Relationship between glycated hemoglobin, Intensive Care Unit admission blood sugar and glucose control with ICU mortality in critically ill patients

Abstract

Background and Aims:

The association between hyperglycemia and mortality is believed to be influenced by the presence of diabetes mellitus (DM). In this study, we evaluated the effect of preexisting hyperglycemia on the association between acute blood glucose management and mortality in critically ill patients. The primary objective of the study was the relationship between HbA₁c and mortality in critically ill patients. Secondary objectives of the study were relationship between Intensive Care Unit (ICU) admission blood glucose and glucose control during ICU stay with mortality in critically ill patients.

Materials and Methods:

Five hundred patients admitted to two ICUs were enrolled. Blood sugar and hemoglobin A₁c (HbA₁c) concentrations on ICU admission were measured. Age, sex, history of DM, comorbidities, Acute Physiology and Chronic Health Evaluation II score, sequential organ failure assessment score, hypoglycemic episodes, drug history, mortality, and development of acute kidney injury and liver failure were noted for all patients.

Results:

Without considering the history of diabetes, nonsurvivors had significantly higher HbA₁c values compared to survivors (7.25 ± 1.87 vs. 6.05 ± 1.22, respectively, P < 0.001). Blood glucose levels in ICU admission showed a significant correlation with risk of death (P < 0.006, confidence interval [CI]: 1.004–1.02, relative risk [RR]: 1.01). Logistic regression analysis revealed that HbA₁c increased the risk of death; with each increase in HbA₁c level, the risk of death doubled. However, this relationship was not statistically significant (P: 0.161, CI: 0.933–1.58, RR: 1.2).

Conclusions:

Acute hyperglycemia significantly affects mortality in the critically ill patients; this relation is also influenced by chronic hyperglycemia.

Introduction

Diabetes of injury has been shown to result in increased mortality rates during critical illness,[1,2,3] necessitating targeted blood glucose control according to numerous international guidelines, irrespective of the presence of diabetes mellitus (DM).[4,5,6] Perioperative blood glucose control also decreases morbidity and mortality in critically ill patients.[7,8] Interestingly, acute hyperglycemia (induced by stress), hypoglycemia, and glycemic variability have been found to be poor prognostic factors in various subsets of critically ill patients including trauma patients. Therefore, it seems that the impact of acute hyperglycemia on outcome depends on premorbid glycemia control. Recent studies indicate that the association between hyperglycemia and outcome could be influenced by the presence of chronic hyperglycemia, and they have suggested that optimal blood glucose concentration might be higher in these patients.[9,10,11] Thus, the risk of decreasing blood sugar with insulin therapy during Intensive Care Unit (ICU) stay could be increased; which is related to the severity of hyperglycemia within the previous 2–3 months. It has been estimated that each 1% augmentation in hemoglobin A₁c (HbA₁c) concentrations is associated with 15–20% greater cardiovascular risks.[12] A₁C reflects longer-term glycemic control and is less influenced by acute stress. Therefore, A₁C levels may provide insight into the relation between chronic glucose control and patient outcomes. Thus, it seems that HbA₁c level on admission might be a more proper reflector of chronic hyperglycemia, which would detect occult diabetes as well. HbA₁c can also be used to indicate the correlation between its higher levels and increased morbidity and mortality in the critically ill patients.[13,14] Yet, there are conflicting reports on the impact of chronic hyperglycemia on the outcome of critically ill patients. Previous studies showed a relationship between HbA₁c and mortality in critically ill patients.[15,16] However, Britton et al. in their cohort of patients with diabetes presenting with acute myocardial infarction, did not observe a J-shaped association between A₁C and mortality.[17]

Hence, we hypothesized that beside ICU admission hyperglycemia, HbA₁c (pre-ICU control of hyperglycemia), and also targeted blood glucose control could be considered as three important factors associated with mortality. Consequently, based on the controversies in this field, in this study, we evaluated our hypothesis with showing the relationship between HbA₁c, ICU admission blood sugar, and glucose control during ICU stay with mortality in critically ill patients.

Materials and Methods

This study was approved by the Ethics Committee of Tabriz University of Medical Sciences. Five hundred patients were enrolled in this study performed in two university-based hospitals (Shohada and Imam Reza) affiliated to Tabriz University of Medical Sciences in Tabriz, Iran, from April 2012 to February 2014. Both ICUs were surgical type ICU. Written informed consents were obtained from the patients or their surrogates. ICU admission blood sugar and HbA₁c were measured for all patients. Age, sex, history of DM, comorbidities, Acute Physiology and Chronic Health Evaluation (APACHE II) score, sequential organ failure assessment (SOFA) score, hypoglycemic episodes, type of nutrition, drug history (aspirin, corticosteroids, beta blockers, metformin, and glibenclamide), mortality, and development of acute kidney injury and liver failure were noted for all patients [Table 1]. Targeted glucose control (150 mg/dl in septic patients and 140mg/dl in other patients) was performed in all patients. All patients received intravenous insulin based on their blood glucose values. Glucose measurement was performed every 1 h with glucometers (Omron Hea-230 Glucometer). If four consecutive samples were within the normal range, the intervals for glucose measurement would be increased to every 2 h. Then, if three consecutive samples were in normal range, glucose sampling would be performed every 4 h. All patients had arterial lines for blood glucose sampling and measurement was performed by equal glucometers in two hospitals. Maximum interval for glucose sampling was 4 h. If the patient was hypoglycemic, insulin was held and DW50% in the volume of (100-BS) ×0.4 mL was infused to the patient and sampling was performed every 30 min until he/she was euglycemic. Time-weighted average of glucose concentration was calculated based on the study of Finney et al.[18]

Table 1.

Patients’ characteristics during ICU stay

Statistical analysis

Data were presented as percent and mean ± standard deviation. The primary outcome of the study was mortality. Analysis was performed with SPPS version 16. A two-sided P = 0.05 was considered to be statistically significant. We used student t- test for comparison of mean changes between two parameters and ANOVA for comparison of changes between groups in quantitative variables. Chi-square was used for the analysis of qualitative variables between groups. We analyzed the effects of ICU admission glucose, HbA₁c, and hypoglycemic episode as independent variables on mortality with logistic regression analysis model.

Results

Five hundred patients were enrolled in this study. The demographic characteristics of the patients have been presented in Table 1. The most common reasons for ICU admission were multiple trauma, brain malignancy, and disc herniation. The most common comorbidity was DM (18.8%). There was previous history of taking aspirin, corticosteroids, or beta blockers in 242 patients (48.4%). In 18.8% of patients, a history of oral glucose control agents was noted. Ninety-eight percent of patients received metabolic support as enteral nutrition, and the rest received parenteral nutrition. ICU mortality was 9.2%, and hypoglycemic episodes were reported in 10% of patients. The mean score for APACHE was 20.7 ± 5.8 and mean SOFA score was 9.7 ± 2.7. We compared variables in two groups: Survivors and nonsurvivors. Nonsurvivors had significantly higher HbA1c values compared to survivors, without considering the history of diabetes (7.25 ± 1.87 vs. 6.05 ± 1.22, respectively, P < 0.001). An almost weak correlation was observed between HbA₁c and ICU admission glucose levels (P < 0.001, r: 0.45). ICU admission blood glucose showed a significant correlation with the risk of death (P < 0.001, confidence interval [CI]: 1.004–1.02, relative risk [RR]: 1.01). Hypoglycemic event increased the risk of death by 20%; however, this relation was not statistically significant (P: 0.221, CI: 0.927–1.58, RR: 1.2). Logistic regression analysis showed that HbA1c increased the risk of death; with each increase in HbA1c level, the risk of death doubled. However, this relation was not statistically significant (P: 0.161, CI: 0.933–1.58, RR: 1.2).

Discussion

The concept of harmful glucose levels bears two ends: One end is being too high while the other is being too low. Our study showed that ICU admission blood glucose levels (acute hyperglycemia) and HbA₁c (preexistent hyperglycemia) were correlated with patients mortality. These findings showed that acute hyperglycemia has a correlation with mortality, which might be influenced by previous hyperglycemia. Although HbA1c values have been widely investigated as an index of long-term blood glucose control and outcome predictors in diabetic patients, their predictive value in the critically ill patients admitted to ICU has received little attention.

Egi et al. showed that the higher mean glucose levels (>180 mg/dl) during ICU stay is related to lower hospital mortality in patients with higher HbA₁c (>7%) compared to low HbA₁c (<7%). Zhang et al.[19] showed that HbA1c should be assessed only in patients without the history of DM to determine whether they suffer from hyperglycemia. A study by Plummer et al.[20] in critically ill patients showed that acute hyperglycemia is associated with increased mortality in patients with adequate previous glycemic control, but not in patients with abnormal premorbid hyperglycemia. After adjusting for age, body mass index (BMI), APACHE II, and admission type, the predicted mortality curves were no longer significant for patients with HbA₁c <6% and between 6% and 7%, probably reflecting the dominant association between APACHE II and mortality. Unfortunately, they did not assess the hypoglycemic events and only used peak blood glucose during first 48 h after ICU admission. There are numerous studies suggesting strong associations between hyperglycemia and mortality in critically ill patients, but patients with unrecognized diabetes have not been identified in these studies;[11,21,22,23,24,25] placing them as "not having diabetes" which is also the limitation of our study. Zaghla et al. showed that elevated admission glucose level is a strong predictor of short-term adverse outcome in patients with acute coronary syndrome; however, the prognostic value of HbA₁c levels in these patients is still undefined.[26] The results of ADVANC study showed that a less marked reduction in HbA1c levels (i.e., from 7.5% to 6.5%) was associated with a significant reduction of major vascular events and a nonsignificant beneficial trend in mortality.[27] On the other hand, the Action to Control Cardiovascular Risk in Diabetes study showed that a higher baseline level of HbA₁c and greater reduction in its levels appeared to be associated with increased risk in the ambulatory setting, similar to the acute setting.[28] A study by Farrugia et al. suggests that HbA₁c may not be a proper predictor of outcome in the general ICU population but may be of predictive value in diabetic ICU Patients. On the other hand, in a similar study, blood glucose levels on admission were introduced as a predictor of mortality, while no association between HbA1c and cortisol levels on admission was found.[29] A risk adjusted analysis of 120 traumatic patients by Karen et al. showed that patients with HbA₁c of more than 6% were 4.5 times more likely to have poor outcomes compared with those with HbA₁c of <6. However, the authors did not evaluate the impact of glucose control or acute glycemia on the outcome of patients.[30] Results of a study performed on 119 critically ill patients during the first 24 h showed that the risk of the in-hospital mortality significantly increases with HbA₁c levels above 9.3% and BMIs below 24 kg/m²; surprisingly, having no association with mortality in overweight patients, while progressively declining in those with a BMI above 35 kg/m². The weak correlation between HbA₁c and BMI suggests an independent effect of these metabolic parameters as prognostic factors in critically ill patients.[31] Kompoti et al. showed that a cutoff of 6.5% for HA1c could predicte higher ICU mortality in patients without prior history of diabetes; but, in patients with prior history of diabetes a cutoff of 7 for HbA1c had no association with ICU mortality.[32] The possible mechanisms explaining the effects of HbA₁c variability on outcomes might be as followings: Low socioeconomic status which might indicate a suboptimal management of diabetes, insulin resistance, which in itself has been implicated in the pathogenesis of diabetes complications,[33] and homeostatic imbalance; for example, through intracellular sorbitol accumulation.[34] Previous studies suggest that increased HbA₁c levels are a negative prognostic marker in DM patients admitted to ICUs. The results of a cohort study in German showed that known diabetes and HbA₁c-defined undiagnosed diabetes were both associated with an increased risk for all-cause mortality after 11.6 years of follow-up in this nationwide German cohort. However, adults with prediabetes, as defined by HbA₁c, did not differ in their mortality risk from those with normoglycemia. A U-shaped association was found between continuous HbA₁c measures and all-cause mortality, suggesting that not only high but also low HbA₁c levels might be associated with an increased risk of death.[16] As DM is usually asymptomatic at early stages, many patients may not be aware of this condition on ICU admission. Acute insulin resistance is associated with worse outcomes in nondiabetic patients. In addition, critical illness characteristics influence glycemic control and clinical outcome in ICU patients. The effect of preexisting hyperglycemia in addition to acute hyperglycemia on mortality of critical ill patients might be due to the increased inflammatory cytokine concentration or relative neuroglycopenia, or perhaps other yet-unidentified mechanisms. While there has recently been a shift in focus to individualizing HbA1c targets for ambulant patients with Type 2 diabetes, the hypothesis that premorbid glycemic control may modulate the response to hyperglycemia during critical illness has not been tested.

Limitations of study

Important limitations of this study include reliance on single glycated hemoglobin and glucose measurements at baseline and a limited number of fasting glucose measurements during the follow-up period. This study was performed in two ICUs whose most patients were surgical patients.

Conclusion

Acute hyperglycemia secondary to critical illness-associated hyperglycemia or diabetes (both recognized and unrecognized) frequently occurs in critically ill patients and appears to have a complex relationship with mortality. Our results showed that acute hyperglycemia has a significant effect on mortality of critically ill patients, and this relation is influenced by chronic hyperglycemia.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.