Editorial
August 21, 2024
Uncovering Long COVID in Children
Suchitra Rao
JAMA. Published online August 21, 2024. doi:10.1001/jama.2024.13551
The COVID-19 pandemic has had devastating consequences globally, and the immediate and short-term consequences of SARS-CoV-2 infection have been well described. Although most individuals recover, many endure longer-lasting effects, referred to as long COVID, post-COVID condition, or postacute sequelae of SARS-CoV-2 (PASC). PASC is composed of a heterogeneous collection of symptoms and conditions that can affect virtually any organ system, with common manifestations including fatigue, cough, malaise, and pain.1,2 Those at increased risk for long COVID include females, those hospitalized due to acute COVID-19,3,4 and individuals with underlying comorbidities.5-8 Several hypotheses have been proposed to explain the underlying pathophysiology, which is likely multifactorial, including immune dysregulation, autoimmunity and immune printing, microvascular clotting with endothelial dysfunction, and impaired neurological signaling.9,10
Much of the accrued data on long COVID to date have focused on adult populations, with a relative deficit of large, well-designed, prospective studies in children examining outcomes beyond 12 months after infection. The overall prevalence in children varies widely in the literature11 given variation in definitions, study designs, setting, population under study, and follow-up period, as well as challenges with identifying symptoms and self-report in children. Prospective studies with adequate study designs provide estimates of 10% to 20%.7,12,13 The scientific community has acknowledged that we urgently need to understand the presentations, mechanisms, trajectory, and recovery of PASC in children.14 However, progress into this area of study has been complicated by the evolving nature of the COVID-19 pandemic, prioritization of studies in adults, and the heterogeneous nature of postacute sequelae.
In response to this need to characterize long COVID in children, the National Institutes of Health has funded the Researching COVID to Enhance Recovery (RECOVER) Initiative, a multisite study across the US aiming to characterize the natural history of PASC in children and young adults and explore its underlying mechanisms and long-term health effects using both prospective and electronic health record cohorts. The RECOVER-Pediatrics prospective group of the initiative consists of clinical cohorts from more than 100 study sites throughout the US that are prospectively following up children and young adults for up to 4 years, beginning at birth through 25 years of age. Close to 13 000 children, adolescents, and young adults have been enrolled, including those with and without COVID-19 and long COVID. In addition, electronic health record cohorts comprising data from more than 8.9 million inpatient and outpatient records across 40 sites utilize laboratory data, procedure and diagnostic codes, and clinical notes to provide timely assessments of long COVID at scale. Collectively, medical, social, biological, and immunologic data are being collected to characterize long COVID across the early life course, evaluate its impact on preexisting conditions, and study de novo conditions that arise through studies exploring epidemiology, underlying pathophysiology, and patient-reported outcomes.15
In this issue of JAMA, Gross and colleagues provide their sentinel findings from the RECOVER-Pediatrics study,16which explored the most common symptoms and phenotype clusters associated with long COVID, characterized these differences by age, and developed research indices. The study was a combined retrospective and prospective longitudinal study of 4 cohorts spanning 60 sites across the US, comprising 5376 children and adolescents aged 6 to 17 years with long COVID enrolled between March 2022 and December 2023, and including test-negative, seronegative controls.
Surveys assessing prolonged symptoms across 9 domains, along with Patient-Reported Outcomes Measurement Information System (PROMIS) Global Health Scales, were completed at least 90 days after the initial COVID-19 infection. Subsequently, a model using a penalized logistic regression approach was used to develop an empirically derived index for researchers to identify PASC. Next, correlation analyses examined relationships between PASC indices and overall patient-reported outcomes. Finally, symptom patterns were explored to develop distinct PASC symptom profiles.
At least 1 prolonged symptom was identified in 45% of infected children and 33% of uninfected children and in 39% of infected adolescents and 27% of uninfected adolescents. Ten symptoms were identified in school-age children and 8 symptoms in adolescents that were most associated with infection history, which were used to derive the research index and optimal thresholds. The most common symptoms in children associated with infection history included headache (57%), trouble with memory/focusing and trouble sleeping (44%), and abdominal pain (43%). In contrast, the most common symptoms in adolescents associated with infection history were daytime tiredness/sleepiness or low energy (80%), body/muscle/joint pain (60%), headaches (55%), and trouble with memory/focusing (47%). Distinct symptom clusters were identified in children and adolescents. Among children, 4 clusters included a group comprising multiple body system involvement; headache, musculoskeletal pain, and low energy/sleepiness; trouble sleeping, trouble with memory/focusing; and gastrointestinal manifestations. Among adolescents, 3 clusters included multiple body symptoms, musculoskeletal pain and low energy/sleepiness, and change or loss of smell or taste. The multisystem clusters were more strongly correlated with poor health and quality of life than single symptoms. Symptoms most associated with infection were used to develop a research index, from which an index threshold was derived.
The RECOVER-Pediatrics cohort is one of the largest and most geographically diverse study cohorts in the US, and the program represents the type of prospective, large-scale, methodologically rigorous research needed to better characterize long COVID in children. The study team should be commended for including community, patient, and caregiver partners to encourage patient-centeredness, and the symptom and validated quality of life assessments are extensive and comprehensive and take into account the different developmental stages across the early life course. Further, the study serves as a useful companion to the adult RECOVER study published in 2023,17 in which investigators used similar methodology to develop a score and evaluated different symptom clusters. Taken together, the publications highlight the importance of evaluating chronic conditions such as PASC across the life spectrum because there were clear differences in presentations among children, adolescents, and adults. For example, the most predominant symptoms in adults identified from the model in decreasing order of frequency were postexertional malaise, fatigue, brain fog, dizziness, gastrointestinal symptoms, and palpitations. In contrast, musculoskeletal pain featured more predominantly in children and adolescents. Headaches were a common symptom in children, but were less prominent in adolescents and adults. A gastrointestinal symptom cluster was unique to school-age children, whereas gastrointestinal symptoms were more likely to be present in conjunction with other symptoms in adults. Further studies correlating symptom clusters with risk factors, acute COVID-19 illness, comorbidities, and illness trajectories will be the next steps to predicting recovery and targeting prevention and treatment strategies.
The authors caution against using the indices for clinical practice at this time, given that knowledge regarding long COVID is accruing and this condition may continue to evolve with increased population immunity and changing variants. However, in time, future iterations may serve as a useful tool for evaluating illness trajectories and for evaluating the impact of management strategies.
It is worth recognizing the potential for ascertainment bias in this study, based on the recruitment strategies, with a higher proportion of children and adolescents with neurocognitive/behavioral manifestations included in the study, which may explain the difference in findings from other studies that have indicated a higher predominance of respiratory symptoms.6,7 Evaluations were through caregiver report conducted at varied points after initial infection, which contributes to symptom underreporting, as well as recall and measurement bias. Another limitation of the study is the inability to discern whether quality of life scores reflected underlying medical conditions rather than long COVID symptoms for those enrolled after the development of long COVID, because there were no baseline measurements to ascertain preinfection status.
Four years after the initial reports of long COVID, much remains to be discovered regarding the trajectory and strategies for prevention and management, especially in children. The general and scientific communities, in particular those who continue to endure symptoms on a daily basis, eagerly await the subsequent findings from RECOVER-Pediatrics, which will help uncover more about long COVID in children.