Sweet and Sour: Impact of Early Glycemic Control on Outcomes in Necrotizing Soft-Tissue Infections

Abstract

Background:

Necrotizing soft-tissue infection (NSTI) is a devastating disease associated with high rates of morbidity and mortality. Hyperglycemia is associated with poor wound healing; however, there are no studies evaluating glycemic control outcomes in patients with NSTI. The objective of this study was to examine disease progression and death in patients with NSTI who achieved early glycemic control (EGC) compared with patients that did not.

Methods:

A retrospective chart review of patients with NSTI was conducted between November 2011 and August 2017. Early glycemic control was defined as a daily average blood glucose concentration ≤150 mg/dL for a minimum of two consecutive days from admission to hospital day three. The primary outcome of this study was a composite of ≤3 debridement procedures by hospital day 14 and survival to discharge. Secondary outcomes were the total number of debridement procedures, amputation, hospital length of stay (LOS), intensive care unit (ICU) LOS, number of hypoglycemic events throughout hospitalization, and discharge disposition.

Results:

One-hundred five patients were included in the analysis. There were 62% male patients, mean age of 55.3 years, mean weight of 106.9 kg, and 57.1% with diabetes mellitus (DM). The 54 (51.4%) patients with EGC were less likely to have DM (29.6% versus 86.3%; p < 0.001), had a lower median admission glucose concentration (120.5 [97–144] versus 198 [153–295.5] mg/dL; p < 0.001), and had lower median daily glucose values during the first 96 hours after admission (p < 0.001). There was no significant difference in the primary outcome (83.3%% versus 84.3%; p > 0.99) or incidence of hypoglycemia (14.8% versus 23.5%; p = 0.32). Patients with EGC were more likely to return home after discharge (44.4% versus 23.5%; p = 0.039).

Conclusion:

Overall, there was no difference in composite clinical outcomes between patients with EGC and those without, although more patients who achieved EGC were discharged home. Patients with DM were less likely to achieve EGC.

Necrotizing soft-tissue infection (NSTI) is a rapidly progressive infection of the skin and underlying tissues. The infection invades the deeper layers of the skin than is seen in cellulitis, which infects the upper layers of skin such as the dermis primarily and is associated with higher morbidity and mortality rates than other skin and soft tissue infections [1,2]. Involvement of the superficial fascia may lead to greater morbidity because of the major tissue destruction and death caused by the infection as well as the debridement procedures necessary for the treatment of NSTI [2]. Morbidity and death also may result from systemic toxicity such as renal failure and respiratory failure that occur as the disease continues to worsen [1, 2].

The development of NSTI may be caused by a variety of microorganisms, including group A Streptococcus spp. and Clostridium spp. and may be polymicrobial with organisms obtained from the skin and contaminants through differing causes such as non-penetrating trauma, penetrating trauma, and abscess or ulceration. Necrotizing fasciitis is classified into four categories according to microorganisms and underlying insult [2].

The standard of care in the treatment of all types of NSTI is prompt surgical exploration and debridement [2]. Antibiotics are utilized as an adjunct to treatment to prevent continuous microorganism reproduction and infection throughout debridement procedures. Nutritional support, glycemic control, and incision management are critical for healing. Fluctuating glucose concentrations and hyperglycemia often are associated with poor healing, which may prolong utilization of antibiotics and increase the number of necessary debridement procedures [3,4]. Glycemic control improves outcomes in burn, trauma, and critically ill patients [5 –10]. Early glycemic control (EGC) decrease morbidity and mortality rates in these patient populations, which may be attributable in part to improvement in site healing, as these patients may be subject to larger incisions based on injury [5 –10]. Although there appears to be benefits to glycemic control in burn, trauma, and critically ill patients, there are no studies evaluating outcomes of glycemic control in patients with NSTI.

The objective of this study was to determine the difference in the number of debridement procedures and the mortality rate in patients with NSTI who achieved EGC compared with patients who did not.

Materials and Methods

Study design

This was a retrospective cohort study of patients with NSTI treated at an academic medical center and affiliated community hospital. Outcomes related to EGC were collected in patients treated between November 1, 2011, and August 31, 2017. This study was approved by the University Institutional Review Board.

Patients 18 years to 89 years admitted with documented necrotizing fasciitis or Fournier's gangrene per International Classification of Diseases (ICD) code and Current Procedural Terminology (CPT) were eligible for inclusion [11]. Patients were excluded if they were transferred from an outside hospital greater than 24 hours after initial presentation, received the first debridement procedure at an outside hospital, presented in diabetic ketoacidosis or hyperglycemic hyperosmolar syndrome, or died within the first 72 hours after admission.

Patients were categorized according to whether they achieved EGC. Early glycemic control was defined as a daily average blood glucose concentration ≤150 mg/dL for a minimum of two consecutive days, from admission to hospital day three [5]. Demographic and clinical data were extracted from the patient's electronic medical record (EMR). Demographic data collected included age, body mass index (BMI), smoking history, diabetes mellitus (DM) diagnosis prior to admission or during admission, and sequential organ failure assessment (SOFA) score utilizing the most severe data collected within the first 24 hours after admission. Outcomes data included hospital and intensive care unit (ICU) length of stay (LOS), death, and disposition at discharge. Glycemic control was assessed by blood glucose on admission, daily minimum, maximum, and average blood glucose values for the first four days after admission, and number, type, and outcome of hypoglycemic events. Glucose values over time were recorded as time-weight glucose, which was calculated as blood glucose values weighted by the time difference between two consecutive measurements applied to the average of the two consecutive measurements. For example, if the blood glucose concentration at 0500 was 100 mg/dL and the next blood glucose was 200 mg/dL at 1200, the daily blood glucose average would be weighted considering that the blood glucose concentration was 100 mg/dL for seven hours of the total hours in the day, as there was no additional blood glucose measurement recorded between 0500 and 1200. Our institution has a stress-hyperglycemia insulin protocol that is nurse driven. This protocol was initiated in the surgical ICU in 2004 and expanded throughout the other ICUs in 2005 and the non-ICU floors in 2009. From 2004 to 2012, our goal blood glucose range was 110–150 mg/dL, and in 2012, it was changed to a goal range of 120–150 mg/dL, which is still in place today. Hemoglobin A1c data on admission or within three months of presentation were collected to assess glycemic control prior to admission. Infection severity and subsequent wound healing data collected included source of infection; laboratory risk indicator for necrotizing fasciitis (LRINEC) score; necrotizing fasciitis classification; antimicrobial drug selection; duration of antimicrobial administration; presence of concurrent bacteremia; time to the first debridement procedure; number of debridement procedures; number of amputations; initiation, type, and duration of nutritional supplementation; and the need for a diverting colostomy [2].

Outcomes definitions

The primary outcome of this study was a composite of no more than three debridement procedures by hospital day 14 and survival to discharge. A composite primary outcome was selected to provide a more appropriate estimation of outcomes of glycemic control in patients with NSTI where the mortality rate may not be the most common outcome as it is in studies in burn, trauma, and critically ill patients. Additionally, newer trials of patients with NSTI utilize similar composite outcomes to provide a better estimation of mortality and morbidity rates after treatment [12]. Secondary outcomes included the total number of debridement procedures during admission, need for amputation, diagnosis of DM prior to or during admission, hospital LOS, ICU LOS, number of hypoglycemic events throughout the hospitalization, and disposition at discharge.

Data analysis

Descriptive statistics were used for patient demographics. Categorical data are presented as percentages. The Pearson X² test was utilized to determine if there was a difference in the composite clinical outcome between the two cohorts. Additional categorical variables were compared using the Pearson X² or Fisher exact test. Continuous variables are presented as mean ± standard deviation for normally distributed data and compared using the independent Student t-test. Continuous non-normally distributed data are presented as median (25%–75% interquartile range) and compared using the Mann-Whitney test. A p value of <0.05 was considered statistically significant. Analyses were conducted utilizing Statistical Package for the Social Sciences version 24.0 (SPSS, Inc. Chicago, IL).

Results

There were 105 patients included; 48.6% (n = 51) did not achieve EGC. Patients that did not achieve EGC had a mean age of 58.3 ± 12.1 years and a median (interquartile range [IQR]) weight of 106.6 [84.5–128.2] kg compared with EGC patients, who had a mean age of 52.9 ± 14.3 years (p = 0.042) and a median [IQR] weight of 97.3 [75.5–122.6] kg (p = 0.31) (Table 1). The most common classification of NSTI within both the non-EGC and the EGC group was type 1, occurring in 68.6% and 72.2% of patients, respectively (p = 0.48). The median [IQR] LRINEC score for the non-EGC patients was 6 [4 –8], while the median [IQR] LRINEC score for EGC patients was 4 [2 –7] (p = 0.037).

Table 1.

Demographics

	No EGC (n = 51)	EGC (n = 54)	p
Male (%)	32 (62.7)	33 (66.7)	>0.99
Mean age (y) ± SD	58.3 ± 12.1	52.9 ± 14.3	0.042
Median weight (kg) [IQR]	106.6 [84.5–128.2]	97.3 [75.5–122.6]	0.31
Mean height (cm) ± SD	172.5 ± 9.5	172.1 ± 9.8	0.86
Median body mass index [IQR]	35.1 [29.6–42.8]	31.9 [25.6–43.6]	0.37
Smoker (%)	19 (37.2)	20 (37)	>0.99
Type of NSTI (%)
Type 1 NF (non-Fournier's gangrene)	35 (68.6)	39 (72.2)	0.48
Type 2 NF	0	1 (1.9)	>0.99
Type 3 NF	0	0	>0.99
Fournier's gangrene	13 (25.5)	9 (17.6)	0.34
Unknown (no growth in cultures)	3 (5.9)	5 (9.3)	0.72
Median LRINEC score [IQR]	6 [4–8]	4 [2–7]	0.037
Median SOFA score [IQR]	4 [1–9]	2 [0–8]	0.07
Transfer from outside hospital (%)	24 (47.1)	29 (53.7)	0.56
Median time to transfer from outside hospital (h) [IQR]	5.8 [3.9–7.4]	5.6 [4.3–7.1]	0.32
Mechanical ventilation (%)	27 (52.9)	24 (44.4)	0.44
Median duration of mechanical ventilation (d) [IQR]	3.0 [1.4–10.0]	3.0 [0.7–5.0]	0.67
Chronic immunosuppression (%)^a	7 (13.7)	4 (7.4)	0.35
Corticosteroid	3 (42.9)	3 (75)	0.55
Antirejection drug	3 (42.9)	0	0.24
Chemotherapy	1 (14.3)	2 (50)	0.49
Biologic drug	1 (14.3)	0	>0.99
Median day of diet initiation [IQR]	2 [2–3]	2 [1–3]	0.1

Patients may have received more than one type of immunosuppression.

EGC = early glycemic control; IQR = interquartile range; LRINEC = laboratory risk indicator for necrotizing fasciitis; NF = necrotizing fasciitis; NSTI = necrotizing soft tissue infection; SOFA = sequential organ failure assessment; SD = standard deviation.

Patients who did not achieve EGC were significantly more likely to have a diagnosis of DM (86.2% vs. 29.6%; p < 0.001) (Table 2). Additionally, non-EGC patients had significantly elevated blood glucose concentrations from admission to hospital day 4. The median [IQR] blood glucose concentration on admission in the non-EGC group was 198.0 [153.0–295.5] mg/dL compared to 120.5 [97.0–144.0] mg/dL in the EGC group (p < 0.001). The non-EGC patients were significantly more likely to receive insulin in some manner during admission (90.2% vs. 48.1%; p < 0.001) and were more likely to receive an insulin infusion (54.9% vs. 14.8%; p < 0.001). Overall, 19.5% of these patients developed hypoglycemia, and there was no difference in the number of hypoglycemic events between the groups (23.5% vs 14.8%; p = 0.32). Adverse events reported relative to hypoglycemia included confusion/unresponsiveness (2/20; 10.0%) and sweating (1/20; 5.0%), although in most cases, no adverse events were documented in the EMR.

Table 2.

Glycemic Control

	No EGC (n = 51)	EGC (n = 54)	p
Diagnosis of DM (%)	44 (86.3)	16 (29.6)	<0.001
Diabetes prior to admission	41 (80.4)	16 (29.6)	<0.001
Diabetes diagnosed during admission	3 (5.9)	0	0.11
Median hemoglobin A1c (%) [IQR]	10.0 [6.6–10.9]	5.8 [5.4–7.1]	<0.001
Hemoglobin A1c at admission (%)	24 (47.1)	13 (24.1)	0.016
Median blood glucose concentration (mg/dL) [IQR]
At admission	198.0 [153.0–295.5]	120.5 [97.0–144.0]	<0.001
Day 1	187.0 [159.0–226.5]	127.5 [103.0–153.0]	<0.001
Day 2	168.0 [152.5–184.5]	128.0 [104.0–139.0]	<0.001
Day 3	168.0 [151.0–197.5]	118.0 [102.0–130.5]	<0.001
Day 4	158.0 [144.0–195.0]	120.0 [99.0–134.0]	<0.001
Any insulin administered (%)	50 (98)	26 (48.1)	<0.001
Subcutaneous insulin (%)	36 (70.6)	24 (44.4)	0.01
Median time to subcutaneous insulin (h) [IQR]	8.8 [5.8–15.7]	14.7 [11.5–35.2]	0.002
Insulin infusion (%)	28 (54.9)	8 (14.8)	<0.001
Median time to insulin infusion (h) [IQR]	8.1 [5.8–24.9]	16.7 [3.5–34.3]	0.9
Transition of subcutaneous to intravenous insulin (%)	17 (33.3)	5 (9.3)	0.064
Time to transition of subcutaneous to intravenous insulin (h), median [IQR]	7.3 [3.7–27.1]	10.5 [5.5–34.4]	0.88
Hypoglycemia, n (%)	12 (23.5)	8 (14.8)	0.32
Median nadir blood glucose (mg/dL) [IQR]	61 [53–64.5]	61.5 [48–68]	0.72
Median hypoglycemia events per patient [IQR]	1.5 [1.0–3.5]	1.0 [1.0–2.5]	0.52

EGC = early glycemic control; DM = diabetes mellitus.

The primary composite outcome of three or fewer debridement procedures by hospital day 14 and survival to discharge occurred in 84.3% of the non-EGC patients compared with 83.3% of EGC patients (p > 0.99) (Table 3). There was no difference in the number of patients who had three or fewer debridements by day 14 of hospitalization (p = 0.15) or survival to discharge (p = 0.71). The median [IQR] hospital LOS was 13.0 [9.0–23.5] days in the non-EGC group and 11.0 [8.0–23.0] days in the EGC group (p = 0.29) with no difference in the ICU admission rate (p = 0.2). At discharge, significantly more EGC patients were discharged home than non-EGC patients (44.4% vs. 23.5%; p = 0.039).

Table 3.

Primary and Secondary Outcomes

	No EGC (n = 51)	EGC (n = 54)	p
Composite clinical outcome (%)	43 (84.3)	45 (83.3)	>0.99
≤3 debridements by day 14 of admission, n (%)	47 (92.2)	44 (81.2)	0.15
Survival to discharge (%)	47 (92.2)	51 (94.4)	0.71
Median debridements [IQR]	1.0 [1.0–2.0]	2.0 [1.0–2.0]	0.20
Amputation (%)	1 (2)	1 (1.9)	>0.99
Diverting colostomy (%)	6 (11.8)	6 (11.1)	>0.99
Median hospital length of stay (d) [IQR]	13.0 [9.0–23.5]	11.0 [8.0–23.0]	0.29
ICU admission (%)	40 (78.4)	36 (66.7)	0.2
Median ICU length of stay (d)[IQR]	6.5 [3.5–11.5]	5.5 [4.0–10.5]	0.85
Disposition at discharge (%)
Skilled nursing facility	16 (31.4)	14 (25.9)	0.67
Home	12 (23.5)	24 (44.4)	0.039
Long-term acute care hospital	12 (23.5)	11 (20.4)	0.81
Hospice	4 (7.8)	1 (1.9)	0.20
Inpatient rehabilitation	3 (5.9)	1 (1.9)	0.35

Discussion

There was no significant difference in the primary composite outcome between patients who achieved and did not achieve EGC, although patients who achieved EGC were less likely to have DM. Of the 86.3% of non-EGC patients who were found to have DM, the disease was diagnosed prior to admission in the majority (80.4%). The median hemoglobin A1c vale among patients who did not achieve EGC was 10 [6.6–10.9], indicating that their blood glucose values were consistently uncontrolled from a time prior to admission.

It is not surprising that those who did not achieve EGC had a statistically higher LRINEC score, as glucose values >180 mg/dL add points to the score, and these patients had a higher median admission blood glucose concentration [198 mg/dL vs. 120.5 mg/dL; p < 0.001]. A study by Shakya et al. demonstrated that in patients with persistently elevated blood glucose concentrations, the hyaluronan within the pericellular coat of the small vasculature changes [4]. With this change, inflammatory markers become more abundant and gather in the vasculature, which limits blood flow and may prolong incision healing [4]. In a study by Leong et al., similar results regarding increased vasculature inflammation with hyperglycemia are discussed, and the authors concluded that the compilation of decreased blood flow and alterations in cellular function may prohibit or prolong healing [3,4]. Prolonged incision healing may be detrimental in critically ill patients by prolonging hospital stays and thus increasing the risk of further infections and result in higher cost to the patient [3]. In order to prevent adverse effects on healing in patients with hyperglycemia and DM relative, timeliness to initiation of important medications, such as insulin infusions, is crucial. Reducing the frequency of persistent hyperglycemia while the patient is in the hospital might prevent a delay in site healing.

To our knowledge, this is the first study to examine the specific effects of glycemic control in patients with NSTI. Our results differ from those in other patient populations, such as burn, trauma, and critically ill patients. For example, a study by Murphy et al. examined the influence of EGC in burn patients to determine if there was a difference in the hospital mortality rate [5]. Early glycemic control was defined as the achievement of an average blood glucose concentration ≤150 mg/dL for two consecutive days by day three of admission to the hospital [5]. In patients who achieved EGC, there was a higher rate of survival (p = 0.03). Additionally, in patients who did not achieve EGC, there was a 6.75-fold increase in the mortality rate when adjusted for age, burn size, and presence of inhalation injury [5]. A study of trauma patients by Sperry et al. determined that persistently hyperglycemic patients, defined as those having a blood glucose concentration ≥130 mg/dL for at least 48 hours since injury, had a significantly higher crude mortality rate (p = 0.002) [6]. Van den Berghe et al. designed a study to determine whether the normalization of the blood glucose concentration with insulin improves the prognosis of critically ill patients in the surgical ICU [7]. Intensive insulin control was defined as a blood glucose concentration between 80 and 110 mg/dL, while conventional glucose control was defined as maintaining a blood glucose concentration between 180 and 200 mg/dL after the initiation of insulin once the blood glucose value was >215 mg/dL [7]. Intensive insulin control was associated with a 32% decrease in the mortality rate compared with patients who received conventional treatment (p < 0.04) [7]. Although there was no difference in the primary outcome among patients with and without EGC in NSTI patients, it may be in part because both cohorts were small compared with those in the other studies where power was obtained, and a difference was seen. Because of the retrospective nature of this study, we did not look at incision healing, where EGC may be beneficial as an endpoint.

Patients who achieved EGC were statistically more likely to return home rather than to some type of healthcare facility after the hospitalization compared with patients who did not achieve EGC. More patients who did not achieve EGC were discharged to a skilled nursing facility, 31.4% of patients compared with 23.5% who were discharged home. This discovery may help improve patient satisfaction, as they are able to return to a known environment compared with a facility that may be able to care for the patient for only a short time because of insurance restraints. Additionally, if a patient can return home at discharge, costs will likely be decreased for patients as well as the insurance company, as according to one study, the average private room in a skilled nursing facility costs $253 per day, or more than $92,000 annually [13].

Despite the significant results, this study is not without limitations. Although there was an inclusion time of six years, the patient sample was small. Power was not calculated because of the anticipated small sample, and it may be that the study was not able to identify a difference in outcomes despite p values. In one example, the rate of hypoglycemia was around 1.5 times greater in the EGC group despite the p value indicating no difference. As there was no difference in the nadir blood glucose concentration, this finding is likely more clinically than statistically significant. There are multiple components that affect hypoglycemia, such as abrupt discontinuation of continuous enteral or parenteral nutrition, liver insufficiency, or spontaneously with or without a continuous glucose source, and this study chose to examine only insulin therapy [14].

Also, because of the small sample, only univariable analyses were performed, which would not adjust for confounders; however, it was determined that adjustment for confounders would not have altered the results of the study because of the small sample achieving the primary composite outcome. There also may have been documentation inconsistencies within billing codes after admission. In order to limit bias, patients were included only if they met the criteria for the selected ICD and CPT codes to ensure patients were diagnosed consistently and treated appropriately.

When calculating the time-weighted blood glucose values for patients, it was discovered that there may been differences in nursing care and frequency of blood glucose checks among patients. At our institution, while patients are on an insulin infusion, they are to have a point of care blood glucose check completed every hour. For unknown reasons, blood glucose checks were not recorded consistently for all patients on an insulin infusion, which may have affected the time-weighted blood glucose calculation. Additionally, patients receiving subcutaneous insulin are to have their blood glucose check via point of care testing three to four times per day, which also did not occur consistently in all patients for unknown reasons. This study also did not evaluate whether the patient was receiving appropriate insulin coverage for his or her diet orders. For example, if a patient had a diet ordered, it was not evaluated whether he or she was receiving fast-acting insulin for meal coverage.

Lastly, although there was a significant difference in the number of patients who were discharged home depending on whether the patient did or did not achieve EGC, from this retrospective study, it cannot be determined if home was the optimal location for patients. This study did not collect data on whether insurance restrictions or patient preference prevented a patient from being discharged to another location such as a skilled nursing or rehabilitation facility, even if that may have been what was best for the patient at the time of discharge.

Although there were a variety of limitations associated with this study, this is the first research evaluating outcomes of EGC, and it provides new information regarding EGC in NSTI patients. Future steps include targeted education to describe the goals of blood glucose control and the prevention of hyperglycemia and hypoglycemia. Education will ensure that all team members understand the multitude of factors that are incorporated into a decision to target a specific blood glucose goal, such as underlying illness, glycemic variability, and timing and amount of nutrition [14]. In addition to institutional hyperglycemia guidelines, our institution has guidelines in place for the treatment of hypoglycemia that were designed specifically for patients in the ICU but have been adapted to patients on medicine floors as well, which may be utilized to prevent adverse effects of insulin therapy.

Conclusion

Overall, there was no significant difference in the primary composite clinical outcome in patients who did and did not achieve EGC. Patients who did not achieve EGC were significantly more likely to have DM as well as elevated median blood glucose values throughout the first four days of hospitalization. Elevated blood glucoses may have contributed to patient disposition, as patients achieving EGC were more likely to be discharged home than patients who did not achieve EGC. Future studies with larger patient cohorts may be able to identify other differences in outcomes between patients with NSTI who either did or did not achieve EGC.

Footnotes

Author Disclosure Statement

No competing financial interests exist for any of the authors.

References

Steven

, Bisno

, Chambers

, et al. Practice Guidelines for the Diagnosis and Management of Skin and Soft Tissue Infections: 2014 Update by the Infectious Diseases Society of America. Clin Infect Dis, 2014; 59:e10–e52.

Stevens

, Bryant

. Necrotizing soft-tissue infections. N Engl J Med, 2017; 377:2253–2265.

Leong

, Murphy

, Phillips

. Wound healing. In: Townsend CM, Beauchamp RD, Evers BM, Mattox KL, eds. Sabiston Textbook of Surgery. 20^th edition. Philadelphia: Elsevier Mosby;, 2017:130–162.

Shakya

, Wang

, Mack

, Maytin

. Hyperglycemia-induced changes in hyaluronan contribute to impaired skin wound healing in diabetes: Review and perspective. Int J Cell Biol, 2015; 2015:701–738.

Murphy

, Coffey

, Cook

, et al. Early glycemic control in critically ill patients with burn injury. J Burn Care Res, 2011; 32:583–590.

Sperry

, Frankel

, Nathens

, et al. Characterization of persistent hyperglycemia: What does it mean postinjury. J Trauma, 2009; 66:1076–1082.

Van den Berghe

, Wouters

, Weekers

, et al. Intensive insulin control in critically ill patients. N Engl J Med, 2001; 345:1359–1367.

Gibson

, Galiatsatos

, Rabiee

, et al. Intensive insulin therapy confers a similar survival benefit in the burn intensive care unit to the surgical intensive care unit. Surgery, 2009; 146:922–930.

Schwartz

, Rothrock

, Barron-Vaya

, et al. Impact of diabetes on burn injury: Preliminary results from prospective study. J Burn Care Res, 2011; 32:435–441.

10.

Egi

, Bellomo

, Stachowski

, et al. Blood glucose concentration and outcome of critical illness: The impact of diabetes. Crit Care Med, 2008; 36:2249–2255.

11.

Centers for Medicare and Medicaid Services. ICD-10. Available at: https://www.cms.gov/Medicare/Coding/ICD10/index.html. August 31, 2017. Accessed September 18, 2017.

12.

US National Library of Medicine. Phase III efficacy and safety study of AB103 in the treatment of patients with necrotizing soft tissue infections (ACCUTE). Available at: https://clinicaltrials.gov/ct2/show/NCT02469857?cond=Necrotizing+Soft+Tissue+Infection&draw=2&rank=3. June 12, 2015. Accessed September 18, 2017.

13.

Mullin

, Esposito

“How to Pay for Nursing Home Costs.” US News and World Report: Health. 2016. Available at: https://health.usnews.com/wellness/articles/2016-11-16/how-to-pay-for-nursing-home-costs. Accessed May 5, 2018.

14.

Gerlach

, MacDermott

, Newton

, et al. Stamping out hypoglycemia in a surgical intensive care unit: A multidisciplinary approach. IJAM, 2016; 2:154–158.