In a recent study published in The Lancet Diabetes & Endocrinology, a group of researchers investigated the association between coronavirus disease 2019 (COVID-19) and the incidence of type 2, type 1, non-specific, and gestational diabetes, and the effect of COVID-19 vaccination.

Study: Incidence of diabetes after SARS-CoV-2 infection in England and the implications of COVID-19 vaccination: a retrospective cohort study of 16 million people. Image Credit: vectorfusionart / Shutterstock

Background 

At least 700 million people have been infected with Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), with reports indicating a 30-50% excess incidence of type 2 diabetes post-infection, raising public health concerns. Three studies found no link between type 1 diabetes and SARS-CoV-2.

It remains unclear if excess diabetes is due to short-term hyperglycemia or long-term effects. Most studies examined diabetes incidence at fixed time points post-COVID-19. COVID-19 severity and vaccination likely influence outcomes, but only one study assessed vaccination’s impact on post-COVID-19 diabetes. Further research is needed to understand long-term effects, vaccination impact, and underlying mechanisms.

About the study 

The present study analyzed data from individuals aged 18 or older registered with primary care practices using TPP software in England within the Open Secure Analytics Framework for Electronic Health Records (OpenSAFELY) platform.

This included primary care records from 24 million people, linked to SARS-CoV-2 testing data, National Health Service (NHS) hospital admissions, and death registry records, alongside COVID-19 vaccination data. The United Kingdom (UK) COVID-19 vaccine rollout began on December 8, 2020, with eligibility based on clinical vulnerability, age, and occupation, making all adults eligible by June 18, 2021.

Three cohorts were defined: pre-vaccination (follow-up from January 1, 2020, to December 14, 2021), vaccinated (follow-up starting two weeks after the second vaccination), and unvaccinated (starting 12 weeks after vaccination eligibility without receiving the vaccine).

Eligible participants were registered for at least six months before the baseline, were aged 18-110 years, and had available data for region, sex, and area deprivation. Those with prior SARS-CoV-2 infection or COVID-19 diagnosis were excluded.

The study assessed associations between COVID-19 and diabetes diagnoses, examining variations up to two years post-diagnosis by diabetes type, vaccination status, and COVID-19 severity. Statistical analyses estimated incidence rates and hazard ratios, with subgroup analyses by demographics and comorbidities. Data was sorted and analyzed using Python and R, and ethical approval was obtained.

Study results 

Of the 33,404,025 individuals in OpenSAFELY-TPP eligible for the pre-vaccination cohort, 16,699,943 met the study inclusion criteria. Among the 33,023,366 participants alive at the start of the delta era, 12,279,699 were included in the vaccinated cohort and 3,076,951 in the unvaccinated cohort. Follow-up periods were January 1, 2020, to December 14, 2021, for the pre-vaccination cohort and June 1 to December 14, 2021, for the vaccinated and unvaccinated cohorts.

In the pre-vaccination cohort, 916,802 (5.5%) were diagnosed with COVID-19. In the vaccinated cohort, 774,475 (6.3%) had COVID-19, and in the unvaccinated cohort, 153,941 (5.0%) were diagnosed. Mortality within 28 days of diagnosis was 0.1% in the pre-vaccination cohort, 0.2% in the vaccinated cohort, and 0.3% in the unvaccinated cohort.

The unvaccinated cohort was younger, had a higher proportion of men, more individuals from South Asian and Black ethnic backgrounds, and more from the most deprived backgrounds compared to the vaccinated cohort. The median follow-up times were 714 days for the pre-vaccination cohort, 190 days for the vaccinated cohort, and 126 days for the unvaccinated cohort.

During the study, 145,533 people were diagnosed with type 2 diabetes in the pre-vaccination cohort, meanwhile 34,365 in the vaccinated cohort, and 2,781 in the unvaccinated cohort. Type 1 diabetes diagnoses were 2,619 in the vaccinated cohort, 16,047 in the pre-vaccination cohort, and 747 in the unvaccinated cohort.

Incidence rates for type 2 diabetes were higher in the vaccinated cohort compared to the unvaccinated cohort. Age-sex-standardized incidence rates for diabetes post-COVID-19 diagnosis were higher in the unvaccinated cohort than in the pre-vaccination and vaccinated cohorts. Diabetes incidence was greater in those hospitalized with COVID-19.

In the pre-vaccination cohort, 145,323 type 2 diabetes diagnoses had follow-up data, with 61.6% being persistent. Persistent diabetes was slightly higher after hospitalization with COVID-19. Adjusted hazard ratios (aHRs) showed increased type 2 diabetes incidence post-COVID-19, especially in the unvaccinated cohort. In the pre-vaccination cohort, type 2 diabetes incidence remained elevated 53-102 weeks post-diagnosis.

Type 1 diabetes incidence was higher during the first four weeks post-COVID-19 diagnosis across all cohorts, with higher aHRs in the pre-vaccination and unvaccinated cohorts. The pre-vaccination cohort showed elevated type 1 diabetes incidence up to 52 weeks post-diagnosis. No significant increase in gestational diabetes post-COVID-19 was observed. The incidence of other diabetes types was also elevated, particularly after hospitalization with COVID-19. 

Conclusions 

To summarize, in the cohort exposed to COVID-19 before vaccines, type 2 diabetes incidence was four times higher during the first four weeks post-diagnosis and remained elevated by 64% in the second year, especially in hospitalized cases.

The increase was significantly lower in vaccinated individuals (1.6 vs. 8.8 times higher). Type 1 diabetes incidence was elevated only in the first year post-diagnosis, with no apparent increase in gestational diabetes.

This study highlights the importance of vaccination and routine diabetes testing after severe COVID-19 to manage long-term health impacts. Its findings are supported by a large, representative UK population sample.