Abstract
Background and aims
Hyperglycemia, a major side effect of patients receiving total parenteral nutrition (PN), is associated with higher mortality in critically ill patients. The aim of this study was to determine whether elevated blood glucose levels would be associated with worse outcomes in patients receiving PN.
Methods
This retrospective study included postoperative patients admitted to our surgical intensive care unit (SICU) from July 2008 to June 2009. Data collected included blood glucose levels, length of stay, and outcome measures. Correlations among daily average, maximum, and minimum blood glucose levels and outcome measures were calculated.
Results
Sixty-nine patients were enrolled and divided into PN (n = 40) and non-PN (n = 29) groups. The initial mean blood glucose levels were 138.4 ± 63.1 mg/dL and 123.2 ± 41.8 mg/dL for the PN and non-PN groups, respectively. The mean blood glucose concentration was significantly increased (ΔBS = 44.8 ± 57.3 mg/dL; p < 0.001) in the PN group compared with the non-PN group (ΔBS = 39.4 ± 67.0 mg/dL; p = 0.004). The blood glucose concentration was significantly increased and consequently, consumption of insulin was increased on the 2nd day of ICU admission. The risk of mortality increased by a factor of 1.3 (OR = 1.30, 95% CI = 1.07–1.59, p = 0.010) for each 10 mg/dL increase in blood glucose level, when the daily maximum blood glucose level was >250 mg/dL. There were no cases of mortality in the current study when the blood glucose levels were controlled below 180 mg/dL. The mean blood glucose level in patients receiving PN was higher in those with diabetes than in those without diabetes (215.5 ± 42.8 vs. 165.8 ± 42.0 mg/dL, respectively, p = 0.001).
Conclusion
The blood glucose level was associated with patient outcome and should be intensively monitored in critically ill surgical patients. We suggest that blood glucose levels should be controlled below 180 mg/dL in postoperative critically ill patients.
Keywords
critical Illness; hyperglycemia; mortality; parenteral nutrition, total;
1. Introduction
Stress-induced hyperglycemia is a common problem in patients admitted to the ICU,1 even when glucose homeostasis has previously been normal irrespective of a history of diabetes. Critically ill patients are more likely to have a predisposition to hyperglycemia, and a blood glucose level >200 mg/dL is very common in such patients. Hyperglycemia, as a manifestation of the stress response, is most often evident shortly after admission to the ICU, and may resolve as the underlying catabolic illness subsides. Clinically, enteral or parenteral nutrition (PN) can be given to a patient to maintain an adequate nutritional status, and this is essential to maintain a competent immune system. However, overfeeding and inappropriate use of PN can lead to serious complications. Compared with enteral feeding, patients have a higher frequency of hyperglycemia and infections when receiving PN. The more grave the disease severity is, the higher the dosage of insulin is needed to achieve normoglycemia.
The nutritional status of hospitalized patients has been found to correlate with morbidity, mortality, and length of hospital stay.2, 3, 4, 5 Studies have indicated that 30–50% of hospitalized patients would show some degree of malnutrition, a condition that is associated with increased morbidity, and that malnourished patients have a 27% higher complication rate than well-nourished patients.2 Therefore, how to provide patients with an optimal nutritional supply is an important clinical consideration. PN is usually administered when enteral nutrition cannot be initiated within 24 hours after ICU admission.6
PN has been in common use since the 1960s and is accepted as the standard of care for patients with chronic non-functioning gastrointestinal tracts.6 The formula of PN includes optimal nitrogen and caloric requirements, with dextrose as the main source of energy. Patients receiving dextrose infusions such as total PN (TPN), are at the highest risk of developing hyperglycemia. In addition, stress-induced hyperglycemia is the result of increased sympathomimetic activity and increased release of counter-regulatory hormones and proinflammatory cytokines. Counter-regulatory hormones enhance glycogenolysis and gluconeogenesis to increase glucose production.7These are common problems for the critically ill and patients with diabetes. Moreover, being a patient with hyperglycemia in hospital is associated with higher mortality rates, increased infection rates, poorer outcomes, prolonged mechanical ventilation and longer hospital stay.7, 8, 9, 10 The aim of this study was to investigate the profiles of blood glucose and the related outcomes in postoperative critically ill surgical patients.
2. Materials and methods
2.1. Population and data collection
This study was a retrospective review of 198 adult surgical patients from the surgical intensive care unit (SICU) of a medical center in Taipei, Taiwan, from July 2008 to June 2009. The study was approved by the Review Board of the hospital. The inclusion criteria were postoperative patients aged above 18 years and nil per os (NPO) during their ICU stay after admission. The exclusion criteria were patients with missing records of blood sugar levels, length of stay <24 hours, and feeding or mortality within 24 hours after admission to the ICU. The blood glucose was measured, using a glucometer (Roche, Basel, Switzerland), every 4 hours (q4h) when patients were receiving PN treatment. The data collected included patient information, daily blood glucose levels, daily insulin dose, biochemical tests, past medical history, diagnoses, length of ICU stay, and whether or not the patients were still alive.
2.2. Nutritional assessment
The nutritional status of patients was assessed using the Subjective Global Assessment technique, and the total caloric requirement was calculated using the Harris-Benedict equation. The indication for PN was made according to the principle of the American Society for Parenteral and Enteral Nutrition (ASPEN).11 The TPN formula consisted of dextrose (156 g/L), amino acids (50 g/L), vitamins, and trace elements. TPN was usually commenced at 75 mL/hour, and then the rate was adjusted according to the patient’s requirements after 24 hours. Patients at risk of malnutrition and related morbidity and mortality can be identified with the Nutritional Risk Index (NRI). NRI = (1.5 × serum albumin, g/L) + (41.7 × current weight/usual weight). If NRI >100, it indicated that the patient is not malnourished, while 97.5–100 indicated mild malnourishment, 83.5–97.5 indicated moderate malnourishment, and <83.5 indicated severe malnourishment.
2.3. Statistical analysis
Means and standard deviations (SDs) were used to examine and describe the distribution of the data. The Student t test and Chi-square test were used to detect differences in clinical and demographic characteristics between the PN group and the non-PN group. The risk of outcomes for different blood glucose levels was determined by using logistic regression, in which the blood glucose level was used as a continuous variable. All statistical analyses were performed using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). A p value <0.05 was considered to be statistically significant.
3. Results
3.1. Patient characteristics
A total of 69 of the 198 patients were included in the study; 95 patients were excluded due to being non-NPO, 16 patients due to the length of stay being <24 hours, one patient due to being <18 years old, and 17 patients due to a lack of detailed blood glucose measurements. The recruitment and enrollment flowchart of the study patients is shown in Fig. 1.
Download full-size image
The baseline characteristics of the 69 patients who met the eligibility criteria are shown in Table 1. Overall, 57% (n= 40) of the patients received PN, of whom 26 (65%) were males with an average age of 71.3 ± 14.4 years and mean BMI of 23.7 ± 3.1 kg/m2. The initial mean blood glucose levels were 138.4 ± 63.1 mg/dL and 123.2 ± 41.8 mg/dL for the PN and non-PN groups, respectively. There were no significant differences in any baseline characteristics between the groups, except that the non-PN patients had a significantly lower APACHE II score, shorter hospital and ICU stay, and a higher NRI. We ordered α = 0.05, then detected the differences of baseline characteristics between PN (n = 40) and non-PN (n = 29) patients in the 69 patients; the range of statistical power (1-β) in this study was from 84.39% to 99.99% for APACHE II, length of hospital stay, length of ICU stay, albumin, and NRI (Table 1).
3.2. Variation of glucose levels in patients with and without PN treatment
All of the patients had blood glucose level measurements during their ICU stay, including mean, maximum, and minimum glucose levels. There were no significant differences in mean (183.2 ± 48.1 mg/dL vs. 162.6 ± 53.1 mg/dL; p = 0.098) and minimum (157.9 ± 38.4 mg/dL vs. 152.0 ± 46.6 mg/dL; p = 0.57) blood glucose levels between the PN and non-PN groups, respectively. However, there was a significant difference in the maximum blood glucose level between the two groups (210.9 ± 66.5 mg/dL in the PN group vs. 175.3 ± 64.9 mg/dL in the non-PN group; p = 0.03).
The differences in the mean, minimum and maximum blood glucose levels compared to the initial baseline levels (ΔBS) were calculated. We found that the mean blood glucose concentration was significantly increased (ΔBS = 44.8 ± 57.3 mg/dL; p < 0.001) in the PN group and the non-PN group (ΔBS = 39.4 ± 67.0 mg/dL; p = 0.004). The maximum and minimum blood glucose levels were significantly increased in the PN group (Δmax BS = 72.5 ± 68.7 mg/dL; p < 0.001; Δmin BS = 19.52 ± 58.4 mg/dL; p < 0.041). The maximum blood glucose level in the non-PN group was significantly increased (Δmax BS = 52.0 ± 76.7 mg/dL; p = 0.001), but the minimum blood glucose level was not (Δmin BS = 28.8 ± 62.3 mg/dL; p = 0.190).
The data were also analyzed by dividing the mean glucose levels into five quintiles (Q1: 90–110; Q2: 111–150; Q3: 151–200; Q4: 201–250; Q5: >250 mg/dL). There were no significant differences in complications, including infection, septicemia, acute renal failure, and acute respiratory failure among the different blood glucose groups. More patients in the PN group (35%) had a mean blood glucose level >200 mg/dL in comparison with those in the non-PN group (17%). Additionally, hyperglycemia was more frequently noted in patients in the PN group than in the non-PN group (Fig. 2).
Download full-size image
3.3. Correlation between blood glucose and hospital mortality in patients with PN treatment
The characteristics of the patients are shown in Table 1. Of the 69 patients in the study, 61 survived (PN: 53.2% vs. non-PN: 46.8%). There were 40 patients receiving PN, of whom 33 survived and seven did not. There were no significant differences in age, body height, and body weight between the two groups of patients (Table 2). However, there were significant differences between the survivors and non-survivors in APACHE score (p < 0.001), length of ICU stay (p < 0.001) and NRI (p = 0.005). We further divided the patients by APACHE score into three tertiles (score: 0–14, 15–24, >25), and found that five of the seven patients who died had a first day APACHE score >25, and these patients also stayed in the ICU for >5 days (p = 0.003). Consequently, the higher the APACHE score, the higher the mortality rate and the greater the length of ICU stay would be.
On further investigation of the differences in blood glucose levels between the survivors and non-survivors in the PN group, we found that there were no significant differences in baseline (169.2 ± 92.3 mg/dL vs. 131.8 ± 54.8 mg/dL) and minimum blood glucose levels (155.5 ± 40.7 mg/dL vs. 170.2 ± 25.6 mg/dL). However, the mean and maximum glucose levels were significantly higher in the non-survivors compared with the survivors (Table 3). We also compared the blood glucose levels and the dose of insulin in the first 4 days after ICU admission. We found that the blood glucose concentration was significantly increased and consequently the dose of insulin was higher on the 2nd day of ICU admission (Fig. 3). Logistic regression analysis was performed to determine whether a maximum blood glucose level >250 mg/dL was associated with a higher risk of mortality. We found that the risk of mortality increased by a factor of 1.3 (OR: 1.30, 95% CI: 1.07 to 1.59, p = 0.010) for each 10 mg/dL increase in blood glucose level above 250 mg/dL.
Download full-size image
3.4. Relationship of diabetes to glucose values and hospital mortality
We supposed that diabetic patients receiving PN would have higher blood glucose levels and a higher mortality. Therefore, we investigated the relationship between diabetic glucose values and hospital mortality. In our study, 35% of the PN patients had diabetes. There were no significant differences in age (70.1 ± 12.9 years vs. 72.0 ± 15.4 years, p = 0.692), hospital stay (30.5 ± 16.7 days vs. 26.9 ± 13.1 days, p = 0.459), and ICU stay (7.5 ± 12.2 days vs. 6.0 ± 5.7 days, p = 0.606) between patients with and without diabetes, respectively. In addition, there was no difference in mortality rate between the patients with and without diabetes (13.6% vs. 8.5%, respectively, p = 0.67). However, the mean blood glucose level in those who received PN therapy was higher in the patients with diabetes than that in those without diabetes (215.5 ± 42.8 mg/dL vs. 165.8 ± 42.0 mg/dL, p = 0.001).
4. Discussion
The most important finding of this study is that hyperglycemia was associated with a poor outcome in the surgical critically ill patients who received total PN. Although a close correlation between hyperglycemia and adverse outcomes in critically ill patients has been reported,12, 13 our findings suggest that clinicians should take precautions for postoperative critically ill patients.
Compared with the non-PN group, there were more patients in the PN group with a mean blood glucose level >200 mg/dL (35% vs. 17%). We also found that blood glucose levels were associated with the severity of the disease, especially during the first 2 days of ICU stay. Studies have shown that hyperglycemia is associated with high-rate, continuous infusion of TPN dextrose.14, 15Consequently, the guidelines of the ASPEN for stressed patients recommend a TPN dextrose infusion rate of no more than 5–7 mg/kg/minute to prevent an increased risk of hyperglycemia.11 In our study, the infusion rate of TPN dextrose was kept at 1.2 mg/kg/minute, much lower than the recommended rate. However, hyperglycemia still developed, especially when the APACHE II score was >25, suggesting that glucose control should be carefully monitored for postoperative critically ill patients with severe disease.
Tucker16 described the variation in the average length of stay for 795 patients in medical and surgical ICUs. They found that the average length of stay was influenced by many uncontrollable factors (such as the number of diagnoses, expected length of stay, gender, surgery, and height, weight, and age of the patient), and that the day of nutritional intervention was the only controllable factor. In the current study, we found that the severity of disease correlates with the length of ICU and hospital stay. This suggests that multiple factors may influence the average length of stay. In addition, we found that mortality was associated with APACHE score (p < 0.001), length of ICU stay (p < 0.001) and NRI (p = 0.005). The survivors had lower APACHE II scores than did the non-survivors, but gender had no effect on survival in this study. In addition, we could not analyze the impact of preoperative administration of TPN support on complications and length of hospital stay in this retrospective study.
Cheung et al12 reported that for each 18 mg/dL increase in blood glucose level, the risk of death increased by a factor of 1.77 for patients receiving TPN. Lin et al13 found that an increase of 10 mg/dL in mean blood glucose level was associated with a 1.1-fold increased risk of mortality. In this study, all of our patients were surgical patients without receiving any oral intake, which was quite different from Cheung’s study and Lin’s study, in which surgical and non-surgical patients and enteral feeding were inclusive, except enteral feeding. Similarly, we found that the risk of death increased by a factor of 1.3 for each 10 mg/dL increase in blood glucose level. Therefore, hyperglycemia is a risk factor of mortality, and the higher the blood glucose level is. the higher the mortality rate will be.
It has been demonstrated that diabetic patients have a high risk of death and complications.17, 18, 19 In this study, the incidences of complications and rate of death were not significantly higher in the diabetic than in the non-diabetic patients receiving TPN. This might be due to the fact that the sample size of the diabetic patients in this study was not large enough. However, the mean blood glucose level was higher in patients with diabetes than in those without diabetes. Therefore, the blood glucose level should be intensively monitored in diabetic patients receiving TPN in the ICU setting.
We found that, as the blood glucose concentration was significantly increased, consequently the consumption of insulin was increased on the 2nd day of ICU admission. Hence, we hypothesized that either the severity of disease worsened or the sensitivity of insulin decreased on the 2nd day of ICU admission. In this study, the APACHE score was higher on the 2nd day of admission in the non-survivors than in the survivors. This may explain why the consumption of insulin was higher on the 2nd day. It has been demonstrated that hyperglycemia is associated with the outcome of critically ill patients, and good glucose control can improve the morbidity and reduce mortality.7, 8, 9, 10 In our study, logistic regression analysis disclosed that a maximum blood glucose level >250 mg/dL was associated with a higher risk of mortality. We found that the risk of mortality increased by a factor of 1.3 for each 10 mg/dL increase in blood glucose level.
How is the levels of blood glucose control can improve outcomes is still controversial. Van den Berghe et al20, 21 reported that tight glucose control to achieve a glucose level between 80 and 110 mg/dL resulted in the fall of ICU mortality rate from 8.0% to 4.6%. Some meta-analysis studies have reported that tight glucose control is not associated with a significantly reduced hospital mortality, however, it is associated with an increased risk of hypoglycemia.22, 23 The NICE-SUGAR study group concluded that a blood glucose target of <180 mg/dL is optimal, and that this level can reduce mortality in the ICU.24 There were no cases of mortality in the current study when the mean blood glucose levels were controlled <180 mg/dL. This finding supports the finding of the NICE-SUGAR study group, even in postoperative critically ill patients. Therefore, we suggest that keeping a blood glucose target of <180 mg/dL is essential in treating surgical critically ill patients.
The main limitation of this study is its retrospective nature. Other limitations include the small sample size, short duration of observation, and the effect of medication used during ICU admission. Future studies should address these limitations.
In conclusion, blood glucose levels were associated with patient outcomes and should be intensively monitored in critically ill surgical patients. We suggest that blood glucose levels should be controlled below 180 mg/dL in postoperative critically ill patients. In addition, caution should be taken when the blood glucose level is higher than 250 mg/dL, as we found that the risk of death could increase by a factor of 1.3 for each 10 mg/dL increase in blood glucose level.
Acknowledgments
The authors thank Miss Liu Jo-Hsiang (Institutional Review Board of Tri-Service General Hospital) and Mr. Chou Yu-Ching (School of Public Health, National Defense Medical Center) for their kind assistance in statistics. This study was supported in part by grants DOD97-T09-04 and TSGH-C100-064from Tri-Service General Hospital, Taipei, Taiwan.