Diabetic Patients with Comorbidities had Worse Outcomes When Suffered with COVID-19 and Acarbose might have Protective Effects

Background: Previous studies showed that diabetes was a common comorbidity of COVID-19, but the effects of diabetes or antidiabetic drugs on the mortality of COVID-19 have not been well described. To investigate the outcome of different status (with or without comorbidity) and anti-diabetic medication before admission of diabetic patients after SARS-CoV-2 infected, we collected clinical data of COVID-19 patients from Hubei Province and compared between diabetes and non-diabetes. Methods: In this multicenter and retrospective study, we enrolled 1,422 cases of consecutive hospitalized patients from January 21, 2020 to March 25, 2020 at six hospitals in Hubei Province, China. The primary endpoint was in-hospital mortality. Results: Diabetes patients were 10-years older than non-diabetes (p<0.001), had higher prevalence of comorbidities such as hypertension (p<0.001), coronary heart disease (p<0.001), cerebrovascular disease (CVD) (p<0.001), chronic kidney disease (CKD) (p=0.007). The incidence of mortality (p=0.003) were more prevalent among the diabetes group. Further analysis revealed that diabetes patients who took alpha-glucosidase inhibitor (AGI) had lower mortality rate(p<0.01). Multivariable Cox regression showed that male sex, hypertension, CKD, CVD, age were risk factors for the mortality of COVID-19. Survival curve revealed that, compared with diabetes only group, the mortality was increased in diabetes with comorbidities (p=0.009), but had no signicant difference in the non-comorbidity group, p=0.59). Conclusions: Patients with diabetes had worse outcome when suffered with COVID-19, however, it was not associated with diabetes itself but the comorbidities. Furthermore, the administration of AGI could reduce the risk of death in patients with diabetes. or the Mann-Whitney U test or Kruskal-Wallis H test for skewed continuous variables. The Pearson’s chi-square, the Fisher’s exact test were performed for unordered categorical variables. The Mann-Whitney U test or the Kruskal-Wallis H tests were used for ordered categorical variables. To explore the risk factors associated with mortality, multivariable COX regression models were performed. The Kaplan-Meier plot was performed to compare the survival probability for diabetes and non-diabetes, and among the patients with no comorbidity, only diabetes and diabetes with comorbidities by log-rank test. Additionally, we didn’t process the missing data. The statistical analyses were conducted with SPSS (version 25·0). A two-sided p value less than 0·05 was considered statistically signicant.

Systolic blood pressure (SBP) was higher in diabetes group (128mmHg [120-140] vs 124mmHg [119-135]). Moreover, decreased blood oxygen saturation (lower than 93%) occurred more frequently in the diabetes group versus the non-diabetes group (19.8% vs 19.6%) on admission. Chest CT scan revealed that the incidence of bilateral lesions was higher (94% vs 80.1%) in the diabetes group compared to non-diabetes patients.
There were numerous differences of laboratory results between diabetes group and non-diabetes group with COVID-19 (Table2). In addition, between-group comparison with only diabetes group was performed. The baseline characteristic and radiological nding were also summarized in Table 1. Diabetes patients with comorbidities were the oldest among three groups . There was a signi cant difference in blood oxygen saturation, respiratory rate and SBP among three groups but no signi cant differences in the comparison of non-comorbidity group and only diabetes group or the comparison of diabetes only group and diabetes with comorbidities group.
Chest CT scan indicated that diabetes only group had more incidences of bilateral lesions than non-comorbidity group.
Although numerous differences of laboratory ndings among non-comorbidity group, diabetes only group and diabetes with comorbidities group (Table 2), only ten item had statistical signi cance between non-comorbidity group and diabetes only group,   Table 3, 1223 of the total 1331 patients (91.9%) were discharged from the hospital, the rate of mortality of diabetes group was higher than non-diabetes group (7.2% vs 13.6%). Kaplan-Meier survival analysis for all-cause mortality in COVID-19 patients is shown in Figure 2. The overall survival rate was signi cantly lower in the diabetes group (log-rank p< 0.01, Figure 2A).
Compared with non-diabetes group, the incidence of severe cases was more prevalent among the diabetes group (34.6% vs 21.7%).
However, there were no signi cant in ICU admission, length of hospital stay and the duration from admission to ICU between both groups. Diabetes group had higher rate of ARDS (11% vs 5.7%) and hypoproteinemia (15% vs 6.5%). As regard other complications, ACI, AKI, secondary infection, shock, no observable difference was identi ed between diabetes group and non-diabetes group.
Treatment and primary outcome of non-comorbidity group and diabetes only group were no difference (Table 3), results for all-cause mortality were similar in both groups (log-rank p=0.59) ( Figure 2B). As regard the second endpoint, there was no difference between both groups expect hypoproteinemia (5.0% vs 16.9%). Likewise, there was a similar frequency of COVID-19 pharmacological therapy in only diabetes patients versus diabetes with comorbidities patients, however, the latter was more likely to receive mechanical ventilation (10.8% vs 18.3%), had higher incidence of mortality (4.6% vs 18.3%), shock (0 vs 1.6%) and more severe cases (21.5% vs 41.3% ). The overall survival rate was lowest in the diabetes with comorbidities group (log-rank p< 0.0001) ( Figure 2B).
Clinical characteristics, laboratory results, treatment and outcome of diabetic COVID-19 patients using AGI and matched non-AGI user Of 191 diabetic patients with COVID-19, 77 cases using metformin, after sex, age-matching, there was 46 patients using metformin and sex, age-matched 46 non-metformin user in this part. Table S3 showed that the length about onset of symptom to hospital admission was longer in AGI group when compared with matched non-AGI group, which indicated that the symptoms in former patient might relatively mild.. Noteworthy, parts of in ammation response related laboratory results, such as WBC (4.83×10 9 (Table 6). Furthermore, these differences were not related to glucose control as serum level of both groups was comparable.

Independent risk factors for mortality of patient with COVID-19
Among the included 1131 patients, multivariable Cox regression ( the mortality of COVID-19. Age was also the risk factor for mortality of COVID. However, diabetes alone was not the independent risk factor for mortality of patients with COVID-19.

Discussion
Lots of studies had demonstrated that diabetic patients had higher risk of mortality in COVID, as well as developing more severe cases [2,4,11,12]. Guo et al [12] reported that diabetes was a risk factor for the progression and prognosis of COVID-19. However, Shi et al pointed out that diabetes was not independently associated with COVID mortality, while commonalities, such as hypertension, cardiovascular disease played more important roles in contributing to the in-hospital death of COVID-19 patients which was relatively limited-sized [13]. In this study which had relatively rich clinical data, we found that diabetes alone was not the independent risk factors for in-hospital mortality of COVID-19, but comorbidities such as hypertension, CKD were risk factors, this result was consistent with the previous study [13]. Partially consistent with the previous studies, our study found that, compared with non-diabetic patients, diabetic COVID-19 patients were older, had worse outcome including higher rate of mortality, severe cases and ARDS, presented severe in ammation response, lung and coagulation dysfunction [2,11,12,14]. Additionally, diabetic patients had increased level of urea nitrogen and decreased level of albumin. These abnormalities indicated that COVID-19 may be associated with progressive organ injury in patients with diabetes. Pre-existing hypertension, CHD, CVD, and CKD had higher frequencies in the diabetic group. Recent studies reported that patients with hypertension cardiovascular, CKD, CVD were more likely to develop severe cases [4,6,[15][16][17] , so we compare the diabetic COVID patients without comorbidity and COVID atients without any comorbidity in order to identify if diabetes without comorbidity was a risk factor of COVID-19. In our study, there was no difference in the outcome between non-comorbidity group and diabetes only group. Shi et al reported that even thought COVID-19 patients with diabetes had worse outcomes; diabetes was not independently associated with in-hospital death, which was consistent with our results [16]. Besides, most laboratory results were comparable between non-comorbidity group and diabetes only group except CRP, albumin, sodium, urea nitrogen, HDL-C and, of course blood glucose. As we all known, CRP was an in ammatory biomarker, which was related to glucose homoeostasis, obesity and atherosclerosis [18], and it was independently related to insulin sensitivity [19]. Meanwhile, insulin resistance was a main characteristic of type 2 diabetes, since CRP was related to the chronic in ammatory situation, and the level of WBC, NEU, LY, which re ected the acute infection of disease pathogen, were not statistical signi cance, we inferred that diabetes itself didn't increase the degree of the in ammation after SARS-CoV2 infection.
Diabetic patients with comorbidities were more serious when compared with diabetes only group and non-comorbidity group. The mortality was higher in diabetes with comorbidities group, but the difference between both diabetes groups had no relation of FPG, because the serum of FPG in both diabetes groups was comparable. Patients of diabetes with comorbidity was 10 years older than the patients who had no comorbidity expect diabetes, furthermore, age ≥65 was associated with greater risk of development of death [4]. As described above, patients with hypertension, CVD were more likely to develop severe cases [4,16,17,20]. Furthermore, our analysis indicated that age, hypertension, CKD, CVD were the risk factors for mortality of COVID-19. Since diabetes with comorbidities group higher prevalence of hypertension, CKD and CVD, it was no doubt that this group had worse outcome than patients who just had diabetes.
One unanticipated result was that AGI, not metformin, could improve prognosis through decreased in ammation degree, which was independent of blood glucose level. In addition, acarbose accounted for 97% of the glycosidase inhibitors used. Feng et al reported acarbose could effectively block the metastasis of enter virus 71(EV71) from the intestine to the whole body [21]. EV71 was one of the main cause of hand-foot-and-mouth disease (HFMD), its infection relied on the interaction of the canyon region of its virion surface and the glycosylation of SCARB2 protein which was the cellular receptor of EV71 infection. Dang et al found that acarbose not only to inhibited cellular receptors of various glycosylated viruses but also competitively blocked the canyon region of EV71 virion surface [22], s, so it blocked the metastasis of EV71 from the intestine. Angiotensin converting enzyme II (ACE2) was SARS-CoV2 cell entry receptor [23], glycosylation sites play an important role in the combination of SARS-CoV2 and its receptor [24,25].
Chloroquine was reported that blocked SARS-CoV infection through interfering with the glycosylation of cellular receptors [26]. As previously stated, acarbose inhibited glycosylation of EV71 receptors, meanwhile, diabetic COVID-19 patient who was treated with acarbose had better outcome than the patients who wasn't treated, it can thus be suggested that acarbose could improve prognosis of COVID-19 infection by inhibited the glycosylation of ACE2. In addition, compared to the non-AGI group, the lower level of WBC, NLR, and CRP in AGI group which means decreased in ammatory response and further supported the anti-SARS-CoV2 function of acarbose. Furthermore, previous study showed that acarbose could change gut microbiota, then bene cially regulates the body's immune function [27]. Recent study revealed that fetal microbiomes change was happened in COVID-19 patients, characterized by depletion of bene cial commensals, and enrichment of opportunistic pathogens [28]. Therefore, we inferred that acarbose might increase the baseline abundance of microbiota which had inversely correlated with COVID-19.
As previous studies reported that metformin has multiple additional health bene ts in diabetic patients [29][30][31], we anticipated that metformin would improve prognosis after COVID-19 suffering, however, the results was unexpectedly. Scanning the literature, we found that metformin improves ACE2 stability through AMPK [32], which means metformin may increase ACE2 availability. In addition, the median level of FBG was higher in metformin users than non-users, as previous study reported that improving glycemic control substantially reduced the risk of mortality for COVID-19.
The study has some limitations. Firstly, due to the retrospective and multiple-center study design, some information, such as patients' exposure history, diabetes medication and several laboratory items were not available for all patients. Secondly, samples were only from Hubei province, China, thus more studies in other regions even other countries might get a more comprehensive insight of COVID-19. However, this study is one of the largest retrospective and multicenter studies among patients with COVID-19. Additionally, this study is one the rst study to investigate the in uence of diabetes medications in diabetic patients with COVID-19.
The relatively aplenty clinical data and numerous events also strengthen the results. The conclusion will help clinicians to identify high-risk patients and choose suitable diabetes medication for diabetic patients.
In conclusion, patients with diabetes had worse outcome when suffered with COVID-19; however, it was not related to diabetes itself but the comorbidity, such as hypertension, CKD and CVD. Furthermore, the administration of acarbose could reduce the risk of death, ARDS, shock in patients with diabetes. Availability of data and materials: The data used and/or analyzed during this study are available from the corresponding author on request.         neutrophil lymphocyte ratio; Hb, hemoglobin; PLT, platelet; PCT, procalcitonin; CRP, C reactive protein; PT, prothrombin time ; ALB, albumin; ALT, alanine aminotransferase; AST, aspartate aminotransferase ; ALP, alkaline phosphatase ; TBIL, total bilirubin ; BUN, urea nitrogen ; UA, uric acid ; CK, creatine kinase ; LDH, lactate dehydrogenase ; Hs-cTnI, hypersensitive troponin I ; TG, triglyceride; TCH, total cholesterol; LDL-C, low density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; FBG, fasting blood glucose; HbA1C, glycosylated hemoglobin.           Figure 1 Flow chart of patient recruitment.

Figure 2
Kaplan-Meier survival curves of in-hospital mortality among patients with COVID-19. A, Kaplan-Meier survival curves for in-hospital mortality between diabetes and non-diabetes from hospital admission. B, Kaplan-Meier survival curves in-hospital mortality comparison of patients with non-comorbidity, diabetes only and diabetes with comorbidities from hospital admission. Compared patients with non-comorbidity and diabetes only from hospital admission ( Log rank test, p=0.590). Compared patients with onlydiabetes group and diabetes with comorbidities from hospital admission ( Log rank test, p=0.009).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.