|Home | About | Journals | Submit | Contact Us | Français|
Bacteriological culture and real-time PCR (RT-PCR) were used to detect Campylobacter jejuni in fecal samples from a French cohort of 237 patients with Guillain-Barré syndrome (GBS). We provide evidence that diverse serotypes and genotypes of C. jejuni are a major trigger of GBS in France.
Guillain-Barré syndrome (GBS) is a postinfectious neurological disease, with a median estimated annual incidence of 1.3/100,000 in developed countries (7). Campylobacter jejuni is the most frequent triggering agent, accounting for 15 to 40% of GBS cases in North America, Australia, and Europe (6, 14, 16, 20). The majority of cases of C. jejuni-associated GBS in Japan involve serotype O:19 (18), and the majority of cases in South Africa involve serotype O:41 (12). A much greater diversity of serotypes (4, 17) and genotypes (4) has been reported in Europe. The pathogenesis of C. jejuni-associated GBS has been linked to antiganglioside autoantibodies generated as a result of the molecular mimicry displayed by ganglioside-like oligosaccharides present in the lipopolysaccharide of certain C. jejuni strains (21). More recently, it has also been linked to the presence of particular classes of the lipo-oligosaccharide (LOS) biosynthesis locus, with most GBS strains having class A or B loci (10, 17).
Little is known about the role of C. jejuni as a triggering agent of GBS in France. In a study of 263 GBS cases in the greater Paris area, France, between 1996 and 2001, serological evidence of recent C. jejuni infection was shown in 22% of patients (16). However, no study has provided direct bacteriological evidence of the role of C. jejuni in French GBS cases, and there are no data on the C. jejuni strains involved. We thus undertook a prospective study on a French cohort of patients with GBS based on the detection of C. jejuni in fecal samples by culture and real-time PCR (RT-PCR).
The medical intensive care unit at the Raymond Poincaré Hospital (Garches, France) is a regional reference center for the management of adult patients with GBS. Since November 1999, all admitted GBS cases have been tested for the presence of C. jejuni in fecal samples (rectal swab and/or stool samples) by culture and RT-PCR, and anti-C. jejuni antibodies have been detected by the complement fixation test (CFT). All the patients included in this study were patients with GBS (2) who were admitted between November 1999 and December 2005 and who had (i) at least one fecal sample (rectal swab or stool sample) examined for the presence of C. jejuni by culture and RT-PCR and (ii) measurement of anti-C. jejuni antibodies. The clinical data were collected on inclusion in the study as described previously (16). Serum samples and rectal swabs were taken from patients on admission. The patients' first stools during their hospitalization were also sampled and studied for C. jejuni culture and RT-PCR. Our ethics committee waived the requirement for informed consent because both the screening of fecal samples for C. jejuni and the measurement of anti-C. jejuni antibodies are routinely carried out and do not affect decisions concerning treatment.
Stool samples were cultured with blood-containing Campylosel (bioMérieux, Marcy l'Etoile, France) (November 1999 to May 2002) or Butzler medium (Bio-Rad Laboratories, Hercules, CA) (after May 2002) with or without prior enrichment in Preston selective broth (Oxoid, Basingstoke, Hampshire, England) at 42°C. Rectal swabs were immersed in Preston enrichment broth, which was incubated at 42°C for 24 h before culture on selective medium. Campylobacter isolates were identified to the species level at the French National Reference Center for Campylobacter and Helicobacter. Strains were serotyped with the Penner O scheme (13). Genotyping involved sequence analysis of the short variable region of the flaA gene (flaA SVR) (4) and the Campylobacter Fla database (http://pubmlst.org/campylobacter/flaA/; hosted at the University of Oxford). LOS biosynthesis loci were classified as previously described (10). Campylobacter DNA was amplified from stool and rectal samples by a custom-designed RT-PCR technique using the LightCycler system (Roche Diagnostics GmbH, Mannheim, Germany). Bacterial DNA from stool samples was purified by using the QIAamp stool DNA minikit (Qiagen GmbH, Germany), and DNA from rectal samples was purified by using the QIAamp DNA minikit (Qiagen). Purified DNA (2 μl) was subjected to amplification with the 16S DNA primers C16S-F (5′-CTAGCTTGCTAGAACTTAGA-3′) and C16S-R (5′-GTCCACACCTTCCTCCTC-3′). We used the LightCycler FastStart DNA master hybridization probe kit (Roche Diagnostics) with fluorescent hybridization probes LC-Camp16S-A (5′-ACGTATTTAGTTGCTAACGGTTC-fluorescein-3′) and LC-Camp16S-B (5′-LightCycler Red 640-GAGCACTCTAAATAGACTGCCTTC-P-3′), synthesized by Proligo (Paris, France). The specificity of amplification was checked by melting curve analysis (melting peaks for Campylobacter species as follows: C. jejuni and C. coli, 64°C with a shoulder effect at 57°C; C. fetus, 57°C; and C. lari, 54°C). Results are expressed qualitatively, as positive or negative (estimated detection threshold of 2 × 102 CFU/g of feces). Serum antibodies against C. jejuni were assayed with complement fixation tests (CFTs) (Virion\Serion, Würzburg, Germany), using a cutoff titer of 20 (>95% specificity; Virion\Serion). Antibodies (IgM/IgG) against gangliosides GM1, GM2, GD1a, GD1b, and GQ1b were detected by enzyme immunoassays (GanglioCombi; Bühlmann Laboratories AG, Schönenbuch, Switzerland). Statistical analyses were performed with R. 2.8.1 software (The R Foundation for Statistical Computing, Vienna, Austria). Categorical variables were compared using Fisher's exact tests, and quantitative variables were compared using Wilcoxon tests. All tests were two tailed, and P values of <0.05 were considered significant.
We included 237 patients with GBS in this study (Table (Table1).1). All patients were assessed for the presence of C. jejuni in fecal samples by culture and RT-PCR (rectal swabs alone, 133 patients; stool samples alone, 23 patients; both samples, 81 patients) and for the presence of serum anti-C. jejuni antibodies with CFT. Sixteen patients (6.8%) gave positive test results with RT-PCR and/or culture; of these, eight were positive by both methods, six were positive by RT-PCR alone, and two were positive by culture alone. C. jejuni was isolated from nine patients, and Campylobacter coli was isolated from one patient. The CFT was positive in 63 patients (26.6%), which included 15 of the 16 patients with positive RT-PCR and/or culture results. Thus, in total, 64 were shown to be associated with C. jejuni (or C. coli) from the 237 patients studied (27%). These cases were more likely to be male, have prodromal diarrhea, be admitted a few days after GBS onset, and have antiganglioside antibodies and a pure motor form than cases with no evidence of recent C. jejuni infection (n = 173) (Table (Table11).
The 16 cases with positive culture and/or RT-PCR results are shown in Table Table2.2. Most patients tested positive between 2 and 10 days after the onset of neurological symptoms. Cases with positive cultures and/or RT-PCR results were more likely to have prodromal diarrhea than the other C. jejuni-associated cases (11/16 [69%] versus 16/48 [33%]; P = 0.02). Stool samples tested positive more often than rectal swabs did (11/104 [10.6%] versus 8/214 [3.7%]; P = 0.02). RT-PCR was more sensitive than culture, although this difference was not significant (14/237 [5.9%] versus 10/237 [4.2%]; P = 0.53). All but one of the 16 patients who tested positive by culture and/or RT-PCR also gave positive results for the C. jejuni CFT, mostly with titers of ≥160 (12/15). Serotype analysis and analyses of flaA SVR sequence polymorphism showed the C. jejuni strains involved to be diverse (Table (Table2).2). The few strains with the same serotype (O:19, cases 3 and 15; O:2, cases 4 and 5) had different flaA SVR sequences, and the few strains with the same flaA SVR sequence (type 3, allele 161, cases 10 and 15) had different serotypes. The recovered C. jejuni strains had class A (n = 4), B (n = 1), C (n = 3), or D (n = 1) LOS loci. Four strains were resistant to quinolones, with no apparent relationship to the markers tested or the use of antibiotics at the time of prodromic infection (no patients were treated with antibiotics).
This study provides evidence that the prevalence of C. jejuni-associated GBS in mainland France is similar to that of neighboring European countries. The C. jejuni recovery rate from stools in our study (6.6%) is close to that reported in England and Wales (7.8%) (14) and in the Netherlands (9%) (20). The seroprevalence rate of 26.6% found in our study is also in the range of those reported in other European studies (3, 6, 8, 9, 14, 20). Cytomegalovirus, the other infectious agent commonly associated with GBS in Western countries (6, 8), accounts for 15% of GBS cases in France (16). Thus, C. jejuni is the most frequent triggering factor of GBS in France, as previously reported for other neighboring European countries (6, 8). Also, the strains of C. jejuni isolated here were diverse in terms of their serotypes and genotypes, consistent with findings from other neighboring countries (4, 5). A strain of C. coli was the only Campylobacter species isolated from one of our patients, who also presented a highly positive CFT for C. jejuni. Similar findings have been reported previously, leading some authors to consider C. coli to be a rare but possible GBS triggering agent (19).
RT-PCR proved to be only slightly more sensitive than culture (5.9% versus 4.2%, respectively) in our study. Thus, although theoretically this method may have high sensitivity, the chances of detecting C. jejuni DNA from fecal samples in patients with GBS remain small, probably because of the length of time between the onset of infection and the time the first samples were obtained. It must be emphasized that real-time PCR and bacterial culture gave much better results from stool samples than from rectal swabs (10.6% versus 3.7%, respectively). This may largely explain the high positive result rate (19%) reported in a previous Indian study in which RT-PCR was performed from stool samples alone (15). Thus, stool analysis must be used systematically in cases of GBS, even if samples are taken late, and especially if there is prodromic diarrhea (20). Collection of two or three stool samples may also improve C. jejuni detection (11). As reported in previous studies (14, 18, 20), serology was by far the most sensitive technique, enabling two-thirds of C. jejuni-associated cases to be identified. To avoid possible interference from polyvalent gamma globulins or plasma exchange therapy received by the patient, the serological technique used must be able to identify a recent C. jejuni infection from a single serum sample taken before therapy (e.g., CFT  or IgA and/or IgM detection ). Under these conditions, serology appears to be particularly suitable for the epidemiological survey of C. jejuni-associated GBS, in association with characterization of C. jejuni strains isolated from stool samples.
We thank I. Sénégas for help with data management, E. Prunier and M.-H. Canneson for technical assistance, and V. Prouzet-Mauléon for providing Campylobacter strains.
This work was supported by l'Institut Garches, la Fondation BNP Paribas, and the Laboratoire Français de Fractionnement et des Biotechnologies.
Published ahead of print on 21 April 2010.