Association of COVID-19 infection and the risk of new incident diabetes: a systematic review and meta-analysis

Background As the world population recovers from the COVID-19 infection, a series of acute sequelae emerge including new incident diabetes. However, the association between COVID-19 infection and new incident diabetes is not fully understood. We purpose to determine the risk of new incident diabetes after COVID-19 infection. Methods PubMed, Embase, and Cochrane Library were used as databases to search for cohort studies published from database inception to February 4, 2024. Two reviewers independently conducted the study screening, data extraction, and risk of bias assessment. A random-effects model was adopted to pool the hazard ratio (HR) with corresponding 95% confidence intervals (CI). Subgroup analysis was conducted to explore the potential influencing factors. Results A total of 20 cohort studies with over 60 million individuals were included. The pooling analysis illustrates the association between COVID-19 infection and an increased risk of new incident diabetes (HR = 1.46; 95% CI: 1.38-1.55). In subgroup analysis, the risk of type 1 diabetes was HR=1.44 (95% CI: 1.13-1.82), and type 2 diabetes was HR=1.47 (95% CI: 1.36-1.59). A slightly higher risk of diabetes was found in males (HR=1.37; 95% CI: 1.30-1.45) than in females (HR=1.29; 95% CI: 1.22-1.365). The risk of incident diabetes is associated with hospitalization: non-hospitalized patients have an HR of 1.16 (95% CI: 1.07-1.26), normal hospitalized patients have an HR of 2.15 (95% CI: 1.33-3.49), and patients receiving intensive care have the highest HR of 2.88 (95% CI: 1.73-4.79). Conclusions COVID-19 infection is associated with an elevated risk of new incident diabetes. Patients ever infected with COVID-19 should be recognized as a high-risk population with diabetes. Systematic review registration https://www.crd.york.ac.uk/prospero, identifier CRD42024522050.


Introduction
Diabetes is a chronic non-communicable disease characterized by impaired glucose metabolism that results in persistently raised blood glucose in the context of insufficient insulin caused by autoimmune-mediated destruction of pancreatic b-cells or insulin resistance combined with pancreatic b-cell insufficiency (1).Despite significant process has been made in the exploration of risk factors for diabetes and the implementation of prevention programs, there is a globally increasing incidence and prevalence of the disease (2).Early detection and intensive patient-centered management are expected to optimize the prognosis, reducing morbidity and mortality by preventing or delaying complications (2).A previous study has explored the primary risk factors of diabetes, including BMI, genetics, atmosphere, diet habit, drug use, sedentary way of life, lack of physical exercise, smoking, alcoholic beverages, dyslipidemia, hyperinsulinemia, and improved glucagon activity (3).Recently, the bidirectional interaction between coronavirus disease 2019 (COVID-19) and diabetes has been revealed (4)(5)(6)(7).COVID-19 presumably increases the risk of new incident diabetes (8,9).
The pandemic of COVID-19 caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is recognized as the greatest worldwide public health threat of this century (10).Although the World Health Organization (11) has declared that COVID-19 is no longer a public health emergency of international concern in May 2023, it continues to circulate and evolve, and remains a potentially serious risk to public health.Simultaneously, sequelae after the acute phase of COVID-19 (called long COVID) have aroused wild attention in the medical field (12).Patients with long COVID experience lingering symptoms across multiple organ systems, with common new incident conditions such as diabetes (13).Current reviews revealed an association between COVID and increased incidence of diabetes (14)(15)(16)), but Zareini et al. (17) indicated an opposite perspective.Therefore, we systematically reviewed the existing cohort studies to clarify the association between COVID-19 and the risk of new incident diabetes.

Methods
This systematic review and meta-analysis were conducted in accordance with the Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) guidelines (18).The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) platform on March 12, 2024 (CRD42024522050).

Search strategy
We systematically searched PubMed, Embase, and Cochrane Library for studies published up to February 4, 2024.No language restrictions were applied, and the search strategy combined the use of medical subject headings (MeSH) and free text.The search terms were related to COVID-19, Post-Acute COVID-19 Syndrome, Diabetes Mellitus, and risk.The full search strategies are included in Supplementary Tables 1-3.The reference lists of other published meta-analyses were also considered to identify relevant cohort studies.

Eligibility criteria
Original research studies must meet all the following criteria to be included: (1) the study design was a prospective or retrospective cohort study investigating the association between COVID-19 and the risk of new incident all-type diabetes (no prior history of diabetes); (2) COVID-19 and diabetes were defined based on medical records or International Classification of Diseases (ICD) codes; (3) the hazard ratio (HR) or odds ratio (OR) and its corresponding 95% confidence interval (CI) were reported.
The following were excluded: reviews, study protocols, and commentaries.

Study selection
Study selection was performed by two reviewers (JYZ and YZW), independently.Titles and abstracts were first screened to exclude duplicate and irrelevant articles.Thereafter, the full texts were examined to identify all eligible studies.If multiple studies conducted assessments from the same database, we include the one with more adequate data based on its sample size and follow-up duration.Any disagreements were resolved by discussing them with the third reviewer (RLX).

Data extraction
Two reviewers mentioned above (JYZ and YZW) extracted data independently consulting the guidelines on data extraction for systematic reviews and meta-analysis (19).Predesigned forms were used for data extraction, including the first author, year of publication, country, study type, data source, sample size, follow-up duration, mean age, diagnosis criteria of COVID/diabetes, type of diabetes, interval (interval between the first diagnosis of COVID and the onset of diabetes).Disagreements were resolved by consensus with all researchers (JYZ, YZW, and RLX).

Risk of bias
The quality of the included studies was assessed using the Newcastle-Ottawa scale (NOS) (20).A "star system" was used to judge the studies from three broad perspectives: the selection of participants a measurement of exposure, the comparability of the study groups, and the assessment of outcomes and adequacy of follow-up.Each assessment was carried out by two reviewers (JYZ and YZW) separately and repeatedly.Disagreements were solved by discussion with the third reviewer (RLX).

Statistical analysis
For this meta-analysis, we sought to identify HRs and 95% CI to assess the association between COVID-19 and the risk of new incident diabetes.Heterogeneity among the studies was evaluated by the c 2 -test and the I 2 -values.If I 2 > 50%, a random-effects model of analysis was used.We applied a sensitivity analysis by excluding one study each time and rerunning it to verify the robustness of the overall effects.The funnel plot was constructed to inspect and visualize publication bias, and Egger's regression test was conducted to statically assess it.Subgroup analyses were performed if two or more cohorts were identified.p-values < 0.05 were considered to be statistically significant.All analyses were performed using Stata software (Stata Corp V.14, Texas, USA).

Study characteristics
This meta-analysis included 20 cohort studies covering 60,221,176 individuals, which were published between 2021 and 2023.Out of the 20 studies, one was a prospective cohort study, while the other 19 were retrospective studies.Among all the studies included, one reported gestational diabetes mellitus (GDM), five reported type 1 diabetes (T1D), seven reported type 2 diabetes (T2D), and eight reported both T1D and T2D.The follow-up duration of participants ranges from 3 to 84 months.Additional characteristics of the included studies are shown in Table 1.

Quality assessment
According to the NOS criteria, the average score of all included cohort studies was 8, and the score for five trials (21,22,29,34,39) Studies screening flow diagram.

Publication bias
There is no evidence of a significant publication bias in the COVID-19 infection and risk of new incident diabetes revealed from the visual inspection of the funnel plot (Figure 3).Egger's test (P = 0.166) shows no publication bias in our meta-analysis either.

Main findings
We conducted a meta-analysis of 20 cohort studies covering 60,221,176 individuals, which provided a comprehensive evaluation of the association between COVID-19 and new incident diabetes.We find a significant increase in the risk of all-type diabetes among individuals after COVID-19 infection, with an overall 1.46-fold increase in risk.This indicated that COVID-19 infection might be Meta-analysis of the risk of overall diabetes after a history of COVID-19.Barrett [1]: data from IQVIA; Barrett [2]: data from Health Verity.an independent risk factor for new incident diabetes.The importance of screening, prevention, and management of diabetes for patients ever infected with COVID-19 should be emphasized.

Comparison with previous studies
Our analysis demonstrated a consistent result with previous reviews (14-16), showing that COVID-19 infection increased the risk of all-type diabetes.In addition, Li et al. (40) explored the relationship between new-onset diabetes, hyperglycemia, and COVID-19 infection, showing an elevated incidence and risk.In a review that specifically targeted T2D (41), a higher prevalence of diabetes in people with previous COVID-19 was illustrated, which further corroborated our findings.Compared to prior studies, we added more recent studies and analyzed the data in subgroups, to provide stronger evidence for the association between COVID-19 and diabetes.Simultaneously, we only included data from cases with a confirmed diagnosis of diabetes, contributing to reduced clinical heterogeneity and greater reliability.Although the risk variance between T1D and T2D is not significant in this analysis, a previous study found a higher risk for new incident T2D than T1D for all included cohorts (16).They also indicated that males with COVID-19 were associated with a higher risk of diabetes compared to females, which echoed our conclusions.We assessed the risks between subgroups of hospitalization, vaccination status, and incident diabetes for the first time, and found that an increased risk of diabetes was associated with the exacerbation of hospitalization.However, the risk of new incident diabetes in patients who received vaccination was not statistically significant.

Interpretation of findings
So far, the pathophysiological mechanism of the association between COVID-19 and diabetes is not entirely clear.It has been suggested that SARS-CoV-2 specifically induces the damage of bcells, thereby impairing insulin production (42,43).Angiotensinconverting enzyme 2 (ACE2) is the main receptor of SARS-CoV-2 to gain entry into human cells (44) Several studies have found the ACE2 expression in pancreatic b-cells (42,(45)(46)(47), leading to speculation that SARS-CoV-2 may triggers b-cell damage by penetrating the cells using ACE2 (48).In addition to ACE2, other SARS-CoV-2 related entry factors such as TMPRSS2, NRP1, and TRFC are also expressed in pancreatic b-cells, which might play roles in b-cell damage through similar mechanisms (42,49).However, the expressions of ACE2 and TMPRSS2 in pancreatic b-cells were doubted in other studies (50)(51)(52).Therefore, further research is necessary.ACE2 is a key enzyme in the renin-angiotensin system (RAS).Membrane-bound ACE2 is responsible for catalyzing the conversion from Ang II into Ang-(1-7) (53).Down-regulation of ACE2 is found in patients with COVID-19 that enhances activation of the RAS axis, resulting in decreased insulin and glucose delivery to tissues and impairment of insulin signaling pathways, all of which lead to insulin resistance (54,55).Additionally, uncontrolled inflammatory response caused by RAS imbalance might account for the potential role in pancreatic dysfunction (53, 54).
Autopsy tissue from deceased COVID-19 patients showed that local inflammation and infiltration of immune cells were associated with impairment of b-cells, causing various degrees of metabolic dysregulation (50).SARS-CoV-2 triggers a macrophage-mediated cytokine storm in which the overactivation of immune cells and persistently increasing cytokines promote excessive inflammation and further induced b-cell damage (56).SARS-CoV-2 induce a decreased chromatin-modifying enzyme SETDB2, causing increased transcription of inflammatory cytokines which impair the pancreas (56).
Steroids are used to treat COVID-19, but their pharmacological effects pose extra burden on blood glucose control (57).Steroidinduced hyperglycemia in patients with COVID-19 may be associated with an increased risk of new incident diabetes (9).A cohort study revealed a higher risk of diabetes in COVID-19 patients using glucocorticoids compared to those without steroid treatments (58), which corroborates this perspective.
Lockdowns during the COVID-19 pandemic slowed the rate of infection but caused negative mental health consequences and adverse health-related behaviors, including reduced physical activities, unhealthy eating, smoking, and binge drinking, which are risk factors for diabetes (59).Symptoms of long COVID such as fatigue, muscle pain, and dyspnea, limit exercise capacity (60), therefore sedentary lifestyles have become common.These changes of lifestyle have a series of pathophysiological effects, including metabolic consequences represented by insulin resistance, which might increase the risks of new incident diabetes (61).
In the subgroup analysis, males with a history of COVID-19 have a higher risk of new incident diabetes than females.A previous study has shown that males infected with COVID-19 are more susceptible to worse outcomes and death, independent of age (62).From another perspective, a study on rats indicated a gender-related difference of ACE2 expression, that ACE2 content was slightly lower in males compared to females (63).This might be attributed to diabetes-related pathophysiological changes.Considering vaccination has shown a potential effectiveness on improvement in long-COVID symptoms (64), it might be also helpful to prevent new incident diabetes in patients ever infected with COVID-19, which accounted for the insignificant association between COVID-19 and incident diabetes.

Implications and limitations
The pandemic of COVID-19 has placed a tremendous burden on humanity and might co-exist with us for many years.Our metaanalysis summarizes the existing evidence of the association between COVID-19 infection and the risk of new incident diabetes and shows that a history of COVID-19 is a risk factor for all-type diabetes.It suggests that the identification of high-risk groups of diabetes should cover patients with COVID-19, which is conducive to the early detection and management of diabetes.Vaccination is of critical importance for individuals to reduce the risks of adverse outcomes.More studies should be fostered to clarify the potential mechanisms underlying the COVID-related diabetes, given there might be a complex combination of pathophysiological processes behind the COVID-19 infection and new incident diabetes.
Meanwhile, this study has certain limitations.We only included cohort studies of which retrospective cohort studies are the majority.Though there is a broad and deep use of electronic databases based on validated definitions, it still cannot exclude the bias caused by misclassification, particularly for diabetes types.Moreover, the intervals between COVID-19 infection and diabetes diagnosis differ in studies, which might lead to high heterogeneity, and make it hard to discuss the risks of incident diabetes in different phases of COVID-19.Age stratification is diverse among included studies, so we did not pool related data.

Conclusions
Patients ever infected with COVID-19 had an elevated incidence and risk of new incident diabetes.However, more studies are necessary to specify the pathophysiological mechanisms underlying this association.

TABLE 1
Basic characteristics of the included studies.cohortstudieswere of relatively high quality in this meta-analysis.The score of each study is shown in Table1.

TABLE 2
Subgroup analysis for the risk of diabetes in patients with COVID-19.Publication bias of the risk of new incident diabetes caused by COVID-19 infection.