C-peptide levels and glycemic indices in COVID-19 patients

Abstract Type 2 diabetes (T2DM) is a common comorbidity in COVID-19 patients, which could favor disease progression. Our study aimed to evaluate the impact of some glycemic biomarkers on characteristic features of COVID-19. Electronic medical records of patients consecutively admitted to a COVID-19 ward were analyzed. Demographic and anthropometric as well as clinical and laboratory parameters were obtained. T2DM was present in 47% of the patients, with 22.58% of diabetic patients being newly diagnosed at admission. The most important predictors of severe COVID-19 were age (OR 1.214 [1.078-1.366], p = 0.001), creatinine levels (OR 1.018 [1.003-1.034], p = 0.017), glucose above 7.0 mmol/L at admission (OR 7.800 [2.232-27.255], p = 0.001) and HbA1c ≥ 6.5% (OR 4.840 [1.428-16.405], p = 0.011) irrespective of the presence of DM. C-peptide levels correlated positively with age, creatinine level, the severity of hypoxia and ferritin levels of patients (p < 0.05 for all) and appeared to be a significant predictor of leukocytosis. Suboptimal glycemic indices and impaired kidney function might predict COVID-19 disease worsening. Early detection of DM and optimization of glycemic control in diabetic patients should be among the priorities of the public health systems during COVID-19 pandemics.


Introduction
The risk of bacterial and viral infections, infectious complications and infection-related mortality is significantly increased in patients with diabetes mellitus (DM) compared to the non-diabetic population [1][2][3][4][5]. On the one hand, hyperglycemia increases the virulence of some pathogens, thus, aggravating morbidity and mortality in patients [6,7]. On the other hand, it might provoke immunological disturbances in the affected individuals [7]. Different immune alterations, such as decreased interferon secretion, impaired phagocytosis, decreased complement activation and abnormal cytokine secretion, have been associated with hyperglycemia [7,8]. Moreover, the increased risk of infections in many diabetic patients is accompanied by chronic cardiovascular and neurological complications, increasing infection-related mortality [9].
Recent studies have suggested that type 2 diabetes (T2DM) is a very common comorbidity in patients with severe COVID-19 [4,[10][11][12][13]. Similarly, several meta-analyses have shown increased COVID-19 severity and mortality in diabetic patients [14][15][16][17]. According to Corona et al. [18] DM was the most important predictor of mortality in hospitalized COVID-19 patients among different comorbidities. Though many observational studies on the topic have been published, most of them have been conducted in Asia, the USA and South Europe, while data from Eastern Europe are still scarce. Moreover, the impact of some glycemic biomarkers such as HbA 1c and C-peptide levels on the course of COVID-19 is poorly investigated. Therefore, our study aimed to evaluate the association between glycemic indices and characteristic features of COVID-19 in Bulgarian patients with COVID-19.

Ethics statement
All patients consented to the anonymous use of their data for scientific purposes. The Local Ethics Committee approved the study.
Patients with and without endocrine diseases were referred to the COVID-19 ward because of moderate symptoms of COVID-19 and positive COVID-19 laboratory tests previously registered in the national COVID-19 electronic system. Routine demographic and anthropometric (age, sex, weight, height, body mass index [BMI]), clinical parameters (main clinical symptoms of infection, e.g. fever, dyspnea, cough, fatigue, anosmia, tiredness, headache, myalgia, gastrointestinal complaints, vital signs, oxygen saturation [SpO 2 ] measured by pulse oximetry, comorbidities including a history of diabetes mellitus) and laboratory parameters were obtained at admission following the established hospital protocol. The laboratory tests included routine blood count and biochemical parameters, e.g. total cholesterol, HDL-cholesterol, LDL-cholesterol, triglycerides, ALAT, ASAT, creatinine, LDH, and C-reactive protein (CRP), D-dimer, ferritin, along with fasting plasma glucose, HbA 1c and C-peptide level at admission. HbA 1c ≥ 6.5% in patients without a history of diabetes was used to diagnose DM [19]. The clinical characteristics and C-peptide level of the newly-diagnosed DM patients suggested type 2 DM; in just one patient, type 1 DM was suspected, but the immunological markers were negative. The patients were treated with antibiotics, remdesivir, low-molecularweight heparin, corticosteroids and oxygen support by a nasal prong or mask per the established protocols. Specific markers such as lactate dehydrogenase (LDH), D-dimer and ferritin levels were monitored during hospitalization, and their maximal values were included in the analyses. Diabetes therapy at admission included metformin (53.3%), sulphonylurea (33.3%), insulin (6.7%) and others (SgLT2-inhibitors, gLP-1-based therapy, DPP-4-inhibitors) (26.7%). Eleven patients (16.7%) were already on corticosteroid treatment before hospital admission. Patients with progressive worsening of the condition and those receiving corticosteroids (CS) were switched to a basal-bolus insulin regimen during the hospitalization with frequent monitoring of blood glucose levels. All patients recovered; only one was transferred to an Intensive care unit (ICU) and set on invasive mechanical ventilation but survived.
The duration of hospital stay, need for oxygen therapy, leukocytosis, lymphopenia and thrombocytopenia, and elevated inflammatory markers such as D-dimer, LDH and ferritin were considered markers of unfavorable progression of the disease. According to the World Health Organization, severe COVID-19 in adults is defined as the presence of one of the following three criteria: oxygen saturation on room air <90%, respiratory rate > 30 breaths/min or signs of severe respiratory distress [20]. Since all patients with severe symptoms and respiratory rate >30 breaths/min were with oxygen saturation on room air <90%, the last criterion was used as a marker of severe COVID-19 in our group.

Data analysis
The results are presented as mean values with standard deviation [median] for continuous variables or frequency (%) for dichotomous variables. Categorical data were analyzed by the χ 2 test and Fisher's exact test. A two-tailed Spearman correlation analysis was applied to estimate possible associations. Differences between the two groups were established with the Mann-Whitney test, while the Kruskal-Wallis test was used for multiple comparisons. Logistic regression analysis estimated the risk factors associated with COVID-19 severity. Differences were considered statistically significant at the p < 0.05 level. Calculations were made through the software package MedCalc Statistical Software version 19.2.6 (MedCalc Software bv, Ostend, Belgium; https://www.medcalc.org; 2020).

Results
Type 2 diabetes was present in 43.93% (m14/f15) of the patients, while 3% (m1/f1) were with type 1 diabetes. The main characteristics of patients with DM and non-diabetic patients are presented in Table 1.
A total of 7 non-diabetic patients with COVID-19 appeared to be with newly-diagnosed DM at admission based on increased HbA 1c level/22.58% of all diabetic patients/. They presented with significantly higher glucose levels at admission in comparison to patients with pre-existing DM (21.78 ± 12.31 [22.30] mmol/L vs. 11.59 ± 6.26 [9.02] mmol/L, p = 0.033). Additionally, 65.7% of non-diabetic patients showed mild hyperglycemia (plasma glucose ≥ 5.6 mmol/L) at admission with a maximal fasting glucose level of 6.77 mmol/L irrespective of corticosteroid treatment (62.1% of non-diabetic patients without CS vs. 83.3% of non-diabetic patients on CS at admission, p = 0.639).
Creatinine levels were increased in 35.5% of patients with diabetes at admission and in just 5.7% of patients without diabetes (p = 0.004), though the prevalence of pre-existing renal disease did not differ significantly between the groups (14.3% vs. 22.6%, p = 0.525). Thus the increase in creatinine levels in COVID-19 patients was not related to pre-existing kidney disease (p > 0.05). Creatinine levels returned to normal ranges after the improvement of infection in all but two diabetic patients, one of them being with pre-existing chronic kidney failure. The prevalence of different symptoms of COVID-19 did not differ between patients with and without DM (p > 0.05 for all).
The main factors predicting the need for oxygen therapy in COVID- 19  The associations between the glucose levels at admission, HbA 1c , C-peptide levels and markers of disease worsening are shown in Table 2.

Discussion
Our data demonstrate that fasting blood glucose above 7.0 mmol/L at admission might be a crucial sign of pre-existing undiagnosed T2DM in COVID-19 patients. Italian data have shown that 27% of patients with COVID-19 were with T2DM, and in approximately 43% of them, the pre-existing disease had been diagnosed at hospital admission based on elevated HbA 1c levels [21]. However, only a few published studies have clearly distinguished the patients with acute hyperglycemia at admission with normal HbA 1c levels from diabetic patients. In a retrospective single-center cohort study conducted in Wuhan, 12.65% of COVID-19 patients were with hyperglycemia at admission with normal HbA 1c levels, while 36.75% were with DM. Both hyperglycemic groups presented with increased inflammatory markers and a significantly higher prevalence of critically ill patients compared to a non-diabetic group [22]. In contrast to these results, we did not find patients with hyperglycemia at admission among the individuals with HbA 1c < 6.5%. The discrepancy might be explained by the small number of patients in our group and the different severity of the infectious disease at admission/moderate disease in our cohort vs. moderate, severe and critically ill patients in the Asian group/. Our patients with previously undiagnosed DM presented with higher glucose levels than those with known DM, requiring more intensive treatment with a potentially increased risk of complications. Our results support the conclusions of Li et al. [23], who found an increased prevalence of severe infection and pronounced COVID-19 mortality in patients with both known and newly diagnosed DM compared to non-diabetic patients and patients with prediabetes. Moreover, the highest mortality risk was reported in the newly-diagnosed DM patients [23]. The prevalence of newly-diagnosed diabetes in our group was 10.6% (7 of 66 patients), close to the established prevalence of 14.4% reported in a recent meta-analysis, including more than 3700 COVID-19 patients [24]. However, whether the alterations of glucose metabolism that occur with a sudden onset in severe COVID-19 persist or remit when the infection resolves is unclear. It remains to be clarified whether the newly-diagnosed diabetes is the classic type 1 or type 2 diabetes or even a new type of diabetes. To address these issues, a global registry of patients with COVID-19-related diabetes has been set up with the idea to establish the extent and phenotype of the disease, its natural history and appropriate management [25].
In the current study, the suboptimal glycemic control defined as HbA 1c ≥6.5% appeared to be a stronger predictor of severe COVID-19 than the presence of DM per se. Additionally, HbA 1c levels at admission correlated positively with age, hospital stay duration, hypoxia severity, creatinine and maximal D-dimer levels. Interestingly, many studies have investigated the role of HbA 1c in COVID-19 patients with strikingly contradictory results [26][27][28][29][30][31][32][33]. Poor pre-infection glycemic control in DM patients with HbA 1c ≥9.0% has been associated with the increased need for hospitalization for COVID-19 even after adjustment for multiple confounders. Similar results have been obtained using an HbA 1c cut-off of 7% [26]. HbA 1c levels are significantly related to proinflammatory and procoagulant biomarkers that could predict progressive worsening of COVID-19 [27,28]. According to a USA study, the risk for intubation and a fatal outcome within seven days Table 2. associations between fasting glucose levels at admission, hba1c, c-peptide and other main characteristics of diabetic and non-diabetic patients with coViD-19. of hospital admission increased by 19% for every unit increase in HbA 1c value [29]. On the contrary, the large COROnADO study concluded that poor glycemic control did not predict the need for tracheal intubation or mortality on day 7 of hospitalization. These findings could not be generalized because of the significant percentage of patients with missing HbA 1c data and the inclusion of HbA 1c values measured up to 6 months before the COVID-19 hospitalization [30]. nevertheless, other studies and a large meta-analysis confirmed the lack of associations between HbA 1c levels and COVID-19 severity or mortality [31][32][33]. Further studies in diabetic patients from different ethnic groups are needed to evaluate the role of baseline glycemic control in patients with COVID-19. glucose level above 7 mmol/L at admission was a significant risk factor for severe disease progression in our cohort of COVID-19 patients. These results support the findings of several studies and meta-analyses that confirm the crucial role of hyperglycemia at admission for the prognosis of COVID-19 [33][34][35][36][37]. Thus, the early-diagnosed hyperglycemia might be used as a severity predictor in patients with COVID-19. This might be explained by the increased angiotensin-converting enzyme 2 (ACE2) glycosylation in hyperglycemic states, which promotes the SARS-CoV-2 cell entrance and the subsequent disease spread [38,39]. Therefore, a fast normalization of hyperglycemia during COVID-19 may decrease inflammatory cytokines and lower the ACE2 binding capacity for the virus; these two mechanisms consistently might help improve the prognosis of SARS-CoV-2-infected people [39]. In addition, the tight glycemic control and prevention of complications in diabetic patients might be an essential step toward a decrease in COVID-19 susceptibility, severe progression and mortality [40,41]. It would be equally important to focus efforts on the early diagnosis of DM in the general population.

Spearman's correlation coefficients
Hyperglycemia was tightly related to kidney dysfunction in our patients with COVID-19, and elevated creatinine levels predicted the disease worsening. These data support the conclusions of several studies, which have identified acute kidney injury as a decisive risk factor for severe progression of COVID-19 and increased mortality among both diabetic and non-diabetic patients [42,43]. Different mechanisms might cause acute renal impairment in patients with T2DM and COVID-19, e.g. worsening of previously undiagnosed chronic diabetic nephropathy, severe dehydration following the osmotic diuresis by long-standing hyperglycemia, and specific viral influence. SARS-CoV-2 might exert a direct cytotoxic effect on renal tubular cells expressing ACE2, but also it could induce renal damage indirectly through the aberrantly increased cytokine production [44,45]. Interestingly, the most extensive study comparing histological results of kidney biopsies between COVID-19 patients and non-infected individuals with renal abnormalities prior to the pandemic could not prove direct virus expression in the renal tubular cells [46]. Collapsing glomerulopathy and myoglobin cast nephropathy both have been found significantly more often in COVID-19 patients than in the general biopsied population [46]. On the opposite, the prevalence of diabetic nephropathy is lower in COVID-19 patients than in the general population [46], suggesting that the increased prevalence of acute renal impairment in diabetic COVID-19 patients might not result from pre-existing nephropathy. Similar inflammatory and vascular factors probably induce acute renal injury and worsening of hyperglycemia in diabetic patients suffering from severe COVID-19.
The C-peptide levels in our cohort correlated positively with age, creatinine level, BMI, the severity of hypoxia, and ferritin levels of patients. Moreover, C-peptide was a significant predictor of leukocytosis even after adjustment for age, renal dysfunction, and corticosteroid use. nevertheless, the role of C-peptide has not been sufficiently investigated in COVID-19 patients. Iias et al. [47] observed decreased insulin sensitivity in critically ill COVID-19 patients compared to patients with milder infection, but no difference in C-peptide levels between both groups. A study of 24 Italian patients (only one of them with T2DM) found statistically non-significant differences in C-peptide levels between patients with mild to moderate COVID-19 and those with severe COVID-19. However, the C-peptide levels of both groups were reported to be higher than the standard ranges, suggesting increased insulin resistance associated with COVID-19 [48]. On the opposite, a recent Asian study reported lower than average C-peptide levels in non-diabetic COVID-19 patients, which increased significantly in the post-COVID period [49]. Our results support those of the Italian study, since C-peptide levels below the typical ranges of our laboratory were not found in any of our COVID-19 non-diabetic patients, and almost half of them were even with increased C-peptide levels. We speculate that the close association between C-peptide and leukocytosis in COVID-19 patients might reflect increased insulin resistance by a developing secondary infection. Further studies are needed to support or reject this hypothesis.
An obvious limitation of our study is the small number of patients. However, it is the first study investigating the role of glycemic control and C-peptide levels in diabetic and non-diabetic patients with COVID-19 in Eastern Europe.

Conclusions
Our data suggest that age, creatinine level and blood glucose above 7.0 mmol/L at admission are essential predictors of severe COVID-19. Increased C-peptide levels in moderately ill diabetic and non-diabetic COVID-19 patients suggest increased insulin resistance and predict the development of leukocytosis. Most of the hospitalized, not critically ill COVID-19 patients with hyperglycemia at admission are individuals with previously undiagnosed DM, at least in our country. A suboptimal HbA 1c level is a stronger predictor of disease worsening than the presence of T2DM per se. Therefore early detection of DM and optimization of glycemic control in diabetic patients should be among the priorities of the public health system during COVID-19 pandemics.

Disclosure statement
no potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.

Data availability statement
Data are available upon reasonable request.