Thyroid disturbances after COVID-19 and the effect of vaccination in children: a prospective tri-center registry analysis

Abstract

Rapidly evolving clinical data suggest that the novel coronavirus (SARS-CoV-2) and vaccination against COVID-19 might be associated with thyroid disturbances. However, studies remain limited among the pediatric population. Our aim was to assess the prevalence and permanence of thyroid autoimmunity (TA) and dysfunction in children after an acute infection and its potential association with vaccination. A prospective, multicenter registry analysis was performed among 458 children (mean age: 12.4 ± 3,8 years, 45.4% male) with preceding COVID-19. Patient inclusion lasted from 24^th March, 2021 to 23^rd March, 2022 at three pediatric outpatient facilities at Semmelweis University, Budapest. Primary outcomes were the rate of thyroid disturbances assessed by laboratory parameters (thyroid function tests, antithyroglobulin [ATG] and anti-thyroid peroxidase [ATPO] antibodies) and thyroid ultrasound. TA rate among vaccinated and unvaccinated children was determined. Children with newly diagnosed thyroid alterations were followed up for 12.7 ± 4.3 months. Six children had previous thyroid disease. Out of 452 children, 30 cases (6.6%) of newly diagnosed TA (six of them had abnormal thyroid-stimulating hormone [TSH] levels) and eight cases (1.8%) of isolated TSH elevation were observed. Ultrasound-proven autoimmune thyroiditis (AIT) was 4.0%. No association was found between COVID-19 vaccination and thyroid autoimmunity (χ²(1,N = 452) = 0.138, p = 0.815). Among children with TA, 73.3% had long-lasting alterations. Conclusion: Vaccination had no effect on the prevalence of TA. Until further controlled studies state otherwise, children with preceding COVID-19 might benefit from thyroid screening. What is Known: • Numerous case reports implicate that coronavirus disease-2019 (COVID-19) and vaccination against SARS-CoV-2 can be responsible for thyroid disturbances. • Thyroid alterations discovered during acute COVID-19 tend to cease by time and only incidental thyroid autoimmunity (TA) is diagnosed after COVID-19. In adults, no increase in vaccine-related hyper- or hypothyroidism was found. What is New: • TA rate after COVID-19 vaccination among children was not increased. TA had no role in long COVID syndrome. • We discovered a considerable rate of TA (6.6%) and ultrasound-proven autoimmune thyroiditis (AIT) (4.0%) after SARS-CoV-2 infection, and the majority of these alterations remained positive after 6 months.

Keywords: COVID-19 vaccines; Hashimoto disease; Pediatrics; Post-acute COVID-19 syndrome; SARS-CoV-2; Thyroiditis, autoimmune.

Fig. 1

Flowchart of patient inclusion. LC: Long COVID. LC + : children who had one or more new and/or worsened symptoms since their COVID-19 which was still present at the time of their examination (long COVID syndrome). LC-: children who had confirmed COVID-19 but were not experiencing long-lasting symptoms in connection with the previous infection and were complaint-free at the time of their first visit

Fig. 2

Flowchart of thyroid laboratory, ultrasound, and follow-up results. Ultrasound positivity means thyroiditis. Ultrasound outcome covers positive results diagnosed both at initial (n = 16) and during follow-up (n = 2) examinations. ATPO Anti-thyroid Peroxidase, ATG Antithyroglobulin, TSH Thyroid-stimulating Hormone, US Ultrasound, NI No Information (missing data)

Fig. 3

Follow-up of initial positive thyroid autoantibody results. a Children with initially positive antithyroglobulin. b Children with initially positive anti-thyroid peroxidase. Dashed black line: (a) ATG = 1, (t_ATG = log10(1)), (b) ATPO = 1, (t_ATPO = log10(1)). Wide red line: the estimated mean effect line. With intervals (inner: confidence, outer: prediction; both are 95% intervals). x-axis: time (months) y-axis: t_autoantibody = log10(1). For the estimation of the mean effect of time on autoantibody value changes, a generalized least square method was used with linear relation and autoregressive correlation structure assumptions