Long COVID remains poorly characterized at the genomic level. The primary aim of this study was to examine the relationship between viral sequences and the incidence of Long COVID at a tertiary care center in Louisiana between April 2020 and December 2022. A secondary aim was analysis of the Spike protein to identify conserved regions for multivalent vaccine targets.

To estimate Long COVID incidence across variants, we linked 4789 SARS-CoV-2 sequences to 3090 de-identified patient electronic health record information. The base population was defined as any patient with an International Classification of Diseases-10-Clinical Modification COVID-19 diagnosis code (U07.1) based definitions of Long COVID presentation developed by the N3C consortium.

1,554 patients (1,536 Long COVID-negative) met Long COVID definitions, with 56.3% being female, 36.1% self-reported as African American, 5.5% self-reported as Hispanic/Latino, and 54.5% had received at least one vaccine dose 14 days prior to SARS-CoV-2 collection. Long COVID-positive patients were older (mean age 43.1 years) than negative patients (35.9 years; p = 0.0054) and were more likely to be female (p = 0.0001). Among unvaccinated patients, those with Long COVID were significantly younger than their vaccinated counterparts (p < 0.00001). Long COVID incidence varied by PANGO lineage, ranging between 14% in AY.13 to 67.8% in B.1.1.7. Analysis of spike protein diversity revealed eight conserved amino acid regions (Shannon entropy < 0.43), representing potential targets for vaccine design.

Long COVID rates across thousands of annotated SARS-CoV-2 sequences revealed lineage-specific risk and conserved epitopes for future interventions.