(326) Use of the novel antiviral PAXLOVID for COVID-19 during pregnancy: methodological considerations for defining the baseline period across treated and untreated comparator cohorts

Background: PAXLOVID, a combination of nirmatrelvir and ritonavir, was the first FDA-authorized oral COVID-19 antiviral. There are limited data on the safety of PAXLOVID in pregnant women and their offspring; analysis of safety outcomes requires careful definition of the baseline period to balance underlying risk factors across analytic groups. A common approach for determining the baseline period in pregnancy studies is the period prior to the estimated conception date (ECD). However, this approach may not sufficiently characterize cohorts at treatment initiation (for the treated cohorts) or COVID-19 (for the untreated cohort), both of which can occur after ECD.

Objectives: The objective of this study is to describe an approach for defining the baseline risk periods across PAXLOVID treated and untreated comparator cohorts of pregnant women, including the rationale and analytic considerations.

Methods: A baseline period definition was developed for application to an observational cohort study using healthcare claims data from the United States (the Optum Research Database, ORD). Baseline periods were defined for pregnant users of PAXLOVID (Cohort 1), pregnant women with COVID-19 treated with comparator products (Cohort 2), and pregnant women with COVID-19 who were unexposed to treatments for COVID-19 (Cohort 3).

Results: For individuals with treatment for COVID-19 or a first diagnosis of COVID-19 during pregnancy, the baseline period was defined as 6 months prior to ECD plus the time prior to treatment or diagnosis. For example, a patient with either treatment/COVID-19 diagnosis at Day 60 of pregnancy would have 6 months + 60 days as their baseline period. For individuals with eligible treatment or a COVID-19 diagnosis prior to ECD (eg, PAXLOVID initiation 2 days prior to ECD), the baseline period includes 6 months prior to the start of treatment (Cohorts 1 and 2) or COVID-19 diagnosis (Cohort 3). This definition ensures that the baseline period does not include any exposure time after initiation of treatment or COVID-19 diagnosis. There may be variability across individuals in the length of the baseline window, according to the timing of treatment or COVID-19. To address this variability, timing of key covariates relative to ECD will be assessed.

Conclusions: Defining the study baseline period in a pregnancy study where exposure can occur peri-conception is complex, as both the period prior to pregnancy and the period prior to the exposure of interest must be considered. Recognizing these challenges in advance and addressing them through application of relevant baseline covariate windows will aid in the drawing of inferences from the study.