Identifying Risk Factors for Severe Hypoglycemia in Hospitalized Patients with Diabetes

ObjectiveSome of the deleterious effects of hypoglycemia in hospitalized patients include increased rates of mortality and longer length of stay. Our primary objective was to identify the risk factors associated with severe hypoglycemia to identify those patients at highest risk.MethodsThe medical records of 5,026 patients with diabetes mellitus (DM) admitted in 2010 were reviewed to identify those patients that developed severe hypoglycemia (blood glucose [BG] < 40 mg/dL). We performed c2 tests to assess statistical significance. Adjusted logical regression was used to determine the risk factors for hypoglycemia in the hospital.ResultsOut of 5,026 DM patients included in our review, 81 experienced severe hypoglycemia (1.6%). Statistically higher proportions of chronic kidney disease (CKD; 69.1% vs. 46.9%, P < .001), congestive heart failure (CHF; 48.1% vs. 28.5%, P < .001), sepsis (49.4% vs. 12.5%, P < .001), insulin use (45.7% vs. 26.04%, P = .000), type 1 DM (21% vs. 5.1%, P = .000), and cirrhosis (14.8% vs. 7.2%, P = .009) were seen in the severe hypoglycemic group compared to the nonsevere hypoglycemic group. Overall, 84% of patients who experienced an episode of severe hypoglycemia in the hospital (BG < 40 mg/dL) had a previous episode of hypoglycemia (BG < 70 mg/dL). The odds ratios (ORs) for type 1 DM, sepsis, previous hypoglycemia, and insulin use were 3.43 (95% confidence interval [CI] 1.81, 6.49), 2.64 (95% CI 1.6, 4.35), 46.1 (95% CI 24.76, 85.74), and 1.66 (95% CI 1.02, 2.69), respectively.ConclusionPrior episodes of hypoglycemia in the hospital, the presence of type 1 DM, insulin use, and sepsis were identified as independent risk factors for the development of severe hypoglycemia in the hospital. (Endocr Pract. 2014;20:1051-1056)     

Benchmarking Glycemic Control In U.S. Hospitals

ObjectiveReport data on glucose control from 635 U.S. hospitals.MethodsPoint-of-care blood glucose (POC-BG) test data from January through December 2012 from 635 facilities were extracted. Glucose control was evaluated using patient-day–weighted mean POC-BG values. We calculated hypoglycemia and hyperglycemia rates, stratified by presence or absence of intensive care unit (ICU) admission, and we evaluated the relationship between glycemic control and hospital characteristics.ResultsIn total, 51,375,764 POC-BG measurements (non-ICU, 39,197,762; ICU, 12,178,002) from 2,612,966 patients (non-ICU, 2,415,209; ICU, 575,084) were analyzed. The mean POC-BG was 167 mg/dL for non-ICU patients and 170 mg/dL for ICU patients. The prevalence of hyperglycemia (defined as glucose value > 180 mg/dL) was 32.3 and 28.2% in non-ICU and ICU patients, respectively. The prevalence of hypoglycemia (defined as glucose value < 70 mg/dL) was 6.1 and 5.6% in non-ICU and ICU patients, respectively. In non-ICU and ICU settings, the patient-day–weighted mean glucose was highest in the smallest hospitals, in rural hospitals, and in hospitals located in the Northeast (all P < .01). For non-ICU patients, we observed a significant difference in the percentage of patient days with hypoglycemia by geographic region only (P < .001). In ICU patients, the prevalence of hypoglycemia varied significantly by hospital type (P < .03) and geographic region (P < .01).ConclusionIn this largest POC-BG data set analysis conducted to date, glycemic control varied according to hospital characteristics. Our findings remain consistent with previous reports. Among other variables, national benchmarking of inpatient glucose data will need to consider differences in hospital characteristics. (Endocr Pract. 2014;20:876-883)     

Comparison of Basal Insulin Regimens on Glycemic Variability in Noncritically Ill Patients with Type 2 Diabetes

Objective: To evaluate the impact of different subcutaneous basal insulin regimens on glycemic variability (GV) and hospital complications in non-intensive care unit (ICU) patients with type 2 diabetes (T2D).Methods: This study is a post hoc analysis of 279 general medicine and surgery patients treated with either a “Basal Bolus” insulin regimen using glargine once daily and glulisine before meals or a “Basal Plus” regimen using glargine once daily plus correction doses of glulisine before meals for glucose >140 mg/dL. GV was calculated as mean delta (Δ) daily glucose, mean SD, and mean amplitude of glycemic excursions (MAGE).Results: Treatment with Basal Bolus and Basal Plus regimens resulted in similar mean daily glucose, hypoglycemia, length of stay (LOS), and hospital complications (all P>.05). There were no differences in GV between treatment groups by Δ change (72.5 ± 36 vs. 69.3 ± 34 mg/dL), SD (38.5 ± 18 vs. 37.1 ± 16 mg/dL) and MAGE (67.5 ± 34 vs. 66.1 ± 39 mg/dL) (all P>.05). Surgery patients treated with Basal Bolus had higher GV compared to those treated with Basal Plus (Δ daily glucose and SD: P = .02, MAGE: P = .009), but no difference in GV was found between treatment groups for the general medicine patients (P>.05). Patients with hypoglycemia events had higher GV compared to subjects without hypoglycemia (P<.05), but no association was found between GV and hospital complications (P>.05).Conclusion: Treating hospitalized, non-ICU, diabetic patients with Basal Plus insulin regimen resulted in similar glucose control and GV compared to the standard Basal Bolus insulin regimen. Higher GV was not associated with hospital complications.Abbreviations:BG = blood glucoseCV= coefficient of variationGV= glycemic variabilityICU = intensive care unitLOS = length of stayMAGE = mean amplitude of glycemic excursionsSSI = sliding scale insulinT2D = type 2 diabetesTDD =total daily dose     

Daily Inpatient Glycemic Survey (DINGS): A Process to Remotely Identify and Assist in the Management of Hospitalized Patients with Diabetes and Hyperglycemia

Objective: Hyperglycemia, hypoglycemia, and glycemic variability have been associated with increased morbidity, mortality, and overall costs of care in hospitalized patients. At the Stratton VA Medical Center in Albany, New York, a process aimed to improve inpatient glycemic control by remotely assisting primary care teams in the management of hyperglycemia and diabetes was designed.Methods: An electronic query comprised of hospitalized patients with glucose values <70 mg/dL or >350 mg/dL is generated daily. Electronic medical records (EMRs) are individually reviewed by diabetes specialist providers, and management recommendations are sent to primary care teams when applicable. Glucose data was retrospectively examined before and after the establishment of the daily inpatient glycemic survey (DINGS) process, and rates of hyperglycemia and hypoglycemia were compared.Results: Patient-day mean glucose slightly but significantly decreased from 177.6 ± 64.4 to 173.2 ± 59.4 mg/dL (P<.001). The percentage of patient-days with any value >350 mg/dL also decreased from 9.69 to 7.36% (P<.001), while the percentage of patient-days with mean glucose values in the range of 90 to 180 mg/dL increased from 58.1 to 61.4% (P<.001). Glycemic variability, assessed by the SD of glucose, significantly decreased from 53.9 to 49.8 mg/dL (P<.001). Moreover, rates of hypoglycemia (<70 mg/dL) decreased significantly by 41% (P<.001).Conclusion: Quality metrics of inpatient glycemic control improved significantly after the establishment of the DINGS process within our facility. Prospective controlled studies are needed to confirm a causal association.Abbreviations: DINGS = daily inpatient glycemic survey EMR = electronic medical record HbA1c = glycated hemoglobin ICU = intensive care unit VA = Veterans Affairs     

Patients Achieving Good Glycemic Control (HBA1c <7%) Experience A Lower Rate of Hypoglycemia With Insulin Degludec Than with Insulin Glargine: A Meta-Analysis of Phase 3A Trials

Objective: Meta-analysis to compare hypoglycemia rates of basal insulin degludec (IDeg) with insulin glargine (IGlar) in patients with diabetes achieving good glycemic control (hemoglobin A_1c [HbA_1c] <7% at end of trial).Methods: In a preplanned meta-analysis, patient data from 7 randomized, treat-to-target, 26- or 52-week trials in patients with type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM) who administered IDeg (n = 2,899) or IGlar (n = 1,431) once daily were analyzed. Using a negative binomial regression model, this meta-analysis compared hypoglycemia rates in patients achieving HbA_1c <7% at end of trial with IDeg (n = 1,347) and IGlar (n = 697).Results: In all trials, IDeg was noninferior to IGlar in HbA_1c reduction from baseline. At end of trial, 2,044 patients (T2DM, n = 1,661; T1DM, n = 383) achieved HbA_1c <7%. The overall confirmed hypoglycemia rate, defined as plasma glucose <56 mg/dL or severe hypoglycemia if requiring assistance, was significantly lower with IDeg versus IGlar (estimated rate ratio [ERR] IDeg:IGlar, 0.86; 95% confidence interval [CI], 0.76 to 0.98). The nocturnal confirmed hypoglycemia rate, defined as occurring between midnight and 6:00 am, was significantly lower with IDeg (ERR, 0.63; 95% CI, 0.52 to 0.77). In the maintenance period (16 weeks onward when average insulin dose and glycemic levels stabilized), the overall confirmed hypoglycemia rate was significantly lower (ERR, 0.79; 95% CI, 0.68 to 0.92) and the nocturnal confirmed hypoglycemia rate was significantly lower (ERR, 0.57; 95% CI, 0.45 to 0.72) with IDeg versus IGlar.Conclusion: Patients with T1DM and T2DM achieved HbA_1c <7% with significantly lower rates of overall and nocturnal confirmed hypoglycemia with IDeg versus IGlar. The lower hypoglycemia rate with IDeg was more pronounced in the maintenance period.Abbreviations: ERR = estimated rate ratio; HbA_1c = hemoglobin A_1c; IDeg = insulin degludec; IGlar = insulin glargine; NPH = Neutral Protamine Hagedorn; PG = plasma glucose; T1DM = type 1 diabetes mellitus; T2DM = type 2 diabetes mellitus     

Effects of a Computerized Order Set on the Inpatient Management of Hyperglycemia: A Cluster-Randomized Controlled Trial

ObjectiveTo determine the effects of a computerized order set on the inpatient management of diabetes and hyperglycemia.MethodsWe conducted a cluster-randomized controlled trial on the general medical service of an academic medical center staffed by residents and hospitalists. Consecutively enrolled patients with diabetes mellitus or inpatient hyperglycemia were randomized on the basis of their medical team to usual care (control group) or an admission order set built into the hospital’s computer provider order entry (CPOE) system (intervention group). All teams received a detailed subcutaneous insulin protocol and case-based education. The primary outcome was the mean percent of glucose readings per patient between 60 and 180 mg/dL.ResultsBetween April 5 and June 22, 2006, we identified 179 eligible study subjects. The mean percent of glucose readings per patient between 60 and 180 mg/dL was 75% in the intervention group and 71% in the usual care group (adjusted relative risk, 1.36; 95% confidence interval, 1.03 to 1.80). In comparison with usual care, the intervention group also had a lower patient-day weighted mean glucose (148 mg/dL versus 158 mg/dL, P = .04), less use of sliding-scale insulin by itself (25% versus 58%, P = .01), and no significant difference in the rate of severe hypoglycemia (glucose < 40 mg/dL; 0.5% versus 0.3% of patient-days, P = .58).ConclusionThe use of an order set built into a hospital’s CPOE system led to improvements in glycemic control and insulin ordering without causing a significant increase in hypoglycemia. Other institutions with CPOE should consider adopting similar order sets as part of a comprehensive inpatient glycemic management program. (Endocr Pract. 2010;16:209-218)     

Update on Inpatient Glycemic Control in Hospitals in the United States

ObjectiveTo provide data on glucose control in hospitals in the United States, analyzing measurements from the largest number of facilities to date.MethodsPoint-of-care bedside glucose (POC-BG) test results were extracted from 575 hospitals from January 2009 to December 2009 by using a laboratory information management system. Glycemic control for patients in the intensive care unit (ICU) and non-ICU areas was assessed by calculating patient-day-weighted mean POC-BG values and rates of hypoglycemia and hyperglycemia. The relationship between POC-BG levels and hospital characteristics was determined.ResultsA total of 49,191,313 POC-BG measurements (12,176,299 ICU and 37,015,014 non-ICU values) were obtained from 3,484,795 inpatients (653,359 in the ICU and 2,831,436 in non-ICU areas). The mean POC-BG was 167 mg/dL for ICU patients and 166 mg/dL for nonICU patients. The prevalence of hyperglycemia (> 180 mg/ dL) was 32.2% of patient-days for ICU patients and 32.0% of patient-days for non-ICU patients. The prevalence of hypoglycemia (< 70 mg/dL) was 6.3% of patient-days for ICU patients and 5.7% of patient-days for non-ICU patients. Patient-day-weighted mean POC-BG levels varied on the basis of hospital size (P < .01), type (P < .01), and geographic location (P < .01) for ICU and non-ICU patients, with larger hospitals (≥ 400 beds), academic hospitals, and US hospitals in the West having the lowest mean POC-BG values. The percentage of patient-days in the ICU characterized by hypoglycemia was highest among larger and academic hospitals (P < .05) and least among hospitals in the Northeast (P < .001).ConclusionHyperglycemia is common in hospitals in the United States, and glycemic control may vary on the basis of hospital characteristics. Increased hospital participation in data collection may support a national benchmarking process for the development of optimal practices to manage inpatient hyperglycemia. (Endocr Pract. 2011;17:853-861)     

Fasting Versus Postprandial Hyperglycemia as a Treatment Target to Lower Elevated Hemoglobin A1C

Objective: Postprandial hyperglycemia (PPHG) may need addressing when glycemic control cannot be maintained in patients with type 2 diabetes mellitus. We investigated whether glycated hemoglobin A_1c (A_1c) levels ≥7.0% can indicate postprandial defects warranting prandial therapy after optimized basal insulin therapy.Methods: From 6 clinical trials of insulin glargine treatment, data were pooled from 496 patients with A_1c ≥7.0% after 24 weeks. Patient characteristics and clinical outcomes were summarized according to fasting plasma glucose (FPG) target achievement (<130 mg/dL), postprandial blood glucose (PPBG) levels, and PPBG increments (ΔPPBG). Basal and postprandial contributions to hyperglycemia were determined.Results: After 24 weeks of insulin glargine titration, A_1c change from baseline was greater in patients with FPG <130 mg/dL versus ≥130 mg/dL (-1.35% versus -1.11%, respectively; P = .0275), but with increased confirmed hypoglycemia rates (blood glucose <70 mg/dL; 4.06 events/patient-year versus 3.31 events/patient-year; P = .0170). However, increased severe hypoglycemia rates were observed in patients with FPG ≥130 mg/dL. At week 24, postprandial contributions to hyperglycemia increased (>60% regardless of PPBG). Patients with high FPG had lower, but substantial, relative postprandial contributions versus patients achieving FPG target. A similar pattern was observed according to whether patients had a ΔPPBG ≥50 mg/dL after any meal.Conclusion: After optimized basal insulin therapy, elevated A_1c is the most effective indicator of residual PPHG, regardless of existent FPG or PPBG. When confronted with an uncontrolled A_1c after reasonable titration of basal insulin, clinicians should be aware of probable postprandial contributions to hyperglycemia and consider prandial therapy.Abbreviations:A_1c = glycated hemoglobin A_1cAUC = area under the curveAUC_B = area under the curve (basal hyperglycemia)AUC_G = total area under the curve (total glucose)AUC_N = area under the curve (normal glycemic exposure)AUC_P = area under the curve (postprandial hyperglycemia)BHG = basal hyperglycemiaFBG = fasting blood glucoseFPG = fasting plasma glucoseGLP-1 = glucagon-like peptide 1HE = hyperglycemic exposureOADs = oral antidiabetes drugsPPBG = postprandial blood glucoseΔPPBG = change in postprandial blood glucosePPHG = postprandial hyperglycemiaSMBG = self-monitored blood glucoseT2DM = type 2 diabetes mellitus     

Decreased Mortality with Tight Glycemic Control in Critically Ill Patients: A Retrospective Analysis in a Large Community Hospital System

ObjectiveTo measure the efficacy and possible adverse consequences of tight blood glucose (BG) control when compared to relaxed control.MethodsA retrospective, observational study was conducted at a community-based teaching hospital system among adult, nonmaternity hospitalized patients admitted to the intensive care unit (ICU). Tight glycemic control of BG was compared with less strict BG control, and the following outcome measurements were compared: BG, average length of stay (ALOS), severe hypoglycemia, and mortality.ResultsBetween 2008 and 2012, 18,919 patients were admitted to the ICU. The mortality rate was significantly lower (P = .0001) in patients with an average BG between 80 and 110 mg/dL (8%) and 111 and 140 mg/dL (9.4%) than in patients with average BG between 141 and 180 mg/dL (12.9%). Using tight glycemic control (80 to 110 mg/dL), the ALOS in the ICU decreased from 4 to 2.9 days (P < .0001) among all patients, and from 4.2 to 2.1 days (P < .0001) among patients who had undergone coronary artery bypass graft. Comparatively, the ALOS for the hospital decreased from 9.4 to 8 days. The incidence of severe hypoglycemia (BG < 40 mg/dL) was higher (P = .01) in the tight BG control group (4.78%) compared with the relaxed control group (3.5%). This rate was lower than in previously published studies that analyzed the use of tight control.ConclusionTight glycemic control using protocolbased insulin administration resulted in a decrease in mortality and ALOS among all patients in the ICU. The incidence of severe hypoglycemic episodes was slightly higher in the tightly controlled group but remained lower than in previously published studies. (Endocr Pract. 2014;20: 907-918)     

Inpatient Hypoglycemic Events in a Comparative Effectiveness Trial for Glycemic Control in a High-Risk Population

Objective: Inpatient hypoglycemia (glucose ≤70 mg/dL) is a limitation of intensive control with insulin. Causes of hypoglycemia were evaluated in a randomized controlled trial examining intensive glycemic control (IG, target 140 mg/dL) versus moderate glycemic control (MG, target 180 mg/dL) on post–liver transplant outcomes.Methods: Hypoglycemic episodes were reviewed by a multidisciplinary team to calculate and identify contributing pathophysiologic and operational factors. A subsequent subgroup case control (1:1) analysis (with/without) hypoglycemia was completed to further delineate factors. A total of 164 participants were enrolled, and 155 patients were examined in depth.Results: Overall, insulin-related hypoglycemia was experienced in 24 of 82 patients in IG (episodes: 20 drip, 36 subcutaneous [SQ]) and 4 of 82 in MG (episodes: 2 drip, 2 SQ). Most episodes occurred at night (41 of 60), with high insulin amounts (44 of 60), and during a protocol deviation (51 of 60). Compared to those without hypoglycemia (n = 127 vs. n = 28), hypoglycemic patients had significantly longer hospital stays (13.6 ± 12.6 days vs. 7.4 ± 6.1 days; P = .002), higher peak insulin drip rates (17.4 ± 10.3 U/h vs. 13.1 ± 9.9 U/h; P = .044), and higher peak insulin glargine doses (36.8 ± 21.4 U vs. 26.2 ± 24.3 U; P = .035). In the case-matched analysis (24 cases, 24 controls), those with insulin-related hypoglycemia had higher median peak insulin drip rates (17 U/h vs. 11 U/h; P = .04) and protocol deviations (92% vs. 50%; P = .004).Conclusion: Peak insulin requirements and protocol deviations were correlated with hypoglycemia.Abbreviations:DM = diabetes mellitusICU = intensive care unitIG = intensive glycemic controlMELD = Model for End-stage Liver DiseaseMG = moderate glycemic controlSQ = subcutaneous     

Temporal Occurrences and Recurrence Patterns of Hypoglycemia During Hospitalization

Objective: To describe the temporal distribution of hypoglycemia and its rate of recurrence during hospitalization to aid in the development of strategies to prevent hypoglycemia in hospitalized patients.Methods: Retrospective review of hypoglycemia (blood glucose <50 mg/dL) audit data in adult hospitalized patients at 2 academic hospitals. Demographics, timing, and blood glucose values were recorded. Antihyperglycemic medications, number of recurrent events, and change in basal insulin dose following the hypoglycemic event were also extracted.Results: A total of 274 index occurrences of hypoglycemia were analyzed. The mean age of the patients was 53.8 years, with roughly equal gender distributions. Twenty-eight percent of the events occurred in the absence of antihyperglycemic therapy. The incidence of hypoglycemia peaked between midnight and 6 AM. There were 36 instances of recurrent hypoglycemia associated with antihyperglycemic therapy, with 78% (n = 28) cases involving basal insulin. Patients on basal insulin who developed hypoglycemia did not have their dose changed prior to the time of the next administration in 75% of the cases.Conclusion: Hypoglycemia in hospitalized patients may occur with greater frequency overnight. Although cumbersome, routine nocturnal glycemic testing should be considered. Education regarding insulin management in the hospital and improved communication between night and day staff may aid in decreasing subsequent hypoglycemic events.Abbreviations: BG = blood glucose EHR = electronic health record ICU = intensive care unit IV = intravenous     

Proactive Protocol-Based Management of Hyperglycemia and Diabetes in Colorectal Surgery Patients

Objective: The management of diabetic patients undergoing elective abdominal surgery continues to be unsystematic, despite evidence that standardized perioperative glycemic control is associated with fewer postoperative surgical complications. We examined the efficacy of a pre-operative diabetes optimization protocol implemented at a single institution in improving perioperative glycemic control with a target blood glucose of 80 to 180 mg/dL.Methods: Patients with established and newly diagnosed diabetes who underwent elective colorectal surgery were included. The control group comprised 103 patients from January 1, 2011, through December 31, 2013, before protocol implementation. The glycemic-optimized group included 96 patients following protocol implementation from January 1, 2014, through July 31, 2016. Data included demographic information, blood glucose levels, insulin doses, hypoglycemic events, and clinical outcomes (length of stay, re-admissions, complications, and mortality).Results: Patients enrolled in the glycemic optimization protocol had significantly lower glucose levels intra-operatively (145.0 mg/dL vs. 158.1 mg/dL; P = .03) and postoperatively (135.6 mg/dL vs. 145.2 mg/dL; P = .005). A higher proportion of patients enrolled in the protocol received insulin than patients in the control group (0.63 vs. 0.48; P = .01), but the insulin was administered less frequently (median [interquartile range] number of times, 6.0 [2.0 to 11.0] vs. 7.0 [5.0 to 11.0]; P = .04). Two episodes of symptomatic hypoglycemia occurred in the control group. There was no difference in clinical outcomes.Conclusion: Improved peri-operative glycemic control was observed following implementation of a standardized institutional protocol for managing diabetic patients undergoing elective colorectal surgery.Abbreviations: HbA1c = glycated hemoglobin A1c; IQR = interquartile range     

Impact of a Hyperglycemic Crises Protocol

ObjectiveThe purpose of this study was to evaluate the efficacy and safety of an adult hyperglycemic crises protocol based upon the 2009 American Diabetes Association (ADA) consensus statement.MethodsWe performed a retrospective review of patients treated before and after protocol implementation at a university teaching hospital. A total of 256 adult patients met the criteria for diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS) and were treated with an insulin infusion between February 2011 and February 2012 (nonprotocol n = 143, protocol n = 113). Protocol efficacy was evaluated by assessing time to resolution of DKA or HHS, length of stay (LOS) in the intensive care unit (ICU), and LOS in the hospital. Protocol safety was evaluated by assessing the numbers of patients with hypoglycemic and hypokalemic events.ResultsPatients on the hyperglycemic crises protocol experienced a 9.2 hour (95% confidence interval (CI): 4.70-13.70; P<.001) decrease in time to resolution, with nonprotocol patients (n = 143) resolving in 22.78 hours and protocol patients (n = 113) resolving in 13.58 hours. There was no difference in safety outcomes, including the number of patients with moderate hypoglycemia (blood glucose <70 mg/dL), severe hypoglycemia (blood glucose <50 mg/dL), or hypokalemia (K⁺ <3.3 mmol/L).ConclusionImplementation of a hyperglycemic crises protocol decreased times to resolution of DKA and HHS without increasing the rate of hypoglycemia or hypokalemia. (Endocr Pract. 2013;19:953-962)     

Hypoglycemia Rates After Restriction of High-Dose Glargine in Hospitalized Patients

Objective: Hypoglycemia remains one of the main challenges of insulin therapy. To reduce insulin-related hypoglycemia at our institution, we restricted inpatient ordering of high glargine doses (≥0.5 U/kg/day) to endocrine staff in May 2013. This retrospective cohort study assesses its effect on hypoglycemia and glycemic control within 48 hours of admission (ADM).Methods: We identified 692 adult patients hospitalized at Boston Medical Center who received glargine upon ADM from November 1, 2012 through April 30, 2013 as the pre-intervention group, and 651 adult patients admitted between November 1, 2013 and April 30, 2014 as the postintervention group. Demographics, medical history, home insulin regimen, concurrent oral diabetes medications or glucocorticoid administration, ADM serum creatinine, all blood glucose levels (BG) ≤48 hours of ADM, and hemoglobin A1c values ≤3 months were assessed. Hypoglycemia was defined as BG ≤70 mg/dL, and hyperglycemia as BG ≥200 mg/dL. Multivariable regression models assessed potential associations between covariates and incidence of hypoglycemia and average BG ≤48 hours of ADM.Results: Demographics were similar between groups. Significantly less patients received high-dose glargine in the post-intervention group (5.2% vs. 0.3%, P<.001). Incidences of hypoglycemia were significantly lower in the postintervention group (20.9% vs. 17.8%, P<.001 per ADM; 3.4% vs. 2.3%, P = .001 per BG measurements [BGM]). Mean BG levels ≤48 hours of ADM and incidence of hyperglycemia were not significantly different. The adjusted incident rate ratio of hypoglycemia was 0.63 per ADM and 0.74 per BGM in the postintervention group compared to the pre-intervention group (P = .001 and P = .063, respectively).Conclusion: We found that implementation of a restriction on high doses of glargine resulted in lower rates of hypoglycemia without worsening glycemic control.Abbreviations:ADM = admissionBG = blood glucoseBGM = blood glucose measurementsBMC = Boston Medical CenterBMI = body mass indexEMR = electronic medical recordHgbA1c = hemoglobin A1cIRR = incidence rate ratioNPH = neutral protamine HagedornTDD = total daily doseT2D = type 2 diabetes     

Risk Factors for Hypoglycemia in Inpatients with Total Parenteral Nutrition and Type 2 Diabetes: A Post HOC Analysis of the Insupar Study

Objective: Treatment of hyperglycemia with insulin is associated with increased risk of hypoglycemia in type 2 diabetes mellitus (T2DM) patients receiving total parenteral nutrition (TPN). The aim of this study was to determine the predictors of hypoglycemia in hospitalized T2DM patients receiving TPN.Methods: Post hoc analysis of the INSUPAR study, which is a prospective, open-label, multicenter clinical trial of adult inpatients with T2DM in a noncritical setting with indication for TPN.Results: The study included 161 patients; 31 patients (19.3%) had hypoglycemic events, but none of them was severe. In univariate analysis, hypoglycemia was significantly associated with the presence of diabetes with end-organ damage, duration of diabetes, use of insulin prior to admission, glycemic variability (GV), belonging to the glargine insulin group in the INSUPAR trial, mean daily grams of lipids in TPN, mean insulin per 10 grams of carbohydrates, duration of TPN, and increase in urea during TPN. Multiple logistic regression analysis showed that the presence of diabetes with end-organ damage, GV, use of glargine insulin, and TPN duration were risk factors for hypoglycemia.Conclusion: The presence of T2DM with end-organ damage complications, longer TPN duration, belonging to the glargine insulin group, and greater GV are factors associated with the risk of hypoglycemia in diabetic noncritically ill inpatients with parenteral nutrition.Abbreviations: ADA = American Diabetes Association; BMI = body mass index; CV% = coefficient of variation; DM = diabetes mellitus; GI = glargine insulin; GV = glycemic variability; ICU = intensive care unit; RI = regular insulin; T2DM = type 2 diabetes mellitus; TPN = total parenteral nutrition     

Basal-Bolus Regimen With Insulin Analogues Versus Human Insulin in Medical Patients with type 2 Diabetes: A Randomized Controlled Trial in Latin America

Objective: Few randomized studies have focused on the optimal management of non–intensive care unit patients with type 2 diabetes in Latin America. We compared the safety and efficacy of a basal-bolus regimen with analogues and human insulins in general medicine patients admitted to a University Hospital in Asunción, Paraguay.Methods: In a prospective, open-label trial, we randomized 134 nonsurgical patients with blood glucose (BG) between 140 and 400 mg/dL to a basal-bolus regimen with glargine once daily and glulisine before meals (n = 66) or Neutral Protamine Hagedorn (NPH) twice daily and regular insulin before meals (n = 68). Major outcomes included differences in daily BG levels and frequency of hypoglycemic events between treatment groups.Results: There were no differences in the mean daily BG (157 ± 37 mg/dL versus 158 ± 44 mg/dL; P = .90) or in the number of BG readings within target <140 mg/dL before meals (76% versus 74%) between the glargine/glulisine and NPH/regular regimens. The mean insulin dose in the glargine/glulisine group was 0.76 ± 0.3 units/kg/day (glargine, 22 ± 9 units/day; glulisine, 31 ± 12 units/day) and was not different compared with NPH/regular group (0.75 ± 0.3 units/kg/day [NPH, 28 ± 12 units/day; regular, 23 ± 9 units/day]). The overall prevalence of hypoglycemia (<70 mg/dL) was similar between patients treated with NPH/regular and glargine/glulisine (38% versus 35%; P = .68), but more patients treated with human insulin had severe (<40 mg/dL) hypoglycemia (7.6% versus 25%; P = .08). There were no differences in length of hospital stay or mortality between groups.Conclusion: The basal-bolus regimen with insulin analogues resulted in equivalent glycemic control and frequency of hypoglycemia compared to treatment with human insulin in hospitalized patients with diabetes.Abbreviations: BG = blood glucose BMI = body mass index HbA_1c = glycated hemoglobin NPH = Neutral Protamine Hagedorn T2D = type 2 diabetes     

Effect Of A Targeted Glycemic Management Program On Provider Response To Inpatient Hyperglycemia

ObjectiveTo report the results of implementation of a Targeted Glycemic Management (TGM) Service pilot, with the goals of improving clinician awareness of available inpatient glycemic management protocols and improving responsiveness to and frequency of severe hyperglycemia.MethodsPatients with a blood glucose (BG) level ≥ 300 mg/dL who were hospitalized on a general medicine unit during three 12-week periods before, during, and after the TGM pilot were compared for responsiveness by the primary team, percentage of subsequent BG measurements between 80 and 180 mg/dL, and frequency of subsequent severe hyperglycemia (BG levels ≥ 300 mg/dL) and hypoglycemia (BG values < 70 mg/dL).ResultsIn comparison with pre-TGM and post-TGM periods, more patients during the TGM pilot had a modification of their glycemic regimen in response to severe hyperglycemia (49% versus 73% versus 50%, before, during, and after TGM, respectively; P = .044), and the percentage of patients with ≥ 50% of subsequent BG measurements in the desired range (27% versus 53% versus 32%; P = .035) was greatest during the TGM period. The incidence of subsequent severe hyperglycemia (20% versus 9% versus 16%; P = .0004) was lowest during the TGM period; however, the incidence of hypoglycemia was similar in all 3 periods (3.9% versus 3.7% versus 3.7%).ConclusionThese results indicate that a TGM Service can favorably influence glycemic management practices and improve glycemic control, but ongoing intervention is necessary for maintenance of these results. (Endocr Pract. 2011;17:552-557)     

Association of Glycemic Control Parameters with Clinical Outcomes in Chronic Critical Illness

ObjectiveChronic critical illness (CCI) is a term used to designate patients requiring prolonged mechanical ventilation and tracheostomy with associated poor outcomes. The present study assessed the impact of glycemic parameters on outcomes in a CCI population.MethodsA retrospective case series was performed including 148 patients in The Mount Sinai Hospital Respiratory Care Unit (2009-2010). Utilizing a semi-parametric mixture model, trajectories for the daily mean blood glucose (BG), BG range, and hypoglycemia rate over time identified low- (n = 87) and high-risk (n = 61) hyperglycemia groups and low- (n = 90) and high-risk (n = 58) hypoglycemia groups. The cohort was also classified into diabetes (DM, n = 48), stress hyperglycemia (SH, n = 85), and normal glucose (n = 15) groups.ResultsHospital- (28% vs. 13%, P = .0199) and 1-year mortality (66% vs. 46%, P = .0185) rates were significantly greater in the high- versus low-risk hyperglycemia groups, respectively. The hypoglycemia rate (< 70 mg/dL) was lower among ventilator-liberated patients compared to those who failed to liberate (0.092 vs. 0.130, P < .0001). In the SH group, both hospital mortality (high-risk hyperglycemia 48% and low-risk hyperglycemia 15%, P = .0013) and 1-year mortality (high-risk 74% and low-risk 50%, P = .0482) remained significantly different, while no significant difference in the diabetes group was observed. There were lower hypoglycemia rates with SH compared to diabetes (< 70 mg/dL: 0.086 vs. 0.182, P < .0001; < 40 mg/dL: 0.012 vs. 0.022, P = .0118, respectively).ConclusionTighter glycemic control was associated with improved outcomes in CCI patients with SH but not in CCI patients with diabetes. Confirmation of these findings may lead to stratified glycemic control protocols in CCI patients based on the presence or absence of diabetes. (Endocr Pract. 2014;20:884-893)     

Retrospective Study of Glycemic Control Following Transition from the Intensive Care Unit in A National Sample of U.S. Hospitals

Objective: Retrospective study to evaluate glycemic control outcomes after transition from the intensive care unit (ICU) to a non-ICU area in a national sample of U.S. hospitals.Methods: Mean point-of-care blood glucose (POC-BG) data were assessed overall and at 24 hours before and up to 72 hours after the transition. Comparisons in glucose variability (standard deviation of POC-BG data) were assessed. Impact on glycemic control was evaluated after accounting for hospital characteristics through logistic regression analysis.Results: POC-BG data were obtained from 576 hospitals. Overall mean (SD) POC-BG values in ICU versus non-ICU areas were 176 (24) versus 169 (21) mg/dL (P<.01). Mean (SD) of the ICU POC-BG data were 76 (16) versus 73 (16) mg/dL in the non-ICU data (P<.01). However, when comparing values of POC-BG in the last 24-hour ICU period with those from up to 72 hours posttransition, we found no differences, indicative of overall stable glycemic control and variability after transition. Any deterioration of glucose control following the transition was significantly associated with hospital size (P<.01): the smallest hospitals had the highest percentage of these cases. In addition, geographic region showed significant variability (P = .04), with hospitals in the Midwest and West having the highest proportion of cases in which glycemic control worsened following the transition.Conclusion: Glycemic control and variability did not change after transition from the ICU, but outcomes may depend on certain hospital characteristics. Inpatient glycemic control assessment should move beyond just cross-sectional studies and consider the impact of transitioning across inpatient areas. Other statistical approaches to studying this question should be evaluated.Abbreviations: DM = diabetes mellitus ICU = intensive care unit POC-BG = point-of-care blood glucose