Main findings
In our nationwide cohorts of patients with diabetes who were hospitalized due to COVID-19, we found that almost 20% of the patients died within 28 days. T1D patients were at greater risk of dying within 28 days compared to T2D patients after both COVID-19 and influenza hospitalization. The risk of cardiorenal disease events after COVID-19 hospitalization was higher for T1D compared with T2D, and this was driven by HF. We made several key observations.
First, age- and sex-adjusted risks of death and HF were significantly increased for T1D relative to T2D in COVID-19 by about 80% and 160%, respectively. In influenza, there was an 80% greater risk of dying for T1D patients compared with T2D patients. With full adjustment for prior comorbidities, the association of increased mortality in T1D vs. T2D was still significant in both COVID-19 and influenza, but was no longer significant for HF. There were also no significant associations for stroke, MI and CKD. This implies that while T1D does exhibit a generally increased net risk of unfavourable outcomes, this may largely be mediated by the presence of other comorbidities. This is further supported by the fact that very few events occurred in a subgroup free of cardiovascular and renal disease.
Second, prior ischemic heart disease was not associated with higher mortality in neither COVID-19 nor influenza, which seems surprising. Possibly, secondary preventive medications and lifestyle intervention had a protective effect in these patients. The lack of increased mortality risk with obesity is also an unexpected finding. Notably, as prior conditions were captured by diagnoses from hospitals, there is under-reporting of obesity, which may lead to selection bias.
Third, COVID-19 relative to influenza appears to confer greater mortality in both T2D and T1D, though to a somewhat greater extent among T2D patients.
Our findings of increased age- and sex-adjusted COVID-19 mortality in T1D vs. T2D patients seem to differ slightly from some previous research findings. Rawshani et al.[9] examined risks in COVID-19 for T1D and T2D separately, in a Swedish cohort based on data from the National Diabetes Register (NDR), whereas no comparisons were made between T1D and T2D directly. There was a higher risk of death for T1D compared to controls seen in age- and sex-adjusted analyses, but this did not remain after full adjustment for other covariates such as education, income, treatment and comorbidities. McGurnaghan and colleagues[10] found similar risks of COVID-19-related death in T1D and T2D in their Scottish nationwide cohort of some 320,000 patients with diabetes. Body mass index (BMI) was not significantly associated with the risk of having fatal or critical care unit-treated COVID-19. This is unexpected given the well-established higher risk of severe disease and mortality in general COVID-19 populations conferred by obesity [11, 12]. Prior studies in diabetes cohorts likewise did not show that obesity is associated with a higher risk of severe disease outcomes [13]. We did not find obesity to be linked with any adverse outcomes, but our results should be interpreted with caution due to lack of data on BMI in our study and limited coverage of obesity in the utilized hospital-based patient register.
Previous studies have shown advanced age[14] and male sex[15] to be important risk factors of severe COVID-19 in general populations, including death. In our cohort, we found this to hold true as well. Pre-existing cardiorenal disease also appeared to contribute to excess mortality, as did previous pulmonary disease.
Surprisingly, we found no significant association between prior ischemic heart disease and mortality in either COVID-19 or influenza. This finding could be due to intense secondary preventive medication regimen after a diagnosis of myocardial infarction or angina pectoris, although one would also expect patients suffering a stroke to similarly be treated with drugs such as low-dose aspirin and statins. Nonetheless, it is quite plausible that the former patient group is more intensively treated with such drugs. It is also interesting to note that the risk of non-fatal MI and HF in T2D was about 30% lower in COVID-19 as compared with influenza. Influenza has been implicated in the pathogenesis of acute myocardial infarction[16] and this could also result in increased rates of heart failure due to chronic myocardial ischemia.
In our study, we did not have a control group of patients without diabetes with which to compare the risk of death. Overall, 60-day mortality was 19.0% for T1D and 20.9% for T2D. This was higher in 2020 (21.8% T1D, 23.6% T2D) than in 2021 (12.6% T1D, 16.2% T2D; see Additional file 1: Table S4). Strålin et al. [17] examined all patients hospitalized with COVID-19 in Sweden between March and November 2020. They found an average 60-day mortality of 17.4%, decreasing from 24.7% in March 2020 during the first wave to 10.4% later on, in July–September 2020. Data from the Swedish National Board of Health and Welfare showed a slight increase in mortality during the autumn months of 2021 (for example around 10% in October) compared to the summer months, as expected in the case of a respiratory pathogen [18]. It therefore appears that the mortality in our cohort is comparable to, or slightly higher than, that of the general COVID-19 hospitalized population in Sweden. In this regard, it is important to note the findings of Bergman and colleagues, who showed that among patients hospitalized with COVID-19, around 25% had diabetes (as compared to 6% in the general population) [19]. Incidence of hospitalization among patients with T2D appears to be higher than that in the general population, while among T1D it appears to be similar to the general population. While we did not have a control group to compare with, considering the age category of 70–79 years in the general population of Sweden, there were 20,005 admissions from the beginning of the pandemic until June 3, 2022, comprising around 1.9% of the population in this age span[20, 21]. This may be compared with our T2D cohort which had a mean age of 70 years and in which 3.3% were hospitalized. For the age category of 50–59 years, there were 14,014 admissions, comprising around 1.1% of the general population at these ages. In our T1D cohort, mean age was 58 years and around 0.9% were hospitalized.
Our results indicate a trend of decreasing mortality of COVID-19 over time, and even more so after the introduction of the nationwide mass vaccination campaign during the spring of 2021. In addition to the protective effect of vaccination, this trend can probably be explained by improved therapeutic options such as treatment with corticosteroids and monoclonal antibodies. It is also possible that the individuals who died during the first wave were generally frailer and more susceptible to severe COVID-19.
To the best of our knowledge, our study is the first to compare outcomes of COVID-19 and influenza in T1D and T2D. Piroth et al.[22] previously compared a general population of around 89,000 COVID-19 and 45,000 seasonal influenza patients who were admitted to hospital and found a relative risk of death of 2.9 for COVID-19 cases vs. influenza. In basic agreement with these findings, we found an approximately 2.8-fold increase in mortality with COVID-19 compared to influenza among patients with T2D and also a 2.4-fold higher risk in T1D. This would suggest that patients with T2D are somewhat more affected by COVID-19 in comparison to influenza than T1D.