New Approaches to Critical Illness Polyneuromyopathy: High-Resolution Neuromuscular Ultrasound Characteristics and Cytokine Profiling

Intensive care unit acquired weakness (ICUAW) is a frequent problem in intensive care medicine and occurs in up to 67% of patients who are mechanically ventilated for more than 10 days [1]. The defining clinical symptom of ICUAW is a generalized symmetric muscle weakness, leading to a flaccid tetraparesis [2]. This results in prolonged ventilator dependency, prolonged hospitalization and rehabilitation [3] as well as to worse clinical outcome [4]. Causes of ICUAW are critical illness polyneuropathy and critical illness myopathy or combination of both (CIPNM).

ICUAW is a clinical diagnosis based on assessment of muscle strength. However, in analgosedated patients with impairment of consciousness, sensorimotor symptoms are often not reliably determinable [5]. Therefore, electrophysiological measurements are used for diagnosis. Reliability of these measurements on intensive care units is limited [6]: Amplitudes may be difficult to interpret due to peripheral edema and electromagnetic artifacts of the intensive care unit devices. Also, electrophysiological diagnosis based on a single measurement is challenging as other polyneuropathies, i.e., preexisting or acute axonal polyradiculoneuropathy (Guillain-Barré syndrome) variants are difficult to differentiate from axonal polyneuropathy in CIPNM. These difficulties often lead to delayed diagnosis of ICUAW and CIPNM [2] and stress the need for novel approaches in diagnosing CIPNM. Some studies used high-resolution neuromuscular ultrasound for diagnosis of CIPNM [7] with focus on muscular ultrasound, showing muscular atrophy and increased muscle echogenicity as the main findings [7‐9]. Only in one study, morphological changes of peripheral nerves in critical illness polyneuropathy were examined using ultrasound. This study did not show a diagnostic benefit of a single ultrasound examination of the median and fibular nerve in diagnosis of critical illness polyneuropathy [10].

The pathogenesis of critical illness polyneuropathy and polymyopathy remains incompletely understood. Metabolic, ischemic and inflammatory factors as well as oxidative stress are considered as major factors leading to axonal damage and demyelination [1, 11]. Microvascular changes increase vascular permeability and enable penetration of neurotoxic inflammatory factors through the blood–nerve barrier in critical illness polyneuropathy [1]. Proinflammatory mediators like tumor necrosis factor alpha, interleukin-1, procalcitonin, e-selectin, syndecan-1 and interleukin-6, interleukin-8, interleukin-10 and fractalkine/CX3CL1 were described to be involved in pathogenesis in the first days of the disease [12, 13]. On the contrary, other studies showed elevated cytokines and complement activation in both patients with ICUAW and without ICUAW as controls, but not in healthy patients, suggesting that cytokine elevation might be present in intensive care unit patients irrespective of the development of ICUAW [12, 14, 15].

This exploratory study investigated whether longitudinal neuromuscular ultrasound examinations and cytokine analyses in correlation to classical clinical and electrophysiological assessment contribute to the understanding of CIPNM. We examine longitudinal cytokine expression in order to discover specific patterns and enhance the understanding of inflammatory mechanisms in CIPNM. Secondly, we analyzed whether high-resolution neuromuscular ultrasound is helpful in diagnosis and monitoring of CIPNM.

In this study, we found that novel approaches like neuromuscular ultrasound and cytokine analyses reveal signs of inflammation in the longitudinal course of electrophysiologically primarily axonal CIPNM.

At the peak of CIPNM, we found signs of monocyte and macrophage activation in cytokine analyses, which have so far rarely been described [15], but are known in autoimmune inflammatory neuropathies [24]. Simultaneously enlargement of the tibial nerve in ultrasound as well as hypoechogenicity of nerves can be considered as signs of an acute inflammation in ultrasound. Hypoechoic nerves in other inflammatory neuropathies, i.e., chronic inflammatory demyelinating polyneuropathy are supposed to correlate to inflammation, in particular infiltration of inflammatory cells, edema and onion bulbs, whereas chronic irreversible axonal damage with scar tissue or fibrosis is regarded to result in hyperechoic alterations [21, 25]. Therefore, our finding of hypoechoic nerves in CIPNM suggests that edema and inflammation of peripheral nerves may occur within the first weeks of CIPNM.

On follow-up day 14, seven days after CIPNM peak, the cytokine measurements showed signs of a systemic immune response and cytokine profiles associated to neovascularization (VEGFd) and regeneration (BDNF). Regeneration was also observed in the CIPNM severity score in those patients who were available for long-term follow-up a few months after intensive care unit treatment.

SOFA score, as well as common laboratory values for infections and inflammation, did not differ significantly between both groups, so systemic infection should not be the cause for differences in cytokine analyses.

Diagnosis and monitoring of ICUAW is challenging, especially in the early days of an intensive care unit, stay as assessment of neuromuscular symptoms is impaired in analgosedated patients and diagnosis relies on the exclusion of other diseases and electrophysiological measurements. Electrophysiology in our study was helpful to diagnose CIPNM as many patients did not reach a sufficient level of consciousness for muscular assessment in our study. Our data show that in patients, in which actually polyneuropathy-causing preexisting diseases were excluded through inclusion and exclusion criteria, a single nerve conduction study detects reduced amplitudes in several patients. We reckon that this does not always reflect a real polyneuropathy or CIPNM. Prevalence of polyneuropathies in the general population between 60 and 70 years of age was 8% previously reported in a Dutch study with approximately 50% of these cases newly diagnosed [26]. However, this prevalence was observed in the general population, not in ICU patients. At ICU, amplitudes of NCS are influenced through conditions like edema which lead to falsely reduced amplitudes in all patients. Therefore, it is most likely that patients did not actually have an unknown preexisting polyneuropathy at baseline but issues like edema caused low amplitudes. Other studies reported reduced amplitudes in nerve conduction studies in ICU patients compared to healthy controls independently from CIPNM [6]. The reference values for amplitudes for ICU patients were reported lower than that for healthy controls probably due to reduction in muscle or nerve excitability or edema. This is also supported by the fact that reduced amplitudes in the first examination were found equally in CIPNM and in non-CIPNM patients and also by the fact that none of the patients without CIPNM but with reduced amplitudes at baseline had spontaneous activity in needle electromyography as a sign of axonal damage. Moreover, we did not exclude patients with reduced amplitudes at baseline as it is still not known how quickly CIPNM develops in ICU patients. The suspected metabolic, ischemic and inflammatory mechanisms can lead to a reduced function of the peripheral nerves within a few hours and days, i.e., in severe sepsis [13]. As a median of 6 days from admission to inclusion in the study passed, we did not want to miss an already developed early CIPNM at the timepoint of the first examination. Therefore, these patients were followed up in their whole course to confirm or rule out further progression over time. Accordingly, certain diagnosis of CIPNM is only possible if the evolution of a polyneuropathy during intensive care unit stay is proven. Therefore, in our definition of CIPNM we used evolution of nerve conduction studies and needle electromyography parameters during intensive care unit stay.

Supporting the reliability of the diagnostic criteria for CIPNM used in this study, the outcome of the CIPNM group regarding ventilation duration was significantly worse than that of the non-CIPNM patients, which is in line with the literature [2, 3, 27]. Also the incidence of CIPNM (69%) in our study group is similar to those reported in literature [2, 3, 27]. The clinical outcome measured by modified Rankin Scale and the portion of deceased patients was not different between CIPNM and non-CIPNM group. Probably this is due to the fact that all patients were severely ill and had a poor outcome, not just those who additionally developed a CIPNM. For future studies, it would be interesting to also examine less severely ill patients. Also, longitudinal needle electromyography studies would be desirable, but could not be performed in this study due to contraindications.

International electrophysiological consensus criteria for diagnosis of CIPNM do not exist. Axonal degeneration in CIPNM can be demonstrated in fibular nerves more than in others¹⁷; therefore, we used nerve conduction studies of fibular nerve for definition and longitudinal assessment. We developed an electrophysiological score that allows detection of worsening in the nerve conduction studies longitudinally.

A major strength of this study is that our study is the first thorough longitudinal study simultaneously investigating pathophysiological characteristics of CIPNM and diagnostic features through the ‘gold standard’ of nerve conduction studies longitudinally. Knowledge about pathophysiology of CIPNM is a prerequisite for the development of new treatment options. Our findings provide first evidence for further research on anti-inflammatory/immunomodulatory treatment for CIPNM. Until now, immunomodulatory therapies were not found to be effective in CIPNM, although studies specifically looking at the CIPNM are rare [28]. In sepsis patients, steroid therapy has been recommended in patients with adrenergic insufficiency, but less due to the immunomodulatory effects [29‐31]. The TNF alpha activation in our CIPNM patients provides the thesis of a potential beneficial effect of corticosteroids in these patients [32, 33]. Also, anti-inflammatory therapies in combination with neuroprotection and reduction of oxidative stress could be promising.

As clinical impact, our results show that CIPNM can be safely diagnosed only by repeated nerve conduction studies during the first week on intensive care unit stay. This is crucial as the early ‘inflammatory’ phase of CIPNM pathophysiology implies that an early diagnosis and treatment will be essential to reduce CIPNM incidence and severity in the future.

However, this study has some important limitations: An important limitation of our study is the small sample size. Particularly, our findings in long-term course are limited due to small sample size and must be interpreted with caution. Overall, ultrasound and cytokine data were conducted and analyzed as exploratory data; therefore, the results need confirmation in further studies. As no other authors described, the ultrasound alterations in CIPNM with a similar study protocol and similar cytokine analyses, we could not perform sample size calculation. Alterations of cross-sectional area were overall small and could possibly also result from technical or examiner-dependent influences. Especially the cross-sectional area of the sural nerve at the peak of CIPNM was reduced statistically significantly in CIPNM patients, but the cross-sectional area of both groups was within normal values and differences were only small. A reduced cross-sectional area in other neuropathies is regarded resulting from axonal damage, but it is unlikely that nerve atrophy develops within 7 days in CIPNM. Instead, in CIPNM inflammation could occur with different celerity in different nerve sections, i.e., sensory nerves like the sural nerve could be affected by inflammation in the early hours/days and other nerves like the tibial nerve later, which is why we would find different results in both nerves and a cross-sectional area in the sural nerve that already normalized again or even reduced. However, as cross-sectional area alterations were small, we cannot substantiate these considerations with our data sufficiently, so that further studies are necessary to assess this more precisely. Currently, nerve ultrasound cannot be recommended as diagnostic tool in clinical practice, confirming Witteveen et al. [10]. We included patients with central nervous system diseases and spastic hemiparesis or inactivity atrophy could influence nerve conduction studies and muscle ultrasound. Also, central paresis due to central nervous system disease influence muscle strength assessment. Furthermore, as median time from intensive care unit admission to baseline examination was 6 days, we were unable to examine the early phase of the CIPNM. For future studies, cytokine profiling would be interesting in this earlier phase in order to find biomarkers, which would distinguish between CIPNM and non-CIPNM patients. Overall, our study is rather a basis for future research and should not be used to conclude specific generalized diagnostic recommendations.

Exploratory neuromuscular ultrasound and cytokine analyses showed signs of inflammation like macrophage and monocyte activation at the peak of CIPNM followed by a systemic immune response occurring parallel to axonal damage. This underlines the role of both axonal damage and inflammation in pathogenesis of CIPNM. As clinical impact, our study shows that repeated nerve conduction studies should be performed to make a reliable diagnosis of CIPNM.