Background/Objectives: There is wide variability in the available estimates on the burden of Long-COVID (LC), both among adults and children. Preliminary evidence suggests that this inconsistency may be due, at least in part, to the adoption of uneven diagnostic criteria, proposed by several international entities during the course of the pandemic, but no comprehensive evaluation has been performed so far to quantify the extent of variation among children. This systematic review and meta-analysis aims to update the available epidemiological estimates of LC among pediatric subjects, and to systematically appraise the degree of variability that can be attributed to the diagnostic criteria adopted. Methods: A systematic literature search identified cohort studies providing data on LC among children and adolescents with a previous SARS-CoV-2 infection. We computed (a) proportion and (b) head-to-head meta-analyses (a) to quantify the pooled rates of LC and LC-related symptoms, and (b) to assess the likelihood of developing LC based upon selected demographic and/or clinical characteristics, respectively. Results: In total, 52 cohort studies and 960,089 subjects were included. The pooled rate of LC was 18.1% (95% CI: 13.1-23.6), ranging from 16.2% to 21.6%, according to NIH or WHO diagnostic criteria, respectively. Fatigue, respiratory and neurological symptoms were most commonly reported by LC patients. Stratified analyses showed that females, adolescents (vs. children), subjects with comorbidities, and those with a previous severe COVID-19 (as compared to asymptomatic primary infections) were more likely to develop LC (all p < 0.05). In contrast, the likelihood of LC onset did not significantly vary between individuals with ≥1 anti-SARS-CoV-2 vaccine dose, versus the unvaccinated (pooled OR: 0.92; 95% CI: 0.61-1.41). Conclusions: Although the available estimates may largely vary depending on the adopted definition, LC affects a meaningful proportion of the pediatric population; as such, common clinical criteria and diagnostic approaches are urgently needed. Moreover, given the relevant methodological heterogeneity and the generally poor-to-fair quality of a large part of the available literature, high-quality studies adopting standardized methodologies are warranted to identify at-risk populations and guide the implementation of targeted follow-up strategies.
Long Covid