LETTER

Byrne et al. (1) presented data on the superiority of cefoperazone amphotericin teicoplanin (CAT) selective medium over modified cefoperazone charcoal deoxycholate selective medium for the efficient isolation of Campylobacter upsaliensis from stools. There are alternatives to the use of selective media for the isolation of C. upsaliensis. Since 1977 we have routinely isolated campylobacters from the diarrhetic stools of pediatric patients at the Red Cross Children's Hospital, Cape Town, South Africa. In 1990, primarily for cost containment reasons, the use of antibiotic-containing selective media for Campylobacter isolation was discontinued in our laboratory and the Cape Town protocol was introduced. This isolation protocol was the first to combine both membrane filtration onto antibiotic-free blood agar plates and incubation in an H₂-enhanced microaerobic atmosphere (3). With the use of this protocol, the number of stool cultures positive for campylobacteria rose to 21.8% from the 7.1% previously obtained with Skirrow's and other selective media available at that time (3). Since the introduction of the Cape Town protocol we have isolated over 1,200 strains of C. upsaliensis from the diarrhetic and normal stools of pediatric and adult patients and from dogs, cats, and meercats (2). Our laboratory could begin to isolate C. upsaliensis, Campylobacter concisus, Campylobacter curvus, Campylobacter rectus, Campylobacter sputorum biovar sputorum, Campylobacter hyointestinalis, Helicobacter fennelliae, Helicobacter cinaedi, Arcobacter butzleri, and other campylobacteria from the stools of humans and animals only with the introduction of the Cape Town protocol. Some strains of campylobacteria are sensitive to antibiotics commonly used in selective media or have an essential requirement for an H₂-enhanced microaerobic atmosphere.

We have compared the efficacy of the filtration component of the Cape Town protocol with that of CAT selective medium for C. upsaliensis isolation from 300 consecutive diarrhetic stool samples from gastroenteritis patients at the Red Cross Children's Hospital (Table 1). The antibiotic-free filtration and CAT isolation plates were incubated under identical conditions, in an H₂-enhanced microaerobic atmosphere at 37°C. Campylobacter, Helicobacter, and Arcobacter isolates were identified by recognized phenotypic and biochemical criteria. The data in Table 1 indicate that with filtration onto antiobiotic-free plates, 20.3% of the stools were positive for campylobacteria, while with the use of CAT selective plates only 4.7% of the same stools were positive for campylobacteria. Both methods were equally efficient for the isolation of Campylobacter coli and A. butzleri; however, filtration was superior to CAT selective medium for all other campylobacteria isolated. Campylobacter jejuni subsp. doylei, H. fennelliae, C. hyointestinalis, and C. concisus strains were isolated with filtration but were not isolated with CAT media. Sixteen strains of C. jejuni subsp. jejuni were isolated with filtration, whereas nine strains were isolated with CAT medium. Eleven C. upsaliensis strains were obtained with filtration, but only a single C. upsaliensis strain was obtained with CAT medium. Generally, colonies of C. upsaliensis and other campylobacteria on the antibiotic-free blood agar plates used in the Cape Town protocol were larger, more prominent, and faster growing (visible growth after 2 to 4 days) than those on the CAT plates.

Byrne et al. (1) state that membrane filtration is costly and labor intensive. We do not agree, as the Cape Town protocol, which has been in continuous use over the last 11 years, has proved to be a simple, efficient, and cost-effective alternative to the use of antibiotic-containing selective media for the isolation of C. upsaliensis and other campylobacteria from stool. The underdetection of C. upsaliensis and other campylobacteria in the stools of gastroenteritis patients is an important diagnostic problem, and application of the Cape Town protocol may help alleviate this concern.