Gasper Letnar, Klaus Kaae Andersen, Tom Skyhøj Olsen
POCs were not associated with increased stroke risk. Relative to CHCs with ethinylestradiol, POCs were linked to approximately half the risk of IS. However, the absolute number of strokes avoided by choosing POCs over CHCs with ethinylestradiol is small and, for most women-particularly those under 40 years-unlikely to be clinically detectable.
OBJECTIVE: We studied the risk of ischemic (IS) and hemorrhagic stroke (ICH) in users of progestogen-only contraceptives (POCs): levonorgestrel-releasing intrauterine device (LNG-IUD), progestogen-only pill (POP), implant and injection.
STUDY DESIGN: In this nationwide cohort study using Danish registry data for all women aged 18-49 years, 2004-2021, we estimated incidence rate ratios (IRRs) for IS and ICH associated with POC use. Comparisons were made with non-use of hormonal contraception and with combined hormonal contraceptives (CHC) with ethinylestradiol containing equivalent progestogens. Poisson regression models were adjusted for age, ethnicity, education, calendar year, and risk factors.
RESULTS: A total of 1,734,905 non-pregnant women contributed 16,854,953 person-years of follow-up. The most frequently used POCs were LNG-IUD (1,720,931 person-years), followed by desogestrel-only pills (255,404), norethisterone-only pills (104,261), etonogestrel implants (61,756), and medroxyprogesterone injections (27,469). CHCs with ethinylestradiol accounted for 4,431,841 person-years. Compared with non-use of hormonal contraception, no POC type was associated with increased adjusted IRRs for IS or ICH. Compared with CHCs containing desogestrel or levonorgestrel in addition to ethinylestradiol, adjusted IRRs for IS were significantly lower for desogestrel-only pills (IRR 0.56, 95%CI 0.38, 0.82) and LNG-IUD (IRR 0.44, 95%CI 0.36, 0.53). IRRs for ICH were not significantly changed: LNG-IUD 0.85 (95%CI 0.37, 1,92); desogestrel-only pills 0.98 (95%CI 0.44, 2,20).
CONCLUSION: POCs were not associated with increased stroke risk. Relative to CHCs with ethinylestradiol, POCs were linked to approximately half the risk of IS. However, the absolute number of strokes avoided by choosing POCs over CHCs with ethinylestradiol is small and, for most women-particularly those under 40 years-unlikely to be clinically detectable.
IMPLICATIONS: POCs-including POPs, implants, injections, and LNG-IUDs-were not associated with an increased risk of either IS or ICH and support their use as appropriate alternatives to CHCs with ethinylestradiol for women in whom stroke risk is an important clinical consideration.