key: cord-0916342-6z4sltri
authors: Patel, Amy J.; Klek, Stanislaw P.; Peragallo-Dittko, Virginia; Goldstein, Michael; Burdge, Eric; Nadile, Victoria; Ramadhar, Julia; Islam, Shahidul; Rothberger, Gary D.
title: Correlation of hemoglobin A(1c) and outcomes in patients hospitalized with COVID-19
date: 2021-07-18
journal: Endocr Pract
DOI: 10.1016/j.eprac.2021.07.008
sha: 0a447cd54d4d6e0803127673479fe4aa0876de7f
doc_id: 916342
cord_uid: 6z4sltri

BACKGROUND: Diabetes is a known risk factor for severe coronavirus disease 2019 (COVID-19). We conducted this study to determine if there is a correlation between hemoglobin A(1c) (HbA(1c)) level and poor outcomes in hospitalized patients with diabetes and COVID-19. METHODS: This is a retrospective, single-center, observational study of patients with diabetes (as defined by an HbA(1c) ≥ 6.5% or known medical history of diabetes) who had a confirmed case of COVID-19 and required hospitalization. All patients were admitted to our institution between March 3, 2020 and May 5, 2020. HbA(1c) results for each patient were divided into quartiles; 5.1-6.7% (32-50 mmol/mol), 6.8-7.5% (51-58 mmol/mol), 7.6-8.9% (60-74 mmol/mol), and >9% (>75 mmol/mol). The primary outcome was in-hospital mortality. Secondary outcomes included admission to an intensive care unit, invasive mechanical ventilation, acute kidney injury, acute thrombosis, and length of hospital stay. RESULTS: Five hundred and six patients were included. The number of deaths within quartiles 1 through 4 were 30 (25%), 37 (27%), 34 (27%) and 24 (19%), respectively. There was no statistical difference in the primary or secondary outcomes between the quartiles except acute kidney injury was less frequent in quartile 4. CONCLUSIONS: There is no significant association between HbA(1c) level and adverse clinical outcomes in patients with diabetes who are hospitalized with COVID-19. HbA(1c) should not be used for risk stratification in these patients.

Patients with diabetes comprise approximately one-third of confirmed cases of 19. 4 Several studies have demonstrated that patients with diabetes are at increased risk 13 of severe disease and poor outcomes. [6] [7] [8] In addition, elevated levels of glycemia during 14 hospital admission has been demonstrated to correlate with poor outcomes. 9,10 The 15 correlation between hemoglobin A1c (HbA1c) and outcomes are less clear. 11 Several 16 studies reported an association between higher HbA1c values and worse outcomes in 17 COVID-19, [12] [13] [14] [15] although all had small samples sizes and two of these studies defined 18 elevated HbA1c as >6.5% (>48 mmol/mol) without stratifying these patients further. Other 19 studies have found no association between HbA1c and outcomes in patients with 19, although several are limited by incomplete data. [16] [17] [18] [19] Our study was designed 21 to specifically investigate the correlation between overall glycemia prior to This is a retrospective, single-center, observational study of patients with diabetes 1 mellitus who had a confirmed case of SARS-CoV-2 on polymerase chain reaction testing 2 of a nasopharyngeal sample and required hospital admission. Our institutional review 3 board approved this study and the requirement for informed consent was waived. All age would affect outcomes. Results were similar to those from the entire cohort (Table   20 3).

[Insert Table 3 ]

In order to confirm that our observed outcomes were not a function of the quartile ranges 23 that were used for this study, we analyzed HbA1c as a continuous variable and assessed 24 its relationship to outcomes ( Hyperglycemia also has a direct impact on airway epithelia and lung integrity, rendering 8 quicker respiratory decline in this sub-population. 22 Our study demonstrates that long-9 term glycemia, as measured by HbA1c, does not predict worse outcomes in patients 10 hospitalized with COVID-19. Among the total cohort of 506 patients with an 11 overall mortality rate of 25%, there was no statistically significant increase in the 12 frequency of primary or secondary outcomes in higher quartiles of HbA1c.

It is important to note that the mean age of patients in quartile 4 was significantly lower in 15 comparison to quartiles 1 through 3. It has been established that age is a key prognostic 16 determinant of outcomes in COVID-19 due to higher susceptibility to the virus and 17 severity of complications in the elderly population. 23 One may attribute the lack of worse 18 outcomes in patients with the highest HbA1c levels to younger age. However, analysis of 19 patients >65 years old failed to reveal any worse outcomes in patients with higher 20 HbA1c levels.

Previously undiagnosed diabetes is common in hospitalized patients with COVID-19. A 23 recent meta-analysis including more than 3700 patients showed a 14.4% of hospitalized 24 patients had newly diagnosed diabetes. 24 In our cohort, nearly 26% of all patients had 25 previously undiagnosed diabetes, similar to the results reported by Li et al. 25 We noted 26 that patients in quartile 4 were less likely to have a new diagnosis of diabetes compared to patients in quartiles 1 and 2. This may have a protective impact on the outcomes of 1 patients in quartile 4, as patients with newly diagnosed diabetes have worse COVID-19-2 related outcomes than those with pre-existing diabetes. 26, 27 Patients in quartile 4 were 3 also more likely to be on long-term insulin therapy, and this did not impact outcomes 4 despite prior studies reporting long-term insulin use to be associated with worse 5 outcomes in hospitalized patients with Metformin use was also more 6 common in patients in quartile 4, and metformin has been associated with improved 7 survival in patients with diabetes and COVID-19. 28

We observed a lower frequency of AKI in quartile 4 compared to other quartiles, both in 10 the entire cohort and in patients >65 years of age. Patients in quartile 4 were younger and 11 had a lower incidence of CKD compared to other quartiles, making them less prone to 12 developing AKI. However, when adjusted for age and the presence of comorbidities, 13 including CKD, HbA1c >9% (75 mmol/mol) was still associated with a lower risk of AKI 14 (appendix). These results are in contrast to the results of a study by Khalili et al. 29 , who 15 reported HbA1c to be a risk factor for development of AKI in patients hospitalized with 16 COVID-19. Further analysis of our data revealed that CKD was a much stronger risk factor 17 for AKI than HbA1c, and when analyzing patients in quartile 4, CKD was associated with 18 AKI while HbA1c was not (appendix). While our data indicate that HbA1c is negatively 19 associated with AKI, it is likely that our observed rates of AKI in quartile 4 are influenced 20 more by the lower rate of CKD in this group.

The absence of worse outcomes in patients with higher levels of HbA1c are surprising.

Diabetes is a known risk factor for worse outcomes in COVID-19 as well as previous diabetes who had early hyperglycemia (glucose >180 mg/dL) in the first two days of ICU inflammatory cytokines and coagulation factors, theorizing the mechanism behind 1 improvement in outcomes. 39 Although admission glucose levels were recorded in our 2 study, we did not evaluate overall inpatient glycemia and its effect on primary and 3 secondary outcomes. This would be an important focus of future studies.

Our study holds clinical significance when determining treatment plans and assessing 6 mortality risk in hospitalized patients. Our findings are indeed reassuring for patients with 7 a history of diabetes, specifically those with suboptimal chronic glycemia.

Nonetheless, optimal glycemic management during the course of illness remains an 9 important aspect of the treatment of COVID-19 to limit the incidence of poor outcomes.

The main strengths of our study are a robust sample size and the availability 12 of HbA1c levels for all subjects in this study. The latter differentiates our study from the 13 previous studies that found no correlation between HbA1c and outcomes in COVID-19. WBC, x10 9 /L 7.6 (5.8 -10.2) 7.0 (5.5 -9.4) 7.4 (5.7 -9.4) 7.6 (5.8 -10.6) 7.4 (5.7 -9.7) 0.338 ANC, /mm 3 6.0 (4.4 -8.5) 5.4 (3.6 -7.9) 5.6 (4.3 -7.5) 6.1 (4.2 -8.5) 5.8 (4. J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f 

Data are presented as median (interquartile range) or frequency (%). BMI, body mass index; CVD, cardiovascular disease; CKD, chronic kidney disease

P-values are from Kruskal-Wallis test for continuous variables and Chi-Square or Fisher's exact test for categorical variables

ESR, mm/hr 73

pg/mL 13