/TPM (P < 0.04; see Table 2)

In evaluating overall carbohydrate metabolism, the IS-SI is a simple surrogate estimate of β-cell function relative to insulin sensitivity determined using an OGTT. Participants' mean IS-SI score in the EQW and EQW/DAPA treatment groups was significantly improved compared with the mean score after 24 weeks of treatment in the DAPA, DAPA/MET, and PHEN/TPM groups (P = .027; see Table 2). As shown in Figure 2, mean IS-SI improved steadily over 24 weeks in all groups, with EQW and EQW/DAPA showing significantly greater improvement than the other groups over the treatment period.

Figure 2.

Insulin secretion-sensitivity index at baseline and after 24 weeks of treatment with exenatide (EQW), exenatide and dapagliflozin (EQW/DAPA), DAPA, DAPA/extended-release metformin (MET), and phentermine topiramate extended release (PHEN/TPM) (only patients who completed the study are included). Data shown are the mean ± SEM; P less than .015 overall (pretreatment vs posttreatment); †P equals .027 specific treatment effect (EQW and EQW/DAPA vs other groups). Mean insulin secretion–sensitivity index (IS-SI) improved steadily over 24 weeks in all groups, with EQW and EQW/DAPA showing significantly greater improvement than the other groups over the treatment period.

Weight Trajectories

There was a significant decrease in absolute BW and BMI from first to last visit in all treatment groups (P < .0001), with the combination EQW/DAPA and PHEN/TPM therapy more effective in promoting weight loss than DAPA or DAPA/MET (P = .005; see Table 2). Absolute BW steadily decreased in all groups but was reduced significantly more with EQW/DAPA and PHEN/TPM.

Measures of Central Adiposity

All drug treatments resulted in weight loss and favorable changes in a variety of measures of truncal adiposity. While abdominal adiposity measures decreased with all drug therapy, statistical comparisons of WC and WHtR showed a significantly greater reduction in mean WC (P ≤ .035) and mean WHtR (P < .046) after 24 weeks in the EQW/DAPA and PHEN/TPM groups compared with DAPA/MET (see Table 2). There were no significant between-group differences in WHR found (P < .05; see Table 3).

Percentage Change in Anthropometric Measures From Baseline at Study Completion

When mean percentage weight loss from baseline was compared, the EQW/DAPA and PHEN/TPM participants showed significantly greater weight loss when compared to DAPA and DAPA/MET; mean weight loss for EQW/DAPA was 6.9% (± 0.9) and PHEN/TPM was 8% (± 1.3) vs 1.5% (± 1.4) for DAPA and 1.7% (± 1.1) for DAPA/MET (P < .001) as illustrated in Figure 3. This was further supported by similar comparisons of decrease from baseline of WC and WhtR over 24 weeks, which showed a significant decrease with EQW/DAPA and PHEN/TPM compared to DAPA/MET treatment.

Figure 3.

Change in percentage of body weight (%) from baseline to 24 weeks in exenatide (EQW), exenatide and dapagliflozin (EQW/DAPA), DAPA, DAPA)/extended-release metformin (MET), and phentermine topiramate extended release (PHEN/TPM). Data are presented as mean ± SEM. When mean percentage weight loss from baseline was compared, the EQW/DAPA and PHEN/TPM participants showed significantly greater weight loss when compared to DAPA and DAPA/MET; mean weight loss for EQW/DAPA was 6.9% (± 0.9) and PHEN/TPM was 8% (± 1.3) vs 1.5% (± 1.4) for DAPA and 1.7% (± 1.1) for DAPA/MET (P < .001).

Body Fat Composition by Dual-energy X-ray Absorptiometry

With all drug treatments, TBM and TBF% as measured by DXA decreased significantly (P < .0001). Similar to absolute BW and BMI, the combinations of EQW/DAPA and PHEN/TPM therapy were more effective in promoting loss of TBF compared with DAPA or DAPA/MET (see Table 2; TFM kg P < .009, TBF% < .0075). All treatments resulted in a reduction in lean BM (see Table 3; P < .0001).

Levels of abdominal adiposity based on AGR and TLR measured by DXA are similar to WHR, for which body shape is described using a ratio but instead of waist and hip circumferences, body trunk/torso and leg fat are used. With all treatments, there was a decrease in the AGR with a significant decrease between EQW/DAPA vs DAPA treatments (P < .01, Table 2). While a significant decrease with all treatments was not measured with the trunk/leg fat mass ratio, a significant decrease in TLR was found between EQW/DAPA vs DAPA and DAPA/MET treatments (Table 2; P ≤ .035).

Correlation Between Glycemic and Dual-energy X-Ray Absorptiometry Measures

We explored the relationship between insulin sensitivity (HOMA-IR and SIOGTT) and AGR in these obese hyperandrogenic women. Fasting insulin sensitivity, HOMA-IR (r = 0.24, P = .04), was significantly associated with AGR. There was an even stronger negative correlation between peripheral insulin sensitivity (SIOGTT) and central adiposity as defined by the AGR (r = –0.41, P < .0001).

Hormonal and Cardiometabolic Measures

TT (P < .001) and FAI (P < .001) were significantly improved in all treatment groups (see Table 3). Although not statistically consistent, DHEA-S levels decreased in all groups except for the PHEN/TPM group but the drug effect did not reach statistical significance (see Table 3). SHBG levels significantly increased with all drug treatments (P < .001)

Levels of CHOL, HDL-C, LDL-C, and TRG/HDL ratio were not consistently altered with any drug treatment as shown in Table 4. In contrast, TRG concentrations were reduced with EQW/DAPA, which was statistically superior PHEN/TPM therapy (P < .05; see Table 2) after 24 weeks of treatment.

SBP and DBP both were significantly decreased by all treatments after 24 weeks (P < 035; see Table 3).

Adverse Events. No serious AEs were reported during the trial. As expected, gastrointestinal events with mild to moderate severity were the most common side effects reported for all treatment groups with nausea most common in the EQW group and stomach issues in the DAPA/MET group. As seen in Table 4, the most common AEs with EQW and EQW/DAPA were injection site reactions. In agreement with previous trials with DAPA, the most common reported AEs reported were urinary and genital tract infections; we found that the rates of clinically diagnosed infections were higher in all participants receiving DAPA (DAPA, EQW/DAPA, and DAPA/MET). Infections were generally mild to moderate and clinically manageable. The highest loss rate was from the PHEN/TPM group because of unpleasant side effects. Four participants (2 participants on EQW and 2 on PHEN/TPM) exited the trial prior to completion because they became pregnant.