Comparison of a novel microaerobic system with three other gas-generating systems for the recovery ofCampylobacter species from human faecal samples

Three commercial gas-generating systems — CampyGen (Oxoid, UK), Oxoid BR56 (Oxoid, UK), and CampyPak Plus (Becton Dickinson, USA)—and the evacuation replacement technique were compared for the recovery ofCampylobacter spp. from 500 human faecal samples collected from patients with gastroenteritis. Four hundred fifty faecal samples were tested upon receipt in the laboratory. Fifty faecal samples that had been previously found to be positive forCampylobacter spp. were tested retrospectively; these had been stored at 4°C for more than 48 h. A total of 41 (9.1%) of the fresh faecal samples and 41 of 50 (82%) of the stored faecal samples were positive for thermophilic campylobacters. The CampyGen, the Oxoid BR56, the CampyPak Plus, and the evacuation replacement system detectedCampylobacter spp. in 40 (97.6%), 39 (95.1%), 41 (100%), and 41 (100%) of the positive fresh faecal samples and in 37 (90.2%), 40 (97.6%), 39 (95.1%), and 40 (97.6%) of the stored samples, respectively. There was no statistical difference in performance of any of the four gas systems used (p=0.98; chi-square test). Eighty-six percent of the isolates wereCampylobacter jejuni and 14% wereCampylobacter coll. Biotyping and phage typing of the isolates demonstrated that they were of a diverse range of subtypes. This study demonstrates that thermophilic campylobacters can be isolated from human diarrhoeal faecal samples using any of the four microaerobic-atmosphere-generating systems.     

Semiquantitative Polymerase Chain Reaction Enzyme Immunoassay for the Diagnosis of Pertussis

 The two most commonly used targets for diagnosis of pertussis by the polymerase chain reaction have been the pertussis toxin promoter and the repeated insertion sequence IS481. A comparative assessment of these primers was performed on routinely collected nasopharyngeal swabs, stored at –20  °C, using novel semiquantitative enzyme immunoassays. Both sets of primers behaved similarly with bacterial suspensions, and the 17 culture-positive nasopharyngeal swabs were also positive with the pertussis toxin promoter primers, with one exception, which had been subject to prolonged storage. Significantly more of the 69 culture-negative swabs were positive with the pertussis toxin promoter primers (n=36) than with the IS481 primers (n=18). To determine the effect of inhibitors, a comparative assessment of three primer pairs against human DNA (β-globin and glyceraldehyde-3-phosphate dehydrogenase) was also performed.     

Detection of Campylobacter in human and animal field samples in Cambodia

Campylobacter are zoonotic bacteria and a leading cause of human gastroenteritis worldwide with Campylobacter jejuni and C. coli being the most commonly detected species. The aim of this study was to detect Campylobacter in humans and livestock (chickens, ducks, pigs, cattle, water buffalo, quail, pigeons and geese) in rural households by routine culturing and multiplex PCR in faecal samples frozen before analysis. Of 681 human samples, 82 (12%) tested positive by PCR (C. jejuni in 66 samples and C. coli in 16), but none by routine culture. Children were more commonly Campylobacter positive (19%) than adult males (8%) and females (7%). Of 853 livestock samples, 106 (12%) tested positive by routine culture and 352 (41%) by PCR. Campylobacter jejuni was more frequent in chickens and ducks and C. coli in pigs. In conclusion, Campylobacter proved to be highly prevalent by PCR in children (19%), ducks (24%), chickens (56%) and pigs (72%). Routine culturing was insufficiently sensitive in detecting Campylobacter in field samples frozen before analysis. These findings suggest that PCR should be the preferred diagnostic method for detection of Campylobacter in humans and livestock where timely culture is not feasible.     

Serological assessment of synthetic peptides of <Emphasis Type="Italic">Campylobacter jejuni</Emphasis> NCTC11168 FlaA protein using antibodies against multiple serotypes

The flagellum of Campylobacter jejuni is not only responsible for initiating colonization of the gastrointestinal tract of host animals but is also a major antigen that induces protective immune responses. However, protection is limited to the homologous strain and the ability to protect against multiple serotypes has yet to be determined. In this study, we have shown that FlaA is an immmunodominant protein on NCTC11168 CJ1 flagella and we mapped the immunoreactive epitopes on the protein by probing a series of overlapping synthetic peptides spanning the entire sequence with sera against multiple C. jejuni serotypes. Amino acid residues 176–205 (P8), 376–405 (P16) and 501–530 (P21) were immunodominant and cross-reactive. The mucosal IgA in the intestinal secretions of CJ1-infected birds reacted significantly with peptides P16 and P21 indicating that the specificity of the mucosal response is different from the systemic response. Antisera raised against formalin-killed CJ1 cells and purified flagellin showed positive reactivity with a subset of peptides identified by antisera against live C. jejuni. This study provides insight into the specificity of the host immune responses to the FlaA protein of C. jejuni and suggests that these sequences merit further testing for their immunogenicity and potential as subunit vaccine candidates for multiple serotypes.     

Growth rate and trapping efficacy of nematode-trapping fungi under constant and fluctuating temperatures

The effect of temperature on radial growth and predatory activity of different isolates of nematode-trapping fungi was assessed. Four isolates of Duddingtonia flagrans and one isolate of Arthrobotrys oligospora were inoculated on petri dishes containing either corn-meal agar (CMA) or faecal agar and then incubated for 14 days under three different constant and fluctuating temperature regimes. The radial growth was similar on the two substrates at each temperature regime. All fungal isolates showed a higher growth rate at a constant 20 °C. At 10° and 15 °C, all D. flagrans isolates showed very similar patterns of radial growth at both constant and fluctuating temperatures. At 20 °C, they grew significantly faster at constant than at fluctuating temperatures. A. oligospora grew significantly faster than all D. flagrans isolates except when incubated at a fluctuating 20 °C. Spores of each fungal isolate were added to faecal cultures containing eggs of Cooperia oncophora at a concentration of 6250 spores/g faeces. The cultures were incubated for 14 days at the same temperature regimes described above. Control faeces (without fungal material) were also cultured. More larvae were recovered from the fungus-treated cultures incubated at a constant 10° or 15 °C than from those incubated at the respective fluctuating temperatures, except for one D. flagrans isolate. Incubation at 20 °C showed the opposite effect. The general reduction observed in the number of nematode larvae due to fungal trapping was 18–25% and 48–80% for a constant and fluctuating 10 °C, 70–96% and 93–95% for a constant and fluctuating 15 °C, and 63–98% and 0–25% for a constant and fluctuating 20 °C, respectively. Received: 15 December 1998 / Accepted: 16 February 1999     

Confirmed identification and toxin profiling of Campylobacter jejuni using a thermostabilized multiplex PCR formulation

Cytolethal distending toxin (CDT) producing Campylobacter jejuni species are one of the leading causes of human gastroenteritis worldwide. The main intent of the study was to develop a multiplex PCR assay for the confirmed identification and toxin profiling of C. jejuni. The genes targeted were rpo B as genus specific, hip O for species; cdt A, cdt B, cdt C encoding respective subunit proteins of CDT with Internal Amplification Control (IAC). To enhance its application as a pre‐mixed ready‐to‐use format, the master mix of developed mPCR was dried by lyophilization and stability was assessed. Thermostabilized reagents showed stability of 1.5 months at room‐temperature and upto six months at 4 °C without any loss of functionality. The assay was evaluated on a number of presumptive Campylobacter isolates along with biochemical tests. Results obtained indicated the accurate identification of C. jejuni by developed mPCR format in contrast to misconception associated with biochemical assays. The assay was also tested on spiked samples for its real‐time utility. Altogether, the room‐temperature storable and ready‐to‐ use mPCR format developed in this study could be preferred for rapid detection and confirmed identification of toxigenic strains of C. jejuni in place of conventional biochemical assays.     

Artifactual changes in the haematocrit buffy coat of stored anti-coagulated blood samples of farm animals

The buffy coat percentage (BCP) of centrifuged anti-coagulated blood is routinely clinically used to estimate the total leukocyte count (TLC). There had been observations in our clinical laboratory of BCPs that do not correlate with TLC, especially in blood samples in which centrifugation were delayed. This study therefore investigated the artifactual changes that occur in the BCP of stored anti-coagulated blood samples of farm animals. The BCPs of blood samples from a total of 16 cattle, 18 goats, 15 pigs, and 16 chickens were determined immediately upon blood collection to obtain the baseline value (BV). The blood samples were then divided into three parts and stored at 5°C (3–7°C), 30°C (27.5–32.5°C), and 37°C (35.5–38.5°C). Further BCP determinations on the samples were carried out at 24-h intervals for 72 h (3 days). Results showed that there were statistically significant increases (p < 0.05) in BCP of cattle blood samples at all temperatures of storage as from the 48th hour of storage onwards, in goat blood samples stored at 5°C and 37°C at the 72nd hour of storage and as from the 48th hour of storage in the goat blood samples stored at 30°C, in pig blood samples as from the 24th hour of storage onwards for all the storage temperatures, and in chicken blood samples stored at 30°C and 37°C as from the 48th hour of storage onwards. However, clinically significant increases (BCP > 1.5%) occurred only in cattle and chicken blood samples stored at 30°C and 37°C at hour 72 of storage and in pig blood samples stored at 30°C and 37°C as from the 48th hour of storage onwards. It was concluded that statistically significant increases in BCP occur in the blood of cattle, goats, pigs, and chicken during storage, but clinically significant increases that could lead to a false notion of leukocytosis do occur in the BCP of stored anti-coagulated blood of cattle, pigs, and chicken when the blood is stored/kept at 30°C and 37°C.     

Development of a quantitative, competitive polymerase chain reaction–enzyme-linked immunosorbent assay for the detection of Wuchereria bancrofti DNA

A quantitative, competitive polymerase chain reaction (QC-PCR) assay for the sensitive detection of Wuchereria bancrofti DNA was developed. A competitor sequence was constructed by an exchange of nucleotides in the Wuchereria-specific Ssp I repeat. The PCR products were hybridized to specific DNA probes and their amounts, determined by an enzyme-linked immunosorbent assay (ELISA). In laboratory-prepared samples the QC-PCR-ELISA assay was capable of detecting the amount of DNA equivalent to 0.1 microfilaria (mf) added to 200 μl of blood lysate. The assay was also tested on 78 blood samples collected in endemic areas in Egypt. All 28 samples that were positive both for mf and for circulating antigen were also QC-PCR-ELISA-positive. In addition, one mf-negative but antigen-positive sample was also positive as determined by QC-PCR-ELISA. A positive correlation of mf density with the QC-PCR-ELISA was observed. Samples containing 10 or fewer mf/ml had a mean relative amount of Ssp I PCR product of 19.7 units, whereas samples with 11–100 mf/ml had a mean of 36.3 units and those with more than 100 mf/ml had a mean of 84.6 units. Because of the high standard deviation within each group, estimates of worm burdens in infected individuals using the QC-PCR-ELISA are not recommended. However, we present data indicating that the W. bancrofti QC-PCR-ELISA is a powerful new tool for evaluation of parasitic loads for community-based diagnosis of bancroftian filariasis. Received: 15 April 1998 / Accepted: 10 September 1998     

Immunopathology and Th1/Th2 immune response of <Emphasis Type="Italic">Campylobacter jejuni</Emphasis>-induced paralysis resembling Guillain–Barré syndrome in chicken

Immunopathogenesis of Campylobacter jejuni-associated Guillain–Barré syndrome (GBS) is not yet well established probably due to lack of experimental model. Therefore, we studied the Th1/Th2 immune response and pathological changes in C. jejuni-induced chicken model for GBS. C. jejuni (5 × 10⁹ CFU/ml) and placebo were fed to 30 chickens each. Stools of all birds were negative for C. jejuni by culture and PCR before experiment. The birds were regularly assessed for disease symptoms up to 30 days. Sciatic nerves from all chickens were examined at 5 days intervals by histopathology and immunohistochemistry, and also for the expression of Th1/Th2 cytokines. Twenty-two chickens (73.3%) developed diarrhea after C. jejuni infection; 18 (60.0%) experimental chickens developed GBS-like paralytic neuropathy. Pathology in the sciatic nerves of these chickens included perinodal and/or patchy demyelination, perivascular focal lymphocytic infiltration, myelin swelling and presence of macrophages within the nerve fibers on 10th–20th post-infection day (PID). Cytokines (IFN-γ, IL-1β, TNF-α, IL-6 and IL-2) were elevated in early phase (5th–15th PID) and TGF-β2, IL-10 and IL-4 in the recovery phase (25th–30th PID) of the disease. The study provides evidence that C. jejuni infection in the chicken can provide an experimental animal model of GBS.     

Usefulness of culture in the diagnosis ofClostridium difficile infection

Between January and April 1993, culture forClostridium difficile and a faecal cytotoxin assay were performed on 500 selected specimens. Isolates from culture-positive patients from whom faecal samples were cytotoxin negative were also examined in vitro for cytotoxin production. The significance of a positive culture result in the absence of faecal cytotoxin was assessed. Forty-one of the 500 specimens were toxin positive. In only 25 of these wasClostridium difficile examination specifically requested. Six of nine culture-positive cytotoxin-negative patients (11 specimens) had recently received antibiotics. In four of these,Clostridium difficile was considered to be of possible clinical significance. Culture and in vitro determination of toxin production of isolates may aid in the diagnosis of some additional cases, but cytotoxin detection remains the single optimal routine laboratory method for diagnosis.     

Reduced Detection of Chlamydia trachomatis by the Ligase Chain Reaction Assay due to Suboptimal Storage of Urine

Detection of Chlamydia trachomatis by the ligase chain reaction assay was assessed in urine samples that had been stored at 4°C and at ambient temperature for 6–10 days before testing. Six of 67 (9%) ligase chain reaction-positive urine samples stored at 4°C and 5 of 29 (17%) stored at ambient temperature became negative, a difference that is not statistically significant. Most of the urine samples that were negative after storage contained a small number of chlamydial elementary bodies, and almost three-quarters of them were from women. Optimal pretest storage conditions for urine samples should be maintained if the maximum benefit is to be obtained from this highly sensitive assay. Electronic Publication     

Real-time multiplex PCR for simultaneous detection of Campylobacter jejuni, Salmonella, Shigella and Yersinia species in fecal samples

Diarrheal diseases due to notifiable bacterial infections require rapid diagnosis of the causative pathogens. To facilitate detection, a real-time multiplex PCR was developed that identifies common diarrhea-causing bacteria in fecal samples. On the basis of published sequence data, sets of primers and probes were designed that were specific for Campylobacter jejuni, Salmonella, Shigella/enteroinvasive Escherichia coli EIEC, and Yersinia species, suitable for use in a one-tube PCR assay. The assay was assessed using a list of 137 well-defined intestinal bacterial strains or isolates. Furthermore, 393 routine clinical stool samples were analyzed, and the results of real-time multiplex PCR were compared with those obtained by established microbiological methods. The PCR yielded results within 3 h including DNA purification. No false-positive signals or cross-reactions were observed. The analytical sensitivity was 10³ cfu mL⁻¹ for Campylobacter jejuni, 10⁴ cfu mL⁻¹ for Salmonella, and 10⁵ cfu mL⁻¹ for Shigella/EIEC and Yersinia, respectively. Compared with culture, PCR detected 79 out of 81 Campylobacter jejuni (97.5%), 71 out of 74 Salmonella (96%), 8 out of 8 Shigella (100%), and 10 out of 10 Yersinia-positive (100%) clinical samples. In culture-negative samples (n = 192), PCR additionally detected 2 Shigella, 1 Salmonella, and 5 Campylobacter jejuni infections. Thus, the new real-time multiplex PCR provides reliable results within a short time and might be useful as an additional diagnostic tool whenever time is important in the diagnosis of enteropathogenic bacteria.     

Rapid detection ofCampylobacter jejuni andCampylobacter coli isolated from clinical specimens using the polymerase chain reaction

SeventeenCampylobacter strains isolated from 16 children hospitalised with acute diarrhea were analysed by in vitro enzymatic amplification using two sets of oligonucleotide primers specific forCampylobacter jejuni andCampylobacter coli, respectively. Thirteen strains (76 %) were identified asCampylobacter jejuni and four strains (24 %) asCampylobacter coli. Subsequent bacteriological identification confirmed the identity of the same 13Campylobacter jejuni strains and the 4Campylobacter coli strains. Thus, these PCR methods enabled rapid and specific detection of all theCampylobacter jejuni andCampylobacter coli strains without any false-positive or false-negative results.     

Morphological differences in Blastocystiscysts– an indication of different species?

Cyst forms of Blastocystis that show disparate morphology in relation to the previously described cysts were detected in faecal material from animal hosts. Transmission electron microscopy was performed without attempts to isolate or concentrate Blastocystis from the faecal material. Large, multinucleate cyst forms were found in faecal material from Macaca monkeys. These cyst forms measured up to approximately 15 μm in diameter and were often larger than vacuolar forms present in the same samples. Four or more nuclei were frequently seen in the cysts. Multiple individual cysts enclosed by a single fibrillar layer were found in faecal material from domestic chickens. Each individual cyst within the multiple cyst form measured approximately 3–4 μm in diameter and appeared to be uninucleate. Received: 12 August 1996 / Accepted: 6 November 1996     

Paramphistomum daubneyi and Fasciola hepatica: influence of temperature changes on the shedding of cercariae from dually infected Lymnaea truncatula

Dual infections of Lymnaea truncatula with Paramphistomum daubneyi and Fasciola hepatica were performed to determine whether temperature changes in snails (daily water change with spring water at 6°–8 °C, which subsequently increased to room temperature at 20 °C) would influence snail infection and the production of cercariae by both trematodes. At day 30 postexposure the surviving snails were individually placed in petri dishes to constitute two groups. Snails from the first group were maintained at a temperature of 20 °C, and the water in the petri dishes was changed daily. The protocol was identical for the second group of snails except that the water temperature was 6°–8 °C when changed. The interval between exposure and the first shedding of cercariae in snails immersed in cold water for a short period was longer (67–69 days instead of 48–50 days in the 20 °C group). In both groups, snails infected only with F. hepatica or P. daubneyi or with both trematodes were detected. In snails infected only with F. hepatica the frequency of cercaria-shedding snails and the total number of metacercariae were significantly greater in the 20 °C group. Inversely, in snails infected only with P. daubneyi the frequency of cercaria-shedding snails and the number of metacercariae were significantly greater in the 6°–8 °C group. In snails harboring both trematode larval forms, no significant difference in the frequencies of cercaria-shedding snails between the two groups was noted. Metacercariae of both trematodes were obtained from these snails. In the 20 °C group, F. hepatica metacercariae were more numerous, whereas in the 6°–8 °C group the number of P. daubneyi metacercariae was greater. From these results it appears that greater activity of P. daubneyi cercariae occurs in snails subjected to daily temperature changes (from 6° to 20 °C). Received: 24 March 1999 / Accepted: 1 April 1999     

Detection of Campylobacter Species and Arcobacter butzleri in Stool Samples by Use of Real-Time Multiplex PCR

The presence of Campylobacter (or Campylobacter-like) species in stools from patients suspected of infectious gastroenteritis (n = 493) was investigated using real-time PCR for detection of Arcobacter butzleri (hsp60 gene), Campylobacter coli (ceuE gene), Campylobacter jejuni (mapA), five acknowledged pathogenic Campylobacter spp. (C16S_Lund assay), and the Campylobacter genus (C16S_LvI assay). In total, 71.4% of the samples were positive for Campylobacter DNA (n = 352) by a Campylobacter genus-specific (C16S_LvI) assay. A total of 23 samples (4.7%) were positive in the C16S_Lund assay, used for detection of C. jejuni, C. coli, C. lari, C. upsaliensis, and C. hyointestinalis. Subsequent identification of these samples yielded detection frequencies (DF) of 4.1% (C. jejuni), 0.4% (C. coli), and 0.4% (C. upsaliensis). The DF of A. butzleri was 0.4%. Interestingly, sequencing of a subgroup (n = 46) of C16S_LvI PCR-positive samples resulted in a considerable number of Campylobacter concisus-positive samples (n = 20). PCR-positive findings with the C16S_Lund and C. jejuni/C. coli-specific assays were associated with more serious clinical symptoms (diarrhea and blood). Threshold cycle (C_T) values of C. jejuni/C. coli PCR-positive samples were comparable to those of the C16S_Lund PCR (P = 0.21). C_T values for both assays were significantly lower than those of the C16S_LvI assay (P < 0.001 and P < 0.00001, respectively). In conclusion, this study demonstrated that in combination, the C. jejuni/C coli-specific assays and the C16S_Lund assay are both useful for routine screening purposes. Furthermore, the DF of the emerging pathogen C. concisus was at least similar to the DF of C. jejuni.     

Acute hepatitis associated withCampylobacter jejuni bacteraemia

A case of acute hepatitis associated withCampylobacter jejuni bacteraemia is reported. Transaminase levels were increased over 50-fold in a patient with clinical features of enteritis and septicaemia.Campylobacter jejuni was isolated from blood and faecal cultures. Other infective and noninfective causes of acute hepatitis were excluded. The patient's symptoms and liver function values improved after antimicrobial therapy. Hepatitis should be considered as a complication of humanCampylobacter jejuni infection.     

Detection by PCR of eight groups of enteric pathogens in 4,627 faecal samples: re-examination of the English case-control Infectious Intestinal Disease Study (1993–1996)

The English case-control Infectious Intestinal Disease Study (1993–1996) failed to detect an enteric pathogen or toxin in 49% of cases of gastroenteritis. In the present study, polymerase chain reaction (PCR) assays were applied to DNA and cDNA generated from 4,627 faecal samples from cases and controls archived during the original study for the detection of norovirus, rotavirus, sapovirus, Campylobacter spp., Salmonella spp., enteroaggregative Escherichia coli, Cryptosporidium spp., and Giardia spp. The percentage of archived samples from cases and from controls in which at least one agent (or toxin) was detected increased from 53% in the original study to 75% and from 19 to 42%, respectively, after the application of PCR assays. Among cases, the following percentages of enteric pathogens were detected: norovirus 36%, rotavirus A 31%, sapovirus 4%, Salmonella spp. 6%, Campylobacter jejuni 13%, Campylobacter coli 2%, other Campylobacter spp. 8%, enteroaggregative E. coli 6%, Giardia spp. 2%, and Cryptosporidium spp. 2%. The present study provides additional insight into the aetiology of infectious intestinal disease in England and highlights the occurrence of viral infections in cases as well as in asymptomatic individuals. Other notable findings include the frequent presence of Campylobacter spp. other than C. jejuni or C. coli, the high frequency of multiple agents in 41% of cases and in 13% of controls, and the variation in the aetiology and rate of infection found for different age groups. The results demonstrate the greater sensitivity of PCR-based methods compared to current conventional methods.     

Rapid detection ofShigella dysenteriae andShigella flexneri in faeces by an immunomagnetic assay with monoclonal antibodies

A rapid and sensitive method for the detection ofShigella dysenteriae type 1 andShigella flexneri serotypes in faeces based on capture of the bacteria with immunomagnetic particles is described. The particles were coated with either of two different monoclonal antibodies specific for the O-antigens ofShigella dysenteriae type 1 andShigella flexneri serotypes. Captured bacteria were detected by an enzyme immunoassay with O-antigen specific rabbit antiserum. The whole assay required 2 to 3 hours to perform and the sensitivity limit was 10³ cfu/ml as determined by viable cell counting. One hundred and fifty enterobacteria strains, including 100Shigella strains from a strain collection, and 302 fresh faecal samples were used for the study. AllShigella dysenteriae type 1 andShigella flexneri culture-positive faecal samples were positive in the immunomagnetic assay. In addition 18 of 252 culture-negative faecal samples were positive. The immunomagnetic assay was compared with latex agglutination and indirect immunofluorescence using culture as the reference method. The immunomagnetic assay had a sensitivity of 100%, latex agglutination a sensitivity of 72% with 28% false-negative results, and indirect immunofluorescence a sensitivity of 95%. The immunomagnetic assay was superior in sensitivity since it also detectedShigella in faecal samples up to two days after antibiotic therapy had been started, which both latex agglutination and indirect immunofluorescence failed to do. The high sensitivity in detecting live and dead bacteria, and the ease of performance of the immunomagnetic assay render it an attractive method for detection ofShigella.     

Early Identification of Mycobacterium tuberculosis and Mycobacterium avium Using the Polymerase Chain Reaction on Samples Positive by a Rapid Commercial Culture System

 A combination of two methods – a rapid culture method [Mycobacteria Growth Indicator Tube (MGIT); Becton-Dickinson, USA] and a double polymerase chain reaction (PCR) assay – was assessed for the detection and identification of Mycobacterium tuberculosis and Mycobacterium avium from clinical samples. The aim of the study was to evaluate the ability of the system to offer rapid and accurate diagnosis of mycobacterial infections. After decontamination, clinical samples (n=554) were stained and cultured in parallel on solid media and in MGITs following standard procedures. The performance of the two culture systems was compared. Positive MGITs were tested for the presence of Mycobacterium tuberculosis and Mycobacterium avium by PCR of IS6110 (Mycobacterium tuberculosis) and the 16S rRNA gene (Mycobacterium avium). A total of 41 mycobacteria – 27 Mycobacterium tuberculosis isolates, eight Mycobacterium avium isolates, and six other species of mycobacteria – were isolated by one or both culture media. The MGIT system recovered 36 (87.8%) mycobacteria and the solid media 33 (80.4%). The mean time to detection by the two culture systems did not differ overall, but the mean time to detection of Mycobacterium avium from smear-positive specimens was shorter in MGITs than in solid media (5.25 days vs. 16.25 days, P<0.05). The double PCR assay performed on the 36 positive MGITs correctly identified all 24 Mycobacterium tuberculosis-positive MGITs and all six Mycobacterium avium-positive vials. Therefore, application of the PCR assay to positive MGITs may mean that Mycobacterium tuberculosis and Mycobacterium avium can be identified at an earlier stage than with current methods.