Managing Diabetes and Hyperglycemia in the Hospital Setting. Boris Draznin

mg/dL is used,^2,3 and up to 90% with a lower glucose threshold of 150 mg/dL.^3,4 In another study, 51% of patients on PN were found to have at least one blood glucose (BG) measurement >140 mg/dL.⁵

      The development of hyperglycemia during EN via TF/PN increases the risk of complications and mortality. Adverse consequences may include cardiac complications, infections, sepsis, acute kidney injury, and increased mortality. Pasquel et al. found an increased risk of pneumonia (odds ratio [OR] 3.1, 95% confidence interval [CI] 1.4–7.1), acute renal failure (OR 2.3, 95% CI 1.1–5.0), and mortality (OR 2.8, 95% CI 1.2–6.8, P = 0.02) for patients on total parenteral nutrition (TPN) with a BG >180 mg/dL within 24 h of starting PN.⁴ A multicenter study in Spain reported a 5.6-fold risk of mortality after adjusting for multiple factors in patients with a mean BG >180 mg/dL during PN infusion as compared with a mean BG of <140 mg/dL.⁶

      The Endocrine Society’s clinical practice guidelines for the management of hyperglycemia in the non-critical-care inpatient setting include the following recommendations for patients receiving EN or PN, with or without a history of diabetes:⁷

      1. All patients on EN or PN should be monitored with bedside point-of-care (POC) glucose testing every 4–6 h for at least 24–48 h after initiation of the therapy.

      2. POC glucose testing can be discontinued in patients without a prior history of diabetes if BG values are <7.8 mmol/L (140 mg/dL) without insulin therapy for 24–48 h after achievement of desired caloric intake.

      3. Scheduled insulin therapy should be initiated in patients receiving EN or PN who have hyperglycemia (i.e., BG >7.8 mmol/L (140 mg/dL), with and without known diabetes, and who demonstrate a persistent requirement (i.e., 12–24 h) for correction insulin.

      Unfortunately, strong evidence for the effects of good glycemic control on clinical outcomes or for specific glucose-lowering regimens in patients on EN or PN is lacking; well-designed research studies are needed to close these gaps and improve clinical care.

      Blood Glucose Targets in EN and PN

      A recently published systematic review and meta-analysis of randomized control trials comparing intensive insulin therapy (IIT) versus control in an intensive care unit (ICU) setting is relevant to this discussion because a significant proportion of these patients received PN or EN.⁸ The mean BG of patients on IIT was 112 mg/dL, whereas the mean BG on control therapy was 151 mg/dL. Overall, there was no reduction in mortality with IIT at 28 days, and no decrease in incidence of bloodstream infection or need for renal replacement therapy. Unfortunately there was a higher odds ratio for hypoglycemia (OR 7.7). Because of the difference in 28-day mortality between the Leuven IIT Trials and subsequent trials, a metaregression analysis was performed, which showed a significant difference between 28-day mortality and the proportion of calories provided parenterally (P = 0.005). This suggested that the difference in outcome between the two Leuven IIT Trials and the subsequent trials could be related to the use of PN. After excluding the two Leuven trials that included patients receiving a high percentage of calories from PN, the authors found lower mortality in the control group patients. In general, there was less mortality with IIT in PN patients, but increased mortality with IIT in EN patients. No data support the use of an intensive insulin target in patients receiving EN, and a BG target of 150 mg/dL was considered reasonable for patients being fed enterally.

      The American Society for Parenteral and Enteral Nutrition (ASPEN) recently reviewed available trials comparing tight versus standard glycemic control and mortality.⁹ Five retrospective reviews were included in the analysis that found harm associated with PN-associated hyperglycemia. Tight control resulted in lower mortality in one randomized controlled trial (RCT) and two historical control trials, but there was no difference between tight versus standard control in five other RCTs. The largest RCT¹⁰ demonstrated higher mortality with tight glycemic control. Overall, the ASPEN group found strong evidence for the recommendation of a BG target of 140–180 mg/dL in patients receiving nutritional support and strong evidence for defining hypoglycemia in this setting as a BG of <70 mg/dL.

      The authors suggest that patients receiving artificial nutrition in the form of PN or EN are in a postprandial state, and the previous evidence supports a BG target of 140–180 mg/dL in this setting.

      Parenteral Nutrition

      Evidence for Glycemic Control in Patients on PN

      A number of studies have examined the effect of glycemic control on clinical outcomes in patients receiving PN. A few of these studies are summarized in this section.

      In a single-center retrospective analysis of 109 patients receiving TPN, the odds ratio in relation to BG level of developing a number of different complications ranged from 1.36 to 1.77, all of which were statistically significant.¹¹ Patients in the highest quartile of BG level had an odds ratio of 10.9 for mortality or for acute renal failure, and an odds ratio of 4.3 for developing any complication as compared with patients in the lowest quartile of mean BG level, independent of age, sex, or prior diabetes status.

      Another single-center retrospective study of 276 medical and surgical patients receiving TPN examined the likelihood of death and BG levels on admission, pre-TPN, within 24 h of starting TPN, and 2–10 days after starting TPN.⁴ In multiple regression models adjusted for age, sex, and history of diabetes, the likelihood of death was independently predicted by an elevated pre-TPN BG between 121 and 150 mg/dL (OR 2.2, 95% CI 1.1–4.4, P = 0.030), 151 and 180 mg/dL (OR 3.41, 95% CI 1.3–8.7, P = 0.010), >180 mg/dL (OR 2.2, 95% CI 0.9–5.2, P = 0.077), or >180 mg/dL within 24 h of starting TPN, as compared with patients with a mean BG <120 mg/dL. Patients with higher BG levels during TPN had a longer hospital (P = 0.011) and ICU (P = 0.008) length of stay.

      A prospective multicenter study of 605 noncritically ill patients receiving PN in Spain with 58 deaths (9.6%) revealed a 5.6-fold higher risk of mortality for patients with a mean BG >180 mg/dL during TPN infusion after multiple logistic regression analysis.⁶ The total number of grams of carbohydrates infused was associated with risk of in-hospital death. However, the presence of diabetes was not significant. Approximately half of these patients had normal glucose tolerance (mean HbA_1c 5.3%, n = 308, 50.9% of patients), 27.4% of patients had hyperglycemia without diabetes (mean HbA_1c 5.7%, n = 166), 17.9% of patients had known diabetes (mean HbA_1c 6.6%, n = 108), and the remainder had “unknown,” or previously undiagnosed diabetes (mean HbA_1c 7.2%, n = 23). Forty-one (6.9%) patients had at least one episode of hypoglycemia <70 mg/dL at some time during PN administration.

      Hyperglycemia before and during PN is associated with an increased risk for complications and mortality in patients receiving PN. No published RCT data, however, show that glycemic control reduces the risk for complications or mortality in this setting.

      Evidence for Glucose-Lowering Strategies in Patients Receiving PN

      Despite the lack of RCT data showing that glycemic control reduces the risk for complications or mortality in patients receiving PN, the available data show significant harm associated with uncontrolled hyperglycemia in this population. Therefore, clinicians must anticipate and manage hyperglycemia in their patients on PN while avoiding hypoglycemia.

      In the multicenter prospective cohort described earlier, 71% of patients received insulin along with PN, with 55% receiving this as subcutaneous (SQ) insulin only, 36% as a combination of insulin in the PN bag along with SQ insulin, and 9% as an insulin infusion independent of the PN.⁶ Table 6.1 lists selected studies examining specific glucose-lowering regimens for PN. Standard strategies include (1) using a separate intravenous (IV) insulin infusion of regular insulin with an IV insulin protocol, (2) using subcutaneously administered long- or intermediate-acting insulin with correctional insulin, with or without scheduled prandial insulin, and (3) adding regular insulin to the PN bag in combination with subcutaneously administered basal or correctional insulin. Details of a few studies are outlined in the following paragraphs.

      Table 6.1—Selected Studies of Glucose-Lowering Strategies for Patients on PN