Impact of Age and Sex Interaction on Post-Acute Sequelae of COVID-19: An Italian Cohort Study on Adults and Children

Abstract

Identifying factors predisposing individuals to post-acute sequelae of COVID-19 (PASC) would allow for the timely treatment of those vulnerable. Attention on the role of sex and age is growing, but published studies have shown mixed results. Our objective was to estimate the effect modification of age on sex as a risk factor for PASC. We analyzed data from two longitudinal prospective cohort studies on adult and pediatric subjects positive to SARS-CoV-2 infection that were enrolled between May 2021 and September 2022. Age classes (≤5, 6-11, 12-50, >50 years) were based on the potential role of sex hormones on inflammatory/immune and autoimmune processes. A total of 452 adults and 925 children were analyzed: 46% were female and 42% were adults. After a median follow-up of 7.8 months (IQR: 5.0 to 9.0), 62% of children and 85% of adults reported at least one symptom. Sex and age alone were not significantly associated to PASC, but their interaction was statistically significant (p-value = 0.024): the risk was higher for males aged 0-5 (females vs. males HR: 0.64, 95% CI: 0.45-0.91, p = 0.012) and for females aged 12-50 (HR: 1.39, 95% CI: 1.04-1.86, p = 0.025), especially those in the cardiovascular, neurological, gastrointestinal and sleep categories. Further research on PASC with regard to sex and age is warranted.

Keywords: COVID-19; age; post-acute sequelae; sex hormones.

Figure 1

Flow-chart diagram of the cohorts.

Figure 2

Kaplan-Meier estimates of cumulative reported frequency of post-COVID-19 symptoms, by sex, (A) in subjects 0–5 years of age, (B) 6–11 years of age, (C) 12–50 years of age, and (D) >50 years of age. The numbers in parentheses represent the number of symptoms that were reported within each time interval, by sex. The estimate of the hazard ratio (HR) was based on a Cox proportional hazards regression model, adjusted for age.

Figure 3

Forest plot of the risk estimate (hazard ratio, HR) of post-COVID-19 symptoms for females vs. males in each symptom category, stratified by age group. HRs are adjusted for age. Abbreviations: p-inter., p-value for the sex × age interaction term in the Cox model. Notes: HR for sensory symptoms in the 0–5 year group was not estimable because of a lack of events; data are from all of the surveys collected.

Figure 4

Subpopulation Treatment Effect Pattern Plot (STEPP) of the 3-month cumulative incidence rate of the reported symptoms in males (solid black line) and females (dotted grey line). The plot was drawn adopting the sliding window pattern, including n = 100 subjects in each subpopulation and n = 75 subjects in common among consecutive subpopulations, implementing 2500 permutations of the covariate age.