Predictors and Risk Factors of Severity of GBS, the Requirement for Mechanical Ventilation and Poor Short-Term Prognosis of Severe GBS

Background/Objective:It is still an extraordiarily exigent thing to early recognize the severity, therapeutic method and prognisis of Guillain-Barré syndrome (GBS). Our research’s goal is to investigate the clinical predictive factors indicating severity of GBS and the requirement for mechanical ventilation(MV) in severe patients with GBS, and to explore the identication of modiable risk factors for predicting poor short-run outcome of severe GBS. Methods: A total of 155 patients were included in a cohort of GBS patients admitted to the Aliated Hospital of Qingdao University between 2014 and 2020. Demographic, clinical, therapeutical and evolutionary data were collected. Neurological testing which used standardized data collection were carried out by the same investigators on the whole study period. Comparion of the results were implemented by single and multiple regression analysis. Result: Of 155 patients, 66 patients were severe GBS, of which twenty-nine patients (18.7%) needed MV. Signicant clinical predictors for severe GBS were recent history of surgery, time from onset to admission, Medical Research Council(MRC) sum score on admission, autonomic dysfunction, cranial nerve impairment(glossopharyngeal, facial, Oculomotor and/or abducent nerve decits) and elevated liver enzyme levels were signicantly bound up with severity scores(p<0.05). With regard to risk factors of MV in severe GBS, univariate logistic analysis indicated presence of cranial nerve involvement, autonomic dysfunction, MRC score at nadir, elevated liver enzymes and pneumonia were signicantly different(p<0.05); Multivariate analysis determined that MRC score at nadir and autonomic dysfunction were also considered to be predictors for MV in severe GBS(p<0.05).As to prognostic of factors, recent history of surgery, MRC score at nadir, the requirement for MV, MRC score at admission and pneumonia during hospitalization were predictors of poor short-run outcome in severe patients with GBS by univariate logistic analysis, and that the nadir MRC score and the requirement for MV were identically proved as clinic parameters of unfavourable prognosis by multivariate logistic analysis (p<0.05) . Conclusions: Clinical risk factors of severity in GBS, the requirement of MV and unfavourable short-run prognosis in severe GBS were evident. Recent history of surgery is also a predictor for severity in GBS patients.


Introduction
Guillain-Barré syndrome (GBS) is a in ammatory demyelinative polyradiculoneuropathy [1], and its acute phase is characterised by generalised paralysis, bulbar muscle weakness, autonomic dysfunction, respiratory failure, and with or without sensory associated with hypore exia or ar exia, and in absence of cerebrospinal uid pleocytosis.
Approximately 30% of patients with GBS present with respiratory failure, so that they require endotracheal intubation and MV support [1,2]. Thus, respiratory failure is a life-threatening manifestation, which is the leading cause of death for GBS patients [3]. Furthermore, severe GBS patients need to be closely monitored in intensive care unit(ICU) and even arti cial ventilation to save live. These emphasize the necessity of judging severity early and proper guidelines of allocating patients with GBS to the suitble department(common ward or ICU) to decrease the incidence of respiratory distress and mortality.
It has been reported that the requirement for MV was a prominent prognostic predictor of poor result in GBS [4]. Therefore, identi cating patients who may need to be intubated and MV at an earliar stage of hospitalization is of great improtance. The clinical features and prognosis of severe GBS are many and variable. As severe GBS patients are often complicated with severe complications such as pneumonia and sepsis, these patients usually have poor prognosis. Early identi cation of prognostic predictors of poor prognosis in severe GBS patients and early clinical intervention for these interventional factors are expected to improve the prognosis of GBS patients with respiratory failure.
Additionally, due to the prevalence of GBS, the sample size of severe GBS especially in respiratory failure is usually small. Therefore, through a retrospective study, our study analyzed and studied on the clinical risk factors of severity of GBS, respiratory failure and poor prognosis in severe GBS patients, aims to ameliorate the prognosis of severe GBS and have clinical, therapeutic, and nancial implementations in our setup to a great extent.

Study design and setting
The research was con rmed by the ethics committee of the A liated Yantai Yuhuangding Hospital of Qingdao University, Yantai, China. Because of its retrospective nature of our study, informed consent was waived. Subjects were selected from patients who comformed to the diagnostic criteria of GBS [5] and accepted sequential therapy during hospitalization in the Department of Neurology of Yuhuangding Hospital A liated to Qingdao University, during January 2014 to July 2020. Subjects were excluded from the study if they were: aged<18 years; refused the treatment or diagnosed as either bickerstaff encephalitis, critical illness polyneuropathy/mypathy(CIP/CIM), or chronic in ammatory demyelinating polyradiculoneuropathy [6] or Miller Fisher syndrome. Also subjects has been excluded from the patients who were out of hospital within 3 days, because their illness might not reach the worst condition when discharged and they might lose the data. Clinical data from all subjects were analyzed retrospectively including age, gender, place of residence(urban community or countryside), season of disease occurrence, history of antecedent infections (mainly include diarrhea and upper respiratory tract infection), recent history of surgery(GBS symptom onset within 6 weeks), time from onset to admission, hospital stays, clinical severity evaluated by the Hughes Functional Grading Scale (HFGS) score at nadir/admission, muscle strength assessed by MRC sum score at nadir/admission, tendon re ex, sensory disturbances, cranial nerve damage(including glossopharyngeal and vagus nerves, facial nerve, oculomotor and/or abducent nerve), sensory disturbance, whether requiring mechanical ventilation, autonomic nerve dysfunctions(including blood pressure uctuation, tachycardia, bradycardia, and abnormal sweating), abnormal hepatic enzyme, therapeutic method and complication during hospitalization such as pneumonia. Abnormal hepatic enzyme was de ned as abnormal aspartate aminotransferase and/or alanine aminotransferase which were/was 1.5 times higher than the normal value in the 2nd day after admission. In our department, in the treatment of GBS patients, we chose intravenous immunoglobulin (IVIg) as the preferred treatment method, while some of severe patients were prescribed intravenous corticosteroids which was decided by the neurologists. What's more, the patients who refused using IVIg were administrated intravenous corticosteroids or supporting therapy.
Assessment of neurological functional de cit and clinical severity All 155 patients were assessed the neurological functional impairment and clinical severity. The HFGS was applied to evaluated functional impairment which has 6 degrees[7]: 0, normal; 1, mild symptoms and able to run; 2, capable of walking more than ve meters without help with others but can not run; 3, capable of walking more than ve meters with assistance; 4, chairbound or bedridden; 5, requiring mechanical ventilation for breathing; 6, dead. Additionally, MRC score whose total score rangs from 0 to 60 was used to evaluate muscle strength, its score was calculated according to six bilateral muscles in four limbs [8]. The lowest MRC score or the highest HFGS score was de ned as the nadir of GBS.

Assessment of short-run outcome and grouping
All 155 patients were gotten into groups of two on the basis of HFGS, severe GBS group (HFGS ≥4) [9] and non-severe GBS group(HFGS <4). For severe GBS group, 66 patients were further divided into subgroups of two based on the requirement for MV or not for MV, Group MV including the GBS patients requiring MV and Group NV including the patients not requiring MV at the nadir of illness. Generally, the patient was allowed to discharged from the hospital when his condition was improved or stable in our department. Thus, in this study, patients who could walk with help when discharged (HFGS ≥3) were judeged to receive a favourable short-run outcome, and on the contrary, the patients who could not walk even with assistence(HFGS<3) were considered to have a poor short-run outcome. In view of the above, the severe GBS patients were separated into two other groups: Subgroup 1, patients with good short-run outcome; Subgroup 2, patients with poor short-run outcome.

Statistical analysis
In present study, SPSS version 17.0 software and GraphPad prism 8 were use for statistical analysis.
Categorical data were expressed as proportions and tested by Chi-square test; All continuous data accordde with normal distribution and were expressed as means±standard deviations and tested by independent t-tests. Multivariate logistic regression analysis was used to evaluated the independent predictors of severity in GBS patients, the requirement of MV and unfavorable short-run outcome in severe GBS. For all statistical tests, p-values < 0.05 was deemed to be signi cantly different in statistics.

Demographic features of GBS patients
All 155 patients were registered. The average age of onset was 56.2 years, and the majority were males(57.4%). 65(42.6%) patients with HFGS score ≥4 points at nadir were placed into severe GBS group, and the rest of 89 patients were in non-severe GBS group. Difference between severe GBS group and nonsevere GBS group are shown in Table 1.We demonstrated that cranial nerve impairment and autonomic dysfunction were both more common in severe GBS group (56% versus 28.1% and 43.9% versus 4.5% both p<0.05), and so was a longer hospitalization time (Fig.1A 20.5 days versus 12.1days, p<0.01). In severe GBS group,recent history of surgery was higher than non-severe GBS group (25.8% versus 5.6%, p<0.05). As indicated in Fig.1B, Muscle strength was evaluated by MRC score at admission and at nadir indicated more severe disease in severe GBS group(31.8 versus 49.7 and 20.6 versus 48.3, both p<0.05). Time from onset to admission of non-severe GBS group was longer than that of severe GBS group (Fig.1C 9.6 daysversus 5.4 days, p<0.05). Severe GBS group has higher abnormal liver enzymes than non-severe GBS group (Fig.1D 65.5% versus 13.5%, p<0.05) . In addition, age, sex, the season of morbidity and the place of residence had not signi cant difference(p>0.05), and the same as tendon re ex, sensory dysfunction, pain.

Clinical risk factors of severity
In the study, recent history of surgery, time from onset to admission, MRC sum score on admission, autonomic dysfunction, cranial nerve impairment (glossopharyngeal and vagal nerve de cits, facial), elevated liver enzyme level had signi cant correlation with severity scores by univariate analysis(p<0.05).
In addition, there was no clinically characteristic variable associated with severity scores by mutivariate logistic regression analysis.

Clinical predictors of MV
In severe GBS patients, the mean onset age was 60.5±15.4 years with a male preponderance(65.5% vs 38.5%, p<0.05). Thereinto, Group MV included 29 GBS patients requiring MV and Group NV had 36 severe patients not requiring MV. Comparisons between the two groups above are proved in Table 3. We found that sex, the season of morbidity and the place of residence, surgery, MRC at admission, HFGS at nadir had no statistically signi cant difference(p>0.05), and the same as tendon re ex, sensory dysfunction, pain. However, MRC score at nadir of patients in Group MV was signi cantly lower than that of patients in Group NV(11.9 versus 27.1, p<0.05, as shown in Fig.2A). Moreover, the Group MV was measured higher liver enzyme values than the Group NV (Fig.2C, 65.5% versus 16.7%, p<0.05). Cranial nerve involvement(82.8% versus 36.1%, p< 0.05) and autonomic dysfunction(93.1% versus 5.6%, p< 0.05, Fig.2D ) were both more common in Group MV which also had a longer hospitalization time (Fig.2B 30.1 days versus 13.1days, p<0.05).
Through univariate logistic analysis, we found that male, cranial nerve impairment, glossopharyngeal and vagal nerve de cits, autonomic dysfunction, MRC score at nadir and elevated liver enzyme were signi cantly different(p < 0.05), as shown in Table 2. By multivariate logistic analysis, we determined that the nadir MRC score and autonomic dysfunction were identically proved as the risk factors for MV (p<0.05) ( Table 3).
Clinical predictors of poor short-run outcome in severe GBS group In the 65 severe patients, 20 patients who had good prognosis were placed into Subgroup 1, the remaining 45 patients with poor outcome were assigned to Subgroup 2. The differences between the two groups are indicated in Table 4. From it we can see that recent history of recent history of surgery were more common in Subgroup 2 than Subgroup 1(33.3% versus 5%, p<0.05). The time from onset to admission and seasonal distribution was not different (p>0.05). Nevertheless, the MRC score at nadir and on admission were both signi cantly higher in Subgroup 1 compared with Subgroup 2 (32.5 versus 15.1, 41.1 versus 27.6, both p<0.05). In addition, the requirement of mechanical ventilation was signi cant lower in Subgroup1 than Subgroup 2(25% versus 53.3%, p<0.05).Because pneumonia during hospitalization period could in uence the outcome of severe GBS patients, we make a comparison between the two groups of the occur of pneumonia, from which we found that pneumonia occured comparably in Subgroup 1 and Subgroup 2 (57.8% versus 30%, p<0.05). Furthermore, By multivariate logistic analysis, the nadir MRC score and the requirement of MV were proved as the independent poor outcome predictors(p<0.05) ,as indicated in Table 5.

Discussion
In our study, we investigated predictors for severity in GBS patients at early stages, for the requirement of MV and short-run outcome in severe GBS. The risk factors for severe GBS included recent history of surgery, time from onset to admission, MRC sum score at admission, autonomic dysfunction, cranial nerve impairment and elevated liver enzyme levels, which were also the predictors for MV, except recent history of surgery and interval time from onset to admission. As to prognostic factors in severe GBS patients, we proved that a lower MRC score at admission and at nadir , the requirement of MV and complicated with pneumonia had led to poor short-term prognosis. We think these results may help clinicians to allocate GBS patients to the appropriate unit, decide whether or not tracheotomy and ventilator assisted ventilation should be performed, assess the prognosis and develop a clinical prediction model.
Although GBS is a potential acute self-limited disease and amost of patients either have recovery fully or retain minor sequelae. However, severe GBS often leads to unfavorable residual sequelae or mortality; at the same time, many hospitals has limited medical resources, especially intensive care facilities. Thus, identi cation of modi able risk factors for severe GBS in early stage of illness is of great important to help clinician to allocate appropriate unite and develop individualized treatment, which can make intensive care facilities use rationally and may decline the occur of residual sequelae and mortality. We found that recent history of surgery was most signi cant risk factor for severe GBS, it was in accordance with the result of Lei Bao et al. [10,11] which demonstrated that GBS patients induced by surgery presented severe movement disorder and respiratory failure. The possible potential pathophysiological mechanisms of post-surgical GBS are not clearly known yet. It has been reported that clinical and/or subclinical infections secondary to post-surgical short-term immunosuppressive conditions induce GBS [12]. In addition, the breakdown of its innate protective barrier make antigens in blood entry into nervous system intraoperatively, so that the antigens trigger followed autoimmune responses [13].Nonetheless, to clearly clarify the pathogenesis need further research. Notwithstanding, in our study recent history of surgery was not statistically different between MV group and NV group, it was contradictory to the results of Lei Bao et al [10] which reported that acute respiratory failure and the requiement of MV in post-surgical GBS patients was meaningfully more common than non-surgical GBS patients, and came to the conclusion of higher respiratory depression of GBS induced by surgery.This difference might be due to the smaller sample size that was used in the present study.Next, we will enlarge the sample size to further clarify the relationship between surgery and respiratory failure.
It is of prime importance to predict the need of MV early, by reason of 60% of GBS patients with MV occured plenty of complications which increase the mortality, early recognition and intervention of GBS may decrease the occurrence of complications and ameliorate its prognosis [14,15,16]. Heterogeneous studies had been conducted to researched the predictive factors of requiring MV in GBS patients. The present study demonstrated that shorter interval time from onset to admission, elevated liver enzyme levels, bulbar nerve involvement, autonomic dysfunction and lower MRC score at admission were predictors of MV, this was in accord with the result of previous studies [17,18,19,20] and could explain why they are the predictors of severe GBS. The scoring system "The Erasmus GBS Respriatory Insu ciency Score(EGOS)" was developed by Walgaard et.al on the basis of three respiratory insu ciency predictors: facial and/or bulbar weakness at admission, MRC sum score at admission and the time onset to admission [18].This nding was further proved by the multivariate regression analysis, which identi ed lower MRC at admission and autonomic dysfunction were independent predictors for MV in severe GBS. Islam et.al found that severe muscle weakness(MRC sum scores ranging 0 from 20) at study entry was more likely to progress to MV [21]. Autonomic dysfunction was thirteen time more common in MV patients than NV patients in the present study.It was in accordance with result of a cohort study that reported dysautonomia as an independent predictor of respiratory insu ciency [22,23]. In contract, age, facial paresis, antecedent infection, sensory, pain, place of residence and treatment option showed no association with the need for mechanical ventilation. The result of male more statistically signi cant preponderance in present study than female was different from other study [24] because of the smaller simple.
As classically described, poor prognosis is unfavourable in our study; Recent history of surgery, MRC at admission, MRC at nadir, the requirement of MV and pneumonia in hospitalization period are the risk factors which have been reported for poor outcome. Lei Bao et al [10] indicated patients with post-surgical GBS had poorer prognosis comparied with the patients with non-surgical GBS. Walgaard et al [18]. have eatablished a modi ed EGOS (mEGOS) whose primary dissimilitude is the MRC sum score at admission and in the 7th day in place of the GBS disability score [25], they demonstrated that the MRC sum score at hospital entry is more precise to guide the choice of treatment method. Pneumonia was associated with poor short-term prognosis, which is consistent with the result of a previous study that pneumonia was related to duration of mechanical ventilation for severe GBS patients [26].Severe GBS patients requiring MV occur more likely complication, such as pneumona, sepsis, lung collapse, pneumothorax, urinary tract infection. It may explain that the severity of illness prolonged the time of hospitalization, and longer duration of hospitalization might be related to poor short-term outcome with severe GBS. This result was further proved by the multivariate regression analysis, which identi ed MRC at admission and the requirement of MV were independent predictors for unfavourable short-run prognosis in severe GBS patients.
The present study has following limitations. At rst, it was a retrospective analysis and has the monocentric design, especially the prognosis was done mainly on the patients of the hospital, lacking of follow-up observations to study the long term prognosis. Secondly, because of the retrospective nature of our research, a certain of clinical indexes which have been advocated to be risk factors of MV could not obtained including various IgG antiganglioside antibody species,vital capacity, electrophysiological data, and so on. Thirdly, the detail data about autonomic nervous system involvement and complication expect pneumonia for MV patients was not recorded in this study. In addition, the sample size of our study for strati ed analysis is so small that we can not come to a conclusion with strong statistic signi cance. Further prospective study is required to prove these observations.

Conclusion
We conclude from our results that recent history of surgery, time from onset to admission, MRC sum score at admission, autonomic dysfunction, cranial nerve impairment and elevated liver enzyme levels are the risk factors for severity of GBS, which were also the predictors for MV in severe GBS, except recent history of surgery and interval time from onset to admission. And lower MRC score at admission and at nadir, the requirement of MV and complicated with pneumonia are the poor short-term prognostic factors of severe GBS patients. These results should be con rmed by a prospective study or a comparable multicenter retrospective study. Score, mEGOS: modi ed Erasmus GBS Outcome Score, CI: con dence interval Declarations the manuscript, which was critically reviewed by all other authors. All authors read the manuscript for intellectual content and commented on the nal version of the manuscript before submission.

Availability of data and materials
The datasets analysed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
This study was approved by the ethics committee of Yuhuangding Hospital A liated to Qing University, because of its retrospective nature, all participants did not provide written informed consent.

Consent for publication
No applicable

Competing interests
The authors declare that they have no competing interests.     Comparisons between MV group and NV group. A MRC at nadir was lower in MV group than NV GBS group(11.9 versus 27.1, p<0.05). B Hospital stays was signi cantly longer in MV group compared with NV group (30.1 versus 13.1, p<0.05). C the liver enzymes and D autonomic dysfunction were both higher in MV group than NV group(65.5%versus 16.7%, p<0.05 and 93.1%versus 5.6%, p<0.05).