Dexamethasone use and insulin requirements in coronovirus-19 (COVID-19) infection stratified by Hemoglobin A1c

Abstract

Aims

The study aimed to identify weight-based insulin requirements for dexamethasone-induced hyperglycemia in COVID-19 infection stratified by hemoglobin A1c (HbA1c).

Methods

This retrospective study assessed hospitalized patients ≥ 18 years admitted with COVID-19 and receiving ≥ 1 dose of dexamethasone 6 mG. Daily blood glucose (BG) and insulin doses were collected and organized by HbA1c.

Results

Among 45 patients with available HbA1c, 100% [HbA1c ≥ 7%] and 72% [HbA1c < 7%] developed hyperglycemia (BG ≥180 mG/dL). Median daily insulin (Interquartile Range) (units/kG/day) was 0.03 (0, 0.32) [HbA1c 6–6.9%], 0.1 (0.06, 0.36) [HbA1c 7–7.9%], 0.66 (0.39, 0.69) [HbA1c 8–8.9%], and 0.72 (0.63, 0.78) [HbA1c ≥ 9%]. On day 10 of dexamethasone, when majority of patients were at goal BG, patients required 0.07 (0.01, 0.31) [HbA1c 6–6.9%], 0.59 (0.11, 0.75) [HbA1c 7–7.9%], 1.15 (0.95, 1.35) [HbA1c 8–8.9%], and 1.14 units/kG/day [HbA1c ≥ 9%]. Of 24 patients completing 10 days of dexamethasone, 25% experienced hypoglycemia (BG < 70 mG/dL) upon discontinuation.

Conclusion

Patients with higher HbA1c experienced greater dexamethasone-induced hyperglycemia and required higher insulin doses. Inpatient insulin dosing algorithms should take into consideration baseline HbA1c to avoid delays in achieving normoglycemia.

1. Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus that causes coronavirus disease 2019 (COVID-19), has led to over 600 million confirmed cases and over 6 million deaths worldwide [1,2]. Dexamethasone 6 mG daily is the mainstay of therapy for patients hospitalized with COVID-19 infection requiring supplemental oxygen or invasive mechanical ventilation. The duration of treatment is 10 days or until hospital discharge, whichever is sooner [3], [4], [5].

Steroid-induced hyperglycemia occurs in 50–70% of hospitalized patients without known diabetes [6], [7], [8]. Most of these patients experience hyperglycemia within 48 h of starting high potency steroids [6]. Rhou et al. reports dexamethasone-induced hyperglycemia occurring at a rate of 47.6% in COVID-19 patients without diabetes peaking 7–9 h after exposure [9]. High-risk individuals for steroid-induced hyperglycemia are recommended to undergo blood glucose (BG) monitoring for at least 24 to 48 h after initiating steroids. If BG remains > 140 mG/dL, then continued monitoring and therapy should be considered [10]. Besides the potency and duration of steroid therapy, other risk factors for steroid-induced hyperglycemia include pre-diabetes, family history of diabetes, gestational diabetes, older age, length of hospital stay, and elevated c-reactive protein [6,8,9,11]. Inpatient hyperglycemia in patients with and without diabetes can increase mortality in both critically ill and non-critically ill patients, as well as lead to increased length of stay or admission to the intensive care unit [12]. Inpatient steroid-induced hyperglycemia regardless of diabetes has shown increased mortality, cardiovascular events, and infections, as well as adverse effects from hypoglycemia [10,13].

Elevated BG in COVID-19 is associated with disease progression and increased mortality and can occur in patients with and without pre-existing diabetes [14], [15], [16], [17]. Besides dexamethasone-induced impaired glucose metabolism, characteristics unique to COVID-19, including effects on the pancreatic beta cells, cytokine storm, impaired insulin secretion, and insulin resistance, distinguish hyperglycemia in COVID-19 [17]. One retrospective study reported higher mortality rates with worse glycemic control from dexamethasone-induced hyperglycemia in COVID-19 [18]. Additionally, Farzadfar et al. reported higher mortality and higher cumulative insulin requirements to resolve DKA in type 2 diabetes in COVID-19 patients versus non-COVID-19 patients [19].

Clinicians should initiate insulin in hospitalized patients with steroid-induced hyperglycemia though data regarding ideal insulin dosing in COVID-19 infection is largely based on expert opinion and reactionary to blood glucose values [10,[20], [21], [22], [23], [24], [25], [26]]; only one algorithm to our knowledge has been tested and published in Saudi Arabia [18]. While randomized controlled trials have studied steroid hyperglycemia protocols in other patient populations, such data for COVID-19 is lacking [22]. Furthermore, this study aims to report weight-based daily insulin requirements and mean BG levels stratified by HbA1c in patients treated with dexamethasone for COVID-19 infection.

2. Materials and methods

A retrospective, single-arm study was performed evaluating patients being treated with dexamethasone 6 mG daily for COVID-19 infection during November 2020 at a tertiary care teaching hospital located in Queens, NY; this period was the beginning of the second COVID-19 wave with the original strain. Expedited approval was granted by the Northwell Health® Institutional Review Board. Patients were included if they were 18 years of age or older, had a diagnosis of COVID-19 infection confirmed by positive polymerase chain reaction, received at least one dose of dexamethasone 6 mG, and were admitted to the hospital. Patients were excluded if they were on chronic steroids at home for any indication or if they did not have a HbA1c. Insulin prescribing was based on standard inpatient diabetes practice and varied by provider (no protocol was in place to follow). After patients had been identified through a generated list from the electronic medical record, a retrospective chart review was conducted to collect pertinent data. Data collected included pertinent baseline demographics, baseline HbA1c, diabetes classification, home medications for diabetes, and oxygen status prior to dexamethasone use. Descriptive statistics were computed for all variables.

During inpatient dexamethasone therapy, mean daily BG and median and mean daily insulin (units and units/kG/day) were collected and stratified by HbA1c. Dexamethasone's effects on BG were evaluated including time to hyperglycemia (BG > 180 mG/dL) and time to first insulin adjustment. The safety endpoint included hypoglycemia upon discontinuation of dexamethasone, defined as BG < 70 mG/dL.

3. Results

A total of 75 patients were screened and 30 were excluded due to a lack of HbA1c. The total prevalence of hyperglycemia for all COVID-19 patients on dexamethasone during this time frame was 54.7% (41/75). Forty-five patients were included who received dexamethasone 6 mG daily for COVID-19 infection during the inclusion period. Twenty-two patients (48.9%) did not have a history of diabetes and the remaining 23 patients had type 2 diabetes. For patients with pre-existing diabetes, 39.1% were using insulin therapies prior to admission (with or without non-insulin medications). Eighty-four percent of patients were on oxygen therapy at the time of dexamethasone initiation. Fourteen patients (31.1%) eventually required mechanical ventilation and thirteen patients (28.9%) expired. Median length of inpatient stay was 10 days (8, 21). Additional patient demographics and baseline characteristics are listed in Table 1 .

Table 1.

Demographics and Baseline Characteristics.

Variable	Total (N = 45)
Mean age, years (SD)	63.6 ± 14.9
Male sex, n (%)	33 (73.3)
Race, n (%) White Black Asian Other or Multiracial Unavailable, Unknown	11 (24.4) 5 (11.1) 13 (28.9) 15 (33.3) 1 (2.2)
Ethnicity, n (%) Hispanic or Latino	8 (17.8)
Mean weight, kG (SD)	82.2 ± 21.6
Median length of stay, days (IQR)	10 (8, 21)
Diabetes (DM), n (%) Type 1 DM Type 2 DM No DM history	0 (0) 23 (51.1) 22 (48.9)
Mean HbA1c,% (SD)	7.0 ± 1.4
Home medication(s) for DM, n (%) Insulin Non-insulin medication(s) Both (insulin and non-insulin medications) No medications	4 (17.4) 11 (47.8) 5 (21.7) 3 (13.0)
Respiratory support prior to dexamethasone, n (%) No oxygen Nasal cannula Non-rebreather High flow nasal cannula BIPAP Mechanical ventilation	7 (15.6) 24 (53.3) 11 (24.4) 1 (2.2) 1 (2.2) 1 (2.2)
Route of dexamethasone, n (%) IV Oral	41 (91.1) 4 (8.9)
Mean duration of dexamethasone, days (SD)	10 ± 4.7

Abbreviations: BIPAP, bilevel positive airway pressure; DM, diabetes; HbA1c, hemoglobin A1c; IQR, interquartile range; IV, intravenous; kG, kilogram; SD, standard deviation.

Thirty-seven out of 45 patients (82.2%) with a HbA1c experienced inpatient hyperglycemia; eight patients without hyperglycemia had a mean HbA1c of 5.96%. Ninety-six percent (22/23) of patients with a history of diabetes had inpatient hyperglycemia; one patient with diabetes without hyperglycemia had a HbA1c of 5.1%. All patients with HbA1c ≥ 7% (n = 16) and 72.4% of patients with HbA1c between 5 and 6.9% (n = 21) developed hyperglycemia. Mean number of days until hyperglycemia was 0.6 ± 0.8 for patients with HbA1c ≥ 7% and 2.14 ± 2.9 for patients with HbA1c < 7%. The mean number of days to new insulin or change in insulin dosing was 2.6 ± 3.0 for patients with a HbA1c ≥ 7% and 4.1 ± 3.0 for patients with a HbA1c < 7%. Higher HbA1c was associated with faster onset and longer duration of hyperglycemia (Fig. 1 ).

Fig. 1.

Mean Blood Glucose While on Dexamethasone. Bar graph showing daily mean blood glucose trends while on dexamethasone and two days after discontinuing. Abbreviations: HbA1c, hemoglobin A1c.

One out of five patients with a HbA1c between 5 and 5.9% experienced hyperglycemia starting on day 3 and lasting 2 days. This patient required an average of 2.5 units of insulin daily. Eighty-three percent of patients with a HbA1c of 6 to 6.9% experienced hyperglycemia only on day 1 of dexamethasone therapy (Fig. 1). Median daily insulin was 0 (0, 15.6) or 0.03 units/kG/day (0, 0.32). On day 10 of therapy (n = 11), median insulin required was 0.07 units/kG/day (0.01, 0.31) (Figs. 1, 2 , and 3 ).

Fig. 2.

Mean Daily Insulin Requirements. Bar graph showing mean daily insulin stratified by HbA1c while on dexamethasone. Abbreviations: HbA1c, hemoglobin A1c.

Fig. 3.

Median Weight-Based Insulin Requirements. Line graph showing median weight-based insulin stratified by HbA1c while on dexamethasone. Abbreviations: HbA1c, hemoglobin A1c; IQR, interquartile range; kG, kilogram; SD, standard deviation.

Nine patients with HbA1c between 7 and 7.9%, hyperglycemia on average began on day 2 of dexamethasone treatment and BG remained above goal through day 10. Median daily insulin was 8 units (5, 28) or 0.1 units/kG/day (0.06, 0.36). On day 10 of dexamethasone therapy, when BG was closest to goal, median insulin required was 0.59 units/kG/day (0.11, 0.75) (n = 5).

In four patients with a HbA1c of 8 to 8.9%, hyperglycemia started prior to dexamethasone initiation and remained elevated until day 10 of dexamethasone treatment. Median daily insulin required was 34 units (24, 44) or 0.66 units/kG/day (0.39, 0.69). On day 10 of therapy, median insulin required was 1.15 units/kG/day (0.95, 1.35) (n = 2). Lastly, for three patients with HbA1c ≥ 9%, BG was above goal for 8 out of 10 days of dexamethasone treatment. Median daily insulin required was 90 units (62, 105) or 0.72 units/kG/day (0.63, 0.78). On day 10 of therapy, when BG was at goal, median insulin required was 1.14 units/kG/day (n = 1) (Figs. 1, 2, and 3).

Of the 24 patients that completed 10 days of inpatient dexamethasone treatment, 6 patients (25%) experienced hypoglycemia upon discontinuation of dexamethasone. The median time to hypoglycemia was 96 h after dexamethasone discontinuation (range: 1 – 216 h).

4. Discussion

This study contains novel information related to BG trends and insulin requirements in relation to HbA1c for patients being treated with dexamethasone for COVID-19 infection. HbA1c is an important lab value for anyone started on dexamethasone for COVID-19 given the prevalence of hyperglycemia in patients with and without a history of diabetes (54.7% of the total screened population). Rhou et al. had similar findings where 47.6% of patients without diabetes experienced dexamethasone-induced hyperglycemia [9]. This study showed a clear trend in increasing BG and insulin requirements for higher HbA1c.

Insulin requirements are reported as median (IQR) (days 1 through 10) and median (IQR) on day 10 due to wide variability, particularly at lower HbA1c values (Fig. 3). Day 10 requirements may be more accurate than days 1 – 10 combined since BG control was achieved in most of the cohorts by day 10, meaning earlier days of therapy may have underdosed patients leading to hyperglycemia. One limitation of day 10 requirements was that some patients were discharged prior to day 10 of therapy so it includes a smaller sample of patients.

In patients with a HbA1c < 6%, monitoring of BG via point-of-care testing (POCT) and utilization of correction scale insulin should be considered initially until hyperglycemia can be ruled out. Most of these patients did not have diabetes or hyperglycemia; one patient with hyperglycemia was managed with minimal rapid-acting insulin via correction scale. In patients considered to have pre-diabetes or controlled diabetes (HbA1c between 6 and 6.9%), insulin requirements varied greatly as evidenced by the large interquartile range around the median. Median insulin required on day 10 for these patients was 0.07 units/kG/day (0.01, 0.31) (n = 11) meaning some patients required correction scale alone whereas others required basal ± bolus insulin. This patient population should be monitored closely inpatient as they may require insulin support beyond rapid-acting correction scale (e.g., basal insulin plus correction scale) to treat hyperglycemia. One patient on insulin prior to admission had insulin requirements above 1 unit/kG/day; future studies should calculate the percent increase from home to inpatient doses of insulin while on dexamethasone to achieve normoglycemia.

Similarly, patients with a HbA1c between 7 and 7.9% had variable insulin requirements as evidenced by the large interquartile range (Fig. 3). Since BG control was not achieved until dexamethasone was completed, median day 10 requirements may be a more accurate representation of how to dose this cohort which was 0.59 units/kG/day (0.11, 0.75) (n = 5) though a large interquartile range is present.

The greatest daily insulin requirements were in patients with a HbA1c ≥ 8%. For patients with a HbA1c between 8 and 8.9%, median daily weight-based insulin was 0.66 units/kG/day (0.39, 0.69), while requirements on day 10 were 1.15 units/kG (0.95, 1.35) (n = 2). For patients with a HbA1c ≥ 9%, median daily weight-based insulin was 0.72 units/kG/day (0.63, 0.78) and control of hyperglycemia was only achieved on days 8 and 10 of therapy. Only one patient completed 10 days of therapy with a HbA1c ≥ 9% and their day 10 insulin requirement was 1.14 units/kG/day. In Fig. 3, the trend was increasing weight-based requirements during the 10-day course, more pronounced the higher the HbA1c; on day 8, for HbA1c 8 – 8.9% and ≥ 9%, there is a significant decline in the weight-based requirement which can be explained by small sample size and early discharges (meaning the calculated weight-based dose only accounted for part of the day's insulin). Three out of seven patients with a HbA1c ≥ 8% were on non-insulin therapies at home and one was not on any diabetes medication at home. Although this study did not look at home insulin requirements, this could be an area for future studies.

While this sample size is small, our study highlights that the combination of uncontrolled diabetes, COVID-19, and dexamethasone can temporarily cause high BG and insulin requirements. In patients with a HbA1c ≥ 7%, more than a week was needed to achieve glucose control, which may have been due to a lack of provider familiarity with insulin requirements to achieve normoglycemia while treating COVID-19 with dexamethasone. As a result, our health system has since adopted a hyperglycemia management guideline for patients receiving dexamethasone for COVID-19 with directions on how to start and titrate insulin both while on the medication and when it is completed. Currently, the only published insulin dosing protocol for COVID-19 that was tested against a control group reported a decrease in mortality with better glycemic control; however, they did not disclose patient HbA1c or insulin requirements to achieve normoglycemia [18]. This information is vital given our study's findings. Additionally, since it was limited to one healthcare center in Saudi Arabia, the generalizability may be limited. To our knowledge, ours is the first study to present weight-based insulin requirements stratified by HbA1c. This data could be used to develop future studies and inpatient protocols given the limited information available regarding ideal insulin dosing in this patient population [20], [21], [22], [23], [24], [25], [26].

Hypoglycemia was a common side effect seen in patients after discontinuing dexamethasone. Five of the six patients with hypoglycemia required insulin for hyperglycemia associated with dexamethasone. In Fig. 1, average BG was well-controlled as soon as one day after discontinuing dexamethasone in all patients, regardless of HbA1c level at initiation. It could be that providers did not decrease insulin requirements soon enough after discontinuing dexamethasone. Additionally, the long duration of dexamethasone means the medication may still have an effect after discontinuation which could also be why the median time to hypoglycemia was 96 h [24]. Some outliers exist with hypoglycemia occurring 6 and 9 days after dexamethasone discontinuation; it is likely, though unclear, that insulin dosing and other factors contributed to these hypoglycemia events.

Being that this study was retrospective and observational in nature, limitations included lack of documentation, single-arm, small sample size (based on feasibility of the study authors), and other potential confounders (e.g., other hyperglycemia-causing medications). Thirty patients did not have a HbA1c or POCT, so serum glucose values were only available on morning labs. Since steroids have a post-prandial effect on blood glucose, hyperglycemia may have been underrecognized for these patients. Baseline HbA1c and POCT for the first 48 h should be implemented in all patients on high-potency steroids [10]. This study did not break down insulin requirements based on home insulin regimen. Patients who are more insulin resistant at baseline would therefore require higher inpatient insulin doses. It is also possible for patients with controlled HbA1c to have high insulin requirements at baseline which could skew our inpatient insulin data. This study also did not account for abnormalities in renal function which can affect insulin clearance and dosing. Variability in COVID-19 management could have occurred with the addition of travel nurses and redeployed team members unfamiliar with the institution's inpatient glycemic management standard of care, waivers in documentation, and lack of a standardized guideline for insulin titration with dexamethasone. Additionally, this study took place prior to vaccinations and other COVID-19 variants.

5. Conclusion

Dexamethasone 6 mG daily can cause hyperglycemia in patients with COVID-19 and higher HbA1c is associated with increased hyperglycemia and insulin requirements. It is important to start and titrate insulin accordingly to prevent prolonged hyperglycemia. This is the first study to our knowledge that provides insulin requirements based on HbA1c for COVID-19-infected patients on dexamethasone treatment. Application of these results may aid COVID-19 protocols as well as help achieve BG control while patients are treated with dexamethasone, though more large-scale, prospective studies are needed to corroborate this data.

Author contributions

B.K. collected data; B.K., C.G., L.M., D.B., and R.S.R. designed the study, analyzed data, and reviewed the manuscript.

Funding

This research did not receive any grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.