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

Materials and Methods

An open-label, randomized, comparative study in women with PCOS was performed in the Academic Diabetes, Endocrinology and Metabolism research centre at Hull Royal Infirmary. All participants were women, aged between 18 and 45 years, had a body mass index (BMI) ≥25 kg/m2, were diagnosed with PCOS based on the Rotterdam criteria [biochemical hyperandrogenism, as indicated by a free androgen index (FAI) >4, and self-reported oligomenorrhea (cycle length >35 days and 9 or fewer periods per year) or amenorrhoea (absence of menses for a period ≥3 months).[18] Women with differential diagnoses of non-classical 21-hydroxylase deficiency, hyperprolactinaemia, Cushing's disease and androgen-secreting tumours were excluded from participation. Additional exclusion criteria included pregnancy or intention to become pregnant, breastfeeding, documented use of oral hormonal contraceptives and hormone-releasing implants, metformin or other insulin-sensitizing medications, clomiphene citrate or oestrogen modulators, gonadotropin-releasing hormone (GnRH) modulators and Minoxidil, diagnosis of diabetes, history or presence of malignant neoplasms within the last 5 years, pancreatitis (acute or chronic), recurrent urinary tract infections or gastrointestinal tract surgery, ongoing, inadequately controlled thyroid disorder and known hypersensitivity to the investigational medicinal products or any of their excipients. All participants provided their written informed consent. This study was approved by the Medicines and Healthcare Products Regulatory Authority (MHRA) (Ref: 21411/0254/001-0001), the Yorkshire & Humber Health Research Authority and Leeds East Research Ethics Committee (REC reference: 17/YH/0118), registered at www.clinicaltrials.gov (NCT03008551) and conducted in accordance with the Declaration of Helsinki and local regulations.

Women with PCOS were randomized on a 1:1 ratio using an online web-based randomization service (https://www.sealedenvelope.com) to receive either empagliflozin 25 mg (Jardiance) or metformin SR (slow release) 1500 mg (Bolamyn) daily for 12 weeks. The dosage of empagliflozin (25 mg) was chosen to get the maximum metabolic response with comparable duration to metformin treatment group. Metformin group received Metformin SR 1500 mg, which is the standard dose commonly used in patients with PCOS in clinical practice.[19] Metformin SR was preferred over immediate release metformin in view of better gastrointestinal tolerability.[20] All participants were advised to maintain their usual dietary and lifestyle habits during the study.

Participants attended three visits (visits 1-3). During Visit 1, participants were screened against inclusion and exclusion criteria by medical history and clinical examination, routine blood tests (ie full blood count, liver function tests, urea and electrolytes, clotting screen and a pregnancy test), urine pregnancy test and anthropometric measurements. During Visit 2 (baseline) and Visit 3 (12-week follow-up), participants underwent anthropometric (weight, BMI, waist circumference [WC] and hip circumference [HC]) and body composition assessments and an endothelial function measurement. Blood samples were collected at these time points and analysed for reproductive hormones and cardio-metabolic parameters (fasting glucose, fasting insulin, HOMA-IR, total cholesterol, LDL-C, HDL-C, triglycerides [TG], and hs-CRP).


Height and weight were recorded with participants wearing light clothing and no shoes using a weighing scale with attached stadiometer (MS-4202L; Marsden Weighing Machine Group Limited, Rotherham, UK), and BMI was calculated as weight (kg) divided by the square of height (m2). Blood pressure readings were recorded using an automated device (NPB-3900; Nellcor Puritan Bennett, Pleasanton, CA). Three readings were taken at least two minutes apart, and then the average of the readings was obtained. Waist circumference and hip circumference were measured using a tape measure by wrapping it around the patient's waist at the midway point between the top of iliac crest and the bottom of the ribs. Basal metabolic rate, total body fat percentage, fat mass, fat free mass, total body impedance and total body water were measured by using a body composition analyser (BC 418 MA; Tanita Corporation Itabashi-ku, Tokyo, Japan). All these measurements were performed at baseline and 12 weeks after the empagliflozin and metformin treatments.

Endothelial function was assessed using a plethysmographic device Endo-PAT 2000 (Itamar Medical Ltd, Caesarea, Israel). Participants relaxed for at least 15 minutes in a quiet, controlled temperature (22-24°C) room. Endo-PAT biosensors were placed on the index fingers of both hands. During the measurement, participants were instructed to relax and refrain from talking or making any sudden movements. The probes were inflated, and the signals were recorded on the computer according to manufacturer's instructions. This measurement consisted of 5 minutes of baseline recording, followed by blood pressure cuff inflation to a supra-systolic level (at least 60 mm Hg above systolic pressure and no less than 200 mm Hg) sustained for 5 minutes and subsequent blood pressure cuff deflation and recording of Endo-PAT readings over a further 5-minute period. Output variables, namely Reactive Hyperaemia Index (RHI), a measure for endothelial function, and Augmentation Index (AI), a measure for arterial stiffness, were assessed using an automated computer software (EndoPAT2000 version 3.3.2; Itamar Medical Ltd). Compliance with the treatments was calculated by counting the returned medications at the end of the 12-week period.

Blood Sampling and Biochemical Analyses

Following an overnight fast, blood samples were collected both at the baseline and final visit (end of the 12-week period). The fasting venous blood was collected into fluoride oxalate and serum gel tubes. Blood samples were separated by centrifugation at 3500 g for 15 minutes at 5°C, and the aliquots were stored at −80°C within 1 hour of collection.

Serum insulin was assayed using a competitive chemiluminescent immunoassay performed on the manufacturer's DPC Immulite 2000 analyzer (Euro/DPC, Llanberis, UK), with a coefficient of variation (CV) was 6%, and no stated cross-reactivity with proinsulin. The plasma glucose was measured using a Beckman AU 5800 analyser (Beckman-Coulter, High Wycombe, UK) and according to the manufacturer's recommended protocol. Total cholesterol, high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), alanine aminotransferase (ALT), aspartate aminotransferase (AST) and high-sensitivity C-reactive protein (hs-CRP) levels were measured enzymatically using a Beckman AU 5800 analyser (Beckman-Coulter, High Wycombe, UK) with CVs of <4.9%, 0.9%, 1.6% and 8.4%. Low-density lipoprotein cholesterol (LDL-C) was calculated using the Friedewald equation. Serum testosterone and androstenedione were quantified using isotope-dilution liquid chromatography tandem mass spectrometry (LC-MS/MS). Sex hormone-binding globulin (SHBG), oestrogen and dehydroepiandrosterone (DHEAS) were measured using a chemiluminescent immunoassay on the Beckman-Coulter UniCel DxI 800 analyser, applying the manufacturer's recommended protocol. The free androgen index (FAI) was calculated as: (total testosterone/SHBG) × 100.

Statistical Analysis

There are no previous studies on the effect of empaglifozin or any other SGLT-2 inhibitors on women with PCOS. Sample size was based on an independent t test with an arbitrary level of 5% significance (2-tailed) and power of 80%. Assuming a common standard deviation of 16% for radial augmentation index and a 4% reduction as a significant change,[21] a sample size of 16 patients per group allowed us to detect a between-group mean difference of 4%. To allow for loss-to-follow-up, we aimed to recruit 20 patients per group.

All data were checked for normality according to the Shapiro-Wilk test. Mean differences for all parameters expressed as % change from baseline between women with PCOS in the empagliflozin group and the metformin group were analysed with independent t test or Mann-Whitney U-test for normally and non-normally distributed data, respectively. Mean differences between baseline and 12-week follow-up within each treatment group were analysed with a paired t test or a signed-rank test for normally and non-normally distributed data, respectively. Values are presented as mean SD, if the variables were normally distributed, or median and interquartile range, if the variables were skewed. All statistical analyses were performed using IBM-SPSS version 24.0 (Chicago, IL) with P-values ≤0.05 considered to be significant. Linear regression analysis using function ln R was used to confirm the findings of univariate analysis after adjustments for confounders. Since this was a randomized controlled trial, we only adjusted for age and age + BMI, where relevant.