Abstract
Background
Hyperglycemia is a complication of parenteral nutrition (PN), which can be exacerbated in patients with obesity. Limited data exist on glycemic control in this population. This study aims to evaluate the association of obesity and body mass index (BMI) classification with glycemic control in patients who are critically ill initiated receiving PN.
Methods
This is a retrospective study of patients who are critically ill receiving PN from January 2013 to February 2024. The primary outcome was glycemic control in patients with BMI ≥ 30 kg/m2 or BMI < 30 kg/m2 based on hyperglycemic episodes, peak blood glucose, and insulin requirements on the first and second days of therapy. Multivariate analyses were conducted for the occurrence of a hyperglycemic episode. The secondary outcome was to determine the association of BMI classification with glycemic control.
Results
The study included 220 patients with BMI ≥ 30 kg/m2 (n = 91) and BMI < 30 kg/m2 (n = 129). The BMI < 30 kg/m2 group received more total dextrose (1.58 vs 2.36 mg/kg/min; P < 0.0001). There was no difference in the primary outcome during day 1 of PN, but there were increased hyperglycemic episodes (P = 0.0478) and insulin requirements in the BMI ≥ 30 kg/m2 group on day 2 (P = 0.0226). The only difference in the secondary outcome was insulin requirements on day 2 (P = 0.0453).
Conclusion
Patients who are critically ill with BMI ≥ 30 kg/m2 receiving PN received more conservative dextrose infusion rates yet experienced more hyperglycemic episodes and required more insulin on day 2. However, obesity and BMI classification were not independently associated with hyperglycemic episodes within the first 2 days of PN initiation.
Keywords: critical care, hyperglycemia, insulin, obesity, parenteral nutrition
INTRODUCTION
Obesity has become an increasingly prevalent comorbidity for patients in the United States in the last few decades. 1 The World Health Organization formally recognized obesity as its own disease state at the start of the millennium, defining it as a body mass index (BMI) ≥ 30 kg/m2. Since then, obesity has been divided into three distinct BMI classifications: class I obesity (BMI: 30.0–34.9 kg/m2), class II obesity (BMI: 35.0–39.9 kg/m2), and class III obesity (BMI > 40 kg/m2). 1 , 2 Patients with obesity have been associated with having an increased risk of certain disease states contributing to morbidity and mortality, including coronary heart disease, stroke, dyslipidemia, and type II diabetes. 3 Additionally, patients with obesity may be malnourished because of multifactorial mechanisms, including structural and physiological abnormalities, 3 which complicates assessment and management of nutrition support during hospital admission.
Patients with obesity who are critically ill require an even more specialized approach to care surrounding nutrition needs. 3 The American Society for Parenteral and Enteral Nutrition (ASPEN) and the Society of Critical Care Medicine (SCCM) have published joint recommendations specifically for adult patients with obesity who are critically ill to help guide providers in nutrition support for this specialized patient population. 4 , 5 , 6 ASPEN/SCCM recommend a hypocaloric, high‐protein approach to optimize both enteral nutrition (EN) and parenteral nutrition (PN) in this unique patient population. 4
PN provides intravenous nutrition support in patients that may not be able to tolerate oral nutrition for an extended period of time. A well‐established metabolic complication of PN is hyperglycemia because of the dextrose content. 7 Patients with obesity are already predisposed to hyperglycemic episodes, given the increased rates of insulin resistance and glucose intolerance leading to the higher instances of diabetes that are associated with obesity. 3 In addition, critical illness can further exacerbate hyperglycemia in patients with obesity because of metabolic abnormalities and commonly used medications in critical care, including adrenergic agents such as vasopressors and glucocorticoids. Glycemic control has been extensively studied in the intensive care unit. 8 , 9 Based on evidence demonstrating a reduction in mortality with improved glycemic control, SCCM released guidelines recommending a target blood glucose range of 140–200 mg/dl in patients who are critically ill, with initiation of treatment when concentrations exceed 180 mg/dl. 10
Despite glycemic control being a known issue for patients with obesity who are critically ill receiving PN, data surrounding this patient population are sparse across all BMI classifications. In addition, the ASPEN/SCCM guidelines were derived from older data, necessitating the need for further exploration. 4 , 6 The purpose of this study was to determine the association of obesity (BMI ≥ 30 kg/m2) and BMI classification with glycemic control in patients who are critically ill receiving PN, particularly examining hyperglycemia, peak blood glucose concentrations, and insulin requirements.
METHODS
This study is a retrospective, single‐center, cohort study conducted from January 1, 2013, to February 15, 2024, at Cooper University Health Care, in Camden, NJ, comparing glucose control in patients with obesity (BMI ≥ 30 kg/m2) and patients without obesity (BMI < 30 kg/m2) receiving PN. Patients were included if they were ≥18 years of age, critically ill, and received at least 2 consecutive days of PN. Critical illness was determined based on patient location at the time of PN initiation and included the medical intensive care unit, trauma/surgical intensive care unit, and cardiac coronary unit. Key exclusion criteria included patients <18 years of age, pregnant patients, patients receiving concomitant EN, and patients receiving PN before admission. Data were retrospectively extracted from the electronic health record (EHR), including demographic information, clinical characteristics, laboratory values, and PN prescription details. The institutional review boards at Cooper University Health Care and Saint Joseph's University approved this study.
PN is managed interprofessionally at this institution between advanced practice providers, dietitians, nurses, pharmacists, and physicians. At Cooper University Health Care, before initiating PN in the intensive care units, a dietitian consult is placed by the primary team in order for dietitians to provide individualized macronutrient recommendations. Dietitians provide macronutrient recommendations for the first two days of PN initiation, typically starting with approximately 50% of target energy intake on day 1. If the patient demonstrates tolerance, indicated by stable glucose concentrations, electrolyte balance, and absence of refeeding risk, energy delivery is advanced to 100% of the goal by day 2. The institutional guideline offers hypocaloric, high‐protein recommendations for patients with obesity as a reference, but a specific protocol or method is not required. Energy provided by other sources, such as propofol (1.1 kcal/ml), are taken into account with macronutrient recommendations. The micronutrients, including electrolytes and trace elements, are determined by the ordering clinician with guidance from decision support from the order set in the EHR and pharmacist recommendations. At this institution, PN is outsourced, and administration is standardized to begin at 10:00 PM to ensure adequate time for compounding and delivery of the PN formulation. Accordingly, the data for this study were collected in 24‐h intervals based on the PN hang time, with each day defined from 10:00 PM to 9:59 PM the following day. At this institution, insulin is not routinely included within the PN formulation. Most patients initiated receiving PN have a regular human insulin sliding scale ordered, and blood glucose is typically monitored every 4 to 6 h with correctional insulin initiated at blood glucose concentrations >150 mg/dl. If indicated, long‐acting insulin may be ordered by the primary team or with assistance from the endocrinology service when consulted.
The primary outcome of glycemic control, included individual components of the number of hyperglycemic episodes (defined as a blood glucose >200 mg/dl), peak glucose concentration (mg/dl), and daily insulin requirements (units/day) on day 1 (day of PN initiation) and day 2 (day after PN initiation). The association of additional factors known to influence glycemic control, including dextrose administered externally to PN (eg, intravenous fluids, medication diluent), corticosteroid therapy, and vasopressor use, were evaluated as potential confounding variables. The secondary outcome was to determine the association of BMI classification with the primary outcome. BMI was classified based on Centers for Disease Control and Prevention recommendations: underweight (BMI < 18.5 kg/m2), healthy weight (18.5 to <25 kg/m2), overweight (25 to <30 kg/m2), obesity class I (30 to <35 kg/m2), obesity class II (35 to <40 kg/m2), and obesity class III (≥40 kg/m2).
Statistical analyses were performed using Excel and SAS software, version 9.4 (SAS Institute, Inc). Categorical variables were analyzed using either chi‐squared test or Fisher exact test depending on sample size and reported as counts (percentages). Quantitative variables were analyzed using the Wilcoxon rank sum test and reported as median and interquartile range. Clinical outcomes were assessed using a bivariate analysis. Multivariate analyses were also conducted for the occurrence of a hyperglycemic episode on day 1 and day 2 using a stepwise logistic regression model to account for statistically and clinically significant confounding variables. Statistical significance was defined as a P value < 0.05.
RESULTS
A total of 220 patients were included in the study analysis (91 patients in the BMI ≥ 30 kg/m2 group and 129 patients in the BMI < 30 kg/m2 group). Baseline characteristics, including age and comorbidities, are summarized in Table 1. There was a significantly higher number of patients with baseline diabetes without end‐organ damage, cerebrovascular disease, and metastatic disease in the BMI ≥ 30 kg/m2 group. The number of patients receiving corticosteroids and total hydrocortisone equivalents in milligrams per day were similar between the groups. Actual body weight and BMI were significantly higher in the BMI ≥ 30 kg/m2 group, with a total breakdown based on BMI classification reported in Tables 2 and 3.
Table 1.
Patient characteristics (n = 220).
| Patient characteristics | BMI ≥ 30 kg/m2(n = 91) | BMI < 30 kg/m2 (n = 129) | P |
|---|---|---|---|
| Age, median (IQR), years | 62 (49–70) | 61 (50–75) | 0.5377 |
| Race, n (%) | |||
| African American/Black | 19 (20.9) | 22 (17.1) | |
| Asian | 0 (0) | 3 (2.3) | |
| Hispanic/Latino | 4 (4.4) | 14 (10.9) | |
| Native American | 1 (1.1) | 0 (0) | |
| White | 63 (69.2) | 89 (68.9) | |
| Unknown | 4 (4.4) | 1 (0.8) | |
| Hospital location at PN initiation, n (%) | 0.3176 | ||
| Coronary care unit | 6 (6.6) | 14 (10.9) | |
| Medical ICU | 50 (55) | 59 (45.7) | |
| Trauma ICU | 35 (38.5) | 56 (43.4) | |
| Female, n (%) | 41 (45.1) | 59 (45.7) | 0.9204 |
| Poor nutrient intake prior to PN (days), median (IQR) | 5 (3–8) | 5 (3–9) | 0.3131 |
| SOFA score, median (IQR) | 7 (4–11) | 6 (3–10) | 0.0504 |
| APACHE II score, median (IQR) | 26 (20–32) | 23 (18–29) | 0.0513 |
| Diabetes without end‐organ damage, n (%) | 28 (30.8) | 22 (17.1) | 0.0168 |
| Diabetes with end‐organ damage, n (%) | 8 (8.8) | 4 (3.1) | 0.0777 |
| Cerebrovascular disease | 13 (14.3) | 8 (6.2) | 0.0445 |
| Metastatic solid tumor | 2 (2.2) | 16 (12.4) | 0.0065 |
| Serum creatinine, median (IQR), mg/dl | 1.1 (0.71–2.29) | 1.1 (0.58–1.79) | 0.1913 |
| CrCl, median (IQR), ml/min | 61.9 (38.6–118.2) | 60.4 (25.8–107.3) | 0.3331 |
| AKI, n (%) | 43 (47.3) | 51 (39.5) | 0.2544 |
| CKD/ESKD, n (%) | 19 (20.9) | 21 (16.3) | 0.3836 |
| IHD, n (%) | 5 (5.5) | 6 (4.7) | 0.7647 |
| CVVHD, n (%) | 9 (9.9) | 7 (5.4) | 0.2093 |
| Liver disease, n (%) | 19 (20.9) | 18 (14) | 0.1762 |
| Patients receiving corticosteroids, n (%) | |||
| Day 0 | 11 (12.1) | 15 (11.6) | 0.9171 |
| Day 1 | 9 (9.9) | 14 (10.9) | 0.8182 |
| Day 2 | 10 (11.0) | 12 (9.3) | 0.6813 |
| Total HCT equivalents, median (IQR), mg | |||
| Day 0 | 150 (150–210) | 200 (200–400) | 0.2071 |
| Day 1 | 200 (150–210) | 225 (200–320) | 0.2142 |
| Day 2 | 105 (50–200) | 175 (137.5–200) | 0.2695 |
| Hyperglycemic episodes, day 0, median (IQR)a | 0 (0–0) | 0 (0–0) | 0.3588 |
| Peak glucose, day 0, median (IQR), mg/dla | 135 (115–175) | 126 (104–153) | 0.0500 |
| Insulin requirements, day 0, median (IQR), unitsa | 0 (0–4) | 0 (0–0) | 0.3001 |
Abbreviations: AKI, acute kidney injury; APACHE, Acute Physiology and Chronic Health Evaluation; BMI, Body mass index; CKD, chronic kidney disease; CrCl, creatinine clearance; CVVHD, continuous veno‐venous hemodialysis; EKSD, end‐stage kidney disease; HCT, hydrocortisone; ICU, intensive care unit; IHD, intermittent hemodialysis; IQR, Interquartile range; SD, standard deviation; SOFA, Sequential Organ Failure Assessment.
Day 0: baseline measurement day prior to PN initiation
Table 2.
Patient characteristics related to weight (N = 220).
| Patient characteristics | BMI ≥ 30 kg/m2, n = 91 | BMI < 30 kg/m2, n = 129 | P value |
|---|---|---|---|
| Actual body weight, median (IQR), kg | 103.6 (90.4–117) | 66.9 (55–79.7) | <0.0001 |
Abbreviations: BMI, Body mass index; IQR, interquartile range.
Table 3.
Patient characteristics related to BMI classification (N = 220).
| Classification | BMI, kg/m2 | n (%) | BMI, median (IQR), kg/m2 |
|---|---|---|---|
| Underweight | <18.5 | 18 (8.2) | 17.05 (15.6–17.6) |
| Healthy weight | 18.5 to <25 | 58 (26.4) | 22.42 (20.79–23.67) |
| Overweight | 25 to <30 | 53 (24.1) | 27.42 (26.31–28.62) |
| Obesity class I | ≥30 to <35 | 38 (17.3) | 31.87 (31.05–32.7) |
| Obesity class II | ≥35 to <40 | 28 (12.7) | 36.77 (35.94–38.14) |
| Obesity class III | ≥40 | 25 (11.4) | 44.14 (41.57–53.13) |
Abbreviations: BMI, body mass index; IQR, interquartile range.
PN characteristics are summarized in Table 4. Patients in the BMI ≥ 30 kg/m2 group received numerically more energy (1990 vs 1680 kcal/kg/day; P < 0.0001) but, when normalized for body weight, received significantly less energy provision (18.6 vs 25 kcal/kg/day; P < 0.0001) and lower dextrose infusion rates (1.6 vs 2.4 mg/kg/min; P < 0.0001) than patients without obesity. Similarly, patients with a BMI ≥ 30 kg/m2 received higher energy provision from protein and fat injectable emulsion compared with those with a BMI < 30 kg/m2 (500 kcal vs 400 kcal for protein [P < 0.0001]; 530 vs 465 kcal/day for fat injectable emulsion [P = 0.0003]). However, when normalized for actual body weight, patients with a BMI ≥ 30 kg/m2 received lower amounts (1.2 vs 1.5 g/kg/day for protein [P < 0.000]; 0.5 vs 0.7 g/kg/day for fat injectable emulsion [P < 0.0001]). In contrast, protein provision based on ideal body weight was higher for the BMI ≥ 30 kg/m2 group (2.1 vs 1.6 g/kg/day; P < 0.0001). The median number of days to reach goal PN did not differ significantly between the groups (2 days in both groups; P = 0.3675). Of the 220 patients included, 92.3% of patients had a dietitian consult before PN initiation, and 96.8% were initiated on central PN. Number of days of poor nutrient intake (defined as <50% of daily requirement) was similar between the groups with a median of 5 days (P = 0.3131). Total dextrose infusion rate, accounting for both dextrose within the PN and external sources, was significantly higher in the BMI < 30 kg/m2 group across all days (P < 0.0001).
Table 4.
PN characteristics (N = 220).
| Bivariate analyses | BMI ≥ 30 kg/m2 (n = 91) | BMI < 30 kg/m2 (n = 129) | P value |
|---|---|---|---|
| Indication, n (%) | |||
| Failure to achieve enteral goals in 7 days | 14 (15.4) | 11 (8.5) | — |
| Gut ischemia | 7 (7.7) | 7 (5.4) | — |
| Inaccessible GI tract | 40 (44) | 58 (45) | — |
| Intolerance to EN | 4 (4.4) | 13 (10.1) | — |
| Nonoperative mechanical bowel obstruction or pseudoobstruction | 6 (6.6) | 20 (15.5) | — |
| Refractory diarrhea or vomiting | 1 (1.1) | 1 (0.8) | — |
| Severe paralytic ileus | 8 (8.8) | 7 (5.4) | — |
| Short bowel syndrome | 1 (1.1) | 1 (0.8) | — |
| Other | 10 (11) | 11 (8.5) | — |
| Dietitian consult prior to PN, n (%) | 86 (94.5) | 117 (90.7) | 0.2976 |
| Central PN, n (%) | 90 (98.9) | 123 (95.35) | 0.2438 |
| Time to goal PN, median (IQR), days | 2 (2–3) | 2 (2–3) | 0.3675 |
| Consecutive days of PN, median (IQR) | 7 (5–13) | 8 (4–13) | 0.9983 |
| Goal order: total energy, median (IQR)a | |||
| kcal/day | 1990 (1662–2118.5) | 1680 (1375–1920) | <0.0001 |
| kcal/kg/day | 18.6 (15.7–21.2) | 25 (22.4–26.9) | <0.0001 |
| Goal order: amino acids, median (IQR)a | |||
| kcal/day | 500 (440–600) | 400 (320–480) | <0.0001 |
| g/kg/day | 1.2 (1–1.4) | 1.5 (1.4–1.8) | <0.0001 |
| Goal order: fat injectable emulsion, median (IQR)a | |||
| kcal/day | 530 (422.9–677.3) | 465 (345–530) | 0.0003 |
| g/kg/day | 0.5 (0.4–0.6) | 0.7 (0.6–0.8) | <0.0001 |
| Goal order: dextrose, median (IQR)a | |||
| kcal/day | 820 (700–995) | 780 (630–935) | 0.1400 |
| mg/kg/minb | 1.6 (1.3–2) | 2.4 (2.2–2.7) | <0.0001 |
Abbreviations: BMI, Body mass index; IQR, Interquartile range; PN, parenteral nutrition.
Goal orders, indicating full energy delivery, are generally started on Day 2 of PN therapy in accordance with institutional practices.
Dextrose infusion rate calculated per kg of actual body weight and includes both dextrose within the PN and external sources.
Hyperglycemic episodes, peak blood glucose concentration, and insulin requirements are reported as median values for day 0, day 1, and day 2. No differences were seen between groups on day 0. When comparing each component of the primary outcome, there were no significant difference in hyperglycemic episodes, peak blood glucose, or insulin requirements on day 1; however, there was a significant increase in both hyperglycemic episodes (P = 0.0478) and insulin requirements (P = 0.0226) in the BMI ≥ 30 kg/m2 group on day 2. Because of the high number of patients with no outcome events, the medians in the two groups are both equal to zero; however, based on the Wilcoxon rank sum test, there is a statistically significant difference in the overall distribution of the data, which is depicted in Figure 1. The majority of patients experienced no hyperglycemic episodes on day 1 (71.8%) and day 2 (69.3%). Full results of the bivariate analyses of the primary outcome measures are reported in Table 5. For the secondary outcome, there was no significant difference between BMI classifications in hyperglycemic episodes or peak blood glucose on any of the days analyzed, but there was a significant difference in insulin requirements on day 2 (P = 0.0453). Full results of the bivariate analyses of the secondary outcome are presented in Table 6.
Figure 1.
Number of hyperglycemic episodes. This graphic depicts the number of patients that experienced each number of hyperglycemic episodes in the BMI > 30 and BMI < 30 groups on both day 1 of PN and day 2 of PN. The last column expresses the total number of patients in each group that experienced a hyperglycemic episode on either day 1 or day 2 of PN.
Table 5.
Bivariate analyses for BMI ≥ 30 kg/m2 or < 30 kg/m2.
| Bivariate analyses | BMI ≥ 30 kg/m2 (n = 91), median (IQR) | BMI < 30 kg/m2 (n = 129), median (IQR) | P value |
|---|---|---|---|
| Total dextrose infusion rate, mg/kg/min,a, b | |||
| Day 0 | 0.2 (0.1–0.7) | 0.5 (0.2–1.2) | 0.0012 |
| Day 1 | 1.2 (0.9–1.5) | 1.9 (1.4–2.4) | <0.0001 |
| Day 2 | 1.8 (1.4–2.2) | 2.6 (2.2–3) | <0.0001 |
| Hyperglycemic episodes, n b | |||
| Day 0 | 0 (0–0) | 0 (0–0) | 0.3588 |
| Day 1 | 0 (0–0) | 0 (0–0) | 0.0502 |
| Day 2 | 0 (0–1) | 0 (0–0) | 0.0478 |
| Number of patients whom experienced hyperglycemic events, %c | |||
| Day 0 | 11 (12.0) | 11 (8.5) | 0.3859 |
| Day 1 | 32 (35.2) | 30 (23.3) | 0.0532 |
| Day 2 | 34 (37.8) | 33 (25.8) | 0.0587 |
| Peak blood glucose, mg/dlb | |||
| Day 0 | 135 (115–175) | 126 (104–153) | 0.0500 |
| Day 1 | 171 (139–216) | 158 (139–199) | 0.2772 |
| Day 2 | 178 (140–242) | 161 (137.5–202) | 0.1378 |
| Insulin requirements, unitsb | |||
| Day 0 | 0 (0–4) | 0 (0–0) | 0.3001 |
| Day 1 | 0 (0–16) | 0 (0–8) | 0.1844 |
| Day 2 | 0 (0–24) | 0 (0–8) | 0.0226 |
Abbreviations: BMI, body mass index; IQR, interquartile range.
Dextrose infusion rates calculated per kg of actual body weight and includes both dextrose within the PN and external sources.
P values determined using Wilcoxon rank sum test.
P values determined using chi‐squared test.
Table 6.
Bivariate analyses by BMI classifications.
| Bivariate analyses | Underweight (n = 18) | Healthy weight (n = 58) | Overweight (n = 38) | Class 1 obesity (n = 38) | Class 2 obesity (n = 28) | Class 3 obesity (n = 25) | P valueb |
|---|---|---|---|---|---|---|---|
| Total dextrose infusion rate, mg/kg/mina | |||||||
| Day 0, median (IQR) | 1.02 (0.12–1.86) | 0.56 (0.16–1.45) | 0.34 (0.17–0.7) | 0.21 (0.06–0.66) | 0.28 (0.08–0.92) | 0.12 (0.03–0.35) | 0.0044 |
| Day 1, median (IQR) | 2.09 (1.62–3.14) | 2.00 (1.35–2.53) | 1.79 (1.41–2.06) | 1.44 (1.02–1.8) | 1.28 (0.97–1.58) | 0.90 (0.76–1.16) | <0.0001 |
| Day 2, median (IQR) | 2.97 (2.45–3.38) | 2.70 (2.2–3.06) | 2.47 (2.07–2.66) | 1.95 (1.58–2.47) | 1.80 (1.53–2.13) | 1.54 (1.18–1.77) | <0.0001 |
| Hyperglycemic episodes, n | |||||||
| Day 0, median (IQR) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0.6448 |
| Day 1, median (IQR) | 0 (0–0) | 0 (0–0) | 0 (0–2) | 0 (0–1) | 0 (0–1.5) | 0 (0–1) | 0.1087 |
| Day 2, median (IQR) | 0 (0–1) | 0 (0–0) | 0 (0–2) | 0 (0–2) | 0 (0–2) | 0 (0–1) | 0.1784 |
| Peak blood glucose, mg/dl | |||||||
| Day 0, median (IQR) | 115 (104–135) | 122 (101–148) | 134 (107–171) | 138.5 (115–185) | 129.5 (118–192.5) | 135 (110–160) | 0.2811 |
| Day 1, median (IQR) | 151 (133–179) | 158.5 (134–194) | 165 (147–212) | 178 (140–220) | 169 (141–228.5) | 177 (137–206) | 0.2582 |
| Day 2, median (IQR) | 155 (131–210) | 154 (134–188) | 176 (146–214) | 168 (140–242) | 164 (138–277) | 194 (146–207) | 0.3411 |
| Insulin requirement, units | |||||||
| Day 0, median (IQR) | 0 (0–0) | 0 (0–0) | 0 (0–2) | 0 (0–3) | 0 (0–2) | 0 (0–4) | 0.6506 |
| Day 1, median (IQR) | 0 (0–6) | 0 (0–2) | 2 (0–11) | 0 (0–18) | 1.5 (0–9.5) | 0 (0–14) | 0.1587 |
| Day 2, median (IQR) | 0 (0–2) | 0 (0–2) | 1 (0–14) | 0 (0–24) | 2 (0–22) | 2 (0–20) | 0.0453 |
Dextrose infusion rates calculated per kg of actual body weight and includes both dextrose within the PN and external sources.
P values determined using Wilcoxon rank sum test.
Abbreviations: BMI, Body mass index; IQR, interquartile range.
The results of the multivariate analysis of hyperglycemic episodes are presented in Table 7. After controlling for significant confounding variables, BMI ≥ 30 kg/m2 was not significantly associated with the occurrence of a hyperglycemic episode on Day 1 or Day 2 of PN. Clinically significant confounding variables on Day 1 of PN included diabetes with or without end‐organ damage and total dextrose infusion rate on day 1. Clinically significant confounding variables on day 2 of PN included age, diabetes with or without end‐organ damage, as well as presence of corticosteroids on day 1.
Table 7.
Multivariate analyses: Hyperglycemic episode occurrence.
| BMI classification, BMI ≥ 30 kg/m2 vs BMI < 30 kg/m2 | Odds ratio | 95% CI | P value |
|---|---|---|---|
| Day 1 a | 1.897 | 0.899–4.005 | 0.0929 |
| Day 2 b | 1.316 | 0.642–2.699 | 0.4536 |
| Day 1 a | 0.3103 | ||
| Underweight vs HW | 0.311 | 0.043–2.275 | |
| Overweight vs HW | 1.611 | 0.611–4.249 | |
| Class 1 obesity vs HW | 2.558 | 0.894–7.322 | |
| Class 2 obesity vs HW | 2.167 | 0.645–7.338 | |
| Class 3 obesity vs HW | 1.716 | 0.457–6.439 | |
| Day 2 a | 0.9203 | ||
| Underweight vs HW | 1.039 | 0.225–4.794 | |
| Overweight vs HW | 1.489 | 0.531–4.171 | |
| Class 1 obesity vs HW | 1.715 | 0.574–5.122 | |
| Class 2 obesity vs HW | 1.724 | 0.492–6.043 | |
| Class 3 obesity vs HW | 1.271 | 0.351–4.599 |
Statistically significant confounders in the logistic regression included the following: age, diabetes without end‐organ damage, diabetes with end‐organ damage, total dextrose infusion rate on day 1.
Statistically significant confounders in the logistic regression included the following: age, diabetes without end‐organ damage, diabetes with end‐organ damage, energy recommendation in nutrition consult, presence of corticosteroids on day 1.
Abbreviations: BMI, Body mass index; CI, confidence interval; HW, healthy weight.
DISCUSSION
There were no significant differences in hyperglycemic episodes, peak blood glucose concentration, or insulin requirements on day 1 of PN initiation; however, there was a significant difference in hyperglycemic episodes on day 2 of PN in the BMI ≥ 30 kg/m2 group compared with the BMI < 30 kg/m2 group. In the multivariate analysis, BMI ≥ 30 kg/m2, as well as individual BMI classification, was not independently associated with hyperglycemic episodes on day 1 or day 2 of PN.
The BMI ≥ 30 kg/m2 group had a higher proportion of patients with diabetes without end‐organ damage and numerically higher peak blood glucose concentrations across all assessed days. Although statistically significant, the clinical relevance of this difference remains uncertain. Despite the higher prevalence of diabetes, there was no clinically or statistically significant difference in median peak blood glucose concentrations at baseline. Both groups maintained blood glucose concentrations between 140 and 200 mg/dl as per the SCCM Guidelines on Glycemic Control for Critically Ill Children and Adults. 10 Additionally, blood glucose remained within target ranges on days 1 and 2, with most patients experiencing no hyperglycemic episodes.
Beyond evaluating dextrose within the PN formulation, our study also assessed total dextrose infusion rates, including external sources. Patients in the BMI < 30 kg/m2 group received significantly higher total dextrose infusion rates across days 0, 1, and 2, possibly reflecting greater provider awareness of glucose control in patients with obesity. This emerged as a significant confounding variable in the multivariate analysis for hyperglycemic episodes on day 1, potentially contributing to the lack of observed differences between groups on that day.
Both groups received conservative dextrose infusion rates that remained well below the maximum rate of glucose utilization. Patients in the BMI ≥ 30 kg/m2 group showed a trend toward higher insulin requirements on day 1 and significantly greater insulin requirements on day 2. These findings suggest conservative dextrose administration and insulin therapy were effective in both cohorts, potentially explaining the lack of significant differences in peak blood glucose concentrations after PN initiation.
The ASPEN/SCCM Guidelines for the Provision of Nutrition Support Therapy in the Adult Critically Ill Patient recommend a hypocaloric, high‐protein, and low‐dextrose EN or PN formulation for patients with obesity who have critical illness. 4 , 5 , 6 Specifically, they suggest providing 11–14 kcal/kg/day of actual body weight for patients with a BMI of 30–50 kg/m2 and 22–25 kcal/kg/day of ideal body weight (IBW) for those with a BMI > 50 kg/m2. 4 Although our study aimed to reflect this approach, the actual energy provision was slightly above the SCCM/ASPEN guideline‐recommended targets. Several factors likely contributed to this, including variability in clinical decision‐making in the absence of standardized institutional guidance for hypocaloric feeding during much of the study period, as well as inclusion of nonnutrition energy from propofol in reported totals. Additionally, hypocaloric feeding strategies may not have been applied to patients with kidney dysfunction, for whom hypocaloric, high‐protein feeding has not been well validated. Notably, 47.3% of patients with obesity in our cohort had acute kidney injury (AKI), which may have influenced prescribing.
Despite these factors, patients with BMI ≥ 30 kg/m2 did receive significantly fewer kcal/kg/day compared with those without obesity, suggesting that a hypocaloric feeding approach was considered in clinical practice. However, their absolute energy intake still exceeded 2016 guideline recommendations and may represent overfeeding. This context is important when interpreting glycemic outcomes because the higher incidence of hyperglycemia episodes and insulin requirements observed on day 2 in patients with obesity may reflect the metabolic consequences of excess energy delivery, rather than a physiologic difference attributable to obesity alone. These findings highlight the importance of aligning PN prescriptions with established energy guidelines to reduce the risk of complications.
In this study, protein provision in patients with obesity was aligned with the recommended 2–2.5 g/kg IBW/day. Finally, patients in the BMI ≥ 30 kg/m2 group received significantly lower PN dextrose infusion rate, which may have contributed to reduced events in the primary and secondary outcomes. Although the updated ASPEN guidelines do not provide obesity‐specific recommendations, they remain applicable to this cohort because the clinical trials informing these guidelines included participants with a reported mean BMI in the overweight or obese range. 5
Despite the strengths of our study, certain limitations should be acknowledged. First, this was a single‐center, retrospective cohort study primarily consisting of older, white men, which may limit the generalizability of our findings to more diverse populations. There was also a small number of patients in each BMI classification for the secondary analysis. Additionally, because our study focused strictly on patients who are critically ill, extrapolation to non‐ICU patients may not be appropriate. Second, patient‐to‐patient variability in results could be influenced by multiple factors. PN recommendations were provided by individual dietitians, introducing potential variation in macronutrient assessment. There was no standardized protocol for hypocaloric feeding in patients with obesity. Furthermore, although all patients were critically ill, they were treated in three different intensive care units specializing in different disease states, which may have influenced nutrition recommendations based on the nature of the patients’ illnesses. Third, the study was limited to evaluation of PN during the first two days of therapy. This narrow timeframe may not fully capture longer‐term trends in glycemic control. Additionally, blood glucose data were derived from episodic rather than continuous or uniformly timed monitoring. As such, transient fluctuations in glucose concentrations between testing intervals may have gone undetected. Fourth, practice patterns in this area likely evolved over the 11‐year study period, particularly after the 2019 update to our PN order set to align with the ASPEN PN Safety Consensus Recommendations. 11 However, this period aligns with the modern era of nutrition support practice, as defined by the 2022 ASPEN guidelines, which apply to studies from 2001 onward to reflect contemporary standards such as routine glycemic control, avoidance of overfeeding, and improved catheter care. 5 Finally, we acknowledge that multiple outcomes were assessment, including three components related to glycemic control, which increases the potential for type i error due to multiplicity. Although formal adjustments for multiple comparisons were not applied, our analyses were hypothesis‐driven and limited to prespecified outcomes.
CONCLUSION
Although patients with obesity (BMI ≥ 30 kg/m2) experienced more hyperglycemic episodes and required more insulin on day 2 of PN therapy compared with those with a BMI < 30 kg/m2, obesity and BMI classification were not independently associated with hyperglycemic episodes within the first 2 days of PN initiation in adults who are critically ill. These findings suggest that BMI alone is insufficient for predicting early PN‐related hyperglycemia, highlighting the need for individualized glycemic management. Implementing conservative dextrose infusion rates and closely monitoring blood glucose concentrations, with insulin therapy as needed, may help reduce the risk of hyperglycemia across BMI classifications. Further research is warranted to fully elucidate the factors influencing PN‐related hyperglycemia and improve clinical outcomes in this high‐risk population.
AUHOR CONTRIBUTIONS
Erika L. Mackie, Laura Pontiggia, and Angela L. Bingham contributed to the acquisition, analysis, and/or interpretation; Erika L. Mackie drafted the manuscript; and Erika L. Mackie, Melissa Dang, Kimberly Sharpe, Justin Delic, James M. Hollands, Song Oh, Stacy Pasciolla, Laura Pontiggia, Diana Solomon, and Angela L. Bingham contributed to the conception and/or design, critically revised the manuscript, agree to be fully accountable for ensuring the integrity and accuracy of the work, and read and approved the manuscript.
CONFLICT OF INTEREST STATEMENT
None declared.
ACKNOWLEDGMENTS
This research was conducted as part of the ASHP Foundation Pharmacist Nutrition Support Patient‐care Impact Program. Dr Mackie participated in the program's inaugural cohort, which was designed to develop leaders with knowledge and skills to optimize safe and effective nutrition support. [Correction added on 25 August 2025, after first online publication: The Acknowledgments section has been updated at the author's request.]
Mackie EL; Dang M; Sharpe K, et al. Association of obesity and body mass index classification with glycemic control in adults who are critically ill receiving parenteral nutrition: a retrospective study. Nutr Clin Pract. 2026;41:74‐84. 10.1002/ncp.70017
[Correction added on 25 August 2025, after first online publication: The last author's professional title has been updated from BCPS to BCNSP.]
REFERENCES
- 1. Centers for Disease Control and Prevention . About adult BMI. Accessed May 1, 2024. https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html.
- 2. Pi‐Sunyer FX, Becker DM, Bouchard C, et al. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. WMJ. 1998;97(9):20‐37. [PubMed] [Google Scholar]
- 3. Dickerson RN, Drover JW. Monitoring nutrition therapy in the critically ill patient with obesity. JPEN J Parenter Enteral Nutr. 2011;35(5 suppl):44S‐51S. [DOI] [PubMed] [Google Scholar]
- 4. McClave SA, Taylor BE, Martindale RG, et al. Guidelines for the provision and assessment of nutrition support therapy in the adult critically ill patient: Society of Critical Care Medicine (SCCM) and American Society for Parenteral and Enteral Nutrition (ASPEN). JPEN J Parenter Enteral Nutr. 2016;40(2):159‐211. [DOI] [PubMed] [Google Scholar]
- 5. Compher C, Bingham AL, McCall M, et al. Guidelines for the provision of nutrition support therapy in the adult critically ill patient: he American Society for Parenteral and Enteral Nutrition. JPEN J Parenter Enteral Nutr. 2022;46(1):12‐41. [DOI] [PubMed] [Google Scholar]
- 6. Choban P, Dickerson R, Malone A, Worthington P, Compher C. ASPEN clinical guidelines: nutrition support of hospitalized adult patients with obesity. JPEN J Parenter Enteral Nutr. 2013;37(6):714‐744. [DOI] [PubMed] [Google Scholar]
- 7. Gosmanov AR, Umpierrez GE. Management of hyperglycemia during enteral and parenteral nutrition therapy. Curr Diab Rep. 2013;13(1):155‐162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. NICE‐SUGAR Study Investigators , Finfer S, Chittock DR, et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med. 2009;360(13):1283‐1297. [DOI] [PubMed] [Google Scholar]
- 9. 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] [PubMed] [Google Scholar]
- 10. Honarmand K, Sirimaturos M, Hirshberg EL, et al. Society of critical care medicine guidelines on glycemic control for critically ill children and adults 2024. Crit Care Med. 2024;52(4):e161‐e181. [DOI] [PubMed] [Google Scholar]
- 11. Ayers P, Adams S, Boullata J, et al. ASPEN parenteral nutrition safety consensus recommendations. JPEN J Parenter Enteral Nutr. 2014;38(3):296‐333. [DOI] [PubMed] [Google Scholar]