Summary

Background
Some epidemiological studies have suggested an increase in incidence of type 1 diabetes during the COVID-19 pandemic, however the mechanism(s) behind such an increase have yet to be identified. In this study we aimed to evaluate the possible role of the SARS-CoV-2 virus in the reported increase in the rate of type 1 diabetes.
Methods
In this observational cohort study using data from the Finnish Pediatric Diabetes Register (FPDR), we assessed the incidence of type 1 diabetes (number of children with newly diagnosed type 1 diabetes per 100 ۰۰۰ person-years during the pandemic and the reference period) during the first 18 months of the COVID-19 pandemic in children in Finland younger than 15 years old compared with a reference period which included three corresponding pre-pandemic periods also obtained from the FPDR. Children with confirmed monogenic diabetes were excluded. We also compared the phenotype and HLA genotype of the disease between these two cohorts, and analysed the proportion of newly diagnosed people with type 1 diabetes testing positive for SARS-CoV-2 antibodies.
Findings
۷۸۵ children and adolescents in Finland were diagnosed with type 1 diabetes from March 1, 2020, to Aug 31, 2021. In the reference period, which comprised three similar 18-month terms (from March 1, 2014, to Aug 31, 2015; March 1, 2016, to Aug 31, 2017; and March 1, 2018, to Aug 31, 2019) 2096 children and adolescents were diagnosed. The incidence of type 1 diabetes was 61·۰ per 100 ۰۰۰ person-years (95% CI 56·۸–۶۵·۴) among children younger than 15 years old during the pandemic, which was significantly higher than during the reference period (52·۳ per 100 ۰۰۰ person-years; 50·۱–۵۴·۶). The incidence rate ratio adjusted for age and sex for the COVID-19 pandemic was 1·۱۶ (۱·۰۶–۱·۲۵; p=0·۰۰۰۶) when compared with the reference period. The children diagnosed during the COVID-19 pandemic had more often diabetic ketoacidosis (p<0·۰۰۱), had a higher HbA1c (p<0·۰۰۱), and tested more frequently positive for glutamic acid debarboxylase antibodies at diagnosis (p<0·۰۰۱) than those diagnosed before the pandemic. There were no significant differences in the distribution of HLA genotypes between the two periods. Only five of those diagnosed during the pandemic (0·۹%) of 583 tested positive for infection-induced SARS-CoV-2 antibodies.
Interpretation
Children and adolescents diagnosed with type 1 diabetes during the pandemic had a more severe disease at diagnosis. The observed increase in type 1 diabetes incidence during the first 18 months could be a consequence of lockdown and physical distancing rather than a direct effect of SARS-CoV-2 infection.
Funding
Helsinki University Hospital Research Funds, EU Horizon 2020 (Versatile emerging infectious disease observatory project), Academy of Finland, Sigrid Jusélius Foundation, Jane & Aatos Erkko Foundation, and Medicinska understödsföreningen Liv och Hälsa.

Introduction

A series of epidemiological studies have reported that the number of people with newly diagnosed type 1 diabetes has increased during the COVID-19 pandemic.1,  ۲ According to one meta-analysis, the global incidence of type 1 diabetes among children increased by 9·۵% from 2019 to 2020.3 However, these findings have been questioned due to methodological weaknesses.4 An increase in diabetic ketoacidosis at diagnosis has also been observed during the pandemic.5,  ۶,  ۷ The reasons behind these findings are unknown. There is an ongoing discussion as to whether these observations are a direct effect of a SARS-CoV-2 infection or a consequence of the lockdown and social isolation due to the pandemic.

A direct effect has been implied, either through an injury to the pancreatic β cells by the SARS-CoV-2 virus or through virus-induced precipitation or acceleration of the disease process leading to type 1 diabetes.8 An increase in diabetic ketoacidosis at diagnosis suggests a possible delayed presentation to health-care services, whereas the indirect effect on the incidence of type 1 diabetes might be mediated through an earlier unmasking of the disease process because of the substantially reduced infection load and decreased physical activity, particularly in children. A decreased infection load might affect the immune system, favouring the development of type 1 diabetes, whereas reduced physical activity increases β-cell stress.9 In April 2020, all schools were closed in Finland and parents with children attending day-care centres were asked to keep their children at home, if possible. The biodiversity hypothesis assumes that reduced microbial exposure in early life increases the risk of immune-mediated diseases.10

Results

۷۸۵ children diagnosed with type 1 diabetes were enrolled during the pandemic period, comprising of 18 months from March 1, 2020, to Aug 31, 2021 (appendix 3 p 5). The reference periods comprised three similar 18-month terms (from March 1, 2014, to Aug 31, 2015; March 1, 2016, to Aug 31, 2017; and March 1, 2018, to Aug 31, 2019). Altogether, 2096 children diagnosed with type 1 diabetes during the combined 54-month reference period joined the FPDR. The baseline characteristics of the participants in the pandemic and reference cohorts are shown in table 1.