Post-acute sequelae of SARS-CoV-2 with clinical condition definitions and comparison in a matched cohort

Abstract

Disease characterization of Post-Acute Sequelae of SARS-CoV-2 (PASC) does not account for pre-existing conditions and time course of incidence. We utilized longitudinal data and matching to a COVID PCR-negative population to discriminate PASC conditions over time within our patient population during 2020. Clinical Classification Software was used to identify PASC condition groupings. Conditions were specified acute and persistent (occurring 0-30 days post COVID PCR and persisted 30-120 days post-test) or late (occurring initially 30-120 days post-test). We matched 3:1 COVID PCR-negative COVIDPCR-positive by age, sex, testing month and service area, controlling for pre-existing conditions up to four years prior; 28,118 PCR-positive to 70,293 PCR-negative patients resulted. We estimated PASC risk from the matched cohort. Risk of any PASC condition was 12% greater for PCR-positive patients in the late period with a significantly higher risk of anosmia, cardiac dysrhythmia, diabetes, genitourinary disorders, malaise, and nonspecific chest pain. Our findings contribute to a more refined PASC definition which can enhance clinical care.

Fig. 1. Unadjusted risk ratios (and 95% confidence intervals) of PASC-related conditions comparing PCR-positive (vs. PCR-negative), in three-time periods anchored on the date of SARS-CoV-2 PCR test result.

A CCS condition risk ratio comparison with a 95% CI plot for PCR-positive population vs PCR-negative population within our study time periods. Risk ratio is the measure of interest comprised of the number of CCS conditions incident in the PCR-positive cohort (n = 28,118) versus CCS conditions incident in the PCR-negative cohort (n = 70,293), with 95% confidence intervals represented by the respective bands. Utilizing 1.0 as the baseline, significant risk ratios (p < 0.05) for the PASC-related conditions can be identified in bold and compared in scale to the other conditions. (*) Asterisk designates that the metric was too large to fit within the scale of the graphic.

Fig. 2. Diagnosis Observation Periods.

Diagnostic observation timeline for CCS conditions in relation to the COVID testing date as the index date. The time periods used in this study were defined as follows: Late: 30–120 days post COVID test date; Acute and persistent 0–30 days post COVID test date and persisted 30–120 days; Pre-existing conditions: four years prior to COVID test date.