Children and Adolescents Long-COVID: Systematic Review and Meta-Analysis

General characteristics of the study

The titles and summaries of 8373 publications have been screened. After the duplicates were removed, the search identified 68 publications after screening titles and summaries, and 47 were excluded because they did not meet the selection criteria. A total of 21 studies were selected for analysis (figure. 1).The characteristics of general research are shown in the table. 1.. Various authors use the terms “after acute COVID,” “long-COVID,” “permanent COVID,” and “permanent COVID symptoms” as synonyms. Most studies evaluated the pre-specified symptoms included in the questionnaire. Eighteen of the selected studies are from the European cohort (Denmark, Russia, Italy, Germany, Tukey, Latvia, UK, France, Sweden, Switzerland), one from Iran, one from Brazil and the other from Australia. It is. Study by Kikkenborg¹⁵ And Borch^{16 16} Included overlapping population (Denmark), as Roge did^{17 17} And Sumane^{18 18} (Latvia).. To avoid duplication of data, long COVIDs were estimated and only the largest sample size studies were included for the results reported in both studies. Still, some results were presented in only one study, so both studies were included in the overall meta-analysis. Four studies included only inpatients, and the remaining studies included all COVID-19 severity (asymptomatic, mild, moderate, severe). The number of patients included in the study ranged from 53 to 57,763 and the age ranged from 0 to 18 years. A total of 80,071 COVID-19 children and adolescents were included in the meta-analysis.The literature reviewed identified more than 40 long-term clinical symptoms associated with COVID-19 (Table). 2).

Meta-analysis of prevalence of long-COVID

The prevalence of long-term COVID in children and adolescents, as defined by the presence of one or more symptoms more than 4 weeks after SARS-CoV-2 infection, is 25.24% (95% CI 18.17–33.02, I).² 99.61%) (Figure 2, 3). For inpatients, the prevalence of long-COVID is 29.19% (95% CI 17.83–41.98, I)² 80.84%). The percentage of prevalence associated with the most common symptoms was mood symptoms (eg, sadness, tension, anger, depression, anxiety) (16.50%; 95% CI 7.37–28.15, I).² 97.49%), fatigue (9.66%; 95% CI 4.45–16.46, I² 99.12%), sleep disorders (eg, insomnia, hypersomnia, poor sleep quality) (8.42%; 95% CI 3.41–15.20, I² 93.49%); Headache (7.84%; 95% CI 4.04–12.70, I² 98.49%), respiratory symptoms (7.62%; 95% CI 2.08–15.78, I² 99.15%), sputum production or stuffy nose (7.53%; 95% CI 3.78–12.36, I² 0%), cognitive symptoms (eg, poor concentration, learning disabilities, confusion, memory loss) (6.27%; 95% CI 4.46–8.35, I² 91.32%), loss of appetite (6.07%; 95% CI 3.95–8.59, I² 93.54%), exercise intolerance (5.73%; 95% CI 0.00–19.38, I² 87.77%), and changes in the sense of smell (eg, hyposmia, anosmia, hypersomnia, parosmia, and hallucinations) (5.60%; 95% CI 3.13–8.69, I² 97.11%).The prevalence of all other symptoms was less than 5.00% (table) 2figure. 2Supplementary figure. 1).

Pooled prevalence of long-COVID by symptoms in children and adolescents. A meta-analysis revealed the prevalence of over 40 long-COVID symptoms in children and adolescents. The presence of one or more symptoms after SARS-CoV-2 infection was 25.24%.

Forest plots of pooled prevalence of overall long-COVID in children and adolescents.

OR meta-analysis (contrast with case)

Contrast with cases of 13 symptoms (mood, malaise, headache, dyspnea, difficulty concentrating, anosmia / ageusia, loss of appetite, rhinitis, muscle pain / joint pain, cough, fever, sore throat) I was only able to perform a meta-analysis of the ORs to compare. , And nausea / vomiting) (Figure. Four). Children with long COVIDs were at increased risk of persistent dyspnea when compared to controls (OR 2.69; 95% CI 2.30–3.14, I)² 0%), anosmia / ageusia (OR 10.68; 95% CI 2.48, 46.03, I² 0%), and / or fever (OR 2.23; 95% CI 1.2–4.07, I² 12%). Four of the 13 meta-analyses had significant heterogeneity (figure). Four). Controls were selected in very different ways between studies, which may have introduced significant heterogeneity. The following were different definitions of controls, children and adolescents: (1) other infections (eg, colds, pharyngotonsillitis, gastrointestinal, urinary tract infections, bacterial pneumonia or unknown cause)^{17 17}(2) No antibody test^{twenty four} Mixed with other infectious diseases^{17 17}(3) Negative antibody test²⁹(4) Negative rtPCR test among symptomatic children³⁵(5) Children who do not have a positive test recorded in the database¹⁵ (Supplementary figure 2). Coordination between studies also varied.Several studies adjusted OR by age, gender, ethnicity, socioeconomic status, and comorbidity³⁵..But age and gender¹⁵ Gender only, age only adjustment^{17 17}By OR, did not adjust, or did not adjust the previous conditions^{twenty four}..

Case-to-control 95% CI pool odds ratio. The size of each box shows the effectiveness of each study by symptom assigned using age and odds ratio by domain (95% CI).

Other findings

The supplementary chart shows the prevalence of symptoms during the course of long-term COVID in cases and controls. 2.. Given the non-uniformity of control definitions and the small number of subjects, no formal statistical comparison was made for rough prevalence. Symptoms that cannot be incorporated into the meta-analysis because presented in a single study include orthostatic intolerance, cold limbs, cracked lips, adenopathy, fainting, finger and toe spasms, chills, and toes. Includes swelling. / Fingers, and hallucinations.

In one study, there were statistically significant differences between clinical cases and controls in systolic blood pressure, left ventricular ejection fraction, relative myocardial wall thickness, and systolic excursion of the tricuspid annulus. It was reported.^{twenty two}.. However, given that these variables were evaluated only in this study, it was not possible to perform a meta-analysis of these results.Studies included in the meta-analysis found specific variables such as age, gender, severe acute COVID-19, obesity, allergic disease, and long-term health.^{19 19,26,28 28,36}, The risk of long-COVID has increased. In addition, two studies evaluated the duration of symptoms.A Danish study reports that symptoms resolved within 1-5 months in at least 54-75% of children and adolescents.^{16 16}..Another person in the United Kingdom reported that 4.4% of children were still symptomatic 4 weeks after the onset of COVID-19, but decreased to 1.8% over 8 weeks.²⁷..

Research quality

The quality of the studies was 7 points or higher (supplementary table). 1).table 3 Here is a list of the methodological strengths or limitations of each study. All studies included laboratory-confirmed COVID-19 infection, rt-PCR, antigen, or antibody testing. Two-thirds of the study included more than 100 children. The six meta-analyses had low heterogeneity (I)²<25%) If you have the following symptoms: vomiting and nausea, stuffy nose, nasal congestion, urinary problems, neurological abnormalities, and dysphagia. Three meta-analyses showed moderate heterogeneity in symptoms such as abdominal pain, menstrual changes, and speech disorders. All other meta-analyses had high non-uniformity (I)²> 75%). Few studies have been stratified by different variables (age, gender, country, past comorbidities, etc.), so there is not enough research to include this information to assess where the heterogeneity occurred. was.