Close X

Type 2 diabetes and respiratory infections: new insights

19th January 2023
3D computer image of human lungs

A recent study discussed at the 2022 EASD Annual Meeting found that modifiable risk factors, including HbA1c and BMI, are consistently associated with severe influenza, pneumonia and COVID-19 infection outcomes in people with type 2 diabetes. Lisa Buckingham reports.

Rhian Hopkins, a PhD student from the University of Exeter, began with what we know – that people with diabetes are twice as likely to be hospitalised or to die from an infection-related death compared with the general population. Diabetes is also one of the most important risk factors for COVID-19.

Population-based research in the UK has identified risk factors for COVID-19 mortality within people with diabetes. However, said Rhian, it’s unclear whether these risk factors are specific to COVID-19 or reflect severe respiratory infection in general. Understanding this may be vital should a new respiratory infection emerge.

The key question they wanted to answer in their research was this: are risk factors for severe COVID-19 outcomes similar to other respiratory infections in people with type 2 diabetes? Their aim was to compare clinical and sociodemographic risk factors for COVID-19 with risk factors for influenza and pneumonia.

They studied large scale, population-based data from the UK, using routinely collected health records from the Clinical Practice Research Datalink (CPRD). It contained largely representative primary care data from the UK population alongside individual level linkage to hospital records, allowing the researchers to identify causes of hospitalisation.

They identified over half a million adults with type 2 diabetes for each of their three cohorts: COVID-19 (pre-vaccination); influenza and pneumonia (both from a period just before the pandemic so that COVID-19 infection would not impact outcomes).

They chose to use infection-specific hospitalisation as their outcome, as opposed to mortality (which many early studies chose as an outcome), and used Cox proportional hazard models to explore patterns of association and identify risk factors for each infection. Their models were multivariable and encompassed a wide range of potential risk factors, including 19 co-morbidities, including diabetes features such as microvascular complications.

However, the presentation focused on their most interesting findings on socio-demographic features and the clinical risk factors of HbA1C and body mass index (BMI).

Overall, they saw over 6000 hospitalisations for COVID-19 during the study period, 1354 for influenza and 13,971 for pneumonia. The vast majority were in the older age groups for all three infections, as expected.

With regard to socio-demographic risk factors, they compared the risk between groups for each infection and looked first at male vs female. They found that being male was associated with a greater risk of hospitalisation from COVID-19, compared with influenza and pneumonia where there was less difference between male and female.

For age, they saw that a 10-year increase in age was associated with a greater increase in hospitalisation for pneumonia compared with COVID-19 and influenza, which were similar.

For ethnicity, they replicated previous studies by finding that non-white ethnicity was a risk factor for severe COVID-19 infection outcomes, but they didn’t see this association for the other respiratory infections.

Their results showed a clear differential association between COVID-19 and pneumonia with non-white ethnicity being a greater risk of hospitalisation for COVID-19 and a lesser risk for pneumonia; for influenza, they saw an association only with South Asian ethnicity.

When looking at deprivation, they saw that an increase in deprivation was more strongly associated with hospitalisation for COVID-19 than for pneumonia and no association was observed with influenza.

Once the differences were established, they looked at clinical risk factors. HbA1c and BMI are potentially modifiable and easy-to-measure clinical features that have previously been associated with poor COVID-19 outcomes in people with diabetes, said Rhian. They found that high HbA1c was independently associated with hospitalisation for COVID-19 with increasing risk with high HbA1c starting at a level of 55 mmol/mol. Around 40% of the cohort had an HbA1c above this level.

High HbA1c was also associated with hospitalisation for pneumonia and influenza. The influenza associations looked very similar to those for COVID-19 but for pneumonia, the point at which risk starts to increase was around 80 mmol/mol. An HbA1c of this level was associated with increased risk of hospitalisation for all three infections, but risk starts to increase at a lower level for influenza and especially for COVID-19.

The same analysis was repeated for BMI and it was found that BMI is independently associated with increased risk of hospitalisation for all three infections and that severe obesity is an important and consistent risk factor across the infections.

With COVD-19, however, the risk rises more rapidly from a BMI of 30, suggesting that risk at lower BMI levels may be more important for COVID-19, particularly when compared with pneumonia.

Overall, modifiable risk factors of HbA1c and BMI are broadly consistent between respiratory infections. In contrast, there were clear differences in sociodemographic associations for the three infections:

  • There was a stronger association of male sex with COVID-19
  • Age was more strongly associated with pneumonia
  • There was an association between non-white ethnicities and COVID-19 but not for pneumonia
  • Deprivation was more strongly associated with COVID-19

The message we can take from this, she said, is that approaches to identify people with diabetes at higher risk of severe COVID-19 outcomes may not be entirely applicable to other respiratory infections, and that good glycaemic control may lower risk of all of the severe respiratory infections outcomes that were studied.

After the presentation, the question of vaccination was addressed as the cohorts studied were not vaccinated and this would be an interesting comparison to make in future studies, especially in light of the fact that vaccination rates vary widely across UK populations.

Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.