Abstract
Objective: Assessment of incidence rates of Guillain-Barré syndrome (GBS) in the Netherlands over a 10-year period; investigation of a relationship between possible seasonality in GBS and the occurrence of preceding infections; and determination of distinctive characteristics in patients with GBS who are only mildly affected (able to walk unaided at nadir).
Method: Records of patients with GBS admitted between 1987 and 1996 from all 45 hospitals in the southwest Netherlands were evaluated, covering a population of 4.2 million inhabitants.
Results: A total of 476 patients met National Institute for Neurological and Communicative Disorders and Stroke criteria for GBS. This resulted in a crude incidence rate (IR) of 1.18/100,000 inhabitants. This IR increased linearly with age (p < 0.001). Men were more frequently affected than women (p < 0.001). No seasonal preponderance for GBS, nor for any of the preceding infections, was found. Patients under 50 years of age (p < 0.001) and men (p = 0.01) were more frequently found in the mildly affected group. In both groups a preceding infection was reported in 70% of the cases. In the severely affected group, serologic evidence for infection with Campylobacter jejuni, cytomegalovirus, Epstein–Barr virus, or Mycoplasma pneumoniae was found more frequently than in the mildly affected group (41% versus 16%, p = 0.001).
Conclusions: Overall IR in the Netherlands are similar to those found in other studies. The incidence increases linearly with age and men are more frequently affected than women. Distinctive characteristics for mildly and severely affected patients were found regarding age, sex, and preceding infections. This suggests that other infectious agents or host factors may be involved in mild forms of GBS.
Knowledge of the epidemiology of Guillain-Barré syndrome (GBS) is limited with respect to recent findings regarding preceding infections, prognostic factors, and treatment options. Overall, 13 studies on GBS using National Institute of Neurologic and Communicative Disorders and Stroke (NINCDS) criteria have shown incidence rates (IR) between 0.8 and 2.0 per 100,000 inhabitants.1-13 Some of the more recent studies have also reported on the epidemiology of preceding infections.4,6,12 Most information on subpatterns and the pathogenesis of GBS, however, still derives from clinical trials, including our own studies.14,15 These treatment studies have been restricted to the more severely affected patients, i.e., patients unable to walk unaided at nadir (point in time of maximum severity of the disease). Therefore, understanding of GBS is mainly based on this subgroup of patients.
Here we report an epidemiologic study retrospectively performed over a period of 10 years in the southwest Netherlands. We assessed the IR of GBS in the Netherlands and its fluctuations over a period of 10 years. Furthermore, we hypothesized a relationship between seasonal distribution of GBS and preceding infections. It has been suggested that the failure to find a seasonal preponderance of GBS is because there is an opposite seasonal occurrence between the most frequently antecedent events.16 Finally, we assessed the clinical characteristics of the complete spectrum of patients from severely to mildly affected. Knowledge of factors limiting the disease may be of importance in unraveling its pathogenesis and in the development of new treatments.
Materials and methods.
The area and time period.
The survey was carried out in the southwest Netherlands, an area comprising 4.2 million inhabitants, which is about 25% of the population of the Netherlands. It consists of eight well-defined health regions and is considered to be representative for the Netherlands. We investigated all GBS cases admitted to the hospital between January 1, 1987, and December 31, 1996.
Case collection.
After obtaining permission from the neurologists in all 45 hospitals, medical records were collected of patients discharged under the international classification of disease (ICD) code for GBS (ICD-9, 357.0). This code covers the following diagnosis: acute infectious polyneuritis, postinfectious polyneuritis, GBS, and Miller Fisher syndrome. To exclude false-positive cases, each case was re-evaluated by a senior neurologist using NINCDS criteria for GBS.17 To screen for false-negative cases, records were reviewed of patients discharged under ICD codes 357.8 (other specified inflammatory and toxic neuropathies) and 357.9 (not specified inflammatory and toxic neuropathies) in three randomly selected hospitals. These codes are considered to be the most reasonable alternatives when patients are erroneously classified with GBS and discharged.
For each patient data were obtained concerning age, sex, antecedent events, duration of admission, symptoms and signs, treatment, days to nadir, complications, and functional grading score (f-score) at fixed points (nadir, 2 weeks, 8 weeks, and 6 months). The f-score is a disability scale often used in clinical trials concerning GBS.18 Patients were considered to be mildly affected when this f-score was <3 (able to walk unaided at nadir) and severely affected when the maximal f-score was ≥3 (unable to walk unaided at nadir). In the analyses of differences between mildly and severely affected patients, those with an f-score of 6 (dead) were left out (n = 16). Inclusion of grade 6 patients in the severe group could lead to distorted results regarding the duration of hospitalization, days to nadir, and other factors related to the progression of symptoms. The scores 1 to 5 indicate a functional state that patients may reach due to GBS. Attainment of an f-score of 6 is not automatically due to progression and can also occur as a complication of the disease. Inclusion of grade 6 patients would not have influenced the results regarding the differences in serologically proved infections between mildly and severely affected patients (see Results).
Definition of events.
We scored the events most frequently associated with GBS—namely, cancer, surgery, vaccination, and pregnancy19-22—as positive when they occurred within 3 weeks before the start of the first symptoms of GBS.
Data on preceding influenza or influenza-like illness, respiratory tract infection, gastroenteritis, or diarrhea, and other infections, were obtained from chart reviews and were considered positive when patients reported symptoms meeting the criteria for these infections according to the Centers for Disease Control and Prevention definitions for nosocomial infections23 and when they occurred within 3 weeks before the start of the first symptoms of GBS. The following infections were scored from stool and serologic studies: Campylobacter jejuni (stool and serology), Salmonella, Shigella, Yersinia, cytomegalovirus (CMV), Epstein-Barr virus (EBV), and Mycoplasma pneumoniae.
Statistics.
We used the Poisson distribution for calculating 95% confidence intervals for the IR. Differences in IR regarding sex, age group, or year of admission were calculated with a χ2 test without continuity correction. Depending on the data, differences between mildly and severely affected patients were univariately tested using the χ2 test without continuity correction, Fisher’s exact test or the Wilcoxon–Mann-Whitney U test, and multivariately using logistic regression. Outcome was adjusted for age and sex simultaneously using logistic regression. All calculations were performed using STATA 5.0 for Windows 95 (Stata Statistical Software, Release 5.0, College Station, TX). A p value < 0.05 was considered to be significant.
Results.
Population and patient figures.
During the period under investigation, the number of inhabitants increased from 3.8 million in 1987 to 4.2 million in 1996 (Statistics Netherlands, Voorburg, 1998). According to the ICD code, 615 patients were discharged with the diagnosis of GBS. After re-evaluation, 494 patients met NINCDS criteria and 116 patients did not. Among the latter group were patients with chronic inflammatory demyelinating polyneuropathy (n = 24) and patients reporting a typical GBS-like pattern with symptoms and signs too weak to meet NINCDS criteria (n = 18); 13 patient files were not available. Eighteen of the 494 patients had to be excluded because their domicile was outside the borders of the predefined area. Under the codes 357.8 and 357.9, no patients were identified meeting the criteria for GBS. Table 1 shows the characteristics of 476 patients analyzed in this study.
Characteristics of patients (n = 476)
Incidence rates.
The overall crude IR was 1.18 per 100,000 inhabitants (95% CI 1.08 to 1.29). The age-adjusted IR, using the European population as a standard,24 was 1.14/100,000 (95% CI 1.04 to 1.24). The crude IR for men was 1.42 (95% CI 1.26 to 1.59) and for women 0.94 (95% CI 0.82 to 1.09) (p < 0.001). The IR per age group shows a linear increase with age (p < 0.001) (figure 1). The IR from 1987 to 1996 shows only a slight (nonsignificant) decline over the years. No clear seasonal distribution was found although there was a slight increase in incidence in June and during most winter months. No clear (opposite) peaks of occurrence for any of the four infection categories were found throughout the year.
Figure 1. Incidence rates per age group.
Mildly versus severely affected patients.
F-score at nadir.
Figure 2 shows the distribution of days until nadir for the total group, with a median of 8 days (range 1 to 30 days). The median for the mildly affected group was 7 days and for the severe group 9 days (p = 0.003).
Figure 2. Days to nadir.
We were able to obtain the f-score at nadir for 436 patients. A total of 121 patients (28%) were mildly affected and 315 patients (72%) were severely affected (table 2). Significantly more mildly affected patients were men (p = 0.01) and below 50 years of age (p < 0.001) (figure 3). In a logistic regression, age and sex both maintained their influence on being “mildly or severely affected at nadir” (age p < 0.001, sex p = 0.02).
Characteristics of mildly and severely affected patients
Figure 3. Mildly and severely affected patients per age group and gender.
Antecedent events and infections.
Table 3 shows the events and infections preceding GBS in the total group. A total of 72% of the cases reported that an infection preceded the initial symptoms of GBS whereas in 33% a stool culture or serology was positive for one or more agents. Among the 47 other clinically reported infections, the following were reported more than once: pneumonia (7), cystitis (5), sinusitis (3) and otitis media (2), herpes zoster (2), and varicella zoster (2).
Antecedent events and infections
There was no significant difference in reported infections for the mildly and severely affected patients (73% and 71%, respectively). This is in contrast with the serologic results, where in the severe group 90 out of 220 (41%) and in the mild group 9 out of 55 (16%) were positive (p = 0.001, table 4). The results regarding C. jejuni stool cultures differed from these findings; in the mild group, 7 of 29 (24%) cultured stools were positive versus 6 of 100 (6%) in the severe group (p = 0.004).
Serologic screening of infections in mildly and severely affected patients
Multivariate analysis on the outcome of being mildly or severely affected now showed that age ≥50 years (p = 0.06), male sex (p = 0.07), presence of diarrhea (p = 0.09), and positive serology for C. jejuni, CMV, EBV, or mycoplasma pneumonia (p = 0.02) are more frequently associated with the more severe clinical course.
Complications and recurrence.
Complications were found in 67 cases including 20 autonomic dysfunction and 29 complications reported during intensive care admission (pneumonia, atelectasis, deep venous thrombosis, sepsis). Sixteen patients died (3.4%). In 11 patients GBS recurred (2.5%), with a mean period of 16 months between the two episodes (2 to 47 months). One patient had three episodes of GBS (once in 1987, twice in 1991). There was no difference between the mildly and severely affected patients in recurrence rate. There appeared to be no relation between the severity of the first period and the chance of being mildly or severely affected during the subsequent episode.
Outcome.
We scored outcome as being able to walk independently at 8 weeks and 6 months after onset of weakness. Mildly affected patients were by definition able to walk independently; comparison of outcome between the mildly and severely affected groups was therefore not possible.
For the total group, outcome at 8 weeks and 6 months was worse for the group aged ≥50 years (both p < 0.001). Women were less frequently able to walk independently after 8 weeks than were men (p = 0.06); this difference was not found at 6 months. In a multivariate analysis testing age and sex, only age had an effect on the ability to walk independently at 8 weeks (p < 0.001) and at 6 months (p < 0.001).
Discussion.
In this study we investigated the IR and seasonal fluctuations of GBS in the Netherlands and assessed data on the entire spectrum of severity of GBS. When evaluating epidemiologic data based on population surveys it is important to consider the method of case ascertainment, the size of the population, and the time period under investigation. We used accepted diagnostic criteria (NINCDS) with the reasonable assumption that all patients with GBS are seen by a general practitioner and referred to a neurologist. We studied 476 cases recruited in a 10-year period resulting in one of the largest epidemiologic studies on GBS worldwide.
The crude IR of 1.18/100,000 inhabitants and age-adjusted IR of 1.14/100,000 inhabitants found in this survey are in agreement with other studies using NINCDS criteria. The gender difference in IR reaching a significant level has only been reported earlier in two studies where NINCDS criteria were used.3,10
In our study we found a linear increase in incidence with age (p < 0.001) without an additional peak around 20 to 30 years as has been described in other studies.1,4,5,9,13,25-27 Of those studies describing a linear increase with age,2,3,6,8,10 only one reached a significant level.2
In most studies, including ours, the reported IR over the years suggests that there is no change in incidence over time.1,25,26,28 Although there is hardly ever a significant difference between seasons reported, there appears to be a slight lean toward studies reporting the occurrence of GBS more often in the autumn and early winter.1,3,8,11,29 Data in our study are also suggestive for the preference for the winter months. We were unable to confirm the suggestion that a clear lack of seasonal preponderance is due to the fact that the most frequent antecedent infections have opposite seasonality.16
Patients ≥50 years of age had a significantly worse outcome than the younger patients; this is consistent with other population studies.4-6,25 A finding that has not been described earlier is that women had a worse outcome at 8 weeks compared with men. In a multivariate analysis only age had influence on the ability to walk independently both at 8 weeks (p < 0.001) and 6 months (p < 0.001).
We found 28% of the patients to be mildly and 72% to be severely affected. In three other studies the percentages of mildly affected patients were, respectively, 19, 23, and 24.6,30,31 In our study, age and sex were related to the maximum severity of the disease and therefore the chance of being mildly or severely affected. No previous studies report on women being more frequently severely affected than men. There are, however, studies on CMV-related GBS that report a more severe course and a higher incidence in young women.32 We tested this in a multivariate analysis and showed that there was no relationship between the dominance of women in our severe group and the presence of CMV infections. Although older age is often described as a prognostic factor for worse outcome,4-6,25 this is the first study also reporting age as a predictor for a more severe course of the disease.
That more than 70% of the patients reported a preceding infection confirms the established consensus concerning the triggering role of infections in GBS.33-35 Although C. jejuni infections are generally associated with a more severe clinical course,14,36,37 we also found positive serology and positive stool cultures in the mildly affected group. Surprisingly, the percentage of positive stool cultures for C. jejuni was higher in the mild group. Additional information showed that in the mild group, stools were cultured on indication (most patients had clinical gastroenteritis) whereas culturing in the severe group was more often part of a standard procedure. Future prognostic studies are needed to show whether different strains of C. jejuni are responsible for the severity of GBS, and whether host factors further determine the course of the disease. Serologic screening of the four most frequently reported antecedent infections resulted in significantly more positive patients in the severe group compared with the mild group (41% versus 16%). Considering the similar incidence of clinically reported infections, one may conclude that the common known infections preceding GBS are of less importance in mildly affected patients.
Acknowledgments
Acknowledgment
The authors thank Anne Tio for her critical review of the manuscript, as well as the following neurologists for their permission to study the patients’ records: Drs. R.B. Alting van Geusau, W.C. Baard, P.I.J.M. Berntsen, C.T.E. Beljaars, A.M. Boon, H.J. van den Brand, J.L. van Doorn, A.H.C. Geerlings, R. Gelsema, R.J. de Graaf, U.W. Huisman, R.P. Kleijweg, N.K.D. Kok, H.J.M. van der Leeuw, W.H.G. Lieuwens, C. Masterbroek, M.A. Meegens-de Letter, B.J.M. van Moll, H.T.J. Niekus, C. Oppelaar, W.F. van Oudenaarden, E.A.C.M. Sanders, R.L.L.A. Schellens, H.A.W. Sinnige, A.C.G.A. van Spreeken, J.S. Straver, A.L. Strikwerda, J.T.J. Tans, I.E. Tans-de Jongh, G.A.M. Verheul, F.H. Vermeij, F. Visscher, A.P. Volker, H.J. Vroon, A.R. Wattendorff, R. Wielaard, E.J. Wieringa, and D.M.H. Zuidgeest.
- Received August 13, 1999.
- Accepted October 2, 1999.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
