Prognosis of patients with Guillain-Barré syndrome requiring mechanical ventilation
Correspondence Address: Source of Support: Support in the form of grants, equipment, etc Nil, Conflict of Interest: None declared.grants, equipment, etc Nil DOI: 10.4103/0028-3886.86545
Source of Support: Support in the form of grants, equipment, etc Nil, Conflict of Interest: None declared.grants, equipment, etc Nil
Introduction: Severe Guillain-Barré syndrome (GBS) is associated with significant morbidity and also mortality. Identification of modifiable risk factors may help in reducing the morbidity and mortality. Objective: To study the prognostic factors in a selected cohort of mechanically ventilated GBS patients. Materials and Methods: Case records of GBS patients requiring mechanical ventilation admitted between 1997 and 2007 were analyzed. All patients satisfied the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) criteria for GBS. Primary outcome parameters included mortality and GBS disability (Hughes) scale score at discharge. Results: During the study period, 173 (118 men and 55 women; mean age of 33.5 ± 21 years) GBS patients were mechanically ventilated. A history of antecedent events was present in 83 (48%) patients. In addition to motor weakness, In all facial palsy was present in 106 (61%), bulbar palsy in 91 (53%), sensory involvement in 74 (43%), and symptomatic autonomic dysfunction in 27 (16%). The overall mortality was 10.4%. On univariate analysis the risk factors for mortality included elderly age (P = 0.014), autonomic dysfunction (P = 0.002), pulmonary complications (P = 0.011), hypokalemia (P = 0.011), and bleeding (P = 0.026). All these factors were significant in multivariate analysis except for bleeding from any site and hypokalemia. In univariate analysis factors associated with Hughes scale score ≤ 3 at discharge included younger age (P = 0.02), presence of bulbar symptoms (P = 0.03) and less severe weakness at admission (P = 0.02), slower evolution of disease over more than 3 days (P = 0.01), electrodiagnostic evidence of demyelinating neuropathy (P = 0.00), and absence of sepsis (P = 0.01), hyperkalemia (P = 0.0001), and anemia (P = 0.02). In multivariate analysis age was the only significant factor. Conclusions: Early identification of modifiable risk factors, such as pulmonary involvement, autonomic dysfunction, hypokalemia, sepsis, bleeding, and nutritional complications, may reduce the mortality and morbidity associated with GBS.
Keywords: Electrophysiology, Guillain-Barrι syndrome, mechanical ventilation, prognostic factors
Of the patients with Guillain-Barré syndrome (GBS), 35% recover completely, 35% have minimal residual motor signs, and 30% have moderate to severe residual paresis.  The optimistic view of complete recovery in majority of patients with GBS may not be true and patients with severe GBS are often left with significant residual deficits.  The reported mortality varied between 2% and 18%, the lower mortality rates were in the treatment trials where patients with serious medical illness have been excluded. Inclusion of less severely affected patients may obscure the analysis of prognostic factors in the more severely ill.  However, the reported mortality in ventilated patients was high, 20%-38.3%. , Apart from improvements in intensive care unit (ICU) care and newer immunomodulatory therapies, other factors that have been found to impact the outcome include age, antecedent events,  rapid progression of the disease,  sensory disturbances, ventilatory requirement,  bulbar dysfunction,  low cerebrospinal fluid protein concentration ,  electrodiagnostic features suggestive of axonopathy. , In a disease, such as GBS with unpredictable and variable course, it will be prudent to identify the factors that may help in assessing the prognosis of severe form of GBS. The present analyzed the prognostic factors in a select cohort of mechanically ventilated patients of GBS admitted to a tertiary care center for neurosciences.
A retrospective review of case records of consecutive patients of GBS requiring mechanical ventilation admitted to neurology ICU of a tertiary care center for neurosciences between July 1997 and June 2007 was done. The study had the approval of the institute ethics committee. Only patients satisfying the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) GBS diagnostic criteria were included.  The patients initially evaluated by the neurology team in the emergency department were carefully monitored for any impending respiratory failure or autonomic dysfunction. At the earliest signs of impending respiratory failure, patients were evaluated by the neuroanesthesia team and shifted to neurology ICU for mechanical ventilation. The criteria for mechanical ventilation were clinical features of hypoxia, vital capacity <15 mL/kg, PaO 2 < 70 mmHg, and PaCO 2 > 45 mmHg.  Immunomodulation treatment modalities included plasmapheresis, immunoglobulin (IVIg) or steroids depending on the availability and the practices at that point of time. Age, hemodynamic status, and affordability were also taken into consideration while choosing immunomodulation treatment modality. All patients were started on prophylactic heparin bedside neurorehabilitative care and were closely monitored for any complications. Demographic details, antecedent events, progression of symptoms, laboratory and investigative data, complications in the ICU, and treatment details were entered in a predesigned proforma. The outcome parameters analyzed included the following: (1) in-hospital mortality; (2) GBS disability scale score (Hughes scale: a scale of 0-6 of progressive disability ventilator dependence and death, a score 3 or below indicates that patient is able to walk 5 m).  Patients who were kept in ICU for only observation and monitoring and with incomplete data in case records were excluded from the study.
Motor weakness: The degree of weakness was calculated for each patient by rating the proximal and distal strength of each limb on the British Medical Research Council (MRC) scale (5 = normal, 4 = opposes resistance, 3 = opposes gravity, 2 = moves joint, 1 = flicker, 0 = absent and accepting the lowest score for each limb). 
Autonomic dysfunction: Parameters noted included cardiac arrhythmias, lability of the blood pressure, abnormal hemodynamic response to drugs, electrocardiographic abnormalities, pupillary dysfunction, sweating abnormalities, urinary retention, and gastrointestinal dysfunction. 
Complications in ICU: (1) Electrolyte abnormalities: hyponatremia: sodium < 135 mmols/L, hypokalemia: K < 3.5 mmols/L; (2) Nutritional: hypoalbuminemia = albumin < 35 g/L, anemia = hemoglobin < 120 g/L (at least on 2 occasions); (3) Pulmonary: pneumonia, collapse of lung, pneumothorax as evidenced by abnormalities on the chest radiograph; (4) Sepsis: fever, tachycardia, increased leukocyte count, bacteremia (blood culture), or central access device contamination as defined by positive pathogenic microbial culture; and (5) any 2 or more of the above features in combination.
Death: Day of death: computed from the onset of illness.
Electrodiagnostic studies: Nerve conduction studies included median, ulnar, common peroneal and sural nerves. Difference of >2 standard deviations from the mean values standardized at the institute's laboratory was considered abnormal. The electrophysiological data were analyzed using 3 criteria. ,,
Statistical analysis was performed using SPSS software (Version 11.0). Univariate analysis was done using analysis of variance (ANOVA) for continuous variables and Chi-square test for categorical variables. Chi-square test or Fischer's exact test was done to compare 2 groups (dead vs alive and Hughes scale ≤3 vs >3). The variables which emerged as significant in univariate analysis were used for multivariate analysis using logistic regression analysis to find out which variables were helpful in predicting the outcomes.
Of the 191 GBS patients admitted to ICU during the 10-year study period, 18 patients were excluded from further analysis, 13 patients were admitted for only observation and in 5 patients the data was inadequate. The remaining 173 patients included 118 men and 55 women (age 1-84 years). Of the 173 patients included in the study, history of antecedent events was present in 83 (48%) patients: fever (68), diarrhea (16), respiratory infection (13), recent vaccination (2), surgery (1), and others (14 specific infections: varicella zoster, hepatitis, enteric fever). The time interval between antecedent events and onset of GBS was 1-30 (mean 6.3 ± 5.5) days. Of the 173 patients mechanically ventilated, only 39 (22.5%) had complained of respiratory difficulty at the time of presentation. Paresthesias were reported by 74 patients and objective sensory impairment was noted only in 40 patients. Motor weakness was severe (MRC grade ≤ 2) in 114 (66%) patients at the time of presentation. Bulbar weakness in the form of swallowing or speech problem was noted in 91 (53%) patients. The time required from the onset of the disease to mechanical ventilation was ≤ 3 days in 38 (22%), 4-5 days in 46 (26.5%), and >5 days in 89 (51.5%) patients.
There were 18 deaths (10.4%). The number of days from the onset of the disease to death ranged from 5 to 196 (mean 39 ±.40.3) days. The risk factors determining death are given in [Table 1]. Logistic regression analysis showed significant regression coefficients for age, pneumonia, and dysautonomia. Age had negative regression coefficient and the remaining two variables had positive coefficient. The odds ratio indicated that the risk of mortality was 4.69 times more when pneumonia was present and 3.14 times more when dysautonomia was present. The odds ratio for age was 0.97 indicating that higher age is associated with higher risk of mortality. At the time of discharge 3 patients were ventilator dependent, one of them was in a vegetative state but had to be discharged against advice. Relatives took the other 2 patients to other haspitals. Hughes scale score (HS) at discharge was available for 152 patients. At the time of discharge 47 (30.9%) patients were ambulant with or without support. Factors likely to influence functional status at discharge are given in [Table 2]. Factors associated with Hughes scale score ≤ 3 at discharge were younger age, presence of bulbar symptoms, less severe weakness at presentation, slower evolution of symptoms over more than 3 days, demyelinating neuropathy on electrophysiological studies, absence of sepsis, hypokalemia, and nutritional complications. Binary logistic regression analysis was carried out for predicting the functional status at discharge (Hughes Scale categories: ≤3 and >3). The analysis did not yield significant coefficients for any of these variables except for age. The Nagelkerke R-square value was 0.162. A lack of correlation among the independent variables can be attributed to this type of result. Different treatment modalities, large or small volume plasmapheresis, or intravenous immunoglobulin, did not influence the functional status or mortality.
Electrodiagnostic studies were done in 122 patients and the mean duration of illness at the time of study was 36.9 ± 29.8 days. All the patients had electrophysiological abnormalities [primary demyelinating (86), primary axonal (1), inexcitable (16), and equivocal (19)]. Among the 18 patients who died only 1 patient had undergone electrophysiological studies and he qualified for primary demyelinating group. There was no significant difference in the mean duration of electrophysiological studies among the various groups (35.2 ± 29.9 for primary demyelinating, 55.7 ± 41.4 for in excitable, 35.6 ± 15.5 for equivocal, P = 0.72).
In this study, independent risk factors determining mortality were found to be elderly age group, autonomic dysfunction, pulmonary complications on logistic regression analysis, whereas hypokalemia and bleeding from any site were found to be the risk factors on univariate analysis. The reported mortality rates in ventilated patients ranged between 20% and 38% , as compared to 10.4% in this study. Lawn and Wijdicks  did a comprehensive audit of in-hospital GBS-related deaths and found that deaths most commonly resulted from ventilator-associated pneumonia. In comparison with 101 other patients with severe GBS admitted to the same ICU, the patients who died were older and more likely to have underlying pulmonary disease. In this audit there was a trend toward a greater incidence of autonomic dysfunction in patients who died, however, this had not reached significance. In our study, the higher incidence of autonomic dysfunction in patients who died is similar to the observations in some of the earlier studies.  The present study found no relationship between the presence of bulbar symptoms at presentation and death, similar to our earlier observations.  This may be because all these patients requiring mechanical ventilation would develop bulbar symptoms and they would be managed better in ICU with assisted feeding.
Earlier studies have not analyzed bleeding and hypokalemia as risk factors for mortality. Of the 3 patients with bleeding who died, one was a 64-year-old woman with bleeding per rectum secondary to ischemic colitis for which she underwent hemicolectomy at another facility and was referred back to our facility in a state of sepsis and multiorgan failure and died on day 36 of illness. The second patient was an 11-year-old boy who sustained anoxic brain damage because of delay ventilation and he had upper gastrointestinal (GI) bleed and died of illness on day 18. The third patient was a 35-year-old man who had pneumonia, pneumothorax, and upper GI bleed and died of sepsis on day 59. In him upper GI bleed might probably be related to steroid treatment. It is obvious from this detailed analysis that the death in these 3 patients was not GBS-related and was due to multiple factors, including significant bleeding. Of the 4 patients with bleeding who survived, 3 had upper GI bleed in the form of Ryles tube aspirate and 1 patient had bleeding from the central venous line, which was managed with local pressure. Determining and appropriately treating the cause of significant bleed is important to reduce the bleeding-related mortality in patients with severe GBS.
Hypokalemia is a well-recognized complication of therapeutic plasma exchange.  Of the 18 patients who died, 5 patients had both dysautonomia and hypokalemia and 4 of these patients had a sudden death. It is possible that hypokalemia might have contributed to sudden death by predisposing to cardiac arrhythmias. Dysautonomia is a well-documented feature of severe GBS with respiratory failure in an ICU setup. Sudden deaths have been attributed to the dysautonomia.  Patients with severe GBS need to be carefully monitored for cardiac dysautonomia and hypokalemia to prevent sudden deaths.
In our study, the factors associated with Hughes scale score ≤ 3 at discharge, on univariate analysis, were younger age, presence of bulbar symptoms, and less severe weakness at presentation, slower evolution of symptoms over more than 3 days, primary demyelinating neuropathy in electrophysiological studies, absence of sepsis, hypokalemia, and anaemia. However, on multivariate analysis the only significant variable was age. The lack of significance with other variables may be related to small number of patients in these categories.
Our study confirms that gender as noted by Raphael  and previous infection Pleasure  have no prognostic value. However, these observations are at variance with the more recent studies. [4,6] In this cohort the incidence of diarrheal illness (9.2%) was similar to the Italian study group (8.7%). However, there was no significant correlation with the outcome. This may be related to small number of cases with this antecedent event in our cohort. It is also unclear if there could be etiologically different organisms in different parts of the world explaining varying patterns of recovery.
In our study, the presence of bulbar palsy was one of the variables associated with Hughes scale score ≤ 3 at discharge. This observation is at variance with the findings in other studies. , It was quite probable that the presence of bulbar dysfunction alerted the clinician for close monitoring and mechanical ventilation, thus reducing the risk of associated complications, such as aspiration and pneumonia, which could contribute to delayed recovery. And it is also to be noted that Hughes scale does not take into consideration the bulbar symptoms in assessing the severity of the GBS.
In patients who deteriorate faster possibly the immune-mediated damage to the nervous system occurs much early and extensively and the reversibility of this process is incomplete with various immunomodulatory therapies, accounting for the delayed functional recovery. The poor outcome in patients with rapid worsening observed in this study is similar with most other studies. , Patients with primary demyelinating neuropathy according to the Hadden's classification were likely to fare better in terms of outcome. In GBS inexcitable of nerves could be due to severe conduction blocks or axonal pathology. This is the basis for analyzing the data using compound muscle action potential amplitudes.  Patients with inexcitable nerves had a poor outcome at discharge. Axonal degeneration is independently associated with a more severe disease and less chance of recovery. ,
This study is probably the largest cohort of mechanically ventilated patients with GBS analyzing the prognostic factors. The major limiting factor is that it is a retrospective review of case records of patients admitted over a 10-year period. Analyzing the records spanning a period of 10 years, which has seen a sea change in the ICU care of these patients and the impact of the same on outcome, may also need to be addressed.
[Table 1], [Table 2]