Available medications used as potential therapeutics for COVID-19: What are the known safety profiles in pregnancy

Background Medications already available to treat other conditions are presently being studied in clinical trials as potential treatments for COVID-19. Given that pregnant women are excluded from these trials, we aimed to investigate their safety when used during pregnancy within a unique population source. Methods Using the population-based Quebec Pregnancy Cohort, we identified women who delivered a singleton liveborn (1998–2015). Taking potential confounders into account including indications for use, the risk of prematurity, low birth weight (LBW), small for gestational age (SGA), and major congenital malformation (MCM) associated with COVID-19 repurposed drug use during pregnancy were quantified using generalized estimation equations. Results Of the 231,075 eligible pregnancies, 107 were exposed to dexamethasone (0.05%), 31 to interferons (0.01%), 1,398 to heparins (0.60%), 24 to angiotensin-receptor blockers (ARB) (0.01%), 182 to chloroquine (0.08%), 103 to hydroxychloroquine (0.05%), 6,206 to azithromycin (2.70%), 230 to oseltamivir (0.10%), and 114 to HIV medications (0.05%). Adjusting for potential confounders, we observed an increased risk of prematurity related to dexamethasone (aOR 1.92, 95%CI 1.11–3.33; 15 exposed cases), anti-thrombotics (aOR 1.58, 95%CI 1.31–1.91; 177 exposed cases), and HIV medications (aOR 2.04, 95%CI 1.01–4.11; 20 exposed cases) use. An increased risk for LBW associated with anti-thrombotics (aOR 1.72, 95%CI 1.41–2.11; 152 exposed cases), and HIV medications (aOR 2.48, 95%CI 1.25–4.90; 21 exposed cases) use were also found. Gestational exposure to anti-thrombotics (aOR 1.20, 95%CI 1.00–1.44; 176 exposed cases), and HIV medications (aOR 2.61, 95%CI 1.51–4.51; 30 exposed cases) were associated with SGA. First-trimester dexamethasone (aOR 1.66, 95%CI 1.02–2.69; 20 exposed cases) and azithromycin (aOR 1.10, 95%CI 1.02–1.19; 747 exposed cases) exposures were associated with MCM. Conclusions Many available medications considered as treatments for COVID-19 are associated with adverse pregnancy outcomes. Caution is warranted when considering these medications during the gestational period.


Introduction
Given the changes to the cardiopulmonary and immune systems during pregnancy, pregnant women are at increased risk for severe COVID-19 [1][2][3]. In June 2020, the US Centers for Disease Control and Prevention reported that pregnant women with COVID-19 were more likely to be hospitalized and at increased risk for intensive care unit (ICU) admission and receipt of mechanical ventilation compared with non-pregnant women of reproductive age [4]. The World Association of Perinatal Medicine Working Group on COVID-19, with data from Europe, the US, South America, Asia and Australia reported a 0.8% rate of mortality and a 11.1% rate of ICU admissions among infected pregnant women; vertical transmission was negligible [5]. Although vaccination is underway in many countries, pregnant women are often advised against it due to missing data on safety and efficacy during pregnancy. Medications already available to treat other conditions are presently being studied in clinical trials as potential treatments for COVID-19. However, given that pregnant women are excluded from these trials, it is important to assess the current safety profile of these drugs in pregnancy. According measures from each patients' charts within the QPC [67]. Prospective follow-up is available from 1 year before LMP, during pregnancy, and until 31 /12/2015 (S2 Fig).
The QPC data sources include the medical claims database ('Régie de l'assurance maladie du Québec' (RAMQ): diagnoses, medical procedures, socio-economic status), Quebec's outpatient prescription drug insurance database (drug name, start date, dosage, duration), hospitalization archives database (MedEcho: in-hospital diagnoses and procedures, gestational age), and the Quebec birth certificates database ('Institut de la statistique du Québec' (ISQ): patient socio-demographics, gestational age, birth weight). Birth weight in ISQ, and MCM and other diagnoses in the RAMQ and MedEcho databases have been found to be valid when compared to patient charts [67,68].
Pregnant women in the QPC were eligible for this study if they were i) more than 18 years old; ii) continuously covered by the Quebec prescription drug insurance for �12 months before pregnancy and during pregnancy; and iii) had given birth to a liveborn singleton. This was done because twin pregnancies are at increased risk of adverse pregnancy outcomes regardless of gestational medication exposures. We also excluded pregnancies exposed to known teratogens as described by Kulaga et al. [69] (S1 Table), and those resulting in minor malformations alone or chromosomal abnormalities in the newborns for analyses on MCM. Minor malformations are selectively identified and do not reflect the true prevalence; chromosomal abnormalities are not related to medication use.
Ethics statement. The study was approved by the Sainte-Justine's Hospital Ethics Committee. The Quebec "Commission d'accès à l'information" authorized database linkages. All data were fully anonymized before we accessed them, and the Ethics Committee of CHU Sainte-Justine as well as the 'Commission d'accès à l'information' waived the requirement for informed consent.

Study medication exposures
Study medications included outpatient filled prescriptions of immunomodulator dexamethasone, interferon for multiple sclerosis (beta-1a, beta-1b, and alfa-2b), antithrombotic heparin and heparin derivatives (enoxaparin, dalteparin, and tinzaparin), ARB (losartan and telmisartan), chloroquine, hydroxychloroquine, azithromycin, oseltamivir, and HIV medications (indinavir, lopinavir/ritonavir, saquinavir and raltegravir). We identified study medication prescription fillings from the Quebec prescription drug insurance database (prescribed overthe-counter medications were also included), using timing of exposure determined by the dispensed date and duration of treatment. Pregnancies were dichotomously defined as exposed within each of the study medication groupings if women had filled at least one study medication during pregnancy or if they had filled a prescription with a duration that overlapped the beginning of pregnancy (yes/no). Pregnant women could use more than one study medications during the gestational period, and thus were considered in each corresponding study medication grouping when that was the case. The exposure time window for analyses on prematurity, LBW and SGA was any time during pregnancy; only first trimester exposure (organogenesis) was considered for analyses on MCM.

Outcomes
Cases of prematurity were identified from the RAMQ and MedEcho databases and defined as deliveries before the 37 th week of gestation.
Cases of LBW were identified from the ISQ database as newborns with birthweight less than 2,500g.
Cases of SGA were identified from the MedEcho database (gestational age) and the ISQ database (birth weight and sex) and were defined as birthweights below the 10 th percentile for newborns of the same gestational age and the same sex, according to population-based Canadian references [71]. Birth weight in ISQ and gestational age in MedEcho have been found to be valid when compared to patient charts [67,68].
Cases of MCM diagnosed in the first 12 months of life were identified from the RAMQ and MedEcho databases and defined according to ICD-9 and ICD-10 codes (S2 Table), which have been validated against patient charts with high PPV (78.1%) and NPV (94.2%) [68]. All organ systems were considered and high PPV (over 80%) have also been reported for specific MCMs [68], Twelve months after birth was needed to allow for late detection, and validation of early diagnoses.

Statistical analyses
Within the identified study cohort, we conducted 4 case-control analyses to quantify the effect of the study medication exposures during pregnancy on the occurrence of prematurity, LBW, SGA, and MCM. Although case-control analyses were performed within the study cohort, we have included all controls (no control sampling has been done), and therefore, odds ratios (OR) give the same estimate measure as relative risks.
Potential confounders considered for all analyses were: 1) sociodemographic variables on LMP including maternal age, welfare recipients (yes/no), area of residence (urban/rural); 2) maternal chronic comorbidities (in the 12 months before pregnancy and during pregnancy identified by a diagnosis code or a medication-specific filling) including diabetes, asthma, thyroid disorders (see S3 Table for diagnostic and medication codes used); 3) Tobacco, alcohol, and illicit drug use (See S3 Table); 4) Health care utilization including hospitalizations or emergency department visits during pregnancy (yes/no), number of general practitioner visits and specialist visits (12 months pre-pregnancy); 5) Pregnancy related variables including folic acid use (prescribed high dose (>5 mg/d) and prescribed over-the-counter (OTC) dosage only) in the 6-months prior to LMP and during pregnancy (S3 Table), and previous pregnancy (spontaneous or planned abortion, delivery) in the year prior to LMP (yes/no). We also considered whether pregnant women were followed by an obstetrician (yes/no), and if other medications were used during pregnancy (besides the study medications and medication used to identify comorbidities).
Finally, to control for potential confounding by indication, we adjusted for the presence of the following indications during pregnancy (a pregnant women could have multiple comor- All study medication groupings were always included in analyses, which ensured that estimates were adjusted for concomitant study medication use. The unit of analysis was a pregnancy. Means and proportions for continuous and dichotomous variables were calculated, respectively. Crude and adjusted odds ratios (aOR) with 95% confidence intervals (95%CI) were calculated for each outcome separately. Multivariable generalized estimating equations were used to estimate the association between the study medications and the risk of prematurity, LBW, SGA, and MCM, independently, accounting for clustering by family (mother). All above mentioned potential confounders and covariables were included in all analyses. All statistical analyses were performed using SAS (SAS Institute Inc., Version 9.2, Cary, NC, USA).
Study medication users were slightly older; welfare recipients (33.3% vs. 22.5% in nonusers); more likely to use high dose (>5 mg/d) folic acid; more likely to have hypertension, diabetes or asthma; and had a higher prevalence of health services utilization including other medication use (Table 1).
Overall MCM were identified in 10,4% (23,991) of pregnancies. Adjusting for potential confounders, dexamethasone (aOR 1.66, 95%CI 1.02-2.69; 20 exposed cases) and azithromycin (aOR 1.10, 95%CI 1.02-1.19; 747 exposed cases) use during pregnancy were statistically significantly associated with an increased risk of MCM (Table 2D). Table 3 presents organ specific defects identified with the use of the study medications. Musculoskeletal defects and circulatory malformations including heart defects were the most

PLOS ONE
Available medications used as therapeutics for COVID-19: What are the known safety profiles in pregnancy

PLOS ONE
Available medications used as therapeutics for COVID-19: What are the known safety profiles in pregnancy   Note: The total number of pregnancies exposed to at least one study medications during pregnancy is equal to 8,213. 1 Heparin, dalteparin, enoxaparin and tinzaparin. 2 Beta-1A, beta-1B, and alfa-2B interferons. 3 Indinavir, lopinavir/ritonavir, raltegravir and saquinavir. 4 Angiotensin-receptor blockers (ARB): Losartan, losartan/HCTZ, telmisartan, telmisartan/HCTZ. 5 Urinary tract infection. 6 Human immunodeficiency viruses. 7 Low birth weight is defined as birthweight <2500 grams. 8 Small for gestational age is defined as birthweight below the 10th percentile for newborns of the same gestational age and same sex. https://doi.org/10.1371/journal.pone.0251746.t001 Table 2. A. Association between study medication exposures during pregnancy and the risk of prematurity. B. Association between study medication exposures during pregnancy and the risk of low birth weight (LBW) (birthweigh <2500 grams). C. Association between study medication exposures during pregnancy and the risk of being born small for gestational age (SGA) (birthweight below the 10th percentile for newborns of the same gestational age and same sex). D. Association between study medication exposures during the first trimester of pregnancy and the risk of overall major congenital malformation.      prevalent in each of the study medication groups. No orofacial defects were found in newborns exposed to dexamethasone in-utero.

Discussion
Using the population-based Quebec Pregnancy Cohort, we quantified the risk of adverse perinatal outcomes associated with available medications presently considered as COVID-19 treatments. Indeed, after adjusting for potential confounders including current indications for use, and concomitant COVID-19 potential therapeutic use, we found that anti-thrombotics, mostly heparins, and HIV medication use during pregnancy were associated with the risk of prematurity, LBW and SGA. Dexamethasone was associated with increasing risks of prematurity and MCM; and azithromycin was associated with the risk of MCM. This study adjusted for all known and measurable potential confounding variables and estimates are comparable given that they emerge from the same source population, health insurance coverage, and access to care.
Our results on dexamethasone are consistent with the literature with regards to prematurity [23,24]. Palmsten et al. [24] showed that oral corticosteroid use during pregnancy was associated with a doubling of the risk of preterm birth in women with rheumatoid arthritis recruited within teratogen information services (MotherToBaby). Early pregnancy corticosteroid use has also been associated with increased risk of MCM [15][16][17][18][19][20] similar to what we have shown. Note: The total number of pregnancies exposed to at least one study medications during pregnancy is equal to 8,213. 1 Adjusted for all variables included in the model. 2 Heparin, dalteparin, enoxaparin and tinzaparin. 3 Beta-1A, beta-1B, and alfa-2B interferons. 4 Human immunodeficiency virus medications (Indinavir, lopinavir/ritonavir, raltegravir and saquinavir) 5 Angiotensin-receptor blockers (ARB): Losartan, losartan/HCTZ, telmisartan, telmisartan/HCTZ.

Study medication exposures during the first trimester of pregnancy, n = 8,213
Major congenital malformation by organ system Non exposed to any of the study drugs Chloroquine Hydroxychloroquine Dexamethasone Anti- Heparin, dalteparin, enoxaparin and tinzaparin. 2 Beta-1A, beta-1B, and alfa-2B interferons. 3 Indinavir, lopinavir/ritonavir, raltegravir and saquinavir. 4 Angiotensin-receptor blockers (ARB): Losartan, losartan/HCTZ, telmisartan, telmisartan/HCTZ. https://doi.org/10.1371/journal.pone.0251746.t003 We found no orofacial defect with dexamethasone use as was reported in more recent pregnancy studies [21,22]. Our findings on anti-thrombotics (mostly heparins) use and pregnancy are different from what has been published recently. The increased prevalence of adverse fetal/infant outcomes including death, prematurity and MCM have been reported following heparin use [72]. However, in another more recent study performed by Shlomol et al. [28], no such associations were found within an Israeli cohort of pregnant women. While heparin does not appear to cross the placenta, it may affect embryo and fetal development through interactions with the trophoblast and placental vasculature [73]. Differences between our findings and those from Shlomol et al. [28] could be partly explained by their lack of adjustment for potential confounders such as gestational hypertension and diabetes, indications for heparin use, and lifestyles such as tobacco and alcohol use.
Our findings on the association between indinavir, lopinavir/ritonavir, raltegravir and saquinavir (HIV drugs) use during pregnancy and the risk of prematurity, LBW and SGA [51]; and on chloroquine, and hydroxychloroquine with regards to prematurity, LBW or MCM are also consistent with the literature [46].
We found that azithromycin use was increasing the risk of MCM. A recent populationbased cohort study using data from the Clinical Practice Research Datalink in the United Kingdom has shown that use of macrolide antibiotics, including erythromycin, clarithromycin, or azithromycin, during pregnancy was associated with an increased risk of overall major congenital malformations in children [74]. Similarly, a population based cohort study using data from the Swedish Medical Birth Register has shown an association between early pregnancy erythromycin use and infant cardiovascular defects [75].

Strengths and potential limitations
Study strengths include the use of population-based prospective pregnancy cohort with linkage of data at the individual level, which minimized selection and recall biases; this also allowed for analyses on a large number of pregnancies with detailed information regarding exposure, outcomes, and potential confounders. The fact that all potential available medications for COVID-19 treatments were studied within a unique and single population allowed for comparative safety assessments. QPC data on filled prescriptions [70], gestational age [67], birth weight [67], and MCM [67] have all been validated. Adjustment on all known and measurable potential confounders for adverse pregnancy outcomes was made, including maternal comorbidities, indications for medication uses, lifestyles including smoking, alcohol, illicit drug use, and high dose folic acid intake; adjustment was also made on health services utilization, which is considered a proxy for severity of diseases.
One potential limitation is missing information on over-the-counter (OTC) medication use, and use of medications during in-hospital deliveries. Other than for ibuprofen and acetaminophen use and non-prescribed folic acid use, all other medications will be prescribed. It is possible that some women took folic acid OTC. However, this would lead to non-differential misclassification as it is unlikely that those exposed to the study medications would differ in terms of prevalence of folic acid OTC compared to those who were not exposed. Since the databases only include pregnant women insured by the Prescription Drug Insurance program, generalizability of results to those insured by private drug insurance could be affected. However, validation studies have shown that publicly insured pregnant women have similar characteristics and co-morbidities than those who have private medication insurance [76]. We considered filled prescriptions and not actual intake, but Zhao et al. [70] have shown that prescription filling data in the QPC were valid when compared to maternal report. Although health services utilization was adjusted for and considered a proxy for disease severity, residual confounding by severity of disease could remain. Our estimates could be slightly biased upwards since we only considered deliveries in our analyses as is done in the majority of studies on medications and pregnancy. Finally, the MCM prevalence of 10.3% is higher than what is routinely reported (3-5%) [77]. This could be partly explained by the Founders' effect in the province of Quebec. [78,79]. It can also be partly explained by the fact that we have included all pregnancies between 1998 and 2015, and we have required that all pregnant women and children be insured by the RAMQ public medication insurance program in order to fully measure medication exposures during pregnancy (we only have medication data on those insured by the RAMQ public medication insurance program). This, in addition to the Founders' effect, could explain the MCM prevalence. Although our baseline prevalence of MCMs is high, it does not differ among our compared groups, and therefore does not invalidate our findings. This, however, could limit the generalizability of our results.

Conclusions
Using the population-based Quebec Pregnancy Cohort, gestational exposure to dexamethasone was associated with an increased risk of prematurity and MCM; azithromycin exposure was associated with the risk of MCM, and exposure to anti-thrombotics (mostly heparins), and indinavir, lopinavir/ritonavir, raltegravir and saquinavir (HIV drugs) use during pregnancy were associated with increased risks of prematurity, LBW and SGA. Although these available medications are being considered as treatments for COVID-19, caution is warranted in pregnancy.