Natural course of post-COVID symptoms in adults and children

Abstract

More than 200 million COVID-19 survivors have lasting symptoms after recovering, but the duration and related risk factors remain uncertain. This study focused on all 6551 patients diagnosed with COVID-19 at a medical institution in Hiroshima from March 2020 to July 2022. In November 2022, a questionnaire survey was conducted regarding post-COVID symptoms and their duration. The prevalence and duration of post-COVID symptoms were illustrated using the Kaplan-Meier method. Risk factors for symptoms lasting over 3 months and interfering with daily life were assessed via multivariate logistic regression. A total of 2421 survivors responded: 1391 adults, 1030 children, median age 34 years (IQR 9-55), 51·2% male, 36·7% hospitalized, median time from infection to the survey was 295 days (IQR 201-538). Upon their initial recovery, the prevalence of post-COVID symptoms was 78·4% in adults and 34·6% in children. Three months later, the rates were 47·6% and 10·8%. After over one year, they were 31·0% and 6·8%. Regarding symptoms interfere with daily life, 304 people (12.6%) experienced symptoms lasting for over three months, with independent risk factors including age, being female, diabetes mellitus, infection during the Delta period, and current smoking. There was no significant association between vaccination history and post-COVID symptoms.

Figure 1

Prevalence of post-COVID symptoms over time in adults and children. Prevalence of post-COVID symptoms of any level of severity (A) and prevalence of post-COVID symptoms that interfere with daily life (B) were shown at different time points after the initial recovery in both adults and children. The symptoms tracked include any symptoms (black), fatigue (red), cough (gray), difficulty concentrating (orange), sleep disorders (blue), and altered smell (green).

Figure 2

Univariate and multivariate analyses to identify risk factors associated with post-COVID symptoms that interfere with daily life for over three months. The response variable was the presence or absence of post-COVID symptoms that interfere with daily life for over three months. In the univariate analysis, the chi-square test was applied with Bonferroni correction for comparisons involving three or more groups. In the multivariate logistic regression analysis, the stepwise variable selection method was used to identify the relevant explanatory variables from a set including sex, age, the epidemic period, severity of COVID-19 at the time of infection, pre-infection vaccination history, current smoking, current drinking, diabetes mellitus, hypertension, and chronic renal diseases. N = 2421, R² = 0.1267, Model p < 0.0001. The variables excluded by the stepwise method are shown with the AOR as blank. OR odds ratio, CI confidence interval, AOR adjusted odds ratio, COVID-19 novel coronavirus disease 2019.