Shaalini Ramanadhan, Katie Hansen, Jillian T Henderson, Megan A Cohen, Robin Paynter, Alison Edelman

The novel coronavirus disease (COVID-19) has led to significant mortality and morbidity, including a high incidence of related thrombotic events. There has been concern regarding hormonal contraception use during the COVID-19 pandemic, as this is an independent risk factor for thrombosis, particularly with estrogen-containing formulations. However, higher estrogen levels may be protective against severe COVID-19 disease. Evidence for risks of hormonal contraception use during the COVID-19 pandemic is sparse. We conducted a living systematic review that will be updated as new data emerge on the risk of thromboembolism with hormonal contraception use in patients with COVID-19. To determine if use of hormonal contraception increases risk of venous and arterial thromboembolism in women with COVID-19. To determine if use of hormonal contraception increases other markers of COVID-19 severity including hospitalization in the intensive care unit, acute respiratory distress syndrome, intubation, and mortality. A secondary objective is to maintain the currency of the evidence, using a living systematic review approach. ⁠⁠⁠⁠⁠⁠We searched CENTRAL, MEDLINE, Embase, CINAHL, Global Index Medicus, Global Health, and Scopus from inception on March 2023, and monitored the literature monthly. We updated the search strategies with new terms and added the database Global Index Medicus in lieu of LILACS. We included all published and ongoing studies of patients with COVID-19 comparing outcomes of those on hormonal contraception versus those not on hormonal contraception. This included case series and non-randomized studies of interventions (NRSI). One review author extracted study data and this was checked by a second author. Two authors individually assessed risk of bias for the comparative studies using the ROBINS-I tool and a third helped reconcile differences. For the living systematic review, we will publish updates to our synthesis every six months. In the event that we identify a study with a more rigorous study design than the current included evidence prior to the planned six-month update, we will expedite the synthesis publication. We included three comparative NRSIs with 314,704 participants total and two case series describing 13 patients. The three NRSIs had serious to critical risk of bias in several domains and low study quality. Only one NRSI ascertained current use of contraceptives based on patient report; the other two used diagnostic codes within medical records to assess hormonal contraception use, but did not confirm current use nor indication for use. None of the NRSIs included thromboembolism as an outcome. Studies were not similar enough in terms of their outcomes, interventions, and study populations to combine with meta-analyses. We therefore narratively synthesized all included studies. Based on results from one NRSI, there may be little to no effect of combined hormonal contraception use on odds of mortality for COVID-19 positive patients (OR 1.00, 95% CI 0.41 to 2.40; 1 study, 18,892 participants; very low-certainty evidence). Two NRSIs examined hospitalization rates for hormonal contraception users versus non-users. Based on results from one NRSI, the odds of hospitalization for COVID-19 positive combined hormonal contraception users may be slightly decreased compared with non-users for patients with BMI under 35 kg/m2 (OR 0.79, 95% CI 0.64 to 0.97; 1 study, 295,689 participants; very low-certainty evidence). According to results of the other NRSI assessing use of any type of hormonal contraception, there may be little to no effect on hospitalization rates for COVID-19 positive individuals (OR 0.99, 95% CI 0.68 to 1.44; 1 study, 123 participants; very low-certainty evidence). We included two case series because no comparative studies directly assessed thromboembolism as an outcome. In a case series of six pediatric COVID-19 positive patients with pulmonary embolism, one (older than 15 years of age) was using combined hormonal contraception. In a second case series of seven COVID-19 positive patients with cerebral venous thrombosis, one was using oral contraceptives. One comparative study and one case series reported on intubation rates, but the evidence for both is very uncertain. In the comparative study of 123 COVID-19 positive patients (N = 44 using hormonal contraception and N = 79 not using hormonal contraception), no patients in either group required intubation. In the case series of seven individuals with cerebral venous thromboembolism, one oral contraceptive user and one non-user required intubation. There are no comparative studies assessing risk of thromboembolism in COVID-19 patients who use hormonal contraception, which was the primary objective of this review. Very little evidence exists examining the risk of increased COVID-19 disease severity for combined hormonal contraception users compared to non-users of hormonal contraception, and the evidence that does exist is of very low certainty. The odds of hospitalization for COVID-19 positive users of combined hormonal contraceptives may be slightly decreased compared with those of hormonal contraceptive non-users, but the evidence is very uncertain as this is based on one study restricted to patients with BMI under 35 kg/m2. There may be little to no effect of combined hormonal contraception use on odds of intubation or mortality among COVID-19 positive patients, and little to no effect of using any type of hormonal contraception on odds of hospitalization and intubation for COVID-19 patients. We noted no large effect for risk of increased COVID-19 disease severity among hormonal contraception users. We specifically noted gaps in pertinent data collection regarding hormonal contraception use such as formulation, hormone doses, and duration or timing of contraceptive use. Differing estrogens may have different thrombogenic potential given differing potency, so it would be important to know if a formulation contained, for example, ethinyl estradiol versus estradiol valerate. Additionally, we downgraded several studies for risk of bias because information on the timing of contraceptive use relative to COVID-19 infection and method adherence were not ascertained. No studies reported indication for hormonal contraceptive use, which is important as individuals who use hormonal management for medical conditions like heavy menstrual bleeding might have different risk profiles compared to individuals using hormones for contraception. Future studies should focus on including pertinent confounders like age, obesity, history of prior venous thromboembolism, risk factors for venous thromboembolism, and recent pregnancy.