Prognostic factors of Guillain-Barré syndrome: a 111-case retrospective review

Background To identify the predictive factors associated with worse prognosis in the Guillain-Barré syndrome (GBS), which can be helpful to fully evaluate the disease progression and provide proper treatments. Methods Clinical data of 111 GBS patients who were diagnosed from 2010 to 2015 were collected and retrospectively analyzed. Results Patients with diabetes (P=0.031), high blood pressure at admission (P=0.034), uroschesis (P=0.028), fever (P<0.001), ventilator support (P<0.001) during hospitalization, disorder of consciousness (p=0.007) and absence of preceding respiratory infection(P=0.016) were associated with worse outcome at discharge, while abnormal sensation, ataxia, weakness and decrease of tendon reflex seemed not correlated with the Medical Research Council(MRC) score at discharge. Compared with the subtype of acute inflammatory demyelinating polyneuropathy, prognosis of Miller-Fisher syndrome (p<0.001) and cranial nerve variant (p<0.038) were better, but prognosis of acute motor axonal neuropathy(AMAN) was worse (p<0.032). Laboratory examinations at admission showed that hyperglycemia (P=0.002), high leukocyte count (P=0.010), hyperfibrinogenemia (P=0.001), hyponatremia (P=0.020), hypoalbuminemia (P=0.005), abnormal hepatic (P=0.048) and renal (P=0.009) functions were associated with poorer prognosis at discharge, while albuminocytologic dissociation in cerebrospinal fluid, GM1 and GQ1b antibody showed no correlation with the MRC score at discharge. γ-Globulin and glucocorticoid therapies showed no difference in the MRC score at the discharge. Conclusions AMAN, diabetes, high blood pressure, uroschesis, high body temperature, ventilator support, consciousness disorder, absence of upper respiratory tract preceding infection, hyperglycemia, hyponatremia, hypoalbuminemia, high leukocyte count, hyperfibrinogenemia, abnormal hepatic and renal function were demonstrated as poor prognostic factors.


Background
Guillain-Barre syndrome (GBS) is a set of clinical syndromes with a common pathophysiological basis, and is usually considered to be an immune-mediated disorder of the peripheral nervous system [1,2]. GBS is usually characterized by symmetrical flaccid paralysis with areflexia, which usually reaches a maximum severity within four weeks [3,4]. However, recent studies suggest that some patients with GBS had normal or hyperreflexia [5,6], and a wide range of motor, sensory and autonomic symptoms could also be found from GBS patients [7,8]. The reported mortality of GBS in whole population ranged from 0.89 to 1.89 cases (median 1.11) per 100,000 people [9], which makes GBS the most common cause of acute flaccid paralysis currently. Intravenous immunoglobulin (IVIG) and plasma exchange (PE) were proven to be effective therapeutic method for GBS [10,11] and are now widely used clinically which lower the mortality rate effectively, making most of the patients a complete functional recovery or with minimal deficits [12]. However, there are still some cases with bad prognosis and sequelae such as decreased mobility, severe long-term fatigue syndrome and chronic pain [12]. The reported mortality in GBS patients now varies between 3% and 7% [13][14][15]. With this study, we aimed to identify the predictive factors associated with worse prognosis in the GBS.

Methods
We retrospectively reviewed 111 patients diagnosed GBS who were treated in the neurology department of Huashan Hospital and Huashan North Hospital affiliated to Fudan University from 2010 to 2015. The patients were diagnosed based on clinical features and electrophysiological findings. Demographic, clinical information of all patients was recorded and reviewed.
The Hughes Functional Grading Scale (HFGS) score was used to assess functional disability, which was defined as follows: 0, healthy state; 1, minor symptoms and capable of running; 2, able to walk 5 m or more without assistance but unable to run; 3, able to walk 5 m across an open space with help; 4, bedridden or chairbound; 5, requiring assisted ventilation for at least part of the day; 6, dead.
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).
At admission, the patients' blood samples were drawn and lumbar puncture was performed to collect the cerebrospinal fluid. Laboratory exam data were collected and analyzed.
Some patients also received the electromyogram examination during hospitalization. According to the electromyogram examination, patients with demyelination and axonal damage were classified.
All the demographic, clinical, laboratory exam and electromyogram exam data were all collected from the department of medical record of Huashan Hospital and Huashan North Hospital affiliated to Fudan University.
The study was approved by the Ethics Committee of Huashan Hospital affiliated to Fudan University.
MRC sum score was regarded as the estimation of prognosis, as previous reports revealed that the MRC sum score has predictive value for prognosis which is more accurate than the GBS disability score [16,17]. We assume that the higher MRC score at discharge means better prognosis.
Statistical analysis was performed by using SPSS software (version 24.0). Classification variables analysis was performed with the use of Kruskal-Wallis and U Mann-Withney with the non-parametric ones. Univariate analysis was used to identify the factors associated with poor prognosis, which were further analyzed by Logistic regression analysis for predictors independently related to the poor prognosis. A p value less than 0.05 was considered statistically significant. When doing the Logistic regression analysis, patients who scored 33 or more in the MRC score at discharge were classified as patients with good prognosis, and the rest were classified into the poor prognosis group.
Several clincial features-related prognosis predictors of patients with GBS (as shown in Table 1) were found. Prognosis of GBS patients with diabetes was worse compared with those without diabetes (p=0.031). GBS patients with high blood pressure at admission have worse outcome at discharge (p=0.034). Different subtypes of GBS have different prognosis. Compared with the AIDP subtype which is the most common subtype in our study, the prognosis of MFS (p<0.001) and CNV (p<0.038) were better, and the prognosis of AMAN was worse (p<0.032). Interestingly, we also discovered that the preceding upper respiratory infection is related to better prognosis (p=0.016). The MRC scores of patients who had preceding upper respiratory tract infection were 35.82±8.0 at discharge, while GBS patients with no preceding infection scored 30.80±11.96.
We also found some symptoms and signs during hospitalization-related prognosis predictors of patient with GBS (as shown in Table 2). Prognosis of GBS patients with urine retention was worse compared with that of normal urinating function (p=0.028). Prognosis of GBS patients with high body temperature (p<0.001), of those with conscious disorder was worse (p=0.007), of who needed mechanical ventilation (p<0.001) and stayed in intensive care unit (ICU) (p<0.001) were worse. What's more, the higher of the MRC score at admission was, the better of the prognosis at discharge would be (p<0.001) according to the MRC score at discharge.
We also tried to find whether different treatments lead to different prognosis (as shown in Table 3). Among all the 111 patients, 40 patients (36.0%) received IVIg only, 11 patients (9.9%) received corticosteroids therapy only, 58 patients (52.3%) received both IVIg and corticosteroids therapy, 1 (0.9%) received all the IVIg, corticosteroids, and plasms exchange therapies, and 1 (0.9%) refused all these treatments. According to the MRC score at discharge, the prognostic difference among different therapeutic modalities were not significant.
Some laboratory examination-related early predictors of a low MRC score at discharge, which was associated with a poor prognosis, were found (as shown in Table 4). All the laboratory examinations whose results analyzed were conducted at admission. Prognosis of GBS patients with hyperglycemia(p=0.002), hyponatremia that serum sodium level lower than 135mmol/L (p=0.020), and serum albumin level lower than 40g/L(p=0.005), were worse. Further logistic regression analyses revealed that higher the white blood cell count was, the lower the MRC score would be, indicating that higher white blood cell count was a possible predicator of poor prognosis. The same result occurred when plasma fibrinogen was taken into logistic regression analyses, that the higher the plasma fibrinogen was, the lower the MRC score would be. Prognosis of GBS patients with abnormal hepatic or renal function were worse. The differences were significant (p=0.048 and p=0.009 respectively). Patients with normal level of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) scored 34.73±8.72, and those with abnormal level of ALT or AST scored 29.52±12.63. Patients with normal level of blood urea nitrogen (BUN) and Serum creatinine (Scr) scored 34.39±9.66, and those with abnormal level of BUN or Scr scored 28.86±11.52. The significant association between albuminocytologic dissociation in cerebrospinal fluid (CSF) with the poor prognosis was not found from our study. Neither did the CSF & serum level of GM1 or GQ1B antibody. No significant MRC score at discharge differences between demyelination and axonal damage which were classified by electrophysiological findings.
Factors associated with poor were further analyzed by Logistic regression analysis. The result showed that low MRC score when admission, abnormal renal function, abnormal blood glucose level, and the subtype of AMAN are independently related to the poor prognosis with statistical differences (as shown in Table 5).

Discussion
Guillain-Barré syndrome is a rapid-onset weakness and numbness disease caused by the immune system damaging the peripheral nervous system. GBS is usually self-limiting, and most patients either recover completely or only retain minor residual symptoms. But there are still several patients who may face severe outcomes including death. In our study, GBS prognosis is  There are many subtypes of GBS, such as AIDP, AMAN, MFS, BBE-GBS, and so on, and the proportion of different subtypes varied significantly among different contries and different regions in a same country [18][19][20][21][22][23]. Our hospital is located in Shanghai which is an east China city, and most our patients are from China, so our study may represent the clinical characteristics and distributions of GBS in China, especially the eastern part of China. In our study, AIDP accounted for 60.4% of all the GBS cases, making it the most common subtype in our hospital. The second one was MFS, accounting for 21.6% of GBS cases, which was similar to 26% in Japan [24] and 25% from the report of Singapore [20]. In sharp contrast, only 7% GBS patients are MFS patients in southwest China [25]. Moreover, AMAN accounted for 8.11% in all GBS patients, CNV accounts for 7.2% and BBE-GBS accounts for only 0.9% of all the patients in our study. Among all the subtypes of GBS, the prognosis of MFS and CNV were the best, and the prognosis of AMAN was worse compared with AIDP in our study, as AMAN is the axonal damage subtype while demyelination is more common among AIDP patients according to  electromyogram examinations. González-Suárez et al reported that GBS patients with axon injury were more likely to suffer from respiratory failure [26], leading to poorer prognosis. The infectious event is described to appear in 40-70% of patients [7,8,[27][28][29][30][31]. In our series up to 61.8% of cases have had the infectious event, and respiratory infection was the most frequent one among all these infections. We discovered that explicit upper respiratory tract preceding infection was a protective predictor which may help patients get better MRC score at discharge compared with those who had diarrhea or had no explicit preceding infection. This may due to that diarrhea is usually caused by Campylobacter jejuni (C. jejuni) and according to a previous study, C. jejuni infections exclusively elicit AMAN in East Asia, the axonal damage subtype which has worse prognosis [32], while upper respiratory tract infection is usually caused by other kinds of pathogens.
In our study, we discovered that MFS accounted for 33.33% patients among those who had preceding upper respiratory infection, and only 11.90% patients among those who didn't have preceding infection. At the same time, CNV accounted for 10.26% patients among those who had preceding upper respiratory infection, and only 7.14% patients among those who didn't have preceding infection. Since MFS and CNV are the two subtypes that have better prognosis, patients with explicit upper respiratory tract preceding infection showed better prognosis compared with those who without preceding infections.
In our study, patients with diabetes and high fasting blood glucose had poorer prognosis. A previous reports also claimed that diabetes mellitus (DM) is an independent poor prognostic factor for the ability to walk unaided at 3 months after symptom onset [33]. The mechanism is still unclear, but there are several assumptions which may explain. First, some laboratory evidence showed that patients with DM are in a state of chronic low-level inflammation: elevation of various inflammatory markers such as C-reactive protein, tumor necrosis factor and interleukin-6 [34], and this chronic low-level inflammation may lead GBS patients to poor prognosis. Another assumption believed that neurovascular mechanism of DM neuropathy cause bad prognosis for GBS patients, as a chronic state of nerve ischemia in DM may induce partial axonal injury or loss even in subclinical DM neuropathy [35].
GBS has a tendency for dysautonomic features such as bladder dysfunction, abnormal body temperature and hypertension among patients [36,37]. The underlying mechanisms of urinary dysfunction appear to involve both hypo-and hyperactive lumbosacral nerves caused by GBS [38], and the lumbosacral nerve involvement may relate with poor prognosis. Autonomic dysfunction in patients with GBS reflects dysfunction of sympathic and/or parasympathic innervation, but the exact immunopathological mechanisms remain to be elucidated. Autonomic dysfunction symptoms, such as tachycardia, hypertension, gastrointestinal dysfunction, and bladder dysfunction, can be serious problems as autonomic dysfunction is a predictive poor prognosis factor that may cause sudden death [39][40][41]. There are two main possible reasons to explain. First, autonomic dysfunction is associated with fatal hypoxia because of respiratory muscle involvement and a long duration of mechanical ventilation and the need for tracheostomy [42,43]. What's more, autonomic dysfunction is related to bad prognosis also because of cardiovascular "collapse," as reported by Clarke et al [44]. However, though related with higher mortality, Samadi M et al found that autonomic dysfunction showed no significant association with motor dysfunction among kids [45], which is different from this report.
According to the previous studies, the incidence of mechanical ventilation in GBS patients is 20~30% according to the previous western reports [46][47][48]. In our study, this incidence was 14.4%,which is similar to 14.8% of the northeast China report [49], but lower than reports from western countries. The main reason which lead GBS patients to death is respiratory failure [50]. The need of mechanical ventilation and ICU is not only a necessity to those who suffer from respiratory failure, but also a significant sign of respiratory muscle involvement and severe conditions. In accordance with our assumption, our result showed that those who used mechanical ventilation and ICU during hospitalization had worse MRC score at discharge.
Hyponatremia is so common in GBS patient,despite that it is not a classical manifestation of GBS; however, there are series in which are described to be present in 21.5 to 48% of the cases; in our review, it was found in 21.6% of our patients, which is similar to the Northern China study and the British study [51][52][53][54]. Hyponatremia is also a predictor of poor prognosis. There may be two possible reasons that make hyponatremia happen: First is the syndrome of inappropriate antidiuretic hormone secretion (SIADH) [53], the possible mechanism could be that SIADH was due to abnormalities of peripheral autonomic afferent fibers arising from vascular stretch receptors or due to increased renal tubular sensitivity to vasopressin. Another reason is cerebral salt wasting syndrome (CSWS) [55], and the possible mechanism involves an inappropriate Brain Natriuretic Peptide secretion upon sympatho-adrenal dysregulation as part of GBS dysautonomia. The treatment principles between SIADH and CSWS are totally different, and the incorrect management of hyponatremia may lead to osmotic demyelinating syndrome. This may explain the poor outcome at discharge of hyponatremia patients compared with those with normal blood sodium level.
Our study also found elevated liver enzyme level indicates poorer prognosis as well. There are two possible reasons. First, liver damage conditions such as infection with hepatitis virus; alcohol abuse; hepatotoxic drugs; recent surgery and so on may influence the systemic health condition significantly, and those who with liver damage may recover more slowly compared with those who have normal hepatic function. What's more, according to Oomes PG et al's report, 38% patients showed a plasma alanine aminotransferase elevation, gamma glutamyl transferase elevation, or both or more than 1.5 times the upper limit of normal, of who most were negative for known causes of liver damage [56], and in that study, IVIg treatment seemed to be associated with mild transient liver function disturbances through an unknown mechanism. When admission, some of our patients might have already received IVIg treatment in other hospitals, thus making their liver enzyme level elevated at adimission. Since referral patients' health conditions are worse than new diagnosed patients generally, the elevated liver enzyme on admission may also be related with worse prognosis in this way.
According to Khajehdehi P et al's research, acute renal failure can occur commonly in cases with severe GBS patients particularly in those with dysautonomia [57]. What's more, it has also been reported that patients with GBS can develop acute glomerulonephritis of immune complex origin associated with deterioration of renal function tests. Acute interstitial nephritis is another possibility for renal function deterioration in GBS patients [58]. Elevated Scr and BUN are signs of renal function deterioration and ARF, so that they are related with poorer prognosis in our study.
Our study also found that poor nutrition condition such as low serum albumin level, high coagulation state as high fibrinogen leval, and infection at early stage of GBS such as high white blood cell count, are also predictors of poorer prognosis.
Our work has several limitations. First, as a retrospective study, some clinical parameters which have been reported to be predictors of GBS prognosis were unavailable in our cohort, such as vital capacity and the Peak Flow-test result [59]. What's more, we didn't collect the MRC score and other clinical features after patients' discharge, making the estimation of the long-term prognosis impossible. Some clinical parameters which has been reported as predictors of prognosis in the previous studies but didn't show statistical differences in our study such as Anti-GQ1b antibody [60] may due to the small sample capacity and the regional/racial differences. Larger and prospective studies will be required.

Conclusion
AMAN, diabetes, high blood pressure, uroschesis, high body temperature, ventilator support, consciousness disorder, absence of upper respiratory tract preceding infection, hyperglycemia, hyponatremia, hypoalbuminemia, high leukocyte count, hyperfibrinogenemia, abnormal hepatic and renal function were demonstrated as poor prognostic factors. In order to identify patients with bad prognosis at the early stage and get a good better outcome, more attention should be paid to these poor prognostic factors. Availability of data and materials The datasets generated during and analyzed during the current study are available from the corresponding author on reasonable request.
Authors' contributions ZY collected the clinical data, participated in analyzing the results and drafted the manuscript. ZY participated in the design of the study and coordination of the mauscript. WY participated in the design of the study and the coordination of the study. All authors read and approved the final manuscript.
Ethics approval and consent to participate This study has approved by the ethics committee of Huashan Hospital affiliated to Fudan University.

Consent for publication
Not applicable.