The Challenge of Studying Long COVID: An Updated Review

Abstract

Accurately determining the risk of long COVID is challenging. Existing studies in children and adolescents have considerable limitations and distinguishing long-term SARS-CoV-2 infection-associated symptoms from pandemic-related symptoms is difficult. Over half of individuals in this age group, irrespective of COVID-19, report physical and psychologic symptoms, highlighting the impact of the pandemic. More robust data is needed to inform policy decisions.

The majority of children with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection have asymptomatic or mild disease.¹ The long-term effects of the infection might therefore have greater weight in coronavirus disease 2019 (COVID-19) vaccination and other policy decisions in this age group. We recently reported that the frequency of persistent symptoms after COVID-19 in children and adolescents is uncertain.^2,3 Almost all of the studies on ‘long COVID’ in this age group have considerable limitations, for example, the inclusion of children without confirmed SARS-CoV-2 infection and a lack of appropriate control groups.^3,4 As the number of published studies on this topic has doubled, we reassessed the current evidence on long COVID in children and adolescents.

We identified 27 studies (13 cross-sectional studies,^5–17 9 prospective cohort studies,^18–26 4 case series^27–30 and 1 retrospective cohort study³¹) investigating persistent symptoms in a total of 34,664 SARS-CoV-2-infected and 38,988 uninfected children and adolescents. The number of children in each study varied from 5 to 30,117 [median 105, interquartile range (IQR) 30–859]. The study findings are detailed in the Table, Supplemental Digital Content, http://links.lww.com/INF/E684. Studies which followed children after a SARS-CoV-2 infection but did not evaluate symptoms of long COVID,^32–34 did not evaluate more than 1 symptom³⁵ and those which did not report separate results for children and adolescents^36–39 were not included.

Nine of the 27 studies included an uninfected control group.^{5–8,18–22} Six studies compared the proportion of children and adolescents with persistent symptoms in those with and without evidence of prior SARS-CoV-2 infection.^5,8,19–22 The difference in proportions varied between Please replace by -0.5% and 13.2% (median 3.0%, IQR 1.4%-3.6%) (Fig. 1). In all but one study,²² the difference was less than 4%. The study which reported a difference of 13.2% had a response rate of only 13.4% and therefore a major risk of sampling bias. A further study reported a difference of 45.2% in persistent symptoms when comparing children after SARS-CoV-2 infection with those after other respiratory infections.⁷ However, it is likely that a large part of this difference is attributable to the considerable difference in age between the two groups (median of 10 vs. 2 years). Two studies did not report the proportion of children affected by long COVID symptoms in the control group.^6,18

FIGURE 1.

Prevalence of persistent symptoms after COVID-19 in children and adolescents and in non-COVID-19 controls.

Although many studies included a control group, they all had other deficiencies meaning their results need to be viewed with caution. Many had low response rates or differences in response and inclusion rates between children with and without SARS-CoV-2 infection (Table, Supplemental Digital Content 1, http://links.lww.com/INF/E684).

The 11 studies which investigated persisting symptoms after SARS-CoV-2 infection without control groups reported prevalences of long COVID symptoms between 7.9% and 58.1% (median 27.1%, IQR 12.5%–41.4%).^{9,13–15,17,23–26,30,31} However, many of these studies included children without laboratory-confirmed infections, studied children at arbitrary time points, relied on self- or parent-reported symptoms without clinical assessment and objective parameters or varied in the proportion of children with preexisting medical conditions.

The large variation in results from studies underlines how difficult it is to accurately determine the risk of long COVID. In addition to the lack of a clear case definition, it is impossible to blind participants to whether they have been infected with SARS-CoV-2 or not. Another unavoidable limitation includes the possibility that the uninfected control group is contaminated by children who have been infected with SARS-CoV-2 but who were not tested or who did not seroconvert.⁴⁰ Further, studies that evaluate a single time point might miss transitory or intermittent symptoms of long COVID. Finally, the range and the number of symptoms sought in studies vary considerably and some studies have been criticized for evaluating certain key symptoms.

In future studies, it is important to collect age-aggregated data, as the incidence and characteristics of long COVID will be different in young children and adolescents. Moreover, more studies are needed to investigate the association between the initial severity of COVID-19 and the number and duration of persistent symptoms. Additionally, other risk factors for long COVID should be identified. It is also important to unravel the mechanisms underlying persistent symptoms after COVID-19 and to identify similarities to and differences from other postviral syndromes. This will help find treatment options and define the role of vaccination in the prevention of long COVID.

The fact that nearly all symptoms reported by children and adolescents infected with SARS-CoV-2 are also reported in similar frequencies in those without evidence of infection highlights that one of the major challenges remains to distinguish long-term SARS-CoV-2 infection-associated symptoms from pandemic-related symptoms. It is worrisome that more than half of children and adolescents, even when they have not had COVID-19, report physical and psychologic symptoms, highlighting how much children and adolescents have suffered from the pandemic.