Original Research

Journal of Intensive Care Medicine

2018, Vol. 33(5) 317-321
Association Between the Use of ª The Author(s) 2016
Reprints and permission:
sagepub.com/journalsPermissions.nav
Long-Acting Insulin and Hypoglycemia DOI: 10.1177/0885066616677030
journals.sagepub.com/home/jic
in Nondiabetic Patients in the Surgical
Intensive Care Unit

Jordan Masse, PharmD1, Christopher Alan Giuliano, PharmD1,2,

Sara Brown, PharmD1, and Renee Alexander Paxton, PharmD, BCPS, BCCCP3

Abstract
Purpose: The purpose of this study was to examine the association between long-acting insulin and hypoglycemia in nondiabetic
surgical intensive care patients. Methods: This single-center, retrospective cohort study evaluated glycemic control in non-
diabetic critically ill surgical patients receiving long-acting insulin plus sliding scale versus those receiving sliding scale alone.
Patients were matched based on sliding scale order and type of surgery. The primary outcome was the proportion of patients who
experienced hypoglycemia (glucose values <70 mg/dL). Secondary outcomes included comparing the distribution of glycemic
events in the 2 groups and describing the proportion of patients transferred out of the intensive care unit on long-acting insulin
who experienced hypoglycemia. Results: One hundred twenty patients met the study criteria. Hypoglycemia was significantly
higher in the long-acting insulin plus sliding scale group compared to those receiving sliding scale alone (17 [28.3%] patients vs 8
[13.3%] patients; P ¼ .047). After adjusting for body mass index, renal failure, age, and Acute Physiology and Chronic Health
Evaluation II, the odds of hypoglycemia were 4.1 times higher for patients receiving long-acting insulin and sliding scale compared
to those receiving sliding scale alone (P ¼ .02). There were more hypoglycemic events (42 vs 20; P ¼ .05) and hyperglycemic
events (313 vs 135; P ¼ .02) in the long-acting insulin group. Conclusion: This study demonstrated higher rates of hypoglycemia
associated with the utilization of long-acting insulin in nondiabetic surgical intensive care patients. Risk of hypoglycemia should be
weighed against possible benefits in this population.

Keywords
long-acting insulin, basal insulin, hypoglycemia, critical care, insulin detemir, insulin glargine

Introduction hypoglycemia (BG < 45 mg/dL) and therefore must be

Hypoglycemia occurs in up to 20% of critically ill patients and
is an important predictor of morbidity and mortality.1,2 Criti-
cally ill patients typically possess a variety of risk factors for
hypoglycemia such as liver dysfunction, renal dysfunction,
sepsis, mechanical ventilation, continuous renal replacement
therapy, vasopressor therapy, changes in nutrition, and changes
in corticosteroid dosage.1,3 Even with multiple risk factors for
hypoglycemia, stress-induced hyperglycemia in the absence of
a preexisting diagnosis of diabetes commonly occurs in criti-
cally ill patients.4 A transient increase in blood glucose (BG)
concentration during acute physiological illness can develop as
a result of severe stress and increased counterregulatory hor-
mones.5 Clinical trials have shown benefits relative to morbid-
ity and mortality from management of inpatient hyperglycemia
with insulin, even in patients without a history of diabetes.6-8
Unfortunately, the management of hyperglycemia in critically
ill patients is associated with a higher incidence of hypoglyce-
mia (BG < 70 mg/dL) and a 5-fold increase in the risk of severe
approached with caution.9
Current guidelines recommend initiating insulin therapy
when BG values are greater than 180 mg/dL in critically ill
patients.10 A BG target of less than 180 mg/dL has shown to
decrease mortality when compared to more strict glucose con-
trol (BG <110 mg/dL) in medical intensive care unit (ICU) and
1
Department of Pharmacy, St. John Hospital and Medical Center, Detroit, MI,
USA
2
Department of Pharmacy Practice, Wayne State University, Detroit, MI, USA
3
Surgical Intensive Care Unit, St. John Hospital and Medical Center, Detroit,
MI, USA
Received July 14, 2016. Received revised September 12, 2016. Accepted
October 11, 2016.
Corresponding Author:
Renee Alexander Paxton, Surgical Intensive Care Unit, St. John Hospital
and Medical Center, 22101 Moross Road, Detroit, MI 48236, USA.
Email: renee.alexander@stjohn.org
318
surgical ICU (SICU) patients.11 Available research examining
normalization of elevated BG primarily involves the use of
continuous insulin infusions.11-15 Although an effective means
of achieving and maintaining euglycemia, continuous insulin
infusions can be labor intensive.16 Subcutaneous sliding scale
insulin (SSI) is an alternative treatment strategy for glucose
management.17 However, SSI is neither effective in meeting
physiologic needs nor an efficient treatment method in the
inpatient setting.18 Sliding scale insulin attempts to correct
hyperglycemia on a reactive basis rather than using a proactive
strategy to achieve improved overall control.19,20 As an alter-
native, basal insulin provides uniform action and can provide
more predictable glucose control compared to repeated SSI
doses. However, the effect of basal insulin typically lasts about
24 hours, and unlike SSI, it cannot be adjusted to reflect
changes in the patient’s clinical condition.21,22
At our institution, practitioners frequently prescribe subcu-
taneous long-acting insulin, combined with SSI, to control BG
in the SICU. Currently, there is a lack of research assessing the
safety and efficacy of subcutaneous long-acting insulin regi-
mens in nondiabetic critically ill patients. Hypoglycemia is a
serious safety concern that could result from this practice due to
the high prevalence of risk factors and changing clinical con-
dition of these patients. Therefore, the purpose of this study is
to assess the incidence of hypoglycemic events in nondiabetic
SICU patients receiving long-acting insulin.
Methods
We conducted a single-center, retrospective cohort study com-
paring glycemic control in nondiabetic adult SICU patients
who received long-acting insulin plus sliding scale to those
who received sliding scale alone. This study was conducted
at St. John Hospital and Medical Center, a 772-bed community
teaching hospital, in Detroit, Michigan. The SICU consists of
20 beds where decisions concerning the management of insulin
were predominantly made by a medical or surgical intensivist.
The institutional review board of St. John Hospital and Medical
Center approved this study.
The electronic medical records of patients admitted between
January 1, 2010, and March 30, 2015, were reviewed in a
consecutive manner. No changes in insulin management pro-
tocols occurred in the institution during this time period. Insu-
lin glargine was the insulin on hospital formulary when the
study began; however, this was changed to insulin detemir
during the study period. Patients were eligible for inclusion if
they were at least 18 years old, were admitted to the SICU for a
minimum of 24 hours, and had 2 consecutive BG values
>180 mg/dL during their ICU stay. Patients were excluded if
they had a history of diabetes, a hemoglobin A1c 6.5%, an
intentional b-blocker, insulin, or insulin secretagogue over-
dose, received subcutaneous insulin Neutral Protamine Hage-
dorn (NPH), R, or 70/30 products to control hyperglycemia,
received agents other than insulin to control BG, received par-
enteral nutrition, or died within 2 days of SICU admission.
Groups were matched on a 1:1 ratio based on the type of sliding
Journal of Intensive Care Medicine 33(5)
scale ordered and the type of surgery. Four distinct SSI orders
were available: medium, high, ICU moderate, and ICU tight
(Appendix A). Sliding scales were ordered every 4 to 6 hours
and were not scheduled around meals. Our institution targets
BG values less than 180 mg/dL for patients in the ICU. Types
of surgery were categorized into 1 of 4 groups including neu-
rological, vascular, trauma/orthopedic, and general.
The primary outcome of this study was to compare the
proportion of patients receiving long-acting insulin plus SSI
who experienced hypoglycemia compared to those receiving
SSI alone. Hypoglycemia was defined as BG values <70 mg/dL.
If patients in either group were managed for any amount of
time using intravenous regular insulin infusion, BG values
from initiation until 2 hours after the end of infusion were not
collected. Secondary outcomes included examining the distri-
bution of hypoglycemic, normoglycemic, and hyperglycemic
events between the 2 groups. The BG ranges of hypoglyce-
mia, normoglycemia, and hyperglycemia were defined as
<70 mg/dL, 70 to 180 mg/dL, and >180 mg/dL, respectively.
Hypoglycemic BG values that were drawn within 1 hour of
the previous hypoglycemic measurement were considered to
be part of the same hypoglycemic event to prevent double
counting events. Finally, we intended to evaluate the propor-
tion of patients transferred out of the ICU on long-acting
insulin who experienced hypoglycemic events.
Data collected included patient demographics, type of sur-
gery, SSI ordered, type of insulin received, total amount of
insulin administered, BG levels, duration of therapy, medical
history (sepsis, liver, and renal disease), and medications pre-
scribed (corticosteroids and vasopressors). The Acute Physiol-
ogy and Chronic Health Evaluation II (APACHE II) score was
also calculated for each patient.
Statistical Analysis
Descriptive statistics were generated to characterize the study
population and clinical factors. For continuous variables, the
mean and standard deviations were computed, and frequency
distributions were used for categorical variables. The Student t
test was used to compare continuous variables, and w2 test was
used to compare nominal variables. Mann-Whitney U test was
used to compare the distribution of hypoglycemic, normogly-
cemic, and hyperglycemic events in patients receiving
long-acting insulin plus SSI to those receiving SSI only. Multi-
variable logistic regression was conducted to examine the
effects of long-acting insulin as a predictor of hypoglycemia.
Confounders considered for inclusion in the regression were
determined through directed acyclic graphs.23,24 All data were
analyzed using SPSS version 22.0, and a P value .05 was used
to indicate statistical significance.
Results
Of 1661 consecutive records screened, 138 patients met the
inclusion and exclusion criteria of the study. The most common
reasons for not meeting criteria were a history of diabetes and
Masse et al 319

Table 1. Baseline Characteristics.a Table 2. Multiple Logistic Regression for Long-Acting Insulin as a
Predictor of Hypoglycemia.
Long-Acting Insulin þ SSI P
(n ¼ 60) SSI (n ¼ 60) Value Odds Ratio (95% CI) P Value

Age, years 62.4 + 14.7 64.0 + 18.4 .60 BMI 0.879 (0.804-0.960) <.01
Female 34 (56.7) 29 (48.3) .36 Age (years) 1.040 (1.002-1.080) .04
BMI 30.2 + 8.2 27.1 + 7.0 .03 APACHE II 1.043 (0.955-1.139) .35
Treatment 6.1 + 5.0 5.8 + 6.3 .73 Renal failure 1.719 (0.432-6.849) .44
duration (days) Long-acting insulin 4.116 (1.321-12.825) .02
Mean BG values 145.8 + 21.1 136.3 + 27.5 .24
(mg/dL) Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II;
BMI, body mass index; CI, confidence interval.
Liver failureb 3 (5.0) 0 –
MELD score 10.2 + 5.7 9.0 + 2.6 .14
Renal failurec 13 (21.7) 4 (6.7) .02 Table 3. Dose of Long-Acting Insulin.
CRRT 2 (2.3) 0 –
Sepsis 28 (46.7) 27 (45.0) .86 Long-Acting Patients Who Developed Number of Hypoglyce-
Vasopressors 10 (16.7) 9 (15.0) .80 Insulin Dose, U Hypoglycemia (n ¼ 17) mic Events (n ¼ 42)
Corticosteroids 20 (33.3) 16 (26.7) .43
APACHE II 14.9 + 6.8 12.8 + 5.6 .06 0-9 2 5
10-19 13 28
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; 20-29 0 0
BG, blood glucose; BMI, body mass index; CRRT, continuous renal replace- 30-39 1 5
ment therapy; SSI, sliding scale insulin. 40 1 4
a
Categorical data are presented as n (%). Continuous data are presented as
mean + standard deviation.
b
Model of end-stage liver disease (MELD) score >24.
c
Creatinine clearance <30 mL/min or SCr 1.7. groups (P ¼ .05). There were similar normoglycemic events
between groups, 1698 events in the long-acting insulin arm and
ICU stay less than 24 hours. After matching based on SSI order 1761 in the SSI-only group. Finally, there were significantly
and surgery type, 60 patients remained in both groups. Baseline more hyperglycemic events in the long-acting insulin plus SSI
characteristics such as age, sex, treatment duration, mean BG arm (313) than in the SSI group (135), and the distribution was
values, sepsis, vasopressors, corticosteroids, and APACHE II significantly different (P ¼ .01). Table 3 lists the number of
score were similar between groups (Table 1). Body mass index hypoglycemic events according to the dose of long-acting insu-
(BMI) and the incidence of renal failure were higher in the lin. Finally, 29 (48.3%) patients continued long-acting insulin
long-acting insulin arm (P ¼ .03 and .02, respectively). Patients therapy and SSI when transferred out of the ICU. Of these 29, 1
most commonly underwent general surgery, followed by neu- patient developed hypoglycemia and that same patient devel-
rological, trauma/orthopedic, and vascular surgery (48.3%, oped a recurrent episode.
31.7%, 11.7%, and 8.3%, respectively). The majority of
patients, 58.3%, received the ICU moderate scale, followed
by medium scale, ICU tight scale, and high scale (15%, Discussion
13.3%, 10%, and 3.3%, respectively). Insulin glargine was the Our research is the first published comparison of glycemic
most frequently used insulin in the long-acting insulin control using basal insulin combined with sliding scale to slid-
group (93.3%). ing scale alone in a population of nondiabetic SICU patients.
Proportion of patients who experienced hypoglycemia was When patients were managed with long-acting insulin plus
significantly higher in the long-acting plus SSI arm compared sliding scale, an association with hypoglycemia was observed.
to SSI only (17 [28.3%] vs 8 [13.3%] patients; P ¼ .047). After This relationship remained after controlling for age, BMI, renal
adjusting for BMI, renal failure, age, and APACHE II, the odds failure, and APACHE II score. Despite the higher risk of hypo-
of hypoglycemia were 4.1 times higher for patients receiving glycemia, the proportion of patients experiencing hyperglyce-
long-acting insulin plus SSI compared to SSI alone (P ¼ .02; mia was also higher in the long-acting insulin plus SSI group.
Table 2). Removal of patients receiving insulin detemir did not At the time of initiation of this study, no other studies eval-
change the increased risk of hypoglycemia (odds ratio: 4.2; P ¼ uating long-acting insulin utilization in a nondiabetic SICU
.016). Mean length of insulin therapy was similar between patients had been published. Recently, a randomized, open-
groups (P ¼ .73), 6.1 days in the long-acting insulin group and label trial comparing the use of long-acting insulin to regular
5.8 days in the SSI-only group. There was no difference in the insulin in nondiabetic, stable neurosurgery ICU, SICU, and
average number of glucose draws in the long-acting plus SSI cardiac ICU patients receiving parenteral nutrition was pub-
group compared to SSI alone (34.2 vs 31.9; P ¼ .708). lished.25 This study included a much narrower patient popula-
Overall, there were 42 hypoglycemic events in the long- tion than our study. Patients in this study had been receiving
acting insulin arm and 20 hypoglycemic events in the SSI alone parenteral nutrition and regular insulin for at least 3 days
arm. The distribution of hypoglyemia was different between before, and those receiving corticosteroids or vasopressors
320 Journal of Intensive Care Medicine 33(5)

were excluded.25 Similar rates of hypoglycemia were seen Appendix A

between the 2 groups, although the results are only general-
izable to patients receiving parenteral nutrition.25 The utility
Sliding Scales Orders
of long-acting insulin in nondiabetic, postsurgical, cardio-
vascular ICU patients being transitioned from continous Medium High
intravenous insulin therapy has also been demon- Glucose level: mg/dL Glucose level: mg/dL
strated.13,14,26 However, studies in this patient population 110-150: 0 U 110-150: 0 U
are also not directly comparable to SICU patients as these 151-200: 2 U 151-200: 2 U
201-250: 3 U 201-250: 4 U
patients have more uniformity in their admitting diagnosis,
251-300: 5 U 251-300: 7 U
and patients are being transitioned from continous intrave-
nous therapy so their 24-hour insulin requirement is more ICU moderate ICU tight
precisely known. The results of this study agree with pre- Glucose level: mg/dL Glucose level: mg/dL
vious literature, which has identified risk factors for hypo- 71-110: 0 U 71-110: 0 U
glycemia and hyperglycemia in critically ill patients. Boucai 111-130: 0 U 111-130: 2 U
131-150: 2 U 131-150: 4 U
and colleagues reported that increasing age was associated 151-170: 4 U 151-170: 6 U
with higher rates of hypoglycemia, and similarly, our study 171-190: 6 U 171-190: 8 U
identified age as a significant predictor of hypoglycemia.27 191-210: 8 U 191-210: 10 U
It has also been shown that higher BMI can reduce the risk
of developing hypoglycemia.28
Managing glucose in the SICU can be challenging, as there Declaration of Conflicting Interests
is a multitude of factors that affect glucose control. Interrup- The author(s) declared no potential conflicts of interest with respect to
tions in nutrition, tapering of steroids or vasopressors, and the research, authorship, and/or publication of this article.
fluctuating stress levels can lead to high variability in BG levels
in these patients. If practitioners choose to use long-acting Funding
insulin, they must first assess the nutrition support status and
The author(s) received no financial support for the research, author-
anticipate changes to concurrent pharmacologic therapies and
ship, and/or publication of this article.
stress levels to anticipate the effect these factors could have on
patients’ BG, as a single episode of hypoglycemia is an inde-
pendent risk factor for morbidity and mortality in this popula- References
tion.1,29 Due to glycemic variability in critically ill patients, 1. Krinsley JS, Grover A. Severe hypoglycemia in critically ill
practitioners must also recognize that hypoglycemia can occur patients: risk factors and outcomes. Crit Care Med. 2007;
regardless of the long-acting insulin dose, which is illustrated 35(10):2262-2267. doi:10.1097/01.CCM.0000282073.98414.4B.
by the results of this study. 2. Hermanides J, Bosman RJ, Vriesendorp TM, et al. Hypoglycemia
This study has some limitations. First, this study was a retro- is associated with intensive care unit mortality. Crit Care Med.
spective cohort study that relied on medical chart review at a 2010;38(6):1430-1434. doi:10.1097/CCM.0b013e3181de562c.
single center. Prospective randomized controlled trials con- 3. Vriesendorp TM, van Santen S, DeVries JH, et al. Predisposing
ducted across multiple sites could provide a more conclusive factors for hypoglycemia in the intensive care unit. Crit Care
answer to this question in the future and improve generaliz- Med. 2006;34(1):96-101. doi:10.1097/01.CCM.0000194536.
ability. Additionally, we were unable to accurately capture the 89694.06.
impact of nutritional status, which may have led to fluctuations 4. Deedwania P, Kosiborod M, Barrett E, et al. Hyperglycemia and
in glucose. Capturing these data is difficult in an intensive care acute coronary syndrome: a scientific statement from the
setting where patient’s nutritional needs commonly fluctuate or American Heart Association Diabetes Committee of the Council
are interrupted and accurate information is not always available on Nutrition, Physical Activity, and Metabolism. Circulation.
from chart review. Finally, because only 4 included patients 2008;117(12):1610-1619. doi:10.1161/CIRCULATIONAHA.
received insulin detemir, the results of this study may not be 107.188629.
generalizable to this population. 5. Inzucchi SE. Clinical practice. Management of hyperglycemia
In conclusion, this study demonstrated that the use of in the hospital setting. N Engl J Med. 2006;355(18):1903-1911.
long-acting insulin is associated with a higher incidence of doi:10.1056/nejmcp060094.
hypoglycemia in nondiabetic SICU patients. Within the ICU 6. Capes SE, Hunt D, Malmberg K, Gerstein HC. Stress hypergly-
setting, there are a multitude of factors that contribute to caemia and increased risk of death after myocardial infarction in
variability of BG. Prospective studies should be conducted patients with and without diabetes: a systematic overview. Lancet.
if long-acting insulin is used in nondiabetic patients. Until 2000;355(9206):773-778. doi:10.1016/S0140-6736(99)08415-9.
such studies have been conducted, clinicians should use cau- 7. Capes SE, Hunt D, Malmberg K, Pathak P, Gerstein HC.
tion when considering the use of basal insulin in nondiabetic Stress hyperglycemia and prognosis of stroke in nondiabetic
critical care patients because of the increased risk of and diabetic patients: a systematic overview. Stroke. 2001;
hypoglycemia. 32(10):2426-2432.
Masse et al
8. Kosiborod M, Rathore SS, Inzucchi SE, et al. Admission glucose
and mortality in elderly patients hospitalized with acute myo-
cardial infarction: implications for patients with and without
recognized diabetes. Circulation. 2005;111(23):3078-3086.
doi:10.1161/CIRCULATIONAHA.104.517839.
9. Jacobi J, Bircher N, Krinsley J, et al. Guidelines for the use of
an insulin infusion for the management of hyperglycemia in
critically ill patients. Crit Care Med. 2012;40(12):3251-3276.
doi:10.1097/CCM.0b013e3182653269.
10. Moghissi ES, Korytkowski MT, DiNardo M, et al. American
Association of Clinical Endocrinologists And American Diabetes
Association consensus statement on inpatient glycemic control.
Diabetes Care. 2009;32(6):1119-1131. doi:10.2337/dc09-9029.
11. The NICE-SUGAR Study InvestigatorsFinfer S, Chittock DR,
Su SY, et al. Intensive versus conventional glucose control in
critically ill patients. N Engl J Med. 2009;360(13):1283-1297.
doi:10.1056/NEJMoa0810625.
12. Schmeltz LR, DeSantis AJ, Schmidt K, et al. Conversion of
intravenous insulin infusions to subcutaneously administered
insulin glargine in patients with hyperglycemia. Endocr Pract.
12(6):641-650. doi:10.4158/EP.12.6.641.
13. Silinskie KM, Kirshner R, Hite MS. Converting continuous insu-
lin infusion to subcutaneous insulin glargine after cardiac surgery
using percentage-based versus weight-based dosing: a pilot trial.
Ann Pharmacother. 2013;47(1):20-28. doi:10.1345/aph.1Q698.
14. Olansky L, Sam S, Lober C, Yared JP, Hoogwerf B. Cleveland
clinic cardiovascular intensive care unit insulin conversion pro-
tocol. J Diabetes Sci Technol. 2009;3(3):478-486.
15. van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin
therapy in critically ill patients. N Engl J Med. 2001;345(19):
1359-1367. doi:10.1056/NEJMoa011300.
16. Schmeltz LR, Desantis AJ, Schmidt K, et al. Conversion of intra-
venous insulin infusions to subcutaneously administered insulin
glargine in patients with hyperglycemia. Endocr Pract. 2006;
12(6):641-650.
17. Umpierrez GE, Smiley D, Jacobs S, et al.Randomized study of
basal-bolus insulin therapy in the inpatient management of patients
with type 2 diabetes undergoing general surgery (RABBIT 2 sur-
gery). Diabetes Care. 2011;34(2):256-261. doi:10.2337/dc10-1407.
18. Dickerson LM, Ye X, Sack JL, Hueston WJ. Glycemic control in
medical inpatients with type 2 diabetes mellitus receiving sliding
scale insulin regimens versus routine diabetes medications:
321
a multicenter randomized controlled trial. Ann Fam Med. 2003;
1(1):29-35. doi:10.1370/afm.2.
19. Pandya N, Thompson S, Sambamoorthi U. The prevalence and
persistence of sliding scale insulin use among newly admitted
elderly nursing home residents with diabetes mellitus. J Am Med
Dir Assoc. 2008;9(9):663-669. doi:10.1016/j.jamda.2008.06.003.
20. Kitabchi AE, Nyenwe E. Sliding-scale insulin: more evidence
needed before final exit? Diabetes Care. 2007;30(9):2409-2410.
doi:10.2337/dc07-1141.
21. ‘‘Glargine’’ [package insert]. Sanofi-aventis US LLC. New Jersey:
Bridgewater; 2009.
22. ‘‘Detemir’’ [package insert]. Novo Nordisk. New Jersey:
Plainsboro; 2015. http://www .novo-pi.com/levemir.pdf.
Accessed October 1, 2014.
23. Howards PP, Schisterman EF, Poole C, Kaufman JS, Weinberg
CR. ‘‘Toward a clearer definition of confounding’’ revisited with
directed acyclic graphs. Am J Epidemiol. 2012;176(6):506-511.
doi:10.1093/aje/kws127.
24. Suttorp MM, Siegerink B, Jager KJ, Zoccali C, Dekker FW. Gra-
phical presentation of confounding in directed acyclic graphs.
Nephrol Dial Transplant. 2015;30(9):1418-1423. doi:10.1093/
ndt/gfu325.
25. Oghazian MB, Javadi MR, Radfar M, et al. Effectiveness
of regular versus glargine insulin in stable critical care
patients receiving parenteral nutrition: a randomized controlled
trial. Pharmacotherapy. 2015;35(2):148-157. doi:10.1002/
phar.1546.
26. Aggarwal A, Rawn JD, Pendergrass ML, Garg RK. Long-acting
subcutaneously administered insulin for glycemic control imme-
diately after cardiac surgery. Endocr Pract. 2011;17(4):558-562.
doi:10.4158/EP10354.OR.
27. Boucai L, Southern WN, Zonszein J. Hypoglycemia-associated
mortality is not drug-associated but linked to comorbidities.
Am J Med. 2011;124(11):1028-1035. doi:10.1016/j.amjmed
.2011.07.011.
28. Tsai T, Lee C, Cheng B, et al. Body mass index—mortality rela-
tionship in severe hypoglycemic patients with type 2 diabetes. Am
J Med Sci. 2015;349(3):192-198.
29. Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM,
Kitabchi AE. Hyperglycemia: an independent marker of
in-hospital mortality in patients with undiagnosed diabetes. J Clin
Endocrinol Metab. 2002;87(3):978-982. doi:10.1210/jc.87.3.978.