Guillain-Barré Syndrome: Natural history and prognostic factors: a retrospective review of 106 cases

The term Guillain-Barré syndrome (GBS) includes a set of clinical syndromes (GBS) with a common pathophysiological basis; an acute inflammatory polyneuropathy with an autoimmune etiology [1‐3]. Although usually characterized by a progressive flaccid paralysis with areflexia a wide range of motor, sensory and autonomic symptoms could be seen [1‐4]. In general, the diagnosis is based on clinical criteria [4‐7]; nevertheless, the presence of suggestive findings in the complementary test as demyelinising changes in the nerve conduction studies (NCS) or albuminocytological dissociation in the cerebrospinal fluid (CSF), help to confirm the diagnosis [1].

The worldwide incidence of GBS is reported to be 0.6-2.4 cases per 100,000 per year [8‐15]. The classic form, the acute inflammatory demyelinating polyradiculoneuropathy (AIDP), is the most frequent subtype in Europe, which accounts for 90% of GBS cases [2]. Other subtypes like the axonal forms or the Miller-Fisher syndrome (MFS) [16, 17] are less common.

The prognosis is usually good, showing a complete functional recovery or with minimal deficits in the 90% of patients 1 year after the onset of illness [13, 18]. Several factors have been identified as predictors of poor outcome [13, 14, 19‐21]. Death rate is described to be between 1-18% [14, 15]. This study aimed to describe the epidemiological, clinical, laboratory, and electrodiagnostic features, as well as to identify the predictive factors of worse prognosis in the GBS or its subtypes.

A retrospective review of the medical records of patients admitted at La Paz University Hospital (Madrid, Spain) with the diagnosis of GBS between 2000–2010 was made. 106 fulfilled levels 1, 2 or 3 of diagnostic certainty for GBS/MFS described by Sejvar et al. [7]. All demographic, clinical, laboratory and electrophysiological data were recorded, as well as outcome and treatment.

Severity at admission was assessed by the Medical Research Council (MRC) sum score, valuing the strength from 0 to 5 in 4 muscles (proximal and distal) in both upper and lower limbs on both sides, so that the score ranged from 40 (normal) to 0 (quadriplegic) and by the GBS disability score advocated by Hughes et al. [22]. Cranial nerve involvement was considered separately by the affectation of oculomotor, facial and bulbar nerves. Respiratory weakness was assessed first by the value obtained at the peak expiratory flow meter (Peak Flow), as well as the need for mechanical ventilation throughout the evolution. Sensory disturbances, autonomic alteration or pain presence were also assessed.

Serological screening for preceding infections was recorded, including Herpes Simplex virus (HSV), Varicella-Zoster (VZV), cytomegalovirus (CMV), Epstein-Barr virus (EBV), Mycoplasma Pneumoniae, B and C hepatitis virus, Haemophilus Influenzae and, in selected cases, stool and Campylobacter Jejuni determination. The CSF was analyzed for cell count, glucose and protein concentration.

Neurophysiological studies were evaluated in accordance with the criteria of Hadden et al. [23, 24]. As in a retrospective study, not in all cases were the same nerves measured. Also, electromyography studies (EMG) with concentric needle electrodes were made to evaluate the axonal loss (fibrillation, positive sharp wave…).

The evaluation of the functional impact was graded by the GBS disability score [22] during the discharge from the neurology or the rehabilitation department, and at the third, sixth and twelfth month in the outpatient clinic.

The study was approved by the Research Ethics Committee of La Paz Hospital, Madrid, Spain.

All statistical analyses were performed using the SPSS 12 for Windows program (Chicago, IL, E.U.A.). For univariate analysis, Chi-square test for dichotomic variables was used. For continuous variables, t-Student test in parametric variables or U Mann-Withney with the non-parametric ones were used.

There was no difference between genders (ratio male/female 1.07), with a mean onset age of 43.7 ± 23 years (range 0–85). Demographic and clinical data are summarized in Table 1. There was a seasonal rebound in winter, when 41% of patients were diagnosed. Most patients had an infectious antecedent preceding the onset of the weakness, being the most frequent respiratory tract infection (38%); at least 30% did not present a previous infectious disease [25]. The mean of days since the start of the infectious sickness until the polyneuropathy debut was 12 ± 8.3 days (range 1–30).

The present work has the limitations of a retrospective study based on hospital case-mix. The incidence is reported to be 0.6-2.4 cases per 100,000 per year [18, 19, 21]. Changes suffered in the last years in the attendance area of the hospital make incidence calculation complex and inaccurate; however, it appears to be of 1.68-2.46 per 100,000 per year.

There is no difference between gender [1, 2]. The bimodal shape wasn’t present in our study [8‐10], as there is a linear increase in the incidence with age [1, 2, 9, 11, 12, 24]. GBS is considered a sporadic illness, without a seasonal cluster [1, 9]; however, a trend to accrue in winter is shown in our series [11, 12]. The infectious event is described to appear in 40-70% of patients [1‐3, 8‐12]. In our series up to 70% of cases have had one, of which respiratory infection was the most frequent.

As in previous series, weakness and hypo/areflexia were the most frequent symptoms, followed by neuropathic pain and numbness. Hyponatremia, as a symptom of SIADH is not a classical manifestation of GBS; however, there are series in which are described to be present in up to 58% of the cases; in our review, it was found in 7% of our patients.

There isn’t a consensus about the neurophysiological values defining GBS and its variants [2, 4, 5, 7, 23, 24, 26‐29]. It is accepted as demyelination parameters: motor conduction velocity (MCV) decrease, prolongation of motor distal latency, conduction blocks, temporal dispersion and increased F-wave latency [27]. It is reported that the first electromyographic changes are the alteration of F-wave and H-reflex response [2, 22], altered both in our NCS reported as normal, probably due to the earliness of the exploration. In MFS, the CNS are normal in most cases; nonetheless, discrete changes in the sensory conduction or in H-reflex may be present [30‐32], some authors postulate a damage in the afferent proprioceptive system as a pathophysiological basis [32].

As classically described, in our study illness prognosis is favourable; 81% of patients presented absent or minimum neurological deficits one year after the onset. Older age, illness severity in the acute phase, prior gastrointestinal infection and axonal injury in the CNS and mechanical ventilation requirement [15, 20, 33, 34] are among the factors that have been advocated for poor prognosis. Van Koningsveld et al. defined a clinical prognostic scoring system for GBS outcome at 6 months, the “Erasmus GBS Outcome Score” or EGOS [15], it was based on the punctuation on the GBS disability score at 2 weeks from the admission, the history of diarrhoea and the age. Recently, Walgaard et al. have validated a modified EGOS (mEGOS) with the main difference being the use of the MRC sum score at admission and in the 7th day instead the GBS disability score [35], they claimed that the MRC sum score is more accurate, and the possibility of being used at admission could predict the future treatments. However, the mEGOS made on the 7th day after the admission show increased predictive value instead the one made on the first day [35]. However, although useful, the mEGOS passed on the first day of admission showed lower predictive ability than the one performed on the 7th day. In our study we demonstrate that, even in the first day of admission, lower scores on the GBS scale are associated with worse outcome and greater disability at discharge, 3 and 6 months. Respiratory distress is the leading cause of death in the acute phase, 20-30% requiring ventilatory support [20]. Many factors have been proposed as predictors of the future need for respiratory support, like forced vital capacity (FVC) < 60%, bulbar dysfunction, rapid progression of the illness, and difficulty raising the head [19, 20]. Van Doorn et al. propose a regularly monitoring of the respiratory function initially every 2-4 h, and then every 6-12 h [1]. Although FVC is considered to be the gold standard test for detecting impaired ventilation it has some disadvantages, the requirement of portable spirometers in the acute phase due to the instability of the patient, the need for a minimum of preparation and knowledge of the technique by medical personnel and the higher cost. Suárez et al. describe a serie of 79 patients with neuromuscular diseases in which the Peak Flow test proved to be useful in the monitoring of expiratory muscle weakness [35]. In our hospital, patients were monitored by the Peak Flow test each 6 hours in the acute phase, being observed that values below 250 L/min predict the posterior need of respiratory support (p < 0.05), independently of the presence of facial palsy that could hinder the use of the test (p < 0.05), making the Peak Flow a safe, inexpensive, and widely-available test in the monitoring of patients with GBS.

Our series is in concordance with those previously published. The seasonal cluster in winter is worth noting on which there is a great controversy. Regarding the outcome, our series reported a worse prognosis in patients with older age, severe deficits at the beginning, injured cranial nerves, requiring MV, and axonal lesion patterns in the NCS. Finally, project the Peak Flow-test as a useful predictive factor of respiratory failure by its availability, and easy management.