Chronic dysglycemia and risk of SARS‐CoV‐2 associated respiratory failure in hospitalized patients

Abstract Background Diabetes is common among patients with severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2)‐induced respiratory failure. We aimed to investigate the relationship between different stages of chronic dysglycemia and development of respiratory failure in hospitalized SARS‐CoV‐2 positive patients. Methods In this retrospective observational study, we included 385 hospitalized SARS‐CoV‐2 positive patients at Karolinska University Hospital, Sweden with an HbA1c test obtained within 3 months before admission. Based on HbA1c level and previous diabetes history, we classified patients into the following dysglycemia categories: prediabetes, unknown diabetes, controlled diabetes, or uncontrolled diabetes. We used multivariable logistic regression analysis adjusted for age, sex, and body mass index, to assess the association between dysglycemia categories and development of SARS‐CoV‐2‐induced respiratory failure. Results Of the 385 study patients, 88 (22.9%) had prediabetes, 68 (17.7%) had unknown diabetes, 36 (9.4%) had controlled diabetes, and 83 (21.6%) had uncontrolled diabetes. Overall, 299 (77.7%) patients were admitted with or developed SARS‐CoV‐2‐induced respiratory failure during hospitalization. In multivariable logistic regression analysis compared with no chronic dysglycemia, prediabetes (OR 14.41, 95% CI 5.27–39.43), unknown diabetes (OR 15.86, 95% CI 4.55–55.36), and uncontrolled diabetes (OR 17.61, 95% CI 5.77–53.74) was independently associated with increased risk of SARS‐CoV‐2‐induced respiratory failure. Conclusion In our cohort of hospitalized SARS‐CoV‐2 positive patients with available HbA1c data, prediabetes, undiagnosed diabetes, and poorly controlled diabetes were associated with a markedly increased risk of SARS‐CoV‐2‐associated respiratory failure.


| INTRODUC TI ON
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has challenged healthcare systems around the world over the past year.
Large case numbers and homogeneity in characteristics and presentation has allowed us to rapidly acquire some understanding of the disease. There remain however considerable gaps of knowledge concerning pathophysiology, treatments, and risk factors for developing a severe and sometimes fatal infection. Finding a marker that can predict the course of the Coronavirus disease 2019 (Covid-19) would be of value for prioritizing patients in the healthcare system and for vaccination. Observational data indicates that the most common comorbidities in severe Covid-19 are several of the diagnoses that make up the metabolic syndrome, that is, hypertension, diabetes mellitus type 2, obesity, and hyperlipidemia. [1][2][3] Moreover, physical activity seems to reduce the risk of severe Covid-19. 2 Clinically, we observed that many patients with severe Covid-19 had undiagnosed metabolic syndrome whereas the patients without metabolic imbalance hospitalized for non-Covid related diagnosis seem to have mild or no signs of Covid-19 when testing positive for SARS-CoV-2 during routine screening. This is supported by recent studies. 3,4 Therefore, we hypothesized that untreated diabetes mellitus might be associated with a higher risk of developing severe Covid-19. Unfortunately, we lack a common biochemical marker for recognition of imbalance in the metabolic homeostasis. Dysglycemic metabolic disorders such as diabetes mellitus and pre-diabetes, constitute themselves a pandemic with high prevalence and a rapidly increasing incidence of cases presenting to healthcare. 5 The glycated hemoglobin (HbA1c) test is used as a reliable marker for glycemic status. The test is used to diagnose diabetes mellitus and to evaluate long-term glycemic control. It has been proposed as a useful marker for early detection of insulin resistance, a major etiological factor for the development of the metabolic syndrome. 6,7 The aim of this exploratory study was to assess the association between chronic dysglycemia, diagnosed by analysis of HbA1c, and SARS-CoV-2 associated respiratory failure in hospitalized patients with a positive PCR-test for SARS-CoV-2. We hypothesized that the presence of undiagnosed or insufficiently controlled chronic dysglycemia would be associated with the development of SARS-CoV-2 respiratory failure.

| ME THODS
The study was approved by the regional ethics committee in Stockholm (2020-01447, 2020-06969) with a waiver for informed consent and was conducted according to the Helsinki declaration.

| Study population
We performed a retrospective observational study including adult patients (≥18 years) admitted between March 9, 2020 and September 6, 2020 (first wave) to the Karolinska University Hospital, Stockholm, Sweden. The hospital is a tertiary referral centre receiving trauma, neurosurgical and cardio-thoracic patients from the larger Stockholm region, covering an urban area of approximately 2.5 million patients. In addition, a large number of patients with Covid-19 respiratory failure have been admitted to the hospital´s medical wards or being transferred directly to the intensive care unit (ICU) from other hospitals during the study period. All patients were screened for SARS-CoV-2 on admission and some patients admitted with a primary diagnosis other than Covid-19 tested positive for the virus on screening, many without displaying respiratory symptoms.
All patients testing positive for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) confirmed by polymerase chain reaction and with a HbA1c test obtained within 3 months prior to and including day of admission were included. We excluded patients who were admitted to ICU following extracorporeal membrane oxygenation therapy due to the risk of hemolysis, which may lower blood HbA1c, and patients admitted to hospital ward from ICU at another hospital.

| Data collection
We included routine clinical data prospectively recorded in the

Editorial Comment
The study findings show that identification of patients with chronic dysglycemia and quantification of their premorbid control is clinically important since patients with undiagnosed dysglycemia or poor diabetes control are at particular risk of developing respiratory complications when infected with SARS-CoV-2. It was also found that that HbA1c measurement identifies a significant proportion of SARS-CoV-2 positive patients with prediabetes or previously unknown diabetes.

| Outcomes
The primary outcome was development of respiratory failure associated with SARS-CoV-2 infection as determined by the treating physicians. The treating physicians' assessment of the presence of and cause for respiratory failure was obtained by careful review of clinical notes, ICD-10 diagnosis codes and discharge summaries. SARS-CoV-2 associated respiratory failure was ruled out by the treating clinician when the patient had no respiratory symptoms or when respiratory failure was caused by conditions other than SARS-CoV-2 (e.g., bacterial pneumonia, pulmonary edema, pneumothorax, or invasive mechanical ventilation due to acute neurological injury) and was reversed by targeted treatment of the underlying cause.
SARS-CoV-2 associated respiratory failure will forthwith be referred to as respiratory failure.

| Statistical analysis
Categorical data are presented as numbers and percentages and compared using the chi-square test. Continuous data are presented as median with interquartile range and compared using the Mann-Whitney U test. The association between chronic glucose control (no diabetes, prediabetes, unknown diabetes, controlled diabetes, and uncontrolled diabetes) and HbA1c strata (<42, 42-47, 48-51, and ≥52), respectively, and primary outcome was assessed using separate multivariable logistic regression analyses. The analyses were adjusted for age, sex, and body mass index (BMI). In a sensitivity analysis we also adjusted for baseline variables with a p value < .1 on univariable analysis. Goodness of fit and model discrimination were determined using the Hosmer-Lemeshow test and area under the receiver operating curve (ROC) respectively. All analysis was performed using STATA SE 14.2 and 15. A two-sided p-value of < .05 was considered statistically significant.

| RE SULTS
The selection of study patients is presented in Figure 1. We included 385 patients in the final analysis. Overall, 299 (78%) patients exhibited respiratory failure and, 86 patients (22%) did not. Baseline characteristics of patients with and without respiratory failure are compared in Table 1

| DISCUSS ION
We conducted a retrospective exploratory study to assess the association between chronic dysglycemia, determined by assessment of HbA1c and diabetes history, and respiratory failure in hospitalized SARS-CoV-2 positive patients. We found that prediabetes, unknown diabetes, and uncontrolled diabetes were independently associated with increased risk of developing respiratory failure when infected with SARS-CoV-2. In contrast, we found no significant association between well-controlled diabetes and risk of SARS-CoV-2 disease progression. Although patients with diseases constituting the metabolic syn-  Non-invasive ventilatory support, n (%) 0 (0) 16 (5.4) .25

| Strengths and limitations
Our study has several strengths. We assess the impact of prediabetes and unknown diabetes on the risk of developing severe Covid-19. By comprehensive review of medical records, we were able to make a detailed assessment of several components of the metabolic syndrome and a thorough evaluation of SARS-CoV-2 infected patient at risk for Covid-19 disease, reducing the risk for misclassification.
Our study has some limitations. HbA1c was not consecutively analyzed in all patients admitted to our hospital, which implies a risk of selection bias. Furthermore, we lack data of blood transfusion prior to HbA1c testing, which could display false low values of the test. Moreover, detailed data regarding respiratory failure using results of blood gas analysis and level of supplementary oxygen were not available in all patients. Additionally, the confidence intervals in our analyses of chronic dysglycemia and HbA1c were wide.
However, the point estimates for prediabetes, unknown diabetes, and uncontrolled diabetes were also high indicating that a type I error is unlikely. Although well-accounted for, residual confounding can never be completely ruled out in this study design. Finally, the single center design limits generalizability.

| Conclusions
In our cohort of hospitalized, SARS-CoV-2 positive patients with available preadmission HbA1c, the presence of prediabetes, undiagnosed diabetes, and uncontrolled diabetes was associated with a markedly elevated risk of respiratory failure. We suggest that HbA1c could be used to detect patients at risk of developing severe Covid-19 and using this information to prioritize vaccination and to guide monitoring and care at hospital admission. Our results support further investigation of the pathophysiological link between chronic dysglycemia, metabolic imbalance, and severe Covid-19.

ACK N OWLED G M ENTS
Not applicable

CO N FLI C T O F I NTE R E S T S
None of the authors have any conflict of interests.