Acceptability of low-volume, induced sputum specimens for diagnosis of Pneumocystis carinii pneumonia.

Two milliliters of induced sputum is the minimum volume recommended for microscopic examination for Pneumocystis carinii by direct fluorescent antibody (DFA) stain. Many specimens received in our laboratory do not meet this criterion. Rejection of these inadequate specimens increases cost and discomfort to the patient because additional specimens must be obtained. To determine whether volumes less than 2 mL were acceptable for microscopic analysis, we examined 177 consecutive induced sputum specimens submitted for P. carinii DFA stain. Eighty-four (47.4%) specimens were less (0.5-1.9 mL) than the 2-mL volume recommended by the manufacturer. Overall, 33 specimens were positive. The positivity rates for specimens 2 mL or more and less than 2 mL were 18% and 19%, respectively. Induced sputum volumes between 0.5 and 2 mL may be acceptable for DFA examination.

Invasive group A streptococcal infections in North Carolina: epidemiology, clinical features, and genetic and serotype analysis of causative organisms.

During 1994 and 1995, an increase in the number and severity of group A streptococcal (GAS) infections was noted in North Carolina. Ninety-six patients had GAS recovered from blood and other sterile body fluids, abscesses, and soft tissue. The overall case fatality rate was 11% but was much higher in patients with toxic shock syndrome (55%) and necrotizing fasciitis (58%). Recent invasive GAS isolates were compared with pre-1994 invasive isolates and temporally related pharyngeal isolates by M protein serotyping, pulsed field gel electrophoresis (PFGE), and polymerase chain reaction amplification of the streptococcal pyrogenic exotoxin A gene. Serotypes M1 and M3 accounted for 50% of recent invasive isolates (1994-1995) and 58% of pharyngeal isolates (1994). The latter isolates demonstrated PFGE patterns that were identical to invasive M1 and M3 strains, suggesting that pharyngeal infections may have served as a reservoir for virulent GAS clones.

Confirmation of the presence of Mycobacterium tuberculosis and other mycobacteria in mycobacterial growth indicator tubes (MGIT) by multiplex strand displacement amplification.

Multiplex strand displacement amplification (mSDA) is capable of amplifying three distinct DNA sequences simultaneously. These include sequences present in most genera of mycobacteria, a sequence specific for Mycobacterium tuberculosis, and an internal control. mSDA was used to detect the presence of these target sequences in 154 (72 positive, 76 negative, and 6 failed) clinical specimens cultured in the mycobacterial growth indicator tube (MGIT) system. A wide variety of specimen types were processed and cultured. Once these cultures were deemed positive by MGIT fluorescence or were deemed negative after 8 weeks of incubation, MGIT culture aliquots were processed for mSDA analyses. A chemiluminescent microwell assay was used to detect the amplified products. The procedure was relatively simple and took less than 6 h to complete. The sensitivity of mSDA for detecting acid-fast bacilli was 96.4% compared to that of MGIT culture. Sensitivity and specificity were 97.2 and 96.1%, respectively, when all clinical criteria were considered. mSDA was shown to be a rapid and effective method for confirming the presence of M. tuberculosis and other mycobacteria in positive MGIT cultures.

Comparison of mycobacteria growth indicator tube with BACTEC 460 for detection and recovery of mycobacteria from clinical specimens.

We compared the Mycobacteria Growth Indicator Tube (MGIT) system with the BACTEC 460 (B460) and Lowenstein Jensen (LJ) systems for the recovery of mycobacteria (acid-fast bacteria [AFB]) from 1,441 clinical specimens. Excluding 13 isolates of Mycobacterium gordonae, 178 significant AFB isolates were recovered from 113 patients. Isolates (119) of the Mycobacterium avium complex (MAC) accounted for 67% of all isolates, while isolates (30) of the Mycobacterium tuberculosis complex (MTB) accounted for 17% of isolates. The MGIT system recovered 98 (82%) MAC and 27 (90%) MTB isolates, while the B460 system recovered 101 (85%) MAC and 28 (93%) MTB isolates and the LJ system recovered 91 (76%) MAC and 25 (83%) MTB isolates. Overall, the MGIT system recovered 152 isolates of AFB (85.4% sensitivity), and the B460 and LJ systems recovered 151 (84.8% sensitivity) and 137 (76.9% sensitivity) AFB isolates, respectively. The recoveries of AFB with combinations of media were as follows: MGIT + LJ, 93.2%; B460 + LJ, 92.1%; and MGIT + B460, 96.6%. Although the sensitivity of MGIT was equivalent to that of B460, MGIT required a longer incubation (median, 11 days) than did B460 (median, 8 days) to become positive (P < 0.05).

Accuracy of four commercial systems for identification of Burkholderia cepacia and other gram-negative nonfermenting bacilli recovered from patients with cystic fibrosis.

Burkholderia cepacia has recently been recognized as an important pathogen in chronic lung disease in patients with cystic fibrosis (CF). Because of the social, psychological, and medical implications of the isolation of B. cepacia from CF patients, accurate identification of this organism is essential. We compared the accuracies of four commercial systems developed for the identification of nonfermenting, gram-negative bacilli with that of conventional biochemical testing for 150 nonfermenters including 58 isolates of B. cepacia recovered from respiratory secretions from CF patients. The accuracies of the four systems for identifying all nonfermenters ranged from 57 to 80%, with the RapID NF Plus system being most accurate. The accuracies of these systems for identifying B. cepacia ranged from 43 to 86%, with the Remel system being most accurate. Depending on the commercial system, from two to seven isolates were misidentified as B. cepacia. The relatively poor performance of the commercial systems requires that identification of certain nonfermenters be confirmed by conventional biochemical testing. These organisms include B. cepacia, Burkholderia sp. other than B. cepacia, and infrequently encountered environmental species (Pseudomonas and Flavobacterium species). In addition, conventional biochemical testing should be done if a commercial system fails to assign an identification to an organism. Confirmatory testing should preferably be performed by a reference laboratory with experience in working organisms isolated from CF patients.

Quality assurance study of bacterial antigen testing of cerebrospinal fluid.

Bacterial antigen testing (BAT) of cerebrospinal fluid (CSF) by latex agglutination is a low-yield procedure in patients whose CSF specimens have normal laboratory parameters. Between August 1992 and August 1994, we evaluated 287 bacterial antigen (BA) test requests to determine whether yields could be improved and whether patient costs could be reduced by cancelling BAT for those patients with normal CSF parameters (cell count, protein, glucose) after consultation with physicians. A total of 171 (68%) BA tests were canceled by this approach. None of these CSF specimens was culture positive for an organism detectable by BAT. Of the remaining 116 CSF specimens tested, only 3 were positive by BAT, one each for Neisseria meningitidis, Streptococcus pneumoniae, and group B streptococcus. Only 43 of the CSF specimens tested had at least two abnormal parameters; the 3 positive CSF specimens were included in this group. In light of the low rate of positivity, the number of BA tests can be further reduced by establishing criteria that must be met before a CSF specimen is accepted for BAT. After review of our data and the literature concerning this topic, we concluded that only specimens with leukocyte counts of > or = 50 cells per mm3 should be tested. Of 287 specimens evaluated in our study, only 36 met this criterion, including the 3 BA-positive specimens. Enacting such a restriction would have reduced the total number of BA tests by 251 (87%) without compromising patient care. A laboratory cost savings of $6,500 per year would have been realized, with a substantial reduction in the cost per positive test. Patient charges would have been reduced by $12,500 per year.

Comparison of antimicrobial susceptibility methods for detection of penicillin-resistant Streptococcus pneumoniae.

We sought to determine if commercially available susceptibility tests were accurate in detecting penicillin resistance and relative resistance in Streptococcus pneumoniae. We compared the reference MIC method with oxacillin disk screening and three commercial tests, E-test (AB Biodisk), JustOne (Radiometer America), and MicroScan Pos MIC Panel Type 6 (Baxter Diagnostics), with 80 selected clinical isolates. Thirty-three additional isolates were tested by the reference method and the E-test to further validate the latter method. Oxacillin screening was effective in detecting all penicillin-resistant and relatively resistant strains of S. pneumoniae. The MicroScan method was not effective in detecting penicillin resistance or relative resistance. The JustOne system classified only 6 (35%) of 17 resistant strains correctly, with 11 resistant strains classified as relatively resistant. The E-test correctly classified 30 (83%) of 36 resistant isolates, with 6 resistant isolates interpreted as relatively resistant. For determining penicillin MICs for S. pneumoniae, the E-test was the most accurate of the commercial systems that we studied.

Evaluation of new monoclonal antibody-based latex agglutination test for detection of cryptococcal polysaccharide antigen in serum and cerebrospinal fluid.

We evaluated the performance of CRYPTO-LEX (Trinity Laboratories, Inc., Raleigh, N. C.), a new mouse immunoglobulin M monoclonal antibody latex agglutination reagent which reacts with the capsular polysaccharide of the four serogroups of Cryptococcus neoformans. This test was compared with CALAS (Meridian Diagnostics, Cincinnati, Ohio) for the ability to detect cryptococcal antigen in serum and cerebrospinal fluid (CSF). A total of 580 clinical specimens (327 serum and 253 CSF samples), primarily from human immunodeficiency virus-infected patients, were tested in this study. Sixty-seven specimens (44 serum and 23 CSF samples) were positive for cryptococcal antigen with both tests, and 511 (282 serum and 229 CSF samples) were negative. The two latex reagents agreed for 326 of 327 serum specimens (44 positives and 282 negatives). One serum specimen with a titer of 1:2 was CALAS positive but CRYPTO-LEX negative. The titer correlation coefficient for the two tests was 0.884 when two highly discordant serum specimens were eliminated from analysis of the data. The two latex tests agreed for 252 of 253 CSF specimens (23 positives and 229 negatives). One specimen with a titer of 1:2 was positive with CALAS and negative by CRYPTO-LEX. The correlation coefficient of the two tests for CSF titers was 0.886. The sensitivity and specificity of CRYPTO-LEX were 97 and 100%, respectively, with a 99.6% correlation with CALAS. These data show that the performance of CRYPTO-LEX is comparable to that of CALAS for detection of cryptococcal antigen in serum and CSF.

Genetic analysis of fructan-hyperproducing strains of Streptococcus mutans.

Fructan polymer, synthesized from sucrose by the extracellular fructosyltransferase of Streptococcus mutans, is thought to contribute to the progression of dental caries. It may serve as an extracellular storage polysaccharide facilitating survival and acid production. It may also have a role in adherence or accumulation of bacterial cells on the tooth surface. A number of clinical isolates of S. mutans which produce large, mucoid colonies on sucrose-containing agar as a result of increased production of fructan have been discovered. By using eight independent isolates, we sought to determine if such fructan-hyperproducing strains represented a genetically homogeneous group of organisms. Restriction fragment patterns of total cellular DNA were examined by using pulsed-field and conventional gel electrophoresis. Four genetic types which appeared to correlate with the serotype of the organism and the geographic site of isolation were evident. Southern blot analysis of several genetic loci for extracellular enzymes revealed some minor differences between the strains, but the basic genomic organizations of these loci were similar. To evaluate whether the excess fructan produced by these strains enhanced the virulence of these organisms in the oral cavity, it was of interest to create mutants deficient in fructosidase (FruA), the extracellular enzyme which degrades this polymer. The fruA gene was inactivated by allelic exchange in two fructan-hyperproducing strains as well as in S. mutans GS5, a strain which does not hyperproduce fructan. All of the fruA mutant strains were devoid of fructan hydrolase activity when levan was used as a substrate. However, the fructan-hyperproducing strains retained the ability to hydrolyze inulin, suggesting the presence of a second fructosidase with specificity for inulin in these strains.

Genetic regulation of fructosyltransferase in Streptococcus mutans.

Streptococcus mutans possesses several extracellular sucrose-metabolizing enzymes which have been implicated as important virulence factors in dental caries. This study was initiated to investigate the genetic regulation of one of these enzymes, the extracellular fructosyltransferase (Ftf). Fusions were constructed with the region upstream of the S. mutans GS5 Ftf gene (ftf) and a promoterless chloramphenicol acetyltransferase (CAT) gene. The fusions were integrated at a remote site in the chromosome, and transcriptional activity in response to the addition of various carbohydrates to the growth medium was measured. A significant increase in CAT activity was observed when glucose-grown cells were shifted to sucrose-containing medium. Sucrose-induced expression was repressed immediately upon addition of phosphoenolpyruvate phosphotransferase system sugars to the growth media. Deletion analysis of the ftf upstream region revealed that an inverted repeat structure was involved in the control of ftf expression in response to carbohydrate. However, the control of the level of ftf transcription appeared to involve a region distinct from that mediating carbohydrate regulation. CAT gene fusions also were constructed with the ftf upstream region from S. mutans V403, a fructan-hyperproducing strain which synthesizes increased levels of Ftf. Sequence analysis of the upstream ftf region in this strain revealed several nucleotide sequence changes which were associated with high-level ftf expression. Comparison of the GS5 and V403 ftf expression patterns suggested the presence of a trans-acting factor(s) involved in modulation of ftf expression in response to carbohydrate. This factor(s) was either absent or altered in V403, resulting in the inability of this organism to respond to the presence of carbohydrate. The sequences of the ftf regions from three additional fructan-hyperproducing strains were determined and compared with that of V403. Only one strain displayed nucleotide changes similar to those of V403. Two additional strains did not have these changes, suggesting that several mechanisms for up-regulation of ftf expression exist.