Introduction

Intensive Care Unit-Acquired Weakness (ICU–AW), one of the most troublesome complications of critical illness, is caused by critical illness polyneuropathy (CIP) or critical illness myopathy (CIM). A mixed form where both muscle and nerve are affected is known as critical illness neuromyopathy (CINM). CINM is thought to be the most frequent cause for ICU-AW [1]. The clinical syndrome is characterized by diffuse symmetrical weakness and failure to wean from the ventilator. The dependency on mechanical ventilation is one of the most important short-term complications of ICU-AW, while ICU-AW is also associated with increased mortality and long-term disability [1]. Prevention or treatment strategies are urgently needed.

Strict glycemic control (SGC) has been suggested as an effective intervention for the prevention of CINM. A Cochrane review reported a relative risk (RR) of 0.60 (95% confidence interval 0.49–0.74) [2]. This review also reported that SGC reduced the duration of mechanical ventilation, possibly due to a reduction of CINM. Studies on SGC which investigated the duration of ventilation as a secondary outcome measure, but did not focus on CINM, were not included in this Cochrane review. Besides this Cochrane review, results from pre-clinical and other clinical studies show the detrimental effect of hyperglycemia on muscle and nerve function, providing a sound theoretical framework for this intervention [3].

Diagnosing ICU-AW, CIP, CIM, and CINM in the ICU can be difficult, and different diagnostic strategies and criteria have been used in studies so far. Uniform diagnostic criteria for these conditions have been proposed recently [1]. The diagnostic uncertainty may influence the investigation of treatment effect of therapeutic studies.

The aim of this report is to evaluate the effect of SGC on CINM and the duration of mechanical ventilation, as a surrogate marker for CINM.

Methods

Data Sources

Critical illness neuromyopathy was used where CIP, CIM, and/or CINM were used in the original studies. A systematic search in Medline was conducted to identify studies investigating the effect of SGC on CINM. Searches were conducted in August 2010.

Study Selection

The SGC for the purpose of this review was regarded as any protocol implemented to regulate glucose at lower levels using any insulin type to a predefined level in patients and to compare the study protocol with a conventional protocol using higher glucose levels.

The following query was used: “(strict glycemic control OR intensive insulin treatment) AND critical illness AND (polyneuropathy OR myopathy)” using the following limits: English language. Publications were screened on title and abstract for inclusion using the following criteria: original data, ICU setting, diagnosis of CINM using electrophysiological studies (EMG) or diagnosis of ICU-AW using clinical examination, implementation of SGC protocol.

To identify studies investigating the effect on duration of mechanical ventilation, the query was: “(strict glycemic control OR intensive insulin treatment) AND (intensive care unit OR critical care unit OR critical illness)” using the following limits: randomized controlled trial (RCT), adults (>19 years), humans and English language. Subsequently, all articles identified were screened for eligibility by title, abstract, or full article. Inclusion criteria were: original data, ICU setting, duration of mechanical ventilation reported and implementation of SGC protocol.

Data Extraction and Statistical Analyses

Data on the diagnostic methods and criteria used for CINM, on the effect of SGC on CINM and on the duration of ventilation were extracted. Differences in the number of EMGs preformed were analyzed using the χ 2 test. Study results were scored as having a beneficial effect, no effect, or a negative effect on CINM. RCT quality was assessed with the Jadad five point scale [4]. Due to the different study methodologies and reporting of the duration of ventilation, no formal meta-analysis was preformed and “positive, negative, or no effect” conclusion was made. Sub studies concerning CINM were consulted for additional data [5, 6].

Results

Figure 1 describes the study flow chart. After screening, three original publications on SGC on CINM and two sub studies, describing the same population from two of the original publications in more detail, were available [5–9]. The search for studies on duration of mechanical ventilation identified one study which measured muscle weakness as a secondary outcome, making a total of four studies available for analysis on the effect of SGC on CINM [10]. Eleven original RCT’s on SGC which presented data on the duration of mechanical ventilation were included [7, 8, 10–18].

Fig. 1
figure 1

Study selection flow diagram. Legend: SGC strict glucose control, CINM critical illness polyneuropathy or myopathy, MV mechanical ventilation, RCT randomized controlled trial, ICU intensive care unit

Full size image

Effect of SGC on CINM

Table 1 describes the study characteristics for the effect of SGC on CINM. Implementation of SGC significantly decreased the percentage of CINM in critically ill patients in three studies (see Table 2) [5–9]. The diagnostic strategy in these studies exclusively used needle EMG in multiple proximal and distal muscles to diagnose CINM; no clinical assessment of strength was gathered or reported. Pathological spontaneous activity (PSA) was considered diagnostic for CINM [5, 6, 9]. In the RCTs, EMGs were preformed in all study patients regardless of clinical assessment, and a significant decrease of PSA was seen in the SGC group [5, 6]. In the retrospective study, which compared the incidence of ICU-AW before and after implementation of the SGC protocol, the results from EMGs ordered by treating physicians because of weakness or difficulty weaning from the ventilator were used [9]. Although a decrease in PSA was found in the SGC group, the total number of EMGs ordered did not decrease after implementation of SGC (P = 0.66 and P = 0.74 for the medical and surgical intensive care unit respectively; see Table 2). The only study using a clinical assessment used the muscular disability rating score (MRDS; ranging from 1: no deficit to 5: severe proximal deficit) to investigate the effect of SGC on ICU-AW [10]. No differences were found between the control and SGC group at 28 days after admission (see Table 2). No differences between surgical and medical populations could be detected in any of the studies.

Table 1 Overview of studies on the effect of strict glucose control (SGC) on critical illness neuropathy and myopathy (CINM)
Full size table
Table 2 Effect of strict glucose control (SGC) on critical illness neuropathy and myopathy (CINM)
Full size table

Effect of Strict Glycemic Control on the Duration of Ventilation

Table 3 describes the effect of SGC on the duration of mechanical ventilation. Of the studies identified, nine involved large groups (range 483–6104 patients) [7, 8, 10–16]. Of these large trials, six showed no beneficial effect while three showed decreased duration of mechanical ventilation due to SGC (see Table 3). Multivariate testing in two studies which investigated CINM and the duration of mechanical ventilation showed that the reduction in duration of mechanical ventilation was only partially explained by a reduction in CINM [5, 6]. All studies showed good quality (Jadad score ≥ 3) besides one RCT [18]. Considerable differences existed in study methodology, regarding inclusion and exclusion criteria, and in glucose protocols and achievement of study endpoints, such as glycemic control (see Table 3).

Table 3 The effect of strict glycemic control (SGC) on duration of mechanical ventilation
Full size table

Discussion

The SGC has a beneficial effect on CINM in three out of four studies in this review, but an inconsistency was found. The observation that the number of EMGs ordered because of weakness or weaning failure did not decrease after SGC implementation in the retrospective study indicates that the reported positive effects might be an overestimation due to the diagnostic methods used. In two RCTs, no data on weakness were reported, only the finding of PSA. The relation between the results of needle EMG and clinical assessment has been studied [19]. The presence of PSA on needle EMG was associated with a sensitivity of 48% and specificity of 93% using the Medical Research Council (MRC) measurement of strength as the reference standard. Though limitations exist in extrapolating diagnostic test characteristics between populations, this also shows that the absence of PSA may not accurately identify patients without weakness. Because no information on weakness was reported in either of the studies, this question cannot be answered reliably. The only study that reported clinical information found no effect of SGC on weakness. This study is limited by two important factors; first, the clinical observation used provides only a general clinical picture relying heavily on proximal weakness. Second, the populations differed between studies, the study by Annane et al. [10], for example, did not exclude patients with previous neuromuscular disorders making it more difficult to ascertain the effect of SGC on CINM.

Assessment of weakness, polyneuropathy, or myopathy in the intensive care setting is difficult. Incomplete understanding of pathophysiology, disorders of consciousness limiting clinical assessment, the absence of a “gold standard,” and technical limitations in performing EMGs have led studies investigating ICU-AW and CINM to adopt a wide variety of different diagnostic criteria and strategies. The recently proposed framework uses the MRC scale to assess weakness and EMGs, consisting of nerve conduction studies, direct muscle stimulation, needle myography, and muscle biopsy to localize the cause of the weakness and exclude other disorders causing acute weakness in the intensive care [1]. The MRC lacks sensitivity for detecting all patients with CINM and fails to distinguish between myopathy and polyneuropathy [1, 20]. On the other hand, MRC examination has been confidently associated with clinically important outcomes as increased mortality and morbidity while EMG studies have not [20–22]. So the combination of clinical and EMG results is important to ascertain that both clinically relevant and pathophysiologically correct diagnostic information are gathered.

Studies investigating the effect of SGC on the duration of ventilation showed no clear beneficial effect although, due to differences between studies, a definitive answer can not be provided. Limitations because of differences in study protocols, changes in standard of care, effectuation, and realisation of the SGC protocol are important reasons for different study results [23]. Differences in populations under study can also be an important explanation for the lack of consistency in findings between studies. Because studies reported heterogeneously on known risks factors for prolonged duration of mechanical ventilation such as, for example, pulmonary or cardiac diseases, the impact of these differences can not be assessed. The observation that, when multivariate testing is used, other variables besides the reduction in CINM explain the reduced duration of ventilation, illustrate that the duration of mechanical ventilation may not be a reliable surrogate marker for CINM.

There are some limitations concerning this report. Because only two RCT’s investigating the duration of mechanical ventilation also investigated CINM, the accuracy of duration of ventilation as a surrogate marker for CINM remains unknown for the other studies. Second, a true meta-analysis was not conducted to assess the effect of SGC on the duration of ventilation so the exact attribution cannot be displayed. Differences among studies, implementation, and achieving SGC precluded any sensible pooling of data.

Conclusion

The SGC is likely to be effective for prevention of CINM as shown by clinical trials and ample theoretical reasoning, but the reported beneficial effect might be overestimated due to the diagnostic criteria used [3]. A beneficial effect of SGC on the duration of mechanical ventilation as a surrogate marker for CINM has not been shown, but duration of mechanical ventilation may not be a reliable surrogate marker for CINM when evaluating a treatment effect.

Future studies should incorporate a reliable and accurate diagnostic strategy for diagnosing ICU-AW and CINM to confirm the beneficial effect of SGC. Based on the recently developed framework, we propose that both clinical assessment and results from electrophysiological or pathological studies should be regarded as the “gold standard,” because this combination provides additive information [1].