Elsevier

Journal of Critical Care

Volume 29, Issue 4, August 2014, Pages 611-617
Journal of Critical Care

Monitoring/Outcomes
Hyperglycemia, hypoglycemia, and glycemic complexity are associated with worse outcomes after surgery,☆☆

https://doi.org/10.1016/j.jcrc.2014.03.014Get rights and content

Abstract

Purpose

The purpose of this study was to determine if glycemic complexity, along with hypoglycemia and hyperglycemia, was associated with worse outcomes after cardiac surgery.

Materials and methods

We conducted a retrospective analysis of 970 patients who had insulin infusions designed to keep blood glucose levels between 80 and 110 mg/dL. Glycemic complexity was calculated using jackknifed approximate entropy. Logistic regression was used to adjust for confounders.

Results

A total of 495 patients (51%) developed complications, and 32 patients (3.3%) died. Along with older age, comorbidities, and complicated surgeries, any hypoglycemia (glucose <71 mg/dL) and the number of glucose values greater than 140 mg/dL were independent predictors of complications. Increased risk of mortality, after adjusting for other risk factors, was associated with older age, longer perfusion time, receiving intraoperative transfusions, and greater jackknifed approximate entropy of the glucose time series.

Conclusion

We found that hypoglycemia (glucose <71 mg/dL) and hyperglycemia (glucose >140 mg/dL) were associated with increased risk of complications, whereas greater complexity of the glucose time series was associated with mortality.

Introduction

Hyperglycemia is common in critically ill patients and is associated with increased morbidity and mortality, longer lengths of stay, and higher costs [1], [2]. High glucose levels interfere with protein function, impair endothelial cell function, inhibit nitric oxide production, and suppress normal immune function [3], [4]. Conversely, hypoglycemia is associated with impaired cell energetics, neuronal death, and mortality in critically ill patients [5], [6]. However, the level of glycemic control is controversial [1], [6], [7], and other studies have suggested that the variability or complexity of the glucose time series is more important than the actual glucose levels [8], [9], [10], [11]. The Society of Thoracic Surgeons guideline states that glucose should be kept less than 180 mg/dL but that studies to determine the optimal glucose levels are needed [12]. The purpose of this study is to determine the associations between glucose levels and control and complications.

Section snippets

Methods

This retrospective cohort study of prospectively collected data in cardiac surgery patients who were treated in the era of tight control was approved by the institutional review board, which waived consent. The primary outcome was composite complications of arrest, atrial fibrillation, bleeding, coagulopathy, gastrointestinal, heart block, infections, multisystem organ dysfunction, myocardial infarction, prolonged mechanical ventilation, renal failure, reoperation, stroke, and tamponade (//www.sts.org/sites/default/files/documents/pdf/trainingmanuals/adult2.61/Section_P_COMPLICATIONS.pdf

Results

A total of 970 patients underwent cardiac surgery between February 1, 2008, and July 30, 2010. Of the 431 patients with diabetes, 220 (51%) were controlled on oral medications; 137 (32%), on insulin; 32 (7%), by diet; and 42 (10%) had no treatment. Average glycemic control was poor (mean hemoglobin A1c, 7.7% ± 2.0%; only 45% had hemoglobin A1c <7.0%). Of the 11 235 glucose measurements, only 4194 (37%) were in the goal range (80-110 mg/dL), 1489 (13%) were below goal (<80 mg/dL), with 667 (5.9%)

Discussion

We found that after correcting for presurgical comorbidities, type of surgery, and intraoperative processes, both hyperglycemia and hypoglycemia were independently associated with increased risks of complications (Table 3). Importantly, we also found that hypoglycemia was associated with longer times (114 ± 124 vs 90 ± 91 minutes; P < .001) between glucose measurements, which suggests that avoiding delays or more frequent measurements in checking glucose levels, particularly when glucose levels

References (27)

  • N. Cueni-Villoz et al.

    Increased blood glucose variability during therapeutic hypothermia and outcome after cardiac arrest

    Crit Care Med

    (2011)
  • J. Hermanides et al.

    Glucose variability is associated with intensive care unit mortality

    Crit Care Med

    (2010)
  • G. Meyfroidt et al.

    Dynamic characteristics of blood glucose time series during the course of critical illness: effects of intensive insulin therapy and relative association with mortality

    Crit Care Med

    (2010)
  • Cited by (17)

    • Expert consensus on the glycemic management of critically ill patients

      2022, Journal of Intensive Medicine
      Citation Excerpt :

      Hyperglycemia also correlates closely with mortality, disability, and length of hospital stay. An observational study of critically ill patients undergoing surgery showed significantly higher fatality (OR=4.8, 95% CI: 1.4–20; P=0.02) among those with a blood glucose level >7.8 mmol/L than those with a blood glucose level of 4.4–7.8 mmol/L.[98] Another observational study found that patients with postoperative glucose >10.0 mmol/L had a higher incidence of postoperative complications, including acute kidney injury (OR=2.58), arrhythmias (OR=2.40), and sepsis (OR=3.86) than those with glucose <10.0 mmol/L.[99] However, postoperative hypoglycemia is also strongly associated with increased postoperative infection rates, poor wound healing, and fatality.[97,100] An observational study of patients undergoing cardiac surgery found that postoperative hypoglycemia (blood glucose <3.9 mmol/L) significantly increased postoperative fatality (OR=5.47, 95% CI: 3.14–9.5%) and disability (OR=4.66, 95% CI: 3.55–6.1%) rates.[101]

    • Classification of glucose records from patients at diabetes risk using a combined permutation entropy algorithm

      2018, Computer Methods and Programs in Biomedicine
      Citation Excerpt :

      Among all these more advanced mathematical methods, those based on signal complexity, regularity, or predictability estimation are gaining momentum due to their ability of capturing the subtle differences among subjects. Approximate Entropy (ApEn) [8], Sample Entropy (SampEn) [9], Fuzzy Entropy (FuzzEn) [10], Dispersion Entropy [11], State–Space Correlation Entropy [12], Bubble Entropy [13], Lempel Ziv Complexity (LZC) [14], Detrended Fluctuation Analysis (DFA) [15], Distribution Entropy (DistEn) [16], and Permutation Entropy (PE) [17], are just a few of these methods that have been applied successfully in the context of biomedical records, including glucose time series in some cases [18–23]. Specifically, Permutation Entropy (PE) [17] is a complexity measure that is receiving a lot of attention in the last years.

    View all citing articles on Scopus

    This study was conducted at Mercy St Vincent Medical Center, Toledo, OH, and the University of Michigan, Ann Arbor, MI.

    ☆☆

    This study was supported by departmental and institutional support.

    View full text