Introduction

Survivors of critical illness may suffer from prolonged disability [1]. Several causes contribute to this so-called “post-intensive care syndrome”, which is defined as new or worsening impairment in physical, cognitive, or mental health status arising after critical illness and persisting beyond acute care hospitalization [2]. Critical illness polyneuropathy affects sensory and motor fibers innervating limb and respiratory muscles [3] and is a prominent cause of persisting physical disability [4], leading to difficulty in weaning the patient from mechanical ventilation and flaccid tetraparesis [5]. Sensory nerve impairment is usually reported after intensive care unit (ICU) discharge [5,6]. Stocking and glove sensory loss, feeling of cool extremities, numbness, tingling, and neuropathic pain symptoms including hyperalgesia and burning have been described at long-term follow-up of ICU survivors [5,6]. In some patients, pain can be extremely intense, precluding effective rehabilitation [7]. In a recent series of 16 patients with acute respiratory distress syndrome evaluated at 6 to 24 months after ICU discharge, 6 patients reported variable combinations of pain, numbness, tingling, diffuse weakness, cool extremities, and impaired exercise ability [6] in the absence of abnormalities at nerve conduction study (NCS). The lack of correlation between clinical and electrophysiological findings suggested a small fiber neuropathy [6]; however, skin biopsy, which is required to document the degeneration of intraepidermal nerve fibers (IENF) and to define the diagnosis [8], was not performed and therefore the hypothesis could not be confirmed.

Degeneration of IENF has been described in small fiber neuropathy (SFN) of different etiology [9,10] and in other neuropathies dominated by neuropathic pain, including those with predominantly large fiber involvement such as Guillain-Barré syndrome [11-14] and chronic demyelinating polyradiculoneuropathy [15]. Most recently, small fiber dysfunction has been also described in fibromyalgia [16]. The diagnosis of SFN is confirmed by the quantification of IENF density at the distal site of the leg, based on age and gender-adjusted normative reference values [17]. When a more diffuse and non-length depended process is suspected, such as in sensory neuronopathies [18], or in acute or chronic demyelinating neuropathies [15,19], the IENF density should be examined also at a proximal site of the body, such as the thigh or the trunk.

We hypothesized that critically ill patients may suffer from small nerve fiber pathology as part of the generalized inflammatory process leading to multiple organ dysfunctions and failures. To address this hypothesis, we prospectively monitorized organ function and performed skin biopsy during the ICU stay, and evaluated the survived patients at follow-up with focused neurological examination. Our results provide the first-ever evidence that critical illness causes also small nerve fiber degeneration since the early phases of the disease.

Materials and Methods

Critically ill neurologic patients admitted to the Neurocritical Care Unit of the University of Brescia at the Spedali Civili of Brescia, Italy, from September 1, 2011 to May 1, 2012 were consecutively recruited. Inclusion criteria were age ≥18 years, ICU stay ≥ 3 days, and need of either invasive or non-invasive ventilatory support. Criteria for exclusion were ascertained after discussion with patient’s family physician and family members, and after careful investigation of the clinical charts of previous hospital admissions. Exclusion criteria included the following: any previously known neuromuscular disease or condition at risk for peripheral neuropathy, such as diabetes, systemic diseases with renal, hepatic or thyroid dysfunction, alcohol abuse, HIV or hepatitis C infection, hematological diseases or monoclonal gammopathy of uncertain significance, vitamin B₁₂ deficiency, connective tissue diseases (e.g. Sjögren syndrome), malignancies, previous or current treatment with neurotoxic drugs (e.g. antineoplastic or anti-arrhythmic drugs). We also excluded patients with chronic pain, those in terminal conditions and those in whom skin biopsy could not be taken or electrophysiological investigations of peripheral nerves and muscles were not feasible (e.g. coagulation disorders, plasters, limb amputation, limb edema, etc.) [20].

Results

During the study period, 246 patients were excluded because the ICU stay was less than 3 days (158 patients), or patients had diabetes (41 patients), malignancy (19 patients), alcohol abuse (7 patients), renal insufficiency (4 patients), liver cirrhosis (4 patients), pre-existing neuromuscular disorders (3 patients), hepatitis C infection (2 patients), patients were in terminal condition (7 patients) or skin biopsy was not feasible (1 patients). Fourteen consecutive patients (5.7%) were eventually recruited.

Demographic characteristics, admission diagnoses, NCS, EMG, and skin biopsy findings are detailed in Table 1. Median duration of mechanical ventilation was 12 days (IQR 8-15). Median ICU stay was 17 days (IQR 11-18). Thirteen patients developed sepsis during the ICU stay, predominantly related to respiratory, urinary, and bloodstream infections. Only one did not have proven infection despite persistent fever.

NCS study and EMG were performed on median ICU day 8 (IQR 7-14 days) and revealed CIM in 8 patients, CIP in 4 patients, and normal findings in 2 patients.

Skin biopsies were obtained on median ICU day 22 (IQR 18-24 days). All patients showed a severe and non-length dependent degeneration of IENF with densities below normal values at both distal leg and proximal thigh (table 1). Epidermal and dermal nerves showed diffuse degenerative changes with fragmented and weak staining. Sweat glands were severely denervated in all except one patient with CIM (Figure 1). There were no complications associated with skin biopsies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Skin biopsy.

Bright field immunohistochemical studies (polyclonal anti-protein gene product 9.5 antibody; Ultraclone, Wellow, Isle of Wight, UK) in sections of skin biopsy taken at the proximal thigh (A) and distal site of the leg (B) with a sweat gland (C) in patient no. 7, and at the proximal thigh (D) and distal site of the leg (E) with a sweat gland (F) in a healthy subject.

Arrows indicate intra-epidermal nerve fibers and arrowheads indicate dermal nerve bundles. The patient, with normal nerve conduction study findings and myopathy at needle electromyography, showed a severe depletion of intra-epidermal nerve fibers and a severe reduction in the density of dermal nerve bundles, whose staining was fragmented and weaker due to axonal degeneration. Sweat gland innervation was also markedly reduced. Original magnification 40x. Bar is 50 µm in all images.

https://doi.org/10.1371/journal.pone.0075696.g001

Two patients died after ICU discharge. At follow-up, 2 patients regained full functional independence, 1 patient had mild disability, 4 patients had moderate-to-severe disability, and 5 patients had severe disability (table 2). Seven patients (no. 2, 5, 7, 8, 11, 12, 14) were available for follow-up examination and three of them complained of sensory loss and burning pain (no. 2, 8 and 12). Pain was diffuse, and in one patient (no. 2) it was so severe to interfere with effective physical rehabilitation. Intriguingly, all these 3 patients had acute myopathy with no evidence of large fiber sensory neuropathy at NCS. Their sensory symptoms likely reflected a small fiber neuropathy as diagnosed by skin biopsy.

Clinical dysautonomia was observed in 3 patients, of whom 1 (no. 9) also had abnormal finger wrinkling test (table 2). The only patient with normal density of sweat gland nerve fibers at skin biopsy (no. 6) did not show any sign of systemic, cardiovascular, and peripheral dysautonomia at clinical evaluation and finger wrinkling test.

Discussion

We report the first-ever observation that critical illness can be associated with a non-length dependent degeneration of somatic and autonomic small nerve fibers. This complication developed in critically ill patients with severe infection, sepsis or MOF, and prolonged mechanical ventilation and ICU stay. In association with the degeneration of IENF, patients showed also electrophysiological evidence of acute axonopathy of large nerve fibers and acute myopathy, confirming that neuromuscular alterations of critical illness complicate the clinical course of the most severely ill patients and are not isolated events, but rather are an integral part of the process leading to multiple organ dysfunctions and failures [3].

IENF degeneration has been described in several chronic conditions, such as diabetic neuropathy [27,28], but also in acute inflammatory neuropathies, such as Guillain-Barré syndrome [11-14]. Patients with SFN typically present with neuropathic pain and reduced thermal and pain sensation but with normal strength, proprioceptive sensation, and deep tendon reflexes reflecting normal large nerve fiber functions [9,10,29,30]. At skin biopsy, reduced IENF density is typically demonstrated [30]. In our series, we found a non-length dependent degeneration of somatic IENF in all the patients, suggesting that critical illness likely affects dorsal root ganglion (DRG) neurons. Patients available for neurological examination at follow-up complained of diffuse pain and thermal and nociceptive sensory loss, a pattern typical of sensory neuronopathies [18]. DRG neurons are surrounded by fenestrated capillaries instead of a tight barrier and are therefore exposed to circulating neurotoxic factors from the bloodstream. Circulating mediators such as endotoxin, other proinflammatory agents (tumor necrosis factor-a, serotonin, and histamine) or neurotoxic factors are probably released during sepsis or systemic inflammatory response syndrome, and have been proposed in the pathogenesis of CIP, but their role remains controversial because CIP can occur in the absence of an identifiable circulating mediator [31]. Future studies should investigate these candidate neurotoxic factors for their effects on the tiny nerve fibers of the skin.

Our analyses of skin biopsies were aimed at describing morphometric and morphologic features of epidermal and dermal nerves, and did not include immunohistochemical assays of cytokines or adhesion molecule expression, which might provide clue to the underlying pathogenetic mechanisms of IENF degeneration. Increased production of pro-inflammatory cytokines has a key role in promoting the organ dysfunction during the early phase of critical illness [32], and may also increase the microvascular nerve permeability leading to formation of endoneurial and axon disruption [3,33]. Since all our patients developed infections and sepsis during the ICU stay, we hypothesize that the degeneration of IENF was caused by the same mechanisms leading to axonal degeneration of large nerve fibers and multiple organ dysfunction. This hypothesis is further supported by the strict inclusion criteria adopted, which allowed us to exclude patients at risk for neuropathy before ICU admission. In addition, we performed all skin biopsies during the ICU stay over a relatively short period of time in order to detect potential risk factors with high accuracy. Future studies should evaluate if skin accumulation of inflammatory cytokines or increased expression of adhesion molecules might play a role in IENF degeneration during critical illness.

Neuropathic pain, defined as pain caused by a disease or a lesion of the somatosensory nervous system [34], was observed in 3 patients who showed thermal and nociceptive sensory loss and burning pain associated with IENF loss. This picture was likely caused by the exclusive degeneration of small nerve fibers, because NCS were normal, therefore excluding CIP or other large fiber neuropathies. Besides degeneration of the somatic IENF, all except one patient showed a profound denervation of skin sweat glands indicating autonomic peripheral involvement. Autonomic dysfunctions were evident in 3 patients, one of whom also had abnormal finger wrinkling test. However, the impact of these abnormalities on patients’ outcome could not be assessed due to severe neurologic conditions. This result confirms previous report of autonomic dysfunction with abnormal cardiac R-R variation and sympathetic skin response in ICU patients with critical illness polyneuropathy or septic encephalopathy [35].

One major limitation of our study is the small sample of patients with no control group and the fact that accurate neurological evaluation of the sensory system at follow-up was possible only in a subset of patients. Future studies should involve larger samples of non-neurologic critical care patients. In future follow-up studies, serial skin biopsies could also be offered to patients to evaluate the natural progression of the disease and the effect of potential treatments [36]. Finally, inclusion of patients without sepsis and MOF would help to confirm if IENF degeneration is part of the process leading to MOF.

Our study could contribute in addressing some issues regarding the long-term outcome of ICU survivors. In particular, chronic pain is common and probably involves multiple pathogenetic mechanisms [6,37,38]. Our findings suggest that the degeneration of somatic and autonomic small nerve fibers is a frequent complication of critical illness and explained the persisting sensory deficits and neuropathic pain observed in some survivors. In such cases, skin biopsy is a safe, almost painless, and cheap technique that can be offered to patients for evaluating small nerve fibers.