In quantifying the risk of mortality due to comorbidity in patients with COVID-19, we showed that DM increased the odds of death significantly in all patients. This finding was concordant with previous studies [4, 10, 11, 15, 16]. On the other hand, the association of CVD with COVID-19 mortality lost its significance in the multivariate model after adjusting for age. It shows the great importance of age in morbidity and mortality of cardiovascular complications, which could be justified by the strong correlation of age and CVD incidence. We also conducted a sensitivity analysis, and we showed that DM remained a significant risk factor for mortality in patients with laboratory-confirmed COVID-19. Zhou F, et al. study showed that odds of in-hospital death was higher in patients with DM and coronary heart disease (CHD), and older age was determined as a risk factor for death in adult patients with COVID-19 [15]. Moreover, we showed that diabetic patients without other comorbidities (only diabetes) were at a higher risk of mortality. This finding was more similar to Ghou et al., study [10]. Using data from 174 COVID-19 patients in China, they found that diabetic patients without other comorbidities (n= 24) are at a higher risk for severe pneumonia, death, as well as release of tissue injury related enzymes, and excessive inflammation responses [10]
Based on our finding, lymphocyte count, creatinine and CRP level were significant predictors for death of COVID-19 in diabetic patients. While lymphocyte count was inversely associated with the death following COVID-19, creatinine and CRP level had direct association with it.
The defect of cellular immune response and cytokine storm may play roles in the development of acute respiratory distress syndrome [17]. Since diabetic patients suffer from a less robust immune system due to chronic hyperglycemic and inflammatory states, DM could be a risk factor for COVID-19 progression and death [17-20] Moreover, there are conflicting results regarding the use of angiotensin-converting enzyme 2 (ACE2)-increasing drugs in COVID-19 patients [20, 21]. Some studies proposed the harmful effects of these drugs on infection severity while the other ones found the drugs useful for preventing pneumonia and the risk of mortality [16, 22]. Fang et al. study suggest that patients who are treated with ACE2-increasing drugs are at higher risk for severe COVID-19 infection and, therefore, should be monitored for ACE2-modulating medications, such as ACE inhibitors [16]. Although in a retrospective cohort study it has been found that there were no association between use of ACE2-increasing drugs and COVID-19 test positivity [23], more studies are needed regarding the effect of these drugs on COVID-19 severity.
Based on our findings, the presence of DM might predispose patients with COVID-19 to develop a more severe form of the disease, leading to the worst consequences and death. It could be explained by the weak immune system in diabetic individuals, especially in those with poor glycemic control [24, 25]. This finding highlights that, in the case of access to a safe and efficient vaccine against the COVID-19 virus, these patients should be put in higher priority as a high-risk group.
In stratified analysis, we reanalyzed the data for gender and age groups, and we showed that DM only in females and patients younger than 65 years increased the risk of death. Previous observations confirmed the age and gender differences in glycemic control and treatment responses in diabetic patients [26-30]. According to the evidence, women with DM are less likely to reach the ideal level of hemoglobin A1c (HbA1c) compared with men [26, 27, 30]. The risk of all-cause mortality was higher in females with DM. Moreover, surveys showed that younger adults have poorer glycemic control compared to older diabetic patients [28, 29]. These issues are of great importance in interpreting the findings of epidemiological studies. Furthermore, Bello-Chavolla OY, et al. declared early-onset diabetes and obesity as risk factors for mortality in COVID-19 patients in Mexico [31]. Therefore, newly diagnosed diabetes mellitus with uncontrolled hyperglycemia may be linked to increased risk of COVID-19 fatality. On the other hand, since ACE2 receptors are expressed in pancreatic beta cells, it is plausible that SARS-CoV-2 cause alterations in glucose metabolism which result in complication of the pathophysiology of preexisting diabetes or new-onset diabetes [31]. Although it has been shown that obesity mediates 49.5% of the COVID-19 lethality which attributed to diabetes in Mexican population [32], we had no data on weight status of patients for adjustment. To date, the exact mechanisms underlying strong association between obesity and COVID-19 severity were not clarified. However, it may be the consequence of low-grade chronic inflammation and suppressed immunity in obese persons [33].
In the present study, out of 2957 adult COVID-19 hospitalized patients, 1412 patients were laboratory-confirmed COVID-19. In our study, the diagnostic criteria of COVID-19 were based on abnormal chest CT scans as well as clinical manifestations of infection, and not all the cases had positive RT-PCR test may be due to low sensitivity of the test. Moreover, we did this test only one time without any repetitions, so we believed that detection according to chest CT and clinical symptoms of infection could be more reliable. It should be noted that a report of 1014 cases in China also showed that chest CT had higher sensitivity for the diagnosis of COVID-19 as compared with initial RT-PCR from swab samples [14].
In the present study, the most common symptoms on admission were cough, followed by shortness of breath and fever. The most common symptoms on admission were also reported fever and cough in Wuhan, China [15]. In comparison with survivors, non-survivor were older, and a higher percentage of them were presented with shortness of breath, O2 saturation <93%, and required invasive mechanical ventilation on admission.
According to the previous investigations, older age has been nominated as an important risk factor for mortality in SARS and Middle East respiratory syndrome (MERS) [33, 34]. Consistent with our observation, studies on COVID-19 have also reported that the increase in age was associated with a high mortality rate [15, 36]. The age-dependent defects in immune cell function and increased production of inflammatory cytokines may cause a poor immune response in the control of viral replication and result in poor outcomes [37].
In this study, CVD and DM were the most common comorbidities seen in 10.6 % and 9.0% of infected patients. In a meta-analysis of eight studies with 46248 infected patients, the most prevalent comorbidities among patients were hypertension (17%), DM (8%), and CVD (5%) [5]. In the present study, 44.5% of non-survivors had at least one of comorbidities while the prevalence of any comorbidities was 23.2% in survivors, raising awareness of the need for earlier monitoring and greater supportive care in this vulnerable group. In accordance with previous observations, we showed that the prevalence of DM and CVD in non-survivors was higher than in survivors. In a retrospective, multicentre cohort study in Wuhan, China, which has been conducted on 191 patients with laboratory-confirmed COVID-19, the risk factors associated with in-hospital death had been explored [15]. The prevalence of DM and CVD was recorded respectively 19% and 8% in hospitalized patients and 31% and 24% in non-survivors [38], indicating a higher prevalence of comorbidities in infected non-survivors in agreement with our findings. Moreover, DM has been declared as one of the most common comorbidities in deceased COVID-19 patients in some of the European countries [18].
The present study is among the first studies with the approach of exploring the risk factors of mortality in COVID-19 patients. We included different diagnostic approaches for COVID-19 diagnosis, firstly based on the clinical symptoms and chest CT scan as well as laboratory confirmation tests. Furthermore, we performed subgroup analyses for gender and age groups. Besides, our study has some limitations. Firstly, due to the retrospective design of the study, some laboratory tests were not collected for all patients, and these missing data might lead to bias of clinical characteristics. Considering that cases with more severe disease hospitalized, the percentage of mortality in our study cannot reflect the true fatality ratio of COVID-19 and might limit the interpretation of our findings. Moreover, the existence of DM and CVD was self-reported data, which should be cautiously interpreted because of probable reporting bias. It should be noted that we had no access to details about diabetic patients such as duration of diabetes and kind of treatments so these issues were not considered in our analysis, although we follow the same protocol for controlling DM all over the province. We also missed the effects of some comorbidities like obesity on the mortality rate of COVID-19. Moreover, HbA1c was not measured in all diabetic patients in our study, and therefore the glycemic control level was unclear in our patients.