Introduction
Coronavirus disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2, earlier known as 2019-nCoV) (1). The outbreak of COVID-19 rapidly caused a global health crisis (2), (3) and the World Health Organization has announced COVID-19 as a pandemic.
Diabetes mellitus (DM) is one of the most frequent comorbidities reported in patients infected with COVID-19. The inflammatory process associated with DM and chronic high levels of glucose in the blood can lead to low immunity response which aggravate infections in diabetic patients (4). There are several evidences of increased incidence and severity of COVID-19 in patients with DM reported with higher risk for intensive care unit (ICU) admission and substantial mortality (5). In addition, obesity has been associated with disease severity and guidelines recommend that obese patients especially those with severe obesity should take extra measures to avoid COVID-19 infection (6). It has been reported that compared with non-ICU patients, ICU patients have a higher proportion of diabetes. Additionally, it is also suggested that DM may affect the clinical manifestations and patient’s disease progression (7).
A recent retrospective study suggested that Type 2 DM and obesity may have contributed to disease severity and mortality in COVID-19 critically ill patients (8). Another retrospective cross-sectional study conducted in United Kingdom reported that hospitalized COVID-19 patients with DM had longer LOS in hospital than patients without DM. Same study also documented that elder COVID-19 patients with DM and patients without diabetic ketoacidosis were less likely to survive compared to younger patients and patients with diabetic ketoacidosis, respectively (9). Similar to these findings, another cohort study conducted in China, retrospectively reviewed 258 consecutive hospitalized COVID-19 patients with or without DM suggested that DM is associated with increased disease severity and a higher risk of mortality in patients infected with COVID-19. In addition, impaired fasting glucose and diabetes at admission were significantly associated with higher risks of adverse outcomes among COVID-19 infected patients (10). Few observational studies suggested that diabetes was significantly associated with higher risks of composite adverse endpoints (death, ICU admission, and mechanical ventilation) (10).
The increased rate of those suffering from diabetes combined with the prevalence of COVID-19 suggest that the care for diabetic patients must be increased in order to reduce any further complications and the risk of death. The poor outcomes of patients with diabetes and COVID-19 indicated that more supervision is required in these patients. Hence with the high global burden of diabetes and the increasing number of COVID-19 cases, there is a need to understand the association of diabetes and COVID-19 related outcomes in COVID-19 infected patients.
To date no meta-analysis conducted to explore possible risk factors and resource utilization in DM patients infected with in COVID-19. Therefore, we aimed to conduct a meta-analysis to find out association between DM and poor outcomes in patients with COVID-19 infection.
Materials and methods
This systematic literature review and meta-analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement (11) and it was prospectively registered on Prospective Register of Systematic Reviews (PROSPERO) (CRD42020213791).
Data sources and search strategy
Databases, the Medical Literature Analysis and Retrieval System Online (MEDLINE)/PubMed, and Web of Science were searched using a comprehensive search strategy to identify eligible studies published in the English language from inception to January 2021. Search terms used were “diabetes mellitus”, “diabetes”, “Coronavirus”, “COVID-19”, and “SARS-CoV-2. Additionally, manual searches were performed on Google Scholar and bibliographies of relevant articles and previous systematic review (12) were also screened to identify other significant studies.

Study selection and data extraction 

Two reviewers independently performed the selection of study based on an initial screening of identified titles and abstracts followed by second screening of full-text articles. Studies were considered eligible if they met the following criteria: (i) studied subjects were diabetic and non-diabetic exposed with COVID-19; (ii) outcome of interest reported comorbidities, clinical manifestations of COVID-19; (iii) studies using observational study design (retrospective cohort, case-control, prospective cohort, and cross-sectional studies).
Two independent authors performed study screening and data extraction judiciously according to inclusion criteria. Any discrepancies were resolved through discussion of the third author until consent was achieved. A standard data extraction format was used to capture the study information, including name of the first author, year of the publication, country, study design, age, gender, settings, country, database used, study population, no. of comorbidities/symptoms reported, and other relevant information.

Quality assessment 

Two authors independently assessed the quality of the methodology for each study using the Newcastle-Ottawa scale (NOS) (13).  The NOS comprised of three scales; selection (maximum of four stars), comparability (maximum of two stars), exposure/outcome (maximum of three stars) and graded out of 9 points (stars). This assessment critically appraised the internal (systematic error) and external validity of the studies.  We consider studies with a score of 6 or greater as high quality. Any disagreement was resolved through consensus.

Data analysis

We carried out meta-analysis to calculate the pooled estimates of odds ratio/mean difference (OR)/(MD) and 95% confidence intervals (CI). Random effects model (DerSimonian and Laird) was used for the overall pooled estimate meta-analysis.
To provide a quantitative estimate of the association of the symptoms/complication of interest with severity outcomes (ICU admission, length of stay (LOS) or need for mechanical ventilation) or deaths were calculated from crude frequency of exposed and non-exposed cases.
Heterogeneity among studies was assessed with the Cochran chi-square (χ2) and quantified with the I2 and tau-square (𝜏2). Statistical heterogeneity was reported using the I2 statistic, with results ranging from 0 to 100% and values of 25, 50 and 75% representing low, moderate and high level of heterogeneity, respectively. All the statistical analysis was conducted using Review Manager 5.3 (Nordic Cochrane Centre, Cochrane Collaboration, 2014). A p values of <0.05 was considered statistically significant.

Results

Literature search and study inclusion

The PRISMA flow diagram summarizing the process of study selection is shown in Figure 1. The electronic search identified 1237 potentially relevant records, of which 1100 were excluded after screening the title and abstract. The remaining 61 full-text articles were assessed for eligibility and 47 were further excluded due no information on COVID-19 infected non diabetes patients, mortality, resource utilization, and study design. Total 14 articles were included for the systematic review and meta-analysis.

Study characteristics and quality assessment

Of 14 included observational studies, all studies were retrospective. The general characteristics of included studies are reported in Table 1. Among included 14 studies, 12 were conducted in China, 1 in USA, and 1 in England. This meta-analysis involved 5697 patients (diabetic=1428 and non-diabetic=3644).
The quality assessment of the included studies is shown in supplementary Table S1. The NOS results showed that the average score was 7.2 (range 6-9) for all included studies. Therefore, the quality of included studies suggested good (14).

Symptoms associated with DM and COVID-19 infection

Out of 14 included studies, total 11 studies reported symptoms including fatigue, fever, headache, myalgia, nausea/vomiting, anorexia, cough, diarrhoea, dyspnoea, palpitation, pharyngalgia, shortness of breath, polypnea, sore throat, sputum, chest pain and expectoration. Fever and cough followed by fatigue and diarrhoea were most commonly reported symptoms across the included studies.
Results from the pooled meta-analysis found non-significant symptoms trends in DM patients infected with COVID-19 compared to non-diabetic COVID-19 patients infected. However, dyspnoea was significantly association with DM [2.30 (1.37, 3.84); p value=0.001;I 2=72%; p value=0.003] compared with non-diabetic patients infected with COVID-19 (Table 2).

Comorbidity associated with DM and COVID-19 infection

Total 11 studies reported comorbidities/complications including hypertension, CLD, COPD, thyroid disease, AKI, cancer, cerebrovascular disease, CKD, digestive disease, septic shock, and AKI. COPD, CVD, and hypertension followed by cerebrovascular disease, cancer and CKD were most commonly reported co-morbidities/complications across the included studies.
From reported co-morbidities, CVD, hypertension, AKI, cerebrovascular disease, AKI and ARDS were significantly associated with DM in COVID-19 infected patients compared to non-diabetic patients (2.91 [2.34, 3.63], p<0.00001; 2.19 [1.39, 3.46], p=0.0008; 3.59 [1.46, 8.84], p=0.005; 2.09 [1.22, 3.61], p=0.008; and 3.01 [1.63, 5.58], p=0.0005; and 3.40 [2.09, 5.55], p<0.00001; respectively). However, no significant association was observed with CLD, COPD, thyroid disease, cancer, CKD, digestive disease, and septic shock in DM patients infected with COVID-19 compared to non-diabetic patients (Figure 2. and supplementary Table S2).

Diabetes and outcome

It includes mortality, recovery, length of stay (LOS), mechanical ventilation, need for ICU, discharge from hospital, and hospitalization (Figure 3. and supplementary Table S3).

Association between mortality/recovery and DM in COVID-19 infected patients

Total 12 and 5 studies reported outcome as mortality and recovery respectively. Pooled results from meta-analysis found significant association between mortality and DM compared to non-diabetic patients infected with COVID-19 (2.46 [1.68, 3.58], p <0.00001;I2 : 62%; p=0.002). However, no significant association was found between recovery with DM compared to non-diabetic COVID-19 infected patients (0.48 [0.21, 1.07], p=0.07;I2 : 85%; p <0.00001) (Figure 3).

Association between healthcare utilization and DM in COVID-19 infected patients

LOS and hospitalization were not significantly differing between diabetes and non-diabetes patients infected with COVID-19 (1.91 [-1.12, 4.94], p= 0.22; I2:91%, p < 0.00001 and 1.45 [0.51, 4.08], p= 0.49; I2: 59%, p=0.12 respectively) (Figure 4). However, discharge from hospital was significantly higher in non-diabetic patients infected with COVID-19 compared to diabetic patients (0.52 [0.34, 0.81], p = 0.004; I2: 6%, p=0.35). On the other hand, ICU admission and use of mechanical ventilation was significantly associated with DM and COVID-19 infected patients compared to non-diabetic patients (2.79 [1.79,4.34], p < 0.00001; I2: 0%, p=0.53 and 3.33 [2.05, 5.42], p < 0.00001; I2: 48%, P=0.09, respectively) (Figure 3).

Discussion

This systematic review and meta-analysis of 14 studies aimed to evaluate the association between DM and COVID-19 infected patients. Results from current study showed that mortality, ICU, ARDS, mechanical ventilation, discharge from hospital, hospitalization and co-morbidities (CVD, hypertension, AKI, cerebrovascular disease, acute cardiac injury) are significantly associated with DM in COVID-19 infected patients.
Diabetes is one of the most frequent comorbidities reported in patients with COVID-19. The prevalence of diabetes among patients with COVID-19 varied in different studies. Studies in Chinese patients reported prevalence rates ranged from 5.3 to 8.2% (3-5), while a recent study involved 5,700 patients reported that 33.8% had diabetes (6). Several studies have reported that diabetes and uncontrolled glycemia were significant predictors of severity and mortality in patients infected with lower respiratory tract infections (10).
Previous studies found increased severity of COVID-19, caused by infection with SARS-CoV-2 in patients with DM. Interacting with other risk factors; hyperglycaemia might modulate immune and inflammatory responses, thus predisposing patients to severe COVID-19 and possible lethal outcomes. Potential pathogenetic links between COVID-19 and DM include effects on glucose homeostasis, inflammation, altered immune status and activation of the renin-angiotensin-aldosterone system (15).
Presence of DM and the individual degree of hyperglycaemia seem to be independently associated with COVID-19 severity and increased mortality (12, 16). Furthermore, the presence of typical complications of DM (CVD, heart failure and CKD) also increases COVID-19 mortality (11,19). In current meta-analysis we found higher mortality rate in DM COVID-19 infected patients compared to non-DM COVID-19 infected patients.
COVID-19 can progress to ARDS, which requires positive pressure oxygen and intensive care therapy (9). Previous systematic review and meta-analysis conducted by Huang et al. (2020) (17) showed that ARDS and disease progression was significantly associated with DM in COVID-19 infected patients. In addition, a retrospective case study conducted by Chen et al. (2020) reported that SARS-CoV-2 can cause both pulmonary and systemic inflammation, leading to multi-organ dysfunction in patients at high risk (18). ARDS, respiratory failure, sepsis, acute cardiac injury, and heart failure were the most common critical complications during exacerbation of COVID-19 (18). In line with previous finding, our meta-analysis also found significant association between DM and ARDS and other co-morbidities in COVID-19 infected patients compared to non-DM COVID-19 infected patients.  In addition, various studies reported that patients with diabetes are more likely than healthy people to develop COVID-19 disease and complications such as ARDS and even death (19-22).
Study reported the overall symptoms of COVID-19 in all patients are sore throat, fever, dry cough, fatigue, and diarrhoea (23, 24). To support our findings, international diabetes federation (IDF) statement reported that symptoms in diabetes patients not different with other COVID-19 patients (25). But there is an agreement that symptoms are more developed among diabetic patients (26, 27).
Another study conducted in Greece reported that Type 2 diabetes and obesity may have contributed to disease severity and mortality in COVID-19 critically ill patients (8). HbA1c is associated with inflammation, hypercoagulability, and low SaO2 in COVID-19 patients, and the mortality rate (27%) is higher in DM patients (28).
In a case-report on a patient with both COVID-19 and diabetes, the patient was discharged from the hospital after 15 days. The survival term of the non-survivors is likely to be within 1–2 weeks after ICU admission. The severity of SARS-CoV-2 pneumonia poses great strain to hospital critical care resources, especially if they are not adequately staffed or resourced (29). In a retrospective cohort of 312 patients with COVID-19, we found that diabetes was associated with higher risks of composite adverse endpoints (mechanical ventilation, admission to ICU, or death) and mortality, and IFG was also associated with higher risk of mortality. Similarly, in the current meta-analysis we also observed that ICU admission, mechanical ventilation, discharge from hospital and hospitalizations are significantly associated with DM COVID-19 infected patients compared to non-DM COVID-19 infected patients. This finding could be a potential outcome to support development of health economics model for the use of diabetic medication in DM COVID-19 infected patients and help decision making process.
Association between DM and COVID-19 was also explored in the previous published systematic review and meta-analysis, but these studies considered limited studies and low sample size (references). Interestingly, no studies yet compared DM COVID-19 infected patients and non-DM COVID-19 infected patients. To best of our knowledge this is the only largest meta-analysis to evaluate risk factors and resource utilization associated with DM COVID-19 infected patients compared to non-DM COVID-19 infected patients.