Abstract
Context:
Stress hyperglycemia in critically ill patients has been a matter of debate for years without any conclusive answer till date regarding glucose management and treatment thresholds.
Aims:
We planned a study with an aim to compare the efficacy of intensive versus conventional insulin therapy in reducing the mortality and morbidity in critically ill patients. The primary objective was to compare mortality between the two groups. The secondary objective was to find out if intensive insulin therapy is better than conventional insulin therapy in terms of various outcomes such as infections and need for inotropes and transfusion requirements.
Settings and Design:
It was a prospective randomized controlled study. The study population included 100 patients who received mechanical ventilation and admitted to the intensive care department of a tertiary care institute.
Subjects and Methods:
Patients were randomly assigned to two groups: intensive insulin therapy (IIT) and conventional insulin therapy (CIT) to receive either intensive or conventional insulin therapy. Insulin infusion was started only when blood glucose levels exceeded 200 mg%. Blood glucose levels were maintained between 80 and 110 mg% in the IIG and between 180 and 200 mg% in the CIG.
Statistical Analysis Used:
The data collected were analyzed separately for both the groups using Student's t-test and Chi-square test.
Results:
The two groups were comparable in terms of baseline demographic data including age, sex, preadmission diabetic status, and HbA1c at the time of admission. The two groups were not comparable in terms of Acute Physiology and Chronic Health Evaluation-II scores, and the difference between them was statistically significant with higher scores in the conventional group. The primary outcome, that is, mortality, was higher in the CIG with 29 patients (58%) versus 3 (6%) in the IIG (P = 0.02). The secondary outcomes were the measures of morbidity including infections, need for inotropic support, and need for blood transfusions, and these were significantly higher in the conventional group (P < 0.05).
Conclusions:
We conclude that tight glycemic control significantly lowers mortality and morbidity in critically ill patients, both surgical and medical. These benefits appear with the maintenance of tight blood glucose control of 80–110 mg.dL − 1 and not due to administration of insulin.
Keywords: Conventional, conventional glycemic control, insulin infusion, insulin therapy, intensive care, mechanical ventilation, stress hyperglycemia
INTRODUCTION
Critically ill patients pose unique challenges for the intensivists and endocrinologists with regard to glycemic control.[1] Stress hyperglycemia is a common problem in patients admitted to the intensive care unit, (ICU) even when glucose levels have previously been normal. No clear guidelines have been set for defining hyperglycemia in a critically ill patient, which explains the wide variations in the reported prevalence of hyperglycemia in critically ill patients, ranging from 3% to 71%.[2] Stress hyperglycemia can be defined as a blood glucose level >140 mg.dL −1 without a previous history of diabetes or glycated hemoglobin (HbA1c) >6.5%.[3] Strict control of the blood glucose concentration by intensive insulin therapy has been shown to reduce mortality and morbidity;[4] however, higher incidence of hypoglycemia in patients with tight glucose control using intensive insulin therapy is a matter of concern.
Hyperglycemia has been shown to be associated with increased morbidity and mortality after burns, surgery, strokes, myocardial infarction, and head trauma.[5] One of the earliest studies on this topic demonstrated better outcome with tight glycemic control (80–110 mg.dL − 1) as compared to conventional control in critically ill surgical patients.[6] A repeat study by the same authors a few years later in the medical ICU demonstrated survival benefits with tight glycemic control only in patients with more than 3 days of ICU admission; however, the number of days on ventilator and days spent in ICU were noted to be decreased in all subsets.[7] A meta-analysis done on 29 randomized controlled trials showed no difference in hospital mortality in patients with tight glycemic control.[8] A retrospective association study in patients admitted to medical and surgical ICU found that intravenous insulin administration is associated with an increased risk of ICU and hospital mortality, after correction for confounding factors and the average insulin use was 0.4 IU/kg/day.[9] Therefore, the optimal glycemic target for critically ill patients remains to be defined. Due to ethnic and racial differences in insulin resistance, glucose metabolism, genetic makeup, and dietary pattern, the guidelines which apply to the Western world may not hold true in Indian and South Asian population.
We planned a study with an aim to compare the efficacy of intensive versus conventional insulin therapy in reducing morbidity and mortality in critically ill patients. The primary objective was to compare mortality between the two groups. The secondary objective was to find out if intensive insulin therapy is better than conventional insulin therapy in terms of various outcomes such as infections and need for inotropes and transfusion requirements.
SUBJECTS AND METHODS
After institutional ethical committee approval, we conducted a prospective randomized controlled study which included 100 medical and surgical patients aged more than 12 years, admitted to the ICU of the department of anesthesiology at our institute which is a tertiary health-care institute. The study was conducted for a period of 1 year. All the patients included received mechanical ventilation. All patients were from the same geographical location. The patients with a diagnosis of renal failure on dialysis and pregnant patients were excluded from the study.
Written informed consent was obtained or delayed consent was obtained from the kith and kin (more than 18-year-old) of the patients or from a legal surrogate.
A power calculation based on the reduction in in-hospital mortality in a previous study[6] showed that 44 patients were needed in each group to show a statistically significant difference between treatments at a power of 80% and alpha = 0.05. We included 50 patients in each group keeping in mind the number of dropouts that could exist.
On admission to the ICU, all mechanically ventilated patients were randomly assigned with the help of a computer-generated random number table to receive either intensive insulin treatment (Group-IIT) or conventional insulin treatment (Group-CIT). Randomization was done by the ICU counselor attached to our ICU.
Baseline demographic and clinical information was obtained by us including information necessary to determine the severity of illness. This included age, sex, preadmission diabetes, and diagnosis at the time of admission to ICU and Acute Physiology and Chronic Health Evaluation (APACHE-II) scores. Glycosylated hemoglobin (HbA1C) levels were measured at the time of admission to determine preadmission glycemic control. Blood cultures were obtained whenever the central body temperature exceeded 38.5°C. Urine cultures and tracheal cultures were taken as and when required. In both the groups, baseline blood glucose levels were measured to see the glycemic state at the time of admission. Insulin infusion was started only when blood glucose levels were more than 200 mg% in the conventional group and more than 110 mg% in the intensive group. Insulin infusion was prepared using 100 units of Human Actrapid in 100 ml of normal saline. Insulin infusion rates were regulated to achieve target blood glucose levels.
Blood glucose monitoring was done at the time of admission and thereafter hourly for the first 4 h, then at 6 and 10 h of admission, and then at least once a day by trained ICU staff using a digital electronic glucometer (”Accu-chek Active” from Roche). Blood was drawn from a central venous line if it was existing or capillary blood was drawn by finger-prick method in case there was no central line. Blood glucose levels were maintained between 80 and 110 mg.dL −1 in the IIT group and between 180 and 200 mg% in the CIT group.
All patients were started on partial nutrition support with mainly intravenous glucose with attempts at starting enteral feeding as early as possible.
All the data were collected as in the protocol. Patient outcome was measured based on APACHE-II score. The primary outcome measure was mortality from any cause. Secondary outcome was measured in terms of infections (culture positive), need for inotropic support, need for blood transfusions, duration of mechanical ventilation, and need for prolonged ICU care. The data were presented as means ± standard deviation.
Statistical analysis
For analysis of continuous variables, independent samples Student's t-test was applied, and for categorical variables, Chi-square test was used. P < 0.05 was considered statistically significant.
RESULTS
The baseline demographic data were comparable between the two groups with respect to age, sex, HbA1c levels, and diagnosis at admission [Table 1].
Table 1.
Patient demographics
| Parameter | Group IIT n (%) | Group CIT n(%) |
|---|---|---|
| Age* (in years) | 49.26±14.2 | 49.88±13.26 |
| female sex | 17 (34) | 14 (28) |
| HbA1c† | 4.61 | 5.01 |
| Diagnosis at admission | ||
| Post- operative events | 12 (24) | 9 (18) |
| Infections | 23 (46) | 25 (50) |
| Poisoning | 5 (10) | 1 (2) |
| Trauma | 3 (6) | 1 (2) |
| APACHE II | 23.72±5.42 | 27.64±5.73 |
| Presence of diabetes | 13 (26) | 22 (44) |
| Presence of serum ketones | 2 (4) | 5 (10) |
| History of diabetes at the time of admission | 13 (26) | 22 (44) |
| Blood glucose levels at the time of admission(mg/dl) | ||
| Diabetic* | 234±55.79 | 286.95±54.34 |
| Non- diabetic* | 219.78±43.86 | 260.21±26.97 |
n(%)= number (percentage) of cases, *=Values are expressed as mean± standard deviation. † = mean APACHE= Acute physiology and chronic health evaluation, S.D= Standard deviation, HbA1c=Glycosylated hemoglobin, HbA1c= Glycosylated hemoglobin
The two groups were not comparable in terms of APACHE-II scores and the difference between them was statistically significant with higher scores in the CIT group, with mean score being 23.72 in the IIT group versus 27.64 in the CIT group. There were more patients with preadmission history of diabetes mellitus (DM) in the CIT group. The mean insulin requirement in the first 6 h of admission to the ICU ranged from 1.5 to 6.5 units h−1 with a mean of 2.98 units h−1 ± 0.90 in the IIT group, whereas in the CIT group, insulin requirement in the first 6 h ranged from 1.5 to 6.6 units h−1 with a mean of 3.16 units h−1 ± 1.30, the difference between the two groups being statistically insignificant (P = 0.43) [Table 2]. The target blood glucose control was achieved within 5.9 h ± 2.48 in the IIT group and 4.6 h ± 2.69 in the CIT group. The difference was statistically significant with longer time needed for reaching target blood glucose levels in the IIT group.
Table 2.
Insulin requirements in first 6 hrs
| Time (hours) | Group IIT- Insulin needed (units.hr-1) | Group CIT-Insulin needed (units.hr-1) |
|---|---|---|
| 00 | 4.16 | 4.18 |
| 01 | 3.68 | 3.8 |
| 02 | 3.01 | 3.26 |
| 03 | 2.25 | 2.87 |
| 04 | 1.91 | 2.71 |
| 05 | 1.63 | 1.81 |
| Mean | 2.98 | 3.16 |
| Standard deviation | 0.9 | 1.3 |
| Range | 1.5-6.5 | 1.5-6.6 |
P value:0.43 (P>0.05= not significant)
The primary outcome, that is, mortality from any cause was higher in the conventional group with 29 patients (58%) versus 3 (6%) in the intensive group. The secondary outcomes were the measures of morbidity, with infections, need for inotropic support and need for blood transfusions, and they were significantly higher in the conventional group. Infections were quantified with the number of positive cultures, blood cultures being positive in 18% of conventional group versus 0% in the intensive group, urine cultures were positive in 12% of patients in conventional group versus 0% in the intensive group and tracheal cultures being positive in 80% of patients in conventional group versus 20% in the intensive group [Table 3]. 14% patients in the intensive group and 56% patients in the conventional group developed respiratory infections during the course of treatment. In the intensive group, 26% patients needed inotropic support while in the conventional group, 48% patients needed inotropic support. In the intensive group, 20% patients needed blood transfusions while in the conventional group, 40% patients needed blood transfusions. More patients (16%) had hypoglycemic episodes in the IIT group versus 14% in the CIT group but the difference between the two groups was not found to be statistically significant (P = 0.09) [Table 4].
Table 3.
Number of cultures for quantification of infections
| Cultures | Group IIT n (%) | Group CIT n (%) |
|---|---|---|
| At admission | ||
| Blood | 0 (0) | 1 (2) |
| Urine | 0 (0) | 0 (0) |
| Tracheal | 3 (6) | 12 (24) |
| During ICU stay | ||
| Blood | 0 (0) | 9 (18) |
| Urine | 0 (0) | 6 (12) |
| Tracheal | 10 (20) | 40 (80) |
n(%)=number (percentage) of cases. ICU=intensive care unit
Table 4.
Comparison of outcomes and mortality/morbidity events
| Outcome/mortality/morbidity event | Group IIT n (%) | Group CIT n (%) | P |
|---|---|---|---|
| Mortality | 3 (6) | 28 (56) | 0.02‡ |
| Mortality in APACHE 11 21-25 | 0 (0) | 7(38.89) | 0.33† |
| Mortality in APACHE 11 26-30 | 0 (0) | 10(62.5) | 0.33† |
| Number of days on ventilator* | 4.06±1.76 | 4.9±2.24 | 0.04‡ |
| Need for prolonged ICU care | 1 (2) | 4 (8) | 0.11† |
| Hypoglycemic episodes | 8 (16) | 7 (14) | 0.09† |
| Need for ionotropic support | 13 (26) | 24 (48) | 0.03‡ |
| Need for blood transfusion | 10 (20) | 20 (40) | 0.04‡ |
| Positive cultures n | 20% | 80% | 0.038‡ |
Chi –square test was applied, ‡ =significant; † =non-significant. n(%)= number (percentage) of cases; * =values are expressed as mean± standard deviation, APACHE=Acute physiology and chronic health evaluation, SD= Standard deviation. ICU= Intensive Care Unit
DISCUSSION
The choice of blood sugar control technique in the ICU has been a subject of ongoing research since a long time during the last twenty years following the publication of the study by van den Berghe et al. in 2001. Intensive glucose control is advocated by many authors, but it may be associated with hypoglycaemia, increased hospital cost and frequent blood sampling/finger pricking.[4] To avoid this, continuous glucose monitoring devices are available, but they cannot be used in patients who are unable to provide feedback of their hypoglycemia symptoms. Our patients were on mechanical ventilation and unable to provide feedback and we did not use continuous glucose monitoring. Our study was designed to compare the efficacy of intensive (blood glucose in the range of 80-110 mg%) versus conventional (blood glucose in the range of 180-200 mg%) insulin therapy in reducing mortality and morbidity in critically ill patients. In our study, the primary outcome, that is, mortality was significantly lower in the intensive group as compared to the conventional group. Nevertheless, with respect to mortality, different researchers have obtained varied results. Van den Berghe and others in a study on surgical patients found an apparent risk reduction of 42% in mortality in the intensive group versus the conventional treatment group.[6] In another study by the same authors on the effect of intensive insulin therapy on medical patients, there was no difference in mortality between both the groups. However, subgroup analysis showed reduction in mortality in patients who remained in the ICU for more than 3 days.[7] However, in the NICE- SUGAR study, 27.5% patients died in the intensive group as compared to 24.9% in the conventional insulin group.[10] Nevertheless, the NICE-SUGAR study has been criticised stating that in this study, different target ranges for blood sugar levels were used in the control and interventional groups, different routes of insulin administration were used, different glucometers were used and there was a difference in nutrition strategies.
One important observation in our study was the presence of higher mean APACHE-II scores in the CIT group as compared to IIT group. The authors of the study on critically ill surgical patients found that the APACHE-II score was falsely lowered at the time of admission as 17% of patients were admitted to ICU after a median delay of 48 h.[6] Although the scores were higher in both the groups in our study, which itself is an indicator of high mortality, we compared patients with comparable APACHE-II scores in both the groups. In patients with APACHE-II scores between 21 and 25, none of the patients died in the IIT group, whereas 38.8% of patients died in the IIT group representing a significant reduction in mortality in the CIT group. Likewise in patients with APACHE-II scores between 26 and 30, no death occurred in the IIT group as compared to 62.5% deaths in the CIT group, again signifying increased mortality in the CIT group.
In our study, the mean number of days spent in the ICU was found to be significantly higher in the CIT group; also, 94% of patients in the IIT group were discharged as they no longer needed ventilatory support, whereas only 44% of patients got discharged in the CIT group. The reduced ventilatory requirement in patients who received intensive insulin therapy can be partly explained by a decreased incidence of critical illness polyneuropathy as insulin deficiency and hyperglycemia both can lead to axonal dysfunction; in fact, hyperglycemia has been identified as an independent risk factor for the development of critical illness polyneuropathy.[11] The mean blood glucose levels at the time of admission were higher in the CIT group in our study population and all the patients needed exogenous insulin. In a study on surgical patients, hyperglycemia in the two groups was comparable at the time of admission, with all patients in the IIT group needing insulin, whereas only 39% of patients in the CIT group needed insulin.[6] However, other studies found hyperglycemia at the time of admission to be a risk factor for increased mortality.[12,13,14] Some researchers have said that in intensive care medicine, the estimation of HbA1c at admission to the ICU allows patients with stress hyperglycemia to be discriminated from those with DM and hyperglycemia; also, HbA1c values >6.5 in those with no history of DM were associated with greater mortality in the ICU.[15] We also recorded the HbA1c values immediately on patient admission to the ICU in our study as one of the baseline demographic data and found that this was comparable between the two groups.
Target blood glucose levels were achieved within first 12 h in both the groups with longer time being needed in the IIT group in our study. The mean insulin requirement was found to be similar in both the groups during the first 6 h, showing that the control of blood glucose levels rather than exogenous insulin administration accounts for mortality benefit associated with intensive insulin therapy.
Although the incidence of hypoglycemia was higher in the IIT group as compared to conventional group in our study, the difference failed to achieve any statistical significance. Any adverse events such as hemodynamic deterioration and neurological damage were averted because of quick detection and correction. This probably could happen in our ICU because of a good staff–patient ratio and vigilant intensivists. Reported incidence of hypoglycemia in critically ill surgical patients showed 5% hypoglycemia in the intensive group and 1% in the CIT group, whereas in medical patients, the incidence of hypoglycemia was reported to be 19% in the IIT group versus 3% in the CIT group.[6] In a study on critically ill neurological patients in the ICU, hypoglycemia (<60 mg.dL−1) was more common in the IIT group and the proportion of deaths was higher in the IIT group.[16]
The higher infection rate in the CIT group in our study could possibly be because of the deleterious effects of hyperglycemia on macrophage or neutrophil function or insulin-induced trophic effects on mucosal and skin barriers.[17] The results of our study were comparable to the first study on this topic, wherein the researchers found a 46% reduction in the risk of infections. Some researchers have found a significantly greater incidence of positive blood cultures in patients with poor glucose control (>140 mg.dL−1) as compared to patients with adequate glucose control.[18]
The proportion of patients requiring inotropic support was found to be significantly higher in our CIT group with 24 (48%) patients needing inotropic support in the CIT group versus 13 (26%) patients in the IIT group; this seems appropriate because the number of patients acquiring infections was also more in the CIT group in our study. In this aspect, our findings differed from those of others; nevertheless, other researchers found the proportion of patients needing inotropic or vasopressor support to be the same in both the groups.[6]
The reduced number of transfusions in the IIT group in our study emphasizes the effect of good glycemic control on improving erythropoiesis and decreasing hemolysis. However, in a study, the number of patients who received red-cell transfusions did not differ significantly between the two groups.[6]
Researchers have said that critically ill patients without DM or those diabetics with adequate prior metabolic control may benefit from tighter control with lower target glycemic levels;[19] nevertheless, our study included more of diabetic patients in the control group compared to the IIT group.
Our study had some limitations. It was not possible to blind the nurses to insulin dose adjustments. Furthermore, we included a mixed category of critically ill patients. The patients differed in the nature and severity of disease and thereby in their insulin and glucose needs. The standards of care vary in different ICUs and they were definitely good in our ICU with a good staff–patient ratio; hence, our results might not apply to less staffed ICUs; nevertheless, we suggest future research on this very debatable topic in subgroups of critically ill patients.
CONCLUSIONS
We conclude from our study that tight glycemic control significantly lowers mortality and morbidity in critically ill patients, both surgical and medical. These benefits appear with the maintenance of tight blood glucose control of 80–110 mg.dL−1 and not due to administration of insulin.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
Acknowledgment
We acknowledge Dr. Valsamma Abraham and Dr. Mary John for their support.
REFERENCES
- 1.Bajwa SJ, Baruah MP, Kalra S, Kapoor MC. Interdisciplinary position statement on management of hyperglycemia in peri-operative and intensive care. J Anaesthesiol Clin Pharmacol. 2015;31:155–64. doi: 10.4103/0970-9185.155141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Capes SE, Hunt D, Malmberg K, Gerstein HC. Stress hyperglycaemia and increased risk of death after myocardial infarction in patients with and without diabetes: A systematic overview. Lancet. 2000;355:773–8. doi: 10.1016/S0140-6736(99)08415-9. [DOI] [PubMed] [Google Scholar]
- 3.Smith FG, Sheehy AM, Vincent JL, Coursin DB. Critical illness-induced dysglycaemia: Diabetes and beyond. Crit Care. 2010;14:327. doi: 10.1186/cc9266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rapsang AG, Devajit CS. Blood sugar control in the intensive care unit: Time to relook. South Afr J Anaesthesia Analgesia. 2014;20:185–9. [Google Scholar]
- 5.Viana MV, Moraes RB, Fabbrin AR, Santos MF, Gerchman F. Assessment and treatment of hyperglycemia in critically ill patients. Rev Bras Ter Intensiva. 2014;26:71–6. doi: 10.5935/0103-507X.20140011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, et al. Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345:1359–67. doi: 10.1056/NEJMoa011300. [DOI] [PubMed] [Google Scholar]
- 7.van den Berghe G, Wilmer A, Hermans G, Meersseman W, Wouters PJ, Milants I, et al. Intensive insulin therapy in the medical ICU. N Engl J Med. 2006;354:449–61. doi: 10.1056/NEJMoa052521. [DOI] [PubMed] [Google Scholar]
- 8.Wiener RS, Wiener DC, Larson RJ. Benefits and risks of tight glucose control in critically ill adults: A meta-analysis. JAMA. 2008;300:933–44. doi: 10.1001/jama.300.8.933. [DOI] [PubMed] [Google Scholar]
- 9.van Steen SC, Rijkenberg S, van der Voort PH, deVries JH. The association of intravenous insulin and glucose infusion with intensive care unit and hospital mortality: A retrospective study. Ann Intensive Care. 2019;9:29. doi: 10.1186/s13613-019-0507-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Finfer S, Chittock DR, Su SY, Blir D, Foster D, Dhingra V, et al. NICE-SUGAR study investigators. Intensive verus conventional glucose control in critically ill patients. N Engl J Med. 2009;360:1283–97. doi: 10.1056/NEJMoa0810625. [DOI] [PubMed] [Google Scholar]
- 11.Zhou C, Wu L, Ni F, Ji W, Wu J, Zhang H. Critical illness polyneuropathy and myopathy: A systematic review. Neural Regen Res. 2014;9:101–10. doi: 10.4103/1673-5374.125337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Woo E, Chan YW, Yu YL, Huang CY. Admission glucose level in relation to mortality and morbidity outcome in 252 stroke patients. Stroke. 1988;19:185–91. doi: 10.1161/01.str.19.2.185. [DOI] [PubMed] [Google Scholar]
- 13.Rovalis A, Kotsou S. The influence of hyperglycemia on neurologic outcome in patients with severe head injury. Neurosurgery. 2000;46:335–43. doi: 10.1097/00006123-200002000-00015. [DOI] [PubMed] [Google Scholar]
- 14.Zindrou D, Taylor KM, Bagger JP. Admission plasma glucose: An independent risk factor in nondiabetic women after coronary artery bypass grafting. Diabetes Care. 2001;24:1634–9. doi: 10.2337/diacare.24.9.1634. [DOI] [PubMed] [Google Scholar]
- 15.Kompoti M, Michalia M, Salma V, Diogou E, Lakoumenta A, Clouva-Molyvdas PM. Glycated hemoglobin at admission in the intensive care unit: Clinical implications and prognostic relevance. J Crit Care. 2015;30:150–5. doi: 10.1016/j.jcrc.2014.08.014. [DOI] [PubMed] [Google Scholar]
- 16.Green DM, O'Phelan KH, Bassin SL, Chang CW, Stern TS, Asai SM. Intensive versus conventional insulin therapy in critically ill neurologic patients. Neurocrit Care. 2010;13:299–306. doi: 10.1007/s12028-010-9417-3. [DOI] [PubMed] [Google Scholar]
- 17.Rayfield EJ, Ault MJ, Keusch GT, Brothers MJ, Nechemias C, Smith H. Infection and diabetes: The case for glucose control. Am J Med. 1982;72:439–50. doi: 10.1016/0002-9343(82)90511-3. [DOI] [PubMed] [Google Scholar]
- 18.Gore DC, Chinkes D, Heggers J, Herndon DN, Wolf SE, Desai M. Association of hyperglycemia with increased mortality after severe burn injury. J Trauma. 2001;51:540–4. doi: 10.1097/00005373-200109000-00021. [DOI] [PubMed] [Google Scholar]
- 19.Aramondi I, Burghi G, Manzavares W. Dysglycaemia in the critically ill patient: Current evidence and future perspectives. Rev Brasileira de Terapia Intensiva. 2017;29:364–72. doi: 10.5935/0103-507X.20170054. [DOI] [PMC free article] [PubMed] [Google Scholar]