Background

Guillain–Barré syndrome (GBS) is an acute polyradiculoneuropathy, characterized by a rapidly progressive, nearly symmetrical, flaccid weakness of the limbs [1, 2]. The clinical presentation, course, and the clinical recovery and outcome of GBS all are variable. In addition, GBS varies considerably between geographical regions [2,3,4]. In the acute phase, about 1/5 of GBS patients require mechanical ventilation and 3–4% of patients die of complications [1]. Despite therapy with intravenous immunoglobulin (IVIG) or plasma exchange, about 10–20% of patients remain severely disabled [2, 5]. The recovery from GBS takes a long time and is highly variable, therefore information on the nature, and early predictors of patient-perceived disability after GBS could be of help for patients, doctors, and caregivers both in the acute and chronic phase of disease.

Previous studies found a relationship between outcome and recent preceding infections (antecedent events), clinical presentation, electrophysiological subtype and laboratory findings [6,7,8,9]. Yao and colleagues studied the clinical presentation of GBS in four regions of China and found that a higher proportion of the axonal subtype in central and southwest China; progression in the latter region was more served at nadir and patients had the longest hospital stay [10]

Zhai and their colleagues [11] in their retrospective study of 294 patients with GBS found that the AMAN subtype was predominant in northern China (40.1%) and had shorter time to nadir, with prolonged hospitalization, and worse prognosis at discharge than AIDP. A higher GDS score on admission was a strong predictor for poor outcome at discharge and short-term follow-up, independent of treatment type or in-hospital management (Ruiz-Sandoval and their colleagues) [12]. Luijten and their colleagues found that clinical predictors of the need for mechanical ventilation (MV) were a shorter time from onset of weakness until admission, the presence of facial muscle weakness, bulbar palsy and neck muscle weakness [13].

However, most of the previous studies were retrospective, and the findings were derived from relatively small series or selected groups of patients with a limited set of suboptimal outcome measures [6,7,8,9].

The current study was designed to address these problems using a prospective study design to identify factors that influence outcomes in a large group of well-defined GBS patients with a 3-month follow-up period.

Methods

Sixty-two cases of GBS were recruited and diagnosed according to criteria of the National Institute of Neurological Disorders and Stroke (NINDS) (revised form 1990) [14] and the Brighton Collaboration in 2014 [15] during the period from October 20th 2020 to September 20th 2021.

Cases were recruited within the first 2 weeks of onset, with progressive bilateral weakness of upper and lower limbs, absent or decreased tendon reflexes in affected limbs (at some point in clinical course) as well as other features that support a diagnosis of GBS with the same inclusion and exclusion NINDS criteria.

At admission each patient underwent the following procedures.

History included age, sex, onset, time elapsed between onset of symptoms and admission, presence of other co-morbidites and preceding antecedent events including upper respiratory tract infection, flu-like symptoms or gastrointestinal tract infection (diarrhea or vomiting) and vaccination.

Clinical examination, vital signs including blood pressure, heart and respiratory rates, as well as arterial blood gases (ABG). Neurological assessment included cranial nerve examination and neck muscle weakness.

Autonomic dysfunction included variation in arterial blood pressure, arrhythmia, attacks of diarrhea, vomiting, and sweating. Hughes Disability Scale (HDS) [5, 16] as this scale has six levels ranges from 0 point (healthy), up to 6 points (dead) depending on the disability of lower limbs (capable of running, walking without assistant, walk with assistance, bedridden, need ventilation, and dead) [5, 16].

Medial Research Council sum score (MRC) [17,18,19] as this scale ranges from zero power up to 60 full powers and is scored as the sum of six muscle power in both upper and lower limbs in points. This scale was assessed at onset, at nadir and 10 days of treatment [17,18,19].

The Erasmus GBS Respiratory Insufficiency Score (EGRIS) is a model developed to estimate the risk of respiratory failure. At admission and according to EGRIS, three clinical factors are explored, the time from onset of weakness to admission, presence of facial and/or bulbar weakness, and severity of muscle weakness [7].

Overall Neuropathy Limitation Scale (ONLS), this scale is used to assess the limitations of patients with immune-mediated peripheral neuropathies; it is reliable, responsive and has construct validity in people with GBS, chronic inflammatory demyelinating polyneuropathy and paraprotein-associated demyelinating neuropathy. [20]

Laboratory investigations at admission, complete blood count (CBC), blood urea, creatinine, serum electrolytes (Na, k, Ca), total protein, albumin, and C-reactive protein (CRP). Arterial blood gases (ABG) included PaO2, PaCo2, pH, and SPo2 were also assessed.

Cerebrospinal fluid analysis (CSF) as 42 patients underwent analysis during the first 3 days of admission.

Treatment options as all patients received treatment immediately after diagnosis either by plasmapheresis (plasma Exchange 5 sessions one every other day) or intravenous immunoglobulin (IVIG) 0.4 g/kg/day for 5 consecutive days.

Patients were assessed on admission, at nadir and 10 days after the end of plasma exchange and IVIG, using the MRC sum score and EGRIS. HDS and ONLS GBS scales were assessed 3 months after the onset of illness. All patients received a conventional neurophysiological assessment (distal latency, nerve conduction study and F-wave of four limbs) within 1–3 days of admission in order to confirm the diagnosis.

Statistical analysis used SPSS (Statistical Package for the Social Science, version 17, IBM, and Armonk, New York). Because the data distributions did not differ statistically from normality (using Shapiro–Wilk test, P > 0.05), continuous variables are given as means ± SD; categorical variables are summarized as counts (percentage). Spearman's correlations were calculated between HDS, and ONLS scales outcome (at 3 months follow-up) and clinical or laboratory variables at baseline. Multiple ordinal and linear regression analysis was used to identify best predictors of outcome graded according to HDS score and ONLS score, respectively.

Results

Sixty-two cases of GBS were recruited with a mean age of 37.15 ± 17.60 years (33/29 male/female ratio). 43 (69.4%) patients recorded an antecedent event before the onset of GBS symptoms, the commonest of which were high grade of fever 38 (61.3%), and upper respiratory tract infection while diarrhea with or without vomiting were observed in 13 cases (21%) (see Table 1). Clinical variants included sensory motor (47 patients 75.8%), pure motor (12 patients 19.4%), paraparetic (2 patients 3.2%), and Miller Fisher (1 case 1.6%). A number of specific clinical features were noted including 40 (64.5%) patients with involvement of cranial nerves, 28 (45%) patients with neck muscle weakness, 13 (21%) patients with papilledema, 19 (30.6%) patients with autonomic affection, and 5 (8.1%) mechanically ventilated patients and in ICU. C-reactive protein was positive in 24 (38.7%) patients, hypoalbuminemia and hyponatremia were observed in 20 (32.3%) and 19 (30.6%) patients, respectively, and cytoalbuminous dissociation in CSF was found in 24 (57.1%) patients. Details of outcome according to HDS are illustrated in Table 1. Table 2 shows correlations between clinical data and laboratory findings at baseline assessment and their relation to outcomes using both HDS and ONLS scales at 3 months.

Table 1 Demographic, clinical, laboratory, and outcome data among studied patients
Full size table
Table 2 Spearman correlations of clinical data and baseline laboratory findings with 3-month outcome using the Hughes Disability Scale (HDS) and Overall Neuropathy Limitation Scale (ONLS)
Full size table

Poor outcomes were associated with older age, comorbidities, antecedent events (especially diarrhea), neck muscle weakness, low MRC sum score at onset, nadir and at the 10th day of treatment, papilledema, impaired cough reflex and dysautonomia (Fig. 1).

Fig. 1
figure 1

Spearman correlations of MRC sum score at the 10th day after treatment and Erasmus Guillain–Barré Respiratory Insufficiency Score (EGRIS) at onset with 3-month outcome measured with the Hughes Disability Scale (HDS) and Overall Neuropathy Limitation Scale (ONLS). Outcomes in both scores were significantly correlated with MRC sum score at the 10th day of treatment and EGRIS at onset

Full size image

Of the laboratory findings, only hyponatremia, hypoalbuminemia (Fig. 2), elevated CRP and cytoalbuminous dissociation in CSF were significantly correlated with poor outcomes (Table 2).

Fig. 2
figure 2

Spearman correlations between serum sodium and albumen and outcomes using both HDS and ONLS at end of 3 months. There are significant negative correlations between serum sodium and albumen at onset with both outcome scores

Full size image

Multiple ordinal regression analysis (Table 3) was used to identify the best clinical and laboratory predictors of poor HDS outcome. These were the presence of an antecedent event particularly diarrhea, neck muscle weakness as well as elevated CRP, and low serum sodium at onset. Linear regression analysis was used to identify the best clinical and laboratory predictors of the ONLS Outcome Scale (Table 4) with cytoalbuminous dissociation and treatment plan as factors in addition to all those predictive of HDS outcome.

Table 3 Multiple ordinal regression analysis to identify the best clinical and laboratory predictors of HDS outcome
Full size table
Table 4 Linear regression analysis to identify the best clinical and laboratory predictors of ONLS outcome
Full size table

The best predictor of clinical rating scales of HDS outcome and ONLS (Table 5) using multivariate ordinal regression analysis and multivariate linear regression analysis, respectively, showed that low MRC sum score at the 10th day of treatment was the best single factor.

Table 5 Multiple ordinal and linear regression between baseline clinical rating scales and HDS and ONLS outcome after controlling for age and sex
Full size table

Discussion

Most previous studies have been retrospective, and the outcomes derived from relatively small series or selected variants of patients with relatively short follow-up and limited outcome measures [6,7,8,9]. The present study was a prospective longitudinal study involving a relatively large group of GBS recruited at an early stage of the disease.

The main findings were significant correlations between HDS and ONLS outcomes at 3 months and age at onset, the presence of antecedent events (especially diarrhea), low MRC sum score at onset, nadir, and the 10th day after treatment, neck muscle weakness, papilledema, impaired cough reflex, and dysautonomia. With regard to laboratory findings; hyponatremia, hypoalbuminemia and elevated CRP were correlated with poor outcomes.

Of these factors, multiple ordinal regressions showed that the best predictors for poor prognosis in both outcome measures were the presence of diarrhea, elevated CRP, and hypoalbuminemia. In addition, a low MRC sum score at the 10th day following admission was a predictor of poor HDS outcomes.

Regarding age, a total of 62 patients with GBS enrolled in our study with an average age of 37.15 ± 17.60 years (10–75) with more than half of them ≤ 40 years (54.8 ⁒). These distributions are consistent with Sudulagunta SR and colleagues, and Abdelkader Tunç [21, 22]. The age of the patients in years was significantly correlated with outcomes of HDS and ONLS scores meaning that older age was associated with poorer outcomes, as reported in several previous studies [8, 22,23,24,25,26,27,28]. There was no influence of sex on HDS or ONLS outcomes.

According to antecedent Events, an antecedent event was recorded in 43 (69.4%) patients of which 38 (61.3%) had fever, 30 (48.4%) had upper respiratory tract infection, and 13 (21%) had gastrointestinal infection including diarrhea with or without vomiting. Siddiqui and their colleagues [29] found a lower frequency of various antecedent events was recorded in 33 patients, including respiratory tract infection in 9 (14%) and diarrhea/vomiting in 13 (21%) patients. In contrast, Sudulagunta and colleagues, 2015 found that 80% of patients had a predisposing infection which was higher than our study; however, gastrointestinal infection was more common in that study (47.25%) than upper respiratory infection (34.73%) [21].

In the present study, the presence of a preceding antecedent event especially diarrhea was correlated with outcomes of HDS, and ONLS and was considered as one of the predictors for poor outcome. Hadden and colleagues, van Koningsveld and colleagues, and Rajabally and colleagues found similar results [24, 26, 27]. In contrast, Tunç and colleagues [22] found no correlations between outcome and the presence or absence of antecedent events although this may be related to the small sample size.

In the current study the duration of the antecedent event was 1.82 ± 1.52 days and the number of days between onset of symptoms and nadir was 3.32 ± 1.64 days. Twenty-seven patients (43.5%) presented with weakness within 3 days of onset; 23 additional patients were weak (37.1%) within 7 days or more. Similarly, Walgaard and colleagues, 2010 found that 35% of patients presented with weakness within 3 days from onset [7].

Comorbidities had little effect on outcome. Fourteen patients (22.6%) had diabetes and/or hypertension, but this was not related to outcome.

Low MRC sum scores at onset, at nadir and 10th day of treatment were associated with poor outcomes. Chiò A and colleagues, and McKhann GM and colleagues reported similar findings [23, 25]. Rajaballys and colleagues found that a low MRC sum score at admission or 7 days after admission predicted the need for mechanical ventilation [27]. Wen and colleagues found that MRC sum scores on admission and at nadir were significantly different between severe GBS and non-severe GBS groups [28], while Singh NK and colleagues found that severe weakness at nadir, and rapid onset of weakness were poor prognostic features [30]. Gonzalez-Suarez and colleagues reported that older age, severe deficits at onset, and cranial nerve involvement were poor prognostic factors [31].

More than three-quarters of the patients (47 patients with 75.8%) presented with the classic sensory–motor variant, while 12 patients (19.4%) presented with a pure motor variant, 2 patients (3.2%) had paraparesis and one patient (1.6%) had Miller Fisher syndrome. The pure motor type had significantly poorer outcome compared with classic sensory motor neuropathy consistent with other studies [15, 32,33,34].

Involvement of cranial nerves was observed in 40 patients (64.5%) with either unilateral or bilateral facial palsy, bulbar palsy or both but neither correlated with outcome. Other studies reported a lower incidence of cranial nerve involvement ranging from 25 to 34% [21, 22, 35, 36], while several previous reports found that involvement of cranial nerves was more common in severe forms of GBS and was associated with poor prognosis [27, 28, 31]. Verma, Chaudhari, Raut and colleagues found that cranial nerve involvement was associated with mechanical ventilation [36].

Neck muscle weakness was observed in 28 patients (45.2%) and was significantly correlated with HDS and ONLS outcomes. Malaga and their colleagues [37] showed that two clinical parameters (bulbar and neck weakness) early at onset are strongly associated with the risk of respiratory failure.

Malaga and their colleagues found that the frequency of respiratory failure in GBS was 14% [37]. In the present study, 32 patients (51.6%) had respiratory muscle involvement and an impaired cough reflex, but only 5 (8.1%) were intubated and mechanically ventilated (MV) which was lower than that reported by Sudulagunta and colleagues who found that 38.5% of patients were mechanically ventilated [21]. The impaired cough reflex and respiratory muscle involvement significantly correlated with poor outcomes using HDS and ONLS (p = 0.01, p = 0.035, respectively). However, neither was considered as predictor of poor prognosis. In contrast, EGRIS risk total score at onset for respiratory insufficiency significantly correlated with poor outcomes of HDS and ONLS at 3 months (p = 0.008, p = 0.000). The Erasmus GBS Outcome Scale also correlated with poor outcomes in HDS and ONLS. Respiratory problems may well arise because bulbar weakness leads to reduced protection of the airway and difficulty in clearing secretions. It can also cause upper airway collapse, which increases airway resistance and respiratory muscle load causing fatigue, and ultimately respiratory failure.

In the present study, an impaired cough reflex was correlated negatively with poor outcomes but was not considered as a predictor. Thirty-five (56.5%) patients were considered high-risk according to EGRIS which is consistent with Rajabally and colleagues. They found 65% patients had a high-risk score, 24% intermediate and 4% low-risk scores [27]. Shangab and their Colleagues found that 20 cases (24.4%) required MV at onset and axonal type was presented in 11 (55%) patients requiring intubation [38].Dysautonomia occurred in 19 patients (30.6%) and was significantly correlated with poorer outcomes in HDS and ONLS (P = 0.001, P = 0.001, respectively). Puyuan Wen and colleagues considered dysautonomia as a risk factor for severity with a significant difference in outcome in the severe GBS and non-severe GBS groups [28]. Islam and colleagues identified autonomic involvement as an important risk factor for mechanical ventilation and poor outcome [39]. Netto and colleagues found that older age, dysautonomia, and pulmonary complications were predictors of mortality in MV patients with GBS [40]. Zhai and their colleague [11] found that HDS (at admission), dysphagia, and dysautonomia were independent risk factors for GBS patients requiring MV.

Hyponatremia occurred in 19 (30.6%) patients and was significantly correlated with HDS and ONLS outcomes (p = 0.002, p = 0.001, respectively) and was considered as a predictor of poor prognosis, consistent with previous studies [41,42,43].

In GBS, inflammation, demyelination, and axonal damage produce reactive oxygen species associated with free radical toxicity [44, 45]. In the present study, 20 patients (32.3%) had hypoalbuminemia which was significantly correlated with poor outcomes of HDS and ONLS (p = 0.002, p = 0.017, respectively) but was not considered as one of the predictors for poor outcome. Tunç and colleagues found that serum albumin levels decreased consistently with the progression of the disease severity whils Su and colleagues demonstrated that high albumin was a protective factor for GBS patients [22, 46].

Elevated C-reactive protein occurred in 24 (38.7%) patients and was significantly correlated with poor outcomes only in HDS (p = 0.024). It was also predictive of poor outcome. Abdulkadir Tunç and colleagues demonstrated the negative impact of higher age and higher CRP levels measured at the end of the first month [22]. Although elevated CRP levels have been reported to be potential biomarkers for some inflammatory diseases [47], the association of GBS and CRP is still lacking [48].

24 (38.7%) of the 42 patients who underwent CSF analysis had cytoalbuminous dissociation and elevated protein in CSF protein, both of which were correlated with poor outcome although not considered as predictors. High CSF protein concentration is thought to indicate disruption of the blood–nerve barrier [49], but few previous studies have demonstrated that it is related to poor prognosis [49,50,51]. DiCapua DB and colleagues 2015 found that elevated levels of CSF protein occur in 50% of patients after the first week and in 80% of patients in the second week [49]. Sevki Sahin and colleagues reported a positive association between low baseline levels of CSF protein and good prognosis [52]. Vidrio-Becerra and colleagues, 2018 reported that CSF can be used as a prognostic indicator of severity, and that proteins greater than 100 translate into a torpid evolution with more complications [53]. Sahin and colleagues, 2017 found that CSF protein level was a prognostic indicator as it was negatively correlated with MRC outcome at 6 months follow-up, and it was an independent factor on regression analysis [52].

In the present study, 49 patients were treated with plasma exchange, while 13 were treated with IVIG according to the availability. There were no significant differences in outcomes of HDS between the two lines of treatment, however better outcome was recorded with IVIG than plasma exchange in relation to ONLS outcome (p = 0.016).

Conclusions

Clinical and laboratory predictors of poor outcome were older age, antecedent events (particularly diarrhea), low MRC sum score at the 10th day, elevated CRP, hyponatremia and cytoalbuminous dissociation.