Detection of non-jejuni and -coli Campylobacter Species from Stool Specimens with an Immunochromatographic Antigen Detection Assay

Abstract

The STAT! Campy immunochromatographic assay for Campylobacter antigen was compared to culture for 500 clinical stool specimens. Antigen was detected in six culture-negative, PCR-positive specimens. C. upsaliensis, a pathogenic species that is traditionally difficult to recover in routine stool cultures, was detected in two of these culture-negative specimens. This study provides evidence that antigen testing may cross-react with at least one additional non-jejuni and -coli Campylobacter species that may be missed by routine culture for campylobacteriosis.

TEXT

Campylobacter spp. are fastidious bacteria that may be difficult to recover in culture due to suboptimal specimen transport and/or storage conditions and lack of universal culture procedures. However, Campylobacter antigens may persist in clinical specimens in the absence of viable organisms. The variable clinical specificity reported for several commercially available Campylobacter antigen detection assays remains unresolved in the literature (1–4). Several studies have reported higher numbers of antigen-positive than of culture-positive specimens (1, 2, 5), but it is unclear whether these are simply false-positive results or are due to Campylobacter spp. in the specimens that are nonviable, viable but nonculturable, or not readily cultivated under commonly employed conditions. This study sought to determine whether culture-negative, antigen-positive specimens, as determined by immunochromatographic assay, actually contained difficult-to-culture Campylobacter spp. which could explain the antigen positivity. In addition, we examined the stability and cross-reactivity profile of the immunochromatographic assay to further define the performance characteristics of the test.

A total of 500 fecal specimens submitted over 4 months (July to September 2011 and June 2012) for routine stool culture or Campylobacter-specific culture were tested with the Immunocard STAT! Campy antigen assay (Meridian Bioscience, Cincinnati, OH). The routine stool culture for Campylobacter was performed by inoculating stool preserved in Cary-Blair transport medium (4°C for ≤72 h) on a Campy CVA plate (Hardy Diagnostics, Santa Maria, CA) and incubated at 42°C in a microaerobic environment (AnaeroPouch-MicroAero, Mitsubishi, Atlanta, GA) for at least 72 h. Colonies resembling Campylobacter were Gram stained for morphology and tested for oxidase and hippurate hydrolysis activities. Oxidase- and hippurate hydrolysis-positive isolates with typical colony and Gram stain morphology were identified as C. jejuni (6). Hippurate hydrolysis-negative isolates were further identified by matrix-assisted laser desorption–ionization time of flight mass spectrometry (7). Antigen-positive stool specimens (Table 1) were defined by the presence of a red-pink band and bands of any other color were interpreted as negative, according to the package insert. Of the 500 specimens analyzed, 5 (1%) samples were culture positive for Campylobacter jejuni, and 15 (3%) were antigen positive, yielding a sensitivity of 80% and specificity of 98% relative to our gold standard of culture (Table 1). Four specimens were both culture and antigen positive, while the remaining culture-positive specimen was antigen negative, and 11 antigen-positive specimens were negative by culture (Table 2). Of the four dual-positive specimens, three were available for stability testing, which was done by repeating culture and antigen detection for 5 consecutive days (the claimed stability period for the antigen assay). All three specimens remained both culture and antigen positive after 5 days of storage at 4°C, indicating that culture and antigen have similar stability characteristics under routine storage conditions. This indicates that the higher observed antigen positivity rate is not likely to be due to instability of culture.

Table 1.

Results for culture versus antigen detection assay

Culture result	No. of samples with STAT! Campy result
	Positive	Negative	Total
No. positive	4	1	5
No. negative	11	484	495
Total no.	15	485	500

Table 2.

Direct Campylobacter spp. PCR and sequencing results for specimens generating discrepant results between antigen detection and culture

Specimen	Result using:
	STAT! Campy	Culture	Stool PCR and sequencinga	C. upsaliensis PCR
CM-508	Positive	C. jejuni	C. jejuni	Negative
CM-105	Positive	C. jejuni	C. jejuni	Negative
CM-168	Positive	C. jejuni	C. jejuni	Negative
CM-44	Positive	C. jejuni	C. jejunib	Positive
CM-216	Positive	Negative	C. gracilisb	Positive
CM-372	Positive	Negative	C. concisusb	Positive
CM-23	Positive	Negative	C. concisus	Negative
CM-136	Positive	Negative	C. concisus	Negative
CM-351	Positive	Negative	C. concisusb	Negative
CM-422	Positive	Negative	C. concisus	Negative
CM-411	Positive	Negative	Negative	Negative
CM-45	Positive	Negative	Negative	Negative
CM-474	Positive	Negative	Negative	Negative
CM-423	Positive	Negative	Negative	Negative
CM-65	Positive	Negative	Negative	Negative
CM-485	Negative	C. jejuni	Negative	Negative

^aIdentifications were based on CLSI guidelines of >99% identity for species-level identification.

^bThese sequences had occasional mixed bases.

Direct stool PCR was performed on antigen- or culture-positive stools. DNA was extracted from stool specimens as previously described (8). Primers targeted a Campylobacter-specific region of the 16S rRNA gene, and PCR products were sequenced for confirmation (9). Of the 15 antigen-positive specimens, 10 were PCR positive for Campylobacter spp. (Table 2). Sequencing these amplicons and analyzing by BLAST using the NCBI refseq_rna database identified the presence of C. jejuni (n = 4), C. gracilis (n = 1), and C. concisus (n = 5). Several of the specimens had potentially mixed sequences, suggesting the presence of more than one species of Campylobacter (Table 2). One specimen was culture positive but antigen and PCR negative. The original CVA culture for this specimen yielded fewer than five colonies and may have been below the limit of detection for the antigen and direct stool PCR, but a commercial Campylobacter real-time PCR confirmed the presence of C. jejuni in this specimen (BioGX, Birmingham, AL, data not shown). Filtration culture was attempted on all antigen-positive specimens (10); however, this method yielded no additional positives beyond the C. jejuni obtained by routine culture. Four of the five false-positive specimens that were negative by PCR and culture but positive by STAT! Campy were noted as having very faint pink bands, which should be interpreted as positive according to the package insert. It was occasionally difficult to clearly define the color of weakly reactive assays, which may lead to some variability in the interpretation of this test (data not shown). The commercial Campylobacter jejuni/coli real-time PCR was performed on these four specimens, and all were negative. The remaining antigen-positive, PCR- and culture-negative specimen showed rapid and strong reactivity immediately after the addition of the specimen. It was tested daily for 10 days with no apparent decrease in reactivity; however, positivity was lost after extended storage at −20°C and multiple freeze-thaw cycles. The source of the strong reactivity in this specimen remains unclear.

The ProSpecT Campylobacter enzyme immunoassay (Remel, Lenexa, KS) was previously shown to detect Campylobacter upsaliensis (3, 4). To assess whether the false positives observed in our study were due to the presence of uncultured C. upsaliensis, species-specific primers were used for direct stool PCR (11). Of the 15 antigen-positive specimens, three were positive for C. upsaliensis. Surprisingly, all three appeared to be coinfected with other Campylobacter spp., including C. jejuni, C. gracilis, and C. concisus (Table 2). Coinfection with multiple species of Campylobacter (particularly C. upsaliensis) has been described previously and is thought to be underreported due to culture procedures that are suboptimal for some Campylobacter spp. (12). To determine the extent of antigen reactivity with non-jejuni and -coli Campylobacter spp., 21 clinical or reference isolates of Campylobacter spp. were suspended in phosphate-buffered saline (3.0 McFarland, approximately 4.3 × 10¹⁰ CFU/ml) and directly tested with the STAT! Campy assay. Of the 10 non-jejuni and -coli Campylobacter species tested, positive results were observed only for C. upsaliensis, and all seven clinical isolates of this species were positive (Table 3). In an attempt to discern whether the reactivity seen in the assay could be clinically relevant, dilutions of C. jejuni and C. upsaliensis cultures were tested. The assay had approximately 10-fold lower sensitivity for C. upsaliensis than for C. jejuni (Table 4). To put this in perspective, the assay is 2.5-fold less sensitive in detecting C. coli, for which it holds FDA clearance, than C. jejuni, according to the manufacturer's package insert. Despite the lower sensitivity, this reactivity for the gastrointestinal pathogen C. upsaliensis may explain a portion of the false positives described in this study. C. upsaliensis may be inhibited by some Campylobacter-selective media, and therefore, it is possible that infections that would otherwise go undiagnosed may be identified using antigen testing. Further prospective studies using culture conditions optimized for the recovery of C. upsaliensis will be required to determine the extent to which this species influences STAT! Campy positivity rates. By PCR and sequencing, we determined that five specimens contained C. concisus; however, our C. concisus isolate failed to cross-react in the antigen detection assay (Table 3). This may indicate that C. concisus does not express the antigen(s) detected by the assay and that its presence in the clinical samples may not be the cause of the positive antigen tests. However, as many as four genomospecies have been proposed among isolates identified as C. concisus (13), and it is possible that other genomospecies could cross-react in vitro. Focused studies involving multiple well-characterized representatives of each genomospecies of C. concisus will be required to better define the presence or extent of antigenic reactivity in this assay. One limitation of this study is that antigen-negative specimens were discarded and thus unavailable for retrospective testing by C. upsaliensis- or C. concisus-specific PCRs. Therefore, we were unable to determine whether any other specimens containing C. upsaliensis or C. concisus were not detected by the STAT! Campy assay.

Table 3.

Analysis of antigenic cross-reactivity for multiple clinical or reference Campylobacter isolates

Campylobacter spp. (no. of isolates)	STAT! Campy result
C. coli (1)	Positive
C. curvus (2)	Negative
C. concisusa	Negative
C. fetus (1)	Negative
C. gracillis (1)	Negative
C. jejuni (1)	Positive
C. lari (2)	Negative
C. rectus (1)	Negative
C. sputorum (1)	Negative
C. upsaliensis (7)	Positive
C. ureolyticus (2)	Negative
C. volurcis (1)	Negative

^aATCC 51562.

Table 4.

Relative sensitivities of antigen detection from C. jejuni and C. upsaliensis cultures

McFarland standard	Result for indicated culture
	C. jejuni	C. upsaliensis
2.0	Positive	Positive
1.0	Positive	Positive
0.5	Positive	Positive
1:10 dilutiona	Positive	Negative
1:100 dilutiona	Negative	NDb

^a0.5 McFarland (1.0 × 10⁸ CFU/ml) was diluted 10-fold and 100-fold, as indicated.

^bND, not determined.

Reactivity with C. upsaliensis was clearly established in this study, but that alone does not account for all false-positive specimens seen. False-positive results may also be explained by variability in test interpretation, as positives may be scored subjectively in terms of whether faint reactivity is truly pink or brown. One published report also suggested that blood in the stool may cause false-positive results (5). Of the 500 specimens tested in our study, 31 (6%) were reported as grossly bloody specimens; however, none of those were positive for Campylobacter antigen, suggesting that bloody stools may not be a significant problem with this assay. Limited ancillary testing was ordered on the majority of these specimens: eight were evaluated by ovum and parasite examination, six by Clostridium difficile PCR, five for Giardia, and two for Cryptosporidium by fecal antigen enzyme-linked immunosorbent assay. Testing for gastrointestinal pathogenic viruses was not performed on any of the specimens. Overall, no causative agent of gastroenteritis was identified by testing at our laboratory for any of the discrepant specimens.

Overall, the STAT! Campy assay may detect infection with Campylobacter spp. other than C. jejuni/coli; however, this attribute alone would not account for all of the perceived false positives encountered in this study. Additional studies using well-characterized and highly sensitive molecular assays capable of identifying a variety of Campylobacter spp. may help clarify the role of antigen testing in clinical diagnostics for campylobacteriosis.