Guillain-Barre syndrome in Sweden : from clinical epidemiology to public-health surveillance

Abstract

The Guillain-Barre syndrome (GBS) is an acute inflammatory demyelinating polyradiculoneuropathy of unknown etiology. A suspected diagnosis of GBS usually leads to immediate hospitalization of the patient for early detection and treatment of respiratory failure or cardiac arrhythmia. Antecedent events, usually an infection, occur in around two thirds of GBS patients during weeks before onset. The presumed pathogenesis for GBS is that diverse agents trigger an immune response directed against myelin or axonal antigens, and the immunopathological mechanisms may be predominantly cell-mediated or humoral, or a combination of the two.

This thesis consists of a set of sequential studies on: 1) the descriptive epidemiology of GBS in the populations of South-West Stockholm (SWS), Stockholm County (SC) and Sweden; 2) the clinical epidemiology of GBS in SWS; and 3) the relationship between risk of GBS and pregnancy or postpartum using a cohort design. In addition, a study aimed to assess the requirement of public health surveillance (PHS) of GBS was undertaken. The mean annual incidence rates of GBS per 100,000 population, age-adjusted to the European population, were 1.81 and 1.84 in SWS and in SC during 1973-1991, and 1.77 in Sweden during 1978-1993. In general, the incidence was higher in men (male/female ratio 1.31), and increased with age, presenting a bimodal shape with peaks at 20-24 and 70-74 years in Sweden. Neither clear time trend nor spatial clustering was observed. Moderate seasonality for GBS among the young population, with a peak in late summer or early autumn, was suggested in the aforementioned three populations. The study of annual incidences in SC and Sweden disclosed statistically significant higher rates in 1978 and 1983 for certain age-groups in both populations.

Based on the results from clinicoepidemiological analysis of incident cases in SWS, three major clinicoepidemiological subgroups of GBS in SWS were identified by a hierarchical ascending procedure using complete linkage. One subgroup, with a rapidly progressive course and benign outcome, consisted of young patients and a high proportion of cases with respiratory infection as preceding events, and with normal motor conduction velocity (CV) and less affected biological parameters. A second subgroup, severely affected clinically and functionally, consisted predominantly of young individuals with a high incidence in autumn. A third subgroup was older and, in general, also severely affected.

Forecasts from time series using autoregressive integrated moving average (ARIMA) models fitted to hospital discharge data in SC during 1973-1992 and in Sweden during 1978- 1992, respectively, detected significant variations of GBS incidences in 1993 and displayed a seasonal pattern. When compared with data in retrospect, certain monthly incidences in Sweden and two-monthly numbers of incident cases in SC in 1978 and 1983, located at time-periods when significantly high risks or the outbreak of zimeldine-induced GBS had been identified, exceeded the upper 95 % confidence interval (CI) of the corresponding predicted values. PHS of GBS in Sweden is feasible for populations over 1.5 million and may be worthwhile. We propose a nation-wide GBS surveillance system based on notification by a sentinel network of neurologists in Sweden.

From the results of this study, it is concluded that: 1) the quality of the register data of GBS diagnoses in Sweden is good for epidemiological research and surveillance; 2) the incidence of GBS in Sweden during the last two decades was in magnitude similar to those reported from other populations; 3) in general, rates were stable over the time with occasional significant variation, and no geographical clustering was found; 4) small epidemics of GBS in 1978 and 1983 might have been overlooked; 5) three clinicoepidemiological subgroups were identified; 6) PHS of GBS in Sweden is feasible and may be worthwhile if supported by a clinical network of neurologists; and 7) risk of GBS seems to be lower during pregnancy and increases after delivery.