Effects of oral contraceptives on metabolic parameters in adult premenopausal women: a meta-analysis

Abstract

Objective To estimate the effect of oral contraceptives (OC) containing different progestins on parameters of lipid and carbohydrate metabolism through a systematic review and meta-analysis. Patients and methods Premenopausal women aged 18 or older, who received oral contraceptives containing chlormadinone, cyproterone, drospirenone, levonorgestrel, desogestrel, dienogest, gestodene or norgestimate, for at least 3 months. Outcome variables were changes in plasma lipids, BMI, insulin resistance and plasma glucose. We searched MEDLINE and EMBASE for randomized trials and estimated the pooled within-group change in each outcome variable using a random-effects model. We performed subgroup analyses by study duration (<12 months vs ≥12 months) and polycystic ovary syndrome (PCOS) status. Results Eighty-two clinical trials fulfilled the inclusion criteria. All progestins (except dienogest) increased plasma TG, ranging from 12.1 mg/dL for levonorgestrel (P < 0.001) to 35.1 mg/dL for chlormadinone (P < 0.001). Most progestins also increased HDLc, with the largest effect observed for chlormadinone (+9.6 mg/dL, P < 0.001) and drospirenone (+7.4 mg/dL, P < 0.001). Meanwhile, levonorgestrel decreased HDLc by 4.4 mg/dL (P < 0.001). Levonorgestrel (+6.8 mg/dL, P < 0.001) and norgestimate (+11.5 mg/dL, P = 0.003) increased LDLc, while dienogest decreased it (–7.7 mg/dL, P = 0.04). Cyproterone slightly reduced plasma glucose. None of the progestins affected BMI or HOMA-IR. Similar results were observed in subgroups defined by PCOS or study duration. Conclusion Most progestins increase both TG and HDLc, their effect on LDLc varies widely. OC have minor or no effects on BMI, HOMA-IR and glycemia. The antiandrogen progestins dienogest and cyproterone displayed the most favorable metabolic profile, while levonorgestrel displayed the least favorable.