The risk of developing incident diabetes following hospital discharge or at least 28 days after COVID-19 diagnosis

By Pooja Toshniwal PahariaReviewed by Danielle Ellis, B.Sc.Jun 7 2022

In a recent meta-analysis published in Primary Care Diabetes, researchers evaluated the risk of developing incident diabetes mellitus (DM) post-acute phase of coronavirus disease 2019 (COVID-19).

Background

Previously conducted systematic reviews have established the diagnosis of DM among a substantial number of patients in the acute phase of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections with or without hospital admissions. However, the persistence of such metabolic disorders in the post-acute phase of COVID-19/ long COVID patients is unclear.

About the study

In the present meta-analysis, researchers provided pooled estimates of incident DM risks after hospital discharge after initial hospitalization or after >28 days of SARS-CoV-2 infection diagnosis compared to controls. In addition, they comparatively assessed the impact of SARS-CoV-2 infections versus non-SARS-CoV-2 acute upper respiratory infections (AURI) and flu on the incident DM risks.

Data were searched by two reviewers in the Public/Publisher MEDLINE (PubMed), Web of Science, and Embase databases up to 2 April 2022, using Emtree and Mesh terms and relevant keywords with interposed Boolean operators. Discrepancies were resolved on discussion with a third reviewer. Observational studies were included if they had a retrospective/prospective and cohort/case-control/ study design and provided the adjusted hazard ratios (HR) of incident DM >28 days after COVID-19 diagnosis or after hospital discharge.

Commentaries, reviewers, pre-prints, and articles written in a language other than English were excluded from the analysis. The preferred reporting items for systematic reviews and meta-analyses (PRISMA) checklist were followed for reporting. The quality of the included studies was evaluated with the help of the Newcastle-Ottawa scale (NOS). Pooling of the adjusted HR estimates of the included studies was performed, and fixed-effects modeling and random-effects modeling were used for the analysis.

Results

Out of seven observational studies, four reported the incident DM risk among long COVID patients compared to controls from pooled data of 5,787,027 individuals. Three studies reported the risk of developing incident DM post-acute phase of COVID-19 compared to severity matched influenza patients from pooled data of 308,613 and 24,090 patients with mild and moderate to severe (hospitalized) patients with SARS-CoV-2 infections.

Pooled analysis showed a 59% higher risk of developing incident diabetes in the post-acute phase of SARS-CoV-2 infections compared to healthy SARS-CoV-2-negative controls (adjusted HR 1.6). Identical estimates were found by considering only studies reporting the post-COVID-19 incident DM risk after >28 days of COVID-19 diagnosis (adjusted HR 1.6).

Compared to patients with moderate to severe (requiring hospitalization) influenza, individuals with moderate to severe SARS-CoV-2 infections demonstrated a substantially greater risk of incident DM (adjusted HR 1.5). Similarly, the risk of incident DM was also substantially greater among patients with mild SARS-CoV-2 infections than individuals with mild AURI/flu (HR 1.2).

Overall, the study findings showed significantly greater risks of developing incident DM after the acute phase of COVID-19 than healthy (SARS-CoV-2-negative) controls and severity-matched influenza, AURI, and flu patients and underscored the need for diabetes screening for long COVID patients. However, further research with prospective studies and longer follow-up periods is required to determine individuals' metabolic health further.

Study limitations

All the included studies had a retrospective study design, and none of the included studies considered the higher degree of COVID-19 surveillance after hospital admissions. Further, in most of the studies, only the international classification of diseases and tenth revision (ICD-10) codes were used for defining DM, with the exception of two studies wherein glycated hemoglobin (HbA1c) levels were also evaluated. This could be accountable for the lower sensitivity of identifying the long COVID-associated DM burden on healthcare systems.

Moreover, several studies were conducted on individuals below 18 years of age and did not report HRs for type 1 DM and type 2 DM separately. The pooled estimates of the present study were highly heterogeneous, which could be due to differences in the demographical characteristics, rates of hospitalization, and differences in COVID-19 severity among the study participants. Subgroup analysis considering essential covariates such as obesity and the presence of pre-diabetes was not performed.