Effects of Empagliflozin on Metabolic Parameters in Polycystic Ovary Syndrome

A Randomized Controlled Study

Zeeshan Javed; Maria Papageorgiou; Harshal Deshmukh; Alan S. Rigby; Unaiza Qamar; Jehangir Abbas; Amer Y. Khan; Eric S. Kilpatrick; Stephen L. Atkin; Thozhukat Sathyapalan


Clin Endocrinol. 2019;90(6):805-813. 

In This Article


Forty-two participants with PCOS were screened; two participants were excluded from participation because they did not meet inclusion criteria, as such 40 participants were randomized to the two treatment arms (empagliflozin, n = 20 and metformin, n = 20; Figure 1). During the intervention phase of the study, one participant in the empagliflozin group was lost to follow-up. Finally, 19 participants in the empagliflozin group (age: 26.0 [8.0] years, BMI: 37.1 ± 6.2 kg/m2) and 20 participants in the metformin group (age: 31.5 [20.0] years, BMI: 38.7 ± 7.8 kg/m2) completed the trial, and their data were included in the final analysis (Figure 1). The baseline characteristics of both groups are presented in Table 1 and Table S1. Compliance was over 90% in both groups. There were no adverse events or serious adverse events in the metformin group. In the empagliflozin group, two patients reported adverse events (headache and dizziness, n = 1; mild rash, n = 1), which were, however, unrelated to the study drug.

Figure 1.

Flow diagram of the study

Anthropometric and Body Composition Parameters

In the empagliflozin group, waist circumference (P = 0.024), hip circumference (P = 0.013), BMR (P = 0.016), fat free mass (FFM) (P = 0.013) and total body water (P = 0.014) at 12 weeks decreased significantly compared to baseline, but no changes were seen in total mass (P = 0.079) or BMI (P = 0.069; Table 1). In the metformin group, body mass (P = 0.019), BMI (P = 0.024) and hip circumference (P = 0.031), total fat percentage (P = 0.015) or fat mass (P = 0.005) significantly increased after 12 weeks of treatment (Table 1).

When data were expressed as percentage change from baseline in each group, significant differences in weight (empagliflozin: −1.4 ± 3.2% vs metformin: 1.2 ± 2.3%; P = 0.006), BMI (empagliflozin: −1.4 ± 3.2% vs metformin: 1.1 ± 2.2%; P = 0.007), waist (empagliflozin: −1.6 ± 2.8% vs metformin: 0.2 ± 2.1%; P = 0.029) and hip circumference (empagliflozin: −2.0 ± 3.0% vs metformin: 1.1 ± 1.9%; P = 0.001) were seen between the treatment groups. Similarly, the percentage changes from baseline in BMR (empagliflozin: −1.8 ± 2.9% vs metformin: 0.1 ± 1.9%, P = 0.024) and fat mass (empagliflozin: −0.7 ± 4.9% vs metformin, 3.2 ± 5.0%; P = 0.023) were significantly different between the empagliflozin and metformin groups (Table 1, Figure 2). The proportion of women with PCOS who experienced (a) a decrease >5%, (|b) an increase >5%, or (c) a change ≤5% from baseline in anthropometric and body composition parameters are presented in Table S2.

Figure 2.

Percentage changes from baseline in anthropometric and body composition parameters after 12 wks with empagliflozin and metformin treatment. *P < 0.05; BMI, body mass index; BMR, basic metabolic rate; FFM, fat free mass; HC, hip circumference; TBW, total body water; WC, waist circumference. Body composition data (BMR, body fat %, fat mass, FFM and TBW) are presented for 18 participants in the empagliflozin group with available data

Hormonal and Metabolic Parameters

In the empagliflozin group, significant increases in SHBG (P = 0.049) and oestradiol levels (P = 0.032) were seen after 12 weeks of treatment (Table 1). There were no other hormonal changes for either group (Table 1). No differences were seen in percentage change from baseline for any of the hormone parameters between groups (Table 1, Figure 2).

There were no changes following 12 weeks of treatment in blood pressure, endothelial function (RHI, AI), insulin sensitivity (insulin, fasting glucose, HOMA-IR), fasting lipid profile or hs-CRP in either treatment arm (Table 1). Between groups, comparisons did not reveal any differences in percentage changes from baseline for any of these metabolic parameters (Table 1, Figure 2).

Regression Analysis Modelling

Table 2 shows the results of the linear regression analysis modelling percentage changes from baseline in anthropometric characteristics, hormonal and metabolic parameters as function of metformin or empagliflozin treatment. The results confirm statistically significant reduction in weight, BMI, WC, HC, fat mass and BMR in those randomized to empagliflozin group as compared to those in metformin group.