Effect of elevated fasting blood glucose level on the 1‐year mortality and sequelae in hospitalized COVID‐19 patients: A bidirectional cohort study

Abstract To observe the predictive effect of fasting blood glucose (FBG) level on the prognosis, clinical sequelae, and pulmonary absorption in hospitalized coronavirus disease 2019 (COVID‐19) patients with and without a history of diabetes, respectively, and to evaluate the correlation between the dynamic changes of FBG and poor prognosis. In this bidirectional cohort study, we enrolled 2545 hospitalized COVID‐19 patients (439 diabetics and 2106 without a diabetic history) and followed up for 1 year. The patients were divided according to the level of admission FBG. The dynamic changes of FBG were compared between the survival and the death cases. The prediction effect of FBG on 1‐year mortality and sequelae was analyzed. The 1‐year all cause mortality rate and in‐hospital mortality rate of COVID‐19 patients were J‐curve correlated with FBG (p < 0.001 for both in the nondiabetic history group, p = 0.004 and p = 0.01 in the diabetic history group). FBG ≥ 7.0 mmol/L had a higher risk of developing sequelae (p = 0.025) and have slower recovery of abnormal lung scans (p < 0.001) in patients who denied a history of diabetes. Multivariable Cox regression analysis showed that FBG ≥ 7.0 mmol/L was an independent risk factor for the mortality of COVID‐19 regardless of the presence or deny a history of diabetes (hazard atio [HR] = 10.63, 95% confidence interval [CI]: 7.15−15.83, p < 0.001; HR = 3.9, 95% CI: 1.56−9.77, p = 0.004, respectively). Our study shows that FBG ≥ 7.0 mmol/L can be a predictive factor of 1‐year all‐cause mortality in COVID‐19 patients, independent of diabetes history. FBG ≥ 7.0 mmol/L has an advantage in predicting the severity, clinical sequelae, and pulmonary absorption in COVID‐19 patients without a history of diabetes. Early detection, timely treatment, and strict control of blood glucose when finding hyperglycemia in COVID‐19 patients (with or without diabetes) are critical for their prognosis.


| INTRODUCTION
Coronavirus disease 2019 , caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has posed a serious threat to global health. 1 According to epidemiological and clinical data, patients with COVID-19 have higher risks of elevated blood glucose levels. COVID-19 patients with hyperglycemia will be more likely to transform into severe and fatal cases. [2][3][4] Cai et al. 5  Organization (WHO), 8,9 FBG in the prediabetic status is defined as 5.6−6.9 and 6.1−6.9 mmol/L, respectively. Therefore, in this investigation, we divided admission FBG into four levels in COVID-19 patients without a diabetic history (≤5.5, 5.6−6.0, 6.1−6.9, and ≥7.0 mmol/L) and three levels in COVID-19 patients with a diabetic history (≤6.1, 6.1−6.9, and ≥7.0 mmol/L), so as to investigate the effect of FBG on 1-year mortality, sequelae, and pulmonary absorption in hospitalized COVID-19 patients with and without diabetic history, respectively. On the other hand, few studies report the dynamic changes of FBG in COVID-19 patients during hospitalization and its predictive effect on their long-term prognosis. 9 To observe the impact of SARS-CoV-2 infection on FBG and avoid the possible confounding factors of diabetes on FBG, we divided COVID-19 patients into the diabetic history group and nondiabetic history group, analyzed the dynamic changes of FBG in patients with different admission FBG levels, and compared the level of FBG in the survival and death cases to study the impact of FBG level on the long-term prognosis in the two groups.

| Study design and participants
In this bidirectional cohort study, we chose 4493 hospitalized COVID-19 patients who were admitted to five designated COVID-19 hospitals from January 1, 2020 to March 18, 2020 in Hubei Province, including the headquarter, west Hospital, and tumor center of Union Hospital, the Central Hospital of Wuhan, and Dongfeng Hospital. 10  (3) complete and traceable clinical information; (4) complete followup and answer of questionnaires. According to the FBG level at admission, COVID-19 patients without a diabetic history were divided into four groups: FBG ≤ 5.5, 5.6−6.0, 6.1−6.9, and ≥7.0 mmol/L, 8 and COVID-19 patients with a diabetic history were divided into three groups: FBG ≤ 6.1 6.1−6.9, and ≥7.0 mmol/L.

| Data collection
All data from the electronic medical records of the 2545 COVID-19 patients were reviewed manually. The demographic characteristics, comorbidities, clinical signs and symptoms, laboratory results, inhospital treatment (including drug and oxygen therapy), and clinical outcome (discharge or death) were recorded using standardized case report forms and entered into a secure online database MySQL. 10

| Case follow-up
All survival discharged patients underwent a series of telephone questionnaires or outpatient services every 3 months, including allcause death, clinical sequelae, and lung condition. Death and its cause were reported by the patient's relative or clinician. The follow-up time was up to March 17, 2021, or the date of death.       Figure 2D).

| Clinical sequelae at 1 year after discharge
Because COVID-19 patients had multiple complications during hospitalization, we focused on the clinical sequelae and lung images after discharge (Table 3). We asked patients about their symptoms through telephone and outpatient follow-up. For patients who denied a history of diabetes, patients in the ≥7.0 mmol/L group had a higher risk of chest tightness, (p = 0.025), but other sequelae such as fatigue, cough, shortness of breath, and low back pain showed no significant difference.
For lung images, patients in the ≥7.0 mmol/L group were more likely to have slower recovery of abnormal lung scans (p < 0.001) (Table 3).
However, among COVID-19 patients with diabetic history, there were no significant differences in clinical sequelae and lung images between patients with different admission FBG levels. 1.56−9.77) were independent risk factors for 12-month postdischarge mortality (Table 5); (Figure 4).

| DISCUSSION
In this bidirectional cohort study, our study shows that the 1-year allcause mortality rate and in-hospital mortality rate of COVID-19 patients   manifestations in the acute phase, but also many complications afterward, and the pathologic changes in the lung caused by COVID-19 will take a long time to restore. Therefore, in addition to the survival of the patients, we also focused on the related complications and pulmonary changes of FBG at 1 year after discharge. We found that, among COVID-19 patients without a diabetic history, those whose admission ≥7.0 mmol/L were more likely to have chest tightness and slower recovery of abnormal lung images at 1 year after discharge.
Roncon et al. reported that COVID-19 patients with hyperglycemia or diabetes were more likely to undergo ICU care. 12 Lazarus et al. reported that hyperglycemia promotes the progression of COVID-19. 3,4,13 Diabetic microvascular lesions in the respiratory tract may impair alveolar gas exchange and lung compliance, resulting in impaired lung function and reduced absorption of lung inflammation. 14 Caruso et al. confirmed that diabetes and hyperglycemia could cause pulmonary remodeling and respiratory restriction. 15  indicator that helps distinguish patients with poor long-term glycemic control from those with stress hyperglycemia. Moreover, most of the COVID-19 patients were from the designated hospitals for critical patients during the pandemic, which might cause a bias of higher mortality rates. Hence, we tried to enlarge the sample size to reduce the bias.
In conclusion, our study shows that FBG