Published ahead of print on November 30, 2006, doi:10.1164/rccm.200605-665OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200605-665OC
Impact of Intensive Insulin Therapy on Neuromuscular Complications and Ventilator Dependency in the Medical Intensive Care Unit1 Medical Intensive Care Unit, Department of Internal Medicine, 2 Department of Intensive Care Medicine, and 3 Department of Physical Medicine and Rehabilitation, University Hospitals Leuven, Katholieke Universiteit Leuven, Leuven, Belgium Correspondence and requests for reprints should be addressed to Prof. G. Van den Berghe, M.D., Ph.D., Department of Intensive Care Medicine, University Hospitals Leuven, Herestraat 49, B-3000 Leuven, Belgium. E-mail: greta.vandenberghe{at}med.kuleuven.be
Rationale: Critical illness polyneuropathy/myopathy causes limb and respiratory muscle weakness, prolongs mechanical ventilation, and extends hospitalization of intensive care patients. Besides controlling risk factors, no specific prevention or treatment exists. Recently, intensive insulin therapy prevented critical illness polyneuropathy in a surgical intensive care unit. Objectives: To investigate the impact of intensive insulin therapy on polyneuropathy/myopathy and treatment with prolonged mechanical ventilation in medical patients in the intensive care unit for at least 7 days. Methods: This was a prospectively planned subanalysis of a randomized controlled trial evaluating the effect of intensive insulin versus conventional therapy on morbidity and mortality in critically ill medical patients. All patients who were still in intensive care on Day 7 were screened weekly by electroneuromyography. The effect of intensive insulin therapy on critical illness polyneuropathy/myopathy and the relationship with duration of mechanical ventilation were assessed.
Measurements and Main Results: Independent of risk factors, intensive insulin therapy reduced incidence of critical illness polyneuropathy/myopathy (107/212 [50.5%] to 81/208 [38.9%], p = 0.02). Treatment with prolonged ( Conclusion: In a subset of medical patients in the intensive care unit for at least 7 days, enrolled in a randomized controlled trial of intensive insulin therapy, those assigned to intensive insulin therapy had a reduced incidence of critical illness polyneuropathy/myopathy and were treated with prolonged mechanical ventilation less frequently.
Key Words: polyneuropathy • blood glucose • myopathy
Critical illness polyneuropathy (CIP) is a primary axonal motor and sensory neural disease that occurs in critically ill patients, predominantly in those with sepsis and multiple organ failure. The reported incidence of CIP in this group is 70 to 82% (1, 2). CIP complicates patient recovery due to varying degrees of limb weakness (3–5) and may also cause weaning difficulties due to phrenic nerve involvement (1, 4–8). CIP lengthens hospital stay (7), and mortality rates in these patients are increased more than seven times (9). In these patients, a coexistent primary muscle disease is often present, termed "critical illness myopathy" (CIM) (5, 8, 10). Several risk factors for this disease have been identified, such as the use of corticosteroids (11), neuromuscular blocking agents (9, 12–14), vasopressors (15, 16), or aminoglycosides (17–19), organ failure (20), and parenteral nutrition (9), renal replacement therapy (9), and duration of intensive care unit (ICU) stay (15). Until recently, the only possible way to affect the incidence of CIP or CIM was to reduce the risk factors (8). However, strict glycemic control in a surgical ICU substantially reduced the incidence of CIP in the subset of patients who stayed in the ICU for at least 7 days. This was accompanied by a lower incidence of prolonged mechanical ventilation (15). In the current study, we examine the effect of the same therapeutic intervention in a medical ICU population, which, by the nature of the diseases for which ICU admission is required, has a greater exposure to the known risk factors for CIP and CIM as compared with surgical patients. Some of the results in this article have been presented as an abstract (21).
This study is a prospectively planned subanalysis of a randomized controlled trial (n = 1,200) of medical ICU patients allocated to either intensive or conventional insulin therapy (Figure 1). Intensive insulin therapy (IIT) was aimed at blood glucose levels of between 80 and 110 mg/dl. The conventional insulin therapy (CIT) approach required insulin only when blood glucose rose above 215 mg/dl and tapering or stopping insulin when blood glucose fell below 180 mg/dl. Details of the entire patient population have been described elsewhere (22). Details on measurement of blood glucose, ventilation, weaning, and feeding strategies are given in the online supplement.
Here, we assessed the impact of IIT on CIP and/or CIM (CIP/CIM), and duration of mechanical ventilation, and compared the incidence of CIP/CIM and its risk factors between medical and surgical patients in the ICU for at least 7 days.
Diagnosis of CIP/CIM
Statistics Time to weaning from mechanical ventilatory support was evaluated using cumulative hazard estimates and proportional hazard regression analysis, corrected for baseline risk factors and censoring for early deaths. The definition of prolonged mechanical ventilation was 14 days or more, which is the same as in our surgical study (15). Time to the first positive ENMG was also evaluated using cumulative hazard estimates and Cox regression analysis. A univariate analysis of outcome was performed for the development of CIP/CIM and prolonged mechanical ventilation, evaluating the effect of treatment allocation, baseline risk factors, and known risk factors occurring during ICU stay. These risk factors are mentioned in the online supplement.
In addition to these univariate analyses of outcome, the effect of treatment allocation on the occurrence of CIP/CIM and prolonged mechanical ventilation was assessed using multivariate logistic regression analysis, correcting for those risk factors that showed at least a trend toward significance (p
Patient Characteristics A total of 443 of 1,200 patients were in the ICU for at least 7 days (Figure 1). ENMG data were unavailable for 23 of the 443 patients either due to technical problems or to preexisting neuromuscular disease. Of the remaining 420 patients, 212 were in the CIT group, and 208 in the IIT group. Demographic data and baseline and outcome characteristics of these patients are shown in Table 1. The CIT and IIT groups were comparable on ICU admission, apart from a slightly younger age, and a slightly lower baseline creatinine level, in the intensive group. Censoring patients in the ICU for fewer than 7 days did not lead to selection bias because outcome in the censored group was not affected by IIT (Table 1).
Glycemic Control As per protocol, median A.M. blood glucose levels were 159 (interquartile range [IQR], 139–172) mg/dl in the conventional group and 102 (IQR, 97–115) mg/dl in the intensive group (p < 0.0001). These levels were achieved with 22 (5–43) units insulin per day in the conventional group and 71 (52–101) units in the intensive group (p < 0.0001).
Neuromuscular Outcome
Prolonged infusion of neuromuscular blocking agents was an independent risk factor for the occurrence of at least one positive ENMG (p = 0.02; OR, 2.01 [1.10–3.99]). Other independent protective factors were age (p = 0.003; OR, 0.98 [0.96–0.99], per year added) and number of days of treatment with glucocorticoids (p = 0.02; OR, 0.97 [0.94–0.99], per added day). Glucocorticoids were given in 71.2% of the patients in the CIT group and 74.5% of the IIT patients. Methylprednisolone was the most frequently used drug and its indications were immunosuppressive or antiinflammatory therapy. Hydrocortisone was given for presumed adrenal failure, and dexamethasone was given for cerebral edema and was rarely used. When this analysis was done for the CIT- and IIT-treated patients separately, the reduced risk of CIP/CIM associated with the use of glucocorticoids was only present in the IIT group (p = 0.8; OR, 0.99 [0.96–1.04], per day added, for the conventional group; and p = 0.002; OR, 0.91 [0.86–0.97], per day added, for the IIT group). When entering mean daily blood glucose level and mean daily insulin dose into the model, neither of these could independently explain the prevention of CIP/CIM with IIT (IIT: p = 0.08; OR, 0.50 [0.23–1.07]; mean daily insulin dose: p = 0.8; OR, 1.00 [1.00–1.00]; mean daily glucose level: p = 0.5; OR, 1.00 [0.99–1.00]).
Prolonged Mechanical Ventilation IIT significantly reduced the incidence of treatment with prolonged mechanical ventilation from 99 of 212 patients (46.7%) in the CIT group to 72 of 208 patients (34.6%) in the IIT group (p = 0.01). IIT remained an independent protective factor against treatment with prolonged mechanical ventilation (p = 0.01; OR, 0.56 [0.36–0.87]) in the multivariate logistic regression model. Independent risk factors were as follows: an on-admission diagnostic category of "respiratory" (p = 0.03; OR, 4.17 [1.12–15.50]) as compared with "cardiovascular," number of days of treatment with glucocorticoids (p = 0.04; OR, 1.05 [1.002–1.09]) per day added, number of days of treatment with norepinephrine (p = 0.0008; OR, 1.10 [1.04–1.16]) per day added, and bacteremia (p = 0.001; OR, 3.53 [1.67–7.48]). Finally, exclusively receiving parenteral nutrition appeared to be a protective factor (p = 0.04; OR, 0.51 [0.27–0.97]). Adding the diagnosis of CIP/CIM to the multivariate logistic regression analysis revealed that IIT remained an independent protective factor (p = 0.02; OR, 0.058 [0.37–0.90]), whereas CIP/CIM is also an independent risk factor (p = 0.009; OR, 1.85 [1.17–2.93]). When mean daily insulin dose and mean A.M. glycemia also were added to the multivariate logistic regression analysis, the effect of IIT in reducing the incidence of prolonged mechanical ventilation independently from the effect on CIP/CIM was attributable to the daily insulin dose (IIT: p = 0.5; OR, 0.7 [0.25–2.03]; mean daily glucose: p = 0.6; OR, 0.99 [0.98–1.01]; mean daily insulin dose: p = 0.01; OR, 0.99 [0.99–0.99]).
Comparison between Medical and Surgical Patients in the ICU for at Least 7 Days
CIP/CIM CIP/CIM is an important and frequent complication in critically ill patients and associated with considerable morbidity and increased mortality. In our surgical ICU population, the occurrence of CIP/CIM was significantly reduced by IIT, which was aimed at blood glucose levels between 80 and 110 mg/dl (15). In the current study, we demonstrated that IIT, in a medical population who stayed in the ICU for at least 7 days, significantly reduced the incidence of CIP/CIM. We noticed a comparable incidence of CIP/CIM in the conventional treatment groups of both the surgical and medical long-stay population in the presence of substantially different risk factor pressure. This suggests that these risk factors per se might be less important than the sustained presence of metabolic disturbances over time in the ICU, which can be at least partially prevented by IIT. Indeed, the protective effect of IIT on the occurrence of CIP/CIM in the medical ICU population confirms the beneficial effect, even in a population that is theoretically considered to be at much higher risk. The mechanisms underlying these observations remain unclear. The peripheral effects of chronic hyperglycemia on the nervous system in diabetes are well known (25). As we have hypothesized before, in critically ill patients the peripheral nervous system might be at increased risk for the accelerated development of damage by hyperglycemia due to passive uptake of glucose and increased generation and/or a deficient scavenging system for reactive oxygen species (26). We hypothesized that avoiding hyperglycemia may exert a similar protective effect on neural mitochondrial structure and function (15), as in hepatocytes (27). Maintaining normoglycemia also reduces organ damage by endothelial protection through a diminished release of nitric oxide (28). As we have published before, insulin therapy itself, however, may also contribute to beneficial effects of the intensive therapy regimen due to endothelial protection, reduced inflammation, and improvement of dyslipidemia (26, 28, 29). However, in this study, multivariate logistic regression analysis showed that neither hyperglycemia nor insulin dose was exclusively or independently responsible for the effect of IIT on CIP/CIM. Interestingly, we found that corticosteroids appeared to exert a protective effect on the occurrence of CIP/CIM. This confirms our previous observation in the surgical population (15). Almost three-quarters of our patients received glucocorticoids. This figure may seem high, but the population studied was not an average ICU population but consisted of the sickest patients who were in ICU for at least 7 days. Data in the literature on the impact of corticosteroids on neuromuscular function have been controversial. Different reports have described an increase in CIM (12, 14) and CIP (13) in patients treated with a combination of corticosteroids and neuromuscular blocking agents. Most reports regarding the effect of steroids alone on CIP (9) or CIM (12, 30, 31) and CIP/CIM (20, 32) did not find such a correlation. In two reports, however, corticosteroids were found to be an important risk factor for the development of neuromuscular disorders in patients in the ICU (11, 33). Our results, however, indicated that corticosteroids, which are known to cause hyperglycemia and insulin resistance (34), may not exert deleterious effects on CIP/CIM when blood glucose levels are kept normal with insulin, and might even be protective. Whether antiinflammatory properties of glucocorticoids might also exert a beneficial effect on the neuromuscular system and explain this finding needs further investigation. Our results also confirm the increased risk for CIP/CIM due to prolonged infusion of neuromuscular blocking agents. An unusual finding, however, was the protective effect of aging on the development of CIP/CIM. We do not have an explanation for this.
Prolonged Mechanical Ventilation Interestingly, the number of days of treatment with glucocorticoids clearly increases the risk for prolonged mechanical ventilation. Glucocorticoids are indeed known to have catabolic effects on the skeletal muscle (36). We did notice that the diagnosis of CIP/CIM was made, and therefore risk factors were analyzed, over a median duration of 8 days. Concerning prolonged mechanical ventilation, risk factors were analyzed for all patients until Day 14. Currently, we can only speculate on a possible important time component of the effect of corticosteroids on the neuromuscular system to explain the beneficial effect on CIP/CIM on the one hand, and the detrimental effect on prolonged mechanical ventilation on the other hand. Further studies are needed to confirm these results and to clarify possible underlying mechanisms. Finally, other risk factors identified for prolonged mechanical ventilation were days of treatment with norepinephrine and bacteremia, which are associated with severity of illness. Exclusively receiving parenteral nutrition paradoxically appeared to be protective for prolonged mechanical ventilation. We think this is due to a selection bias, because these patients had a significantly shorter ICU stay.
Limitations Finally, IIT currently has only been proven to have beneficial effects in two large randomized controlled trials at our center. Results of two subsequent studies, which were ended prematurely due to difficulties in safely achieving blood glucose control, have not been published in peer-reviewed journals. Also, IIT can be complicated by hypoglycemia. In our studies, hypoglycemia did not cause early deaths, only minor immediate and transient morbidity in a minority of patients, and no late neurologic sequelae among hospital survivors (40). Nevertheless, prevention and early detection of hypoglycemia need to be a major concern when implementing this therapy. We conclude that IIT is the only therapy that has been proven to actually decrease the incidence of CIP/CIM, and the incidence of prolonged mechanical ventilation, not only in a surgical population but also in a medical ICU population in which the incidence of other known risk factors for CIP/CIM is much higher. Further studies are needed to fully understand the biological mechanisms underlying this finding, and multicenter trials are needed to confirm these results.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200605-665OC on November 30, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form May 17, 2006; accepted in final form November 30, 2006
Related articles in AJRCCM:
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
14 d) mechanical ventilation was reduced from 99 of 212 (46.7%) to 72 of 208 (34.6%) (p = 0.01). This was statistically only partially explained by prevention of critical illness polyneuropathy/myopathy. 
0.1) in the univariate analysis. To examine whether affecting the incidence of CIP/CIM could explain the effect of IIT on prolonged mechanical ventilation, presence or absence of CIP/CIM was entered into the multivariate logistic regression analysis for prolonged mechanical ventilation. Finally, to assess which component of IIT (blood glucose control or insulin dose) best explained its possible impact on CIP/CIM and prolonged mechanical ventilation, mean blood glucose level and insulin dose were added to the models. Parameters were checked for linearity, and nonlinear parameters were entered into the model as nominal variables. 
