Digital, Health and Care Institute (DHC institute)
research institute & advisory consulting
Joaquim Cardoso MSc
Founder and Chief Researcher, Editor and Advisor
December 12, 2022
Key message:
- After COVID-19, patients of all ages and genders had an elevated incidence and relative risk for a new diagnosis of diabetes.
- This reinforces the need for clinicians to pay attention to patients’ glucose metabolism in the post-acute phase of COVID-19.
- Particular attention should be paid during the first 3 months of follow-up after COVID-19 for new-onset diabetes.
Abstract
Background
- There is growing evidence that patients recovering after a severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection may have a variety of acute sequelae including newly diagnosed diabetes.
- However, the risk of diabetes in the post-acute phase is unclear.
- To solve this question, we aimed to determine if there was any association between status post-coronavirus disease (COVID-19) infection and a new diagnosis of diabetes.
Methods
- We performed a systematic review and meta-analysis of cohort studies assessing new-onset diabetes after COVID-19.
- PubMed, Embase, Web of Science, and Cochrane databases were all searched from inception to June 10, 2022.
- Three evaluators independently extracted individual study data and assessed the risk of bias.
- Random-effects models estimated the pooled incidence and relative risk (RR) of diabetes compared to non-COVID-19 after COVID-19.
Results
- Nine studies with nearly 40 million participants were included.
- Overall, the incidence of diabetes after COVID-19 was 15.53 (7.91–25.64) per 1000 person-years, and the relative risk of diabetes after COVID-19 infection was elevated (RR 1.62 [1.45–1.80]).
- The relative risk of type 1 diabetes was RR=1.48 (1.26–1.75) and type 2 diabetes was RR=1.70 (1.32–2.19), compared to non-COVID-19 patients.
- At all ages, there was a statistically significant positive association between infection with COVID-19 and the risk of diabetes: <18 years: RR=1.72 (1.19–2.49), ≥18 years: RR=1.63 (1.26–2.11), and >65 years: RR=1.68 (1.22–2.30).
- The relative risk of diabetes in different gender groups was about 2 (males: RR=2.08 [1.27–3.40]; females: RR=1.99 [1.47–2.80]).
- The risk of diabetes increased 1.17-fold (1.02–1.34) after COVID-19 infection compared to patients with general upper respiratory tract infections.
- Patients with severe COVID-19 were at higher risk (RR=1.67 [1.25–2.23]) of diabetes after COVID-19.
- The risk (RR=1.95 [1.85–2.06]) of diabetes was highest in the first 3 months after COVID-19.
- These results remained after taking confounding factors into account.
Conclusions
- After COVID-19, patients of all ages and genders had an elevated incidence and relative risk for a new diagnosis of diabetes.
- Particular attention should be paid during the first 3 months of follow-up after COVID-19 for new-onset diabetes.
After COVID-19, patients of all ages and genders had an elevated incidence and relative risk for a new diagnosis of diabetes.
Particular attention should be paid during the first 3 months of follow-up after COVID-19 for new-onset diabetes.
ORIGINAL PUBLICATION (excerpt)
Risk for newly diagnosed diabetes after COVID-19: a systematic review and meta-analysis
BMC Medicine
Ting Zhang, Qimin Mei, Zhaocai Zhang, Joseph Harold Walline, Yecheng Liu, Huadong Zhu & Shuyang Zhang
15 November 2022
Background
Coronavirus disease 2019 (COVID-19) is a complex clinical syndrome caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [ 1].
Despite many large studies leading to the approval of vaccines and antivirals, the global spread of SARS-CoV-2 continues [ 2, 3].
As of June 18, 2022, there have been more than 535,863,950 confirmed cases globally, including 6,314,972 deaths (according to the World Health Organization) [ 4].
Factors associated with poor outcomes, including hospitalization, intensive care unit (ICU) admissions, and mortality in COVID-19 patients, are of considerable interest.
More specifically, health comorbidities and baseline physical activity [ 5] may predispose patients to an increased risk of poor outcomes following COVID-19 infection.
Previous studies have indicated that diabetes mellitus (DM) is associated with an increased risk of severe COVID-19, acute respiratory distress syndrome (ARDS), and in-hospital mortality [ 6, 7, 8].
More intriguingly, a recent meta-analysis has reported that newly diagnosed diabetes is commonly observed in COVID-19 patients [ 9, 10, 11].
The world has raised concerns about a bi-directional relationship between these two health conditions [ 12].
As the COVID-19 pandemic has progressed, there is growing evidence that after the acute phase of the disease, people with COVID-19 can develop lingering sequelae (called “long COVID”) that may involve pulmonary and extrapulmonary organ system manifestations, such as diabetes [ 13].
Follow-up of children with COVID-19 has identified that the incidence of type 1 newly diagnosed diabetes has increased [ 14].
An unregistered meta-analysis [ 15] in PROSPERO also found an increased risk of diabetes among adults with long COVID-19, but it has some flaws in the study design which limit the interpretation and applicability of the individual studies’ findings.
Therefore, there is an urgent need for systematic reviews and meta-analyses of the existing literature, particularly focusing on controlled studies.
This systematic review and meta-analysis was conducted to estimate the prevalence of a new diagnosis of diabetes after COVID-19 compared to non-COVID-19.
Methods and other sections
see original publication (this is an excerpt version)
Discussion
This systematic review and meta-analysis provided comprehensive quantitative estimates of the incidence of diabetes in 10 post-COVID-19 populations.
To our knowledge, this is the largest and most wide-ranging analysis of this kind to date.
With nearly 40 million participants, and nearly 200,000 cases of diabetes reported, we found a post-COVID-19 incidence of diabetes per 1000 person-years of 15.53, and a relative risk of 1.62 compared to non-COVID-19-infected people.
With nearly 40 million participants, and nearly 200,000 cases of diabetes reported, we found a post-COVID-19 incidence of diabetes per 1000 person-years of 15.53, and a relative risk of 1.62 compared to non-COVID-19-infected people.
Subgroup analyses suggested that the risk of developing diabetes was also increased regardless of age, gender, type of diabetes, follow-up time, or level of COVID-19 severity, although undifferentiated diabetes did not have a significant relative risk.
Subgroup analyses suggested that the risk of developing diabetes was also increased regardless of age, gender, type of diabetes, follow-up time, or level of COVID-19 severity, although undifferentiated diabetes did not have a significant relative risk.
These results remained significant even after accounting for the possibility of unmeasured confounding.
Similar results have been reported in patients infected with other viruses, with an increased incidence of diabetes compared with those not infected [ 38, 39, 40].
Our subgroup analysis revealed a 1.2-fold increased risk of developing diabetes after COVID-19 compared to patients with other upper respiratory tract infections and a 1.82-fold increased risk of developing diabetes after COVID-19 compared to the general population.
Our subgroup analysis revealed a 1.2-fold increased risk of developing diabetes after COVID-19 compared to patients with other upper respiratory tract infections and a 1.82-fold increased risk of developing diabetes after COVID-19 compared to the general population.
This reinforces the need for clinicians to pay attention to patients’ glucose metabolism in the post-acute phase of COVID-19.
This reinforces the need for clinicians to pay attention to patients’ glucose metabolism in the post-acute phase of COVID-19.
There is also new evidence regarding the effect of the SARS-CoV-2 virus on pancreatic β-cell function [ 41].
It has been suggested that SARS-CoV-2 may affect the pancreas by acting on the mRNA of angiotensin-converting enzyme 2 (ACE2) in the endocrine and exocrine glands of the pancreas [ 11, 42]. The presence of SARS-CoV-2 antigen has recently been reported in the postmortem pancreas of patients who died from COVID-19 [ 43]. In addition, SARS-CoV-2 can induce a cytokine storm, an exaggerated immune response that produces a broad spectrum of cytokines, thereby establishing a systemic pro-inflammatory environment, which may play a role in promoting insulin resistance and β-cell hyperstimulation, ultimately leading to altered cellular function and the death of β-cells [ 44, 45, 46]. According to our subgroup analysis, there was a 1.48-fold increased risk of developing type 1 diabetes and a 1.7-fold increased risk of type 2 diabetes compared to patients not infected with COVID-19.
In our analysis, the incidence rate per 1000 person-years of follow-up was 3.65 (95% CI, 2.91 to 4.83), RR=1.72 (95% CI, 1.1 to 2.50) in the <18-year-old population, with similar results in adults and in those >65 years old.
Moreover, the relative risk of morbidity was similar across genders. These findings underscore the importance of COVID-19 prevention in all age groups and genders, such as encouraging vaccination of all eligible children and adolescents [ 47].
Although all of the studies we included reduced confounders by adjusting for the risk of associated factors (propensity score matching) [ 48], concerns about possible bias due to uncontrolled confounders (e.g., comorbidities, socioeconomic environment, body mass index [BMI], etc.) remain [ 49].
Our study is the first meta-analysis to consider the E-value as a parameter of unmeasured confounders in examining the association between the COVID-19 post-acute phase and diabetes risk, which represents a new methodological contribution to the study of COVID-19 and diabetes [ 24]. The E-value, a sensitivity analysis of unmeasured confounders, is a relatively new method for measuring the association between exposure and outcome robustness and to assess evidence of causality [ 50]. Our results suggest that an unobserved confounder would need to be associated with a risk ratio of 2.08 for exposure and outcome to fully explain the mean RR of 1.62. In addition, a risk ratio of 3.58 would be required for the confounder to make the risk estimate statistically nonsignificant. Propensity matching was performed in all of the studies we included, and most studies adjusted for at least some clinically important confounding factors, such as patient age, gender, BMI, race, and comorbidities. Therefore, we believe it is implausible that residual confounders exist above and beyond these measured confounders that are sufficient to explain away the above results.
Limitations
Several potential study limitations need to be considered.
First, all included studies used a retrospective design and all studies used the breadth and depth of large electronic healthcare databases [ 29, 30, 32, 33, 35, 37, 51, 52, 53] to construct cohorts and define health characteristics based on validated definitions, which cannot exclude misclassification bias, particularly for diabetes types.
Second, some of the studies used contemporary controls, not excluding the possibility that some individuals may be infected with SARS-CoV-2 and have not been tested, which could bias the results toward the null hypothesis if these individuals were present in large numbers in the contemporary control group.
Third, because the included studies were conducted in different countries and in different regions within the same country, differences in national and regional care policies are expected, for which this meta-analysis could not be adapted or adjusted.
Fourth, the study designs were heterogeneous (prospective cohort and retrospective cohort studies). As the number of available prospective cohort studies on this topic remains small, more high-quality studies are needed to confirm our results.
Conclusions
Patients of all ages and genders recovering from COVID-19 had an elevated incidence and relative risk for developing diabetes.
Particular attention should be paid to potential new-onset diabetes during the first 3 months of follow-up after COVID-19.
References and additional information
See the original publication
About the authors & affiliations
Ting Zhang1,2, Qimin Mei1 , Zhaocai Zhang3 , Joseph Harold Walline4 , Yecheng Liu1* , Huadong Zhu1* and Shuyang Zhang5*
1 Emergency Department, State Key Laboratory of Complex Severe and Rare Diseases, Peking Union Medical College Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, China.
2 Department of Family Medicine & Division of General Internal Medicine, Department of Medicine, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, State Key Laboratory of Complex Severe and Rare Diseases, Beijing, China.
3 Department of Critical Care Medicine, The Second Afliated Hospital, Zhejiang University School of Medicine, Hangzhou, China.
4 Department of Emergency Medicine, Penn State Health Milton S. Hershey Medical Center and Penn State College of Medicine, Hershey, PA, USA.
5 Department of Cardiology, State Key Laboratory of Complex Severe and Rare Diseases, Peking Union Medical College Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, China.
Acknowledgements
The authors thank Dr. Maya B. Mathur from the Quantitative Sciences Unit and Department of Pediatrics, Stanford University, for her assistance in bias analysis on the effect of unmeasured confounding.
Funding
The National Key Research and Development Program of China (2021YFC2501800). National High Level Hospital Clinical Research Funding (2022-PUMCH-B-110). Rui E Emergency Medicine Research Special Fund Project (R2021007). 2019 Discipline Development Project of Peking Union Medical College faculty development program in family medicine (№201920200106).
Cite this article
Zhang, T., Mei, Q., Zhang, Z. et al. Risk for newly diagnosed diabetes after COVID-19: a systematic review and meta-analysis. BMC Med 20, 444 (2022). https://doi.org/10.1186/s12916-022-02656-y
Originally published at https://bmcmedicine.biomedcentral.com on November 15, 2022.
