Evaluation of Pyrosequencing technology for the identification of clinically relevant non-dematiaceous yeasts and related species.

Pyrosequencing was used to identify 133 isolates of clinically relevant non-dematiaceous yeasts. These included 97 ATCC strains (42 type strains), seven UAMH strains, and 29 clinical isolates. Isolates belonged to the following genera: Candida (18 species), Trichosporon (10), Cryptococcus (7), Malassezia (3), Rhodotorula (2), Geotrichum (1), Blastoschizomyces (1), and Kodamaea (1). Amplicons of a hyper-variable ITS region were obtained and analyzed using Pyrosequencing technology. The data were evaluated by a BLAST search against the GenBank database and correlated with data obtained by conventional cycle sequencing of the ITS1-5.8S-ITS2 region. Cycle sequencing identified 78.9% of the isolates to the species level. Pyrosequencing technology identified 69.1%. In 90.1% of all of the strains tested, the identification results of both sequencing methods were identical. Most Candida isolates can be identified to the species level by Pyrosequencing. Trichosporon species and some Cryptococcus species cannot be differentiated at the species level. Pyrosequencing can be used for the reliable identification of most commonly isolated non-dematiaceous yeasts, with a reduction of cost per identification compared to conventional sequencing.

Lack of usefulness of carbon utilization tests for identification of Mycobacterium mucogenicum.

Carbon utilization tests have proven to be useful for the identification of some species of rapidly growing mycobacteria and have been described as one of the few tests useful for the differentiation of Mycobacterium mucogenicum from other rapid growers. We have found the carbon utilization tests to be unreliable for the identification of patient isolates of this species. In this study, using 28 isolates of rapidly growing mycobacteria, we examined several variables which might have an effect on results of citrate, inositol, and mannitol utilization: inoculum concentration, incubation temperature, and medium manufacturer. None of these variables affected results obtained for most species of rapid growers or for ATCC strains of M. mucogenicum. Results for patient isolates of M. mucogenicum were found to be inconsistent regardless of the methodology employed and resulted in an ambiguous identification of these isolates or an incorrect identification as Mycobacterium chelonae. Molecular or cell wall analysis may be the best technique to employ for accurate identification of M. mucogenicum.

A prospective, randomized, double-blind, placebo-controlled trial evaluating the effect of nystatin on the development of oral irritation in patients receiving high-dose intravenous interleukin-2.

Interleukin-2 (IL-2) has been used to treat patients with metastatic melanoma and renal cell cancer for nearly two decades, and much progress has been made in ameliorating its adverse effects. One bothersome adverse effect, oral pain or oral irritation, is usually treated with an oral antifungal antibiotic, nystatin. The authors performed a prospective, randomized, double-blind, placebo-controlled trial involving 64 patients to evaluate the effect of prophylactic administration of nystatin or placebo on the development of oral irritation in patients receiving high-dose intravenous IL-2. No difference was found between patients randomized to receive nystatin or placebo in their rates of development of oral irritation, the severity of IL-2 adverse effects, the duration of their treatment, the rate of development of positive studies for oral yeast, or their pattern of experiencing other adverse effects. Thus, patients who receive high-dose intravenous IL-2 should not be treated prophylactically with nystatin to prevent oral irritation, and clinicians should seek evidence of the presence of oral thrush before using antifungal agents to treat oral pain in these patients.

Multisite reproducibility of Etest for susceptibility testing of Mycobacterium abscessus, Mycobacterium chelonae, and Mycobacterium fortuitum.

A multicenter study was conducted to assess the inter- and intralaboratory reproducibility of the Etest for susceptibility testing of the rapidly growing mycobacteria. The accuracy also was evaluated by comparing Etest results to those obtained by broth microdilution. Ten isolates (four of the Mycobacterium fortuitum group, three of Mycobacterium abscessus, and three of Mycobacterium chelonae) were tested against amikacin, cefoxitin, ciprofloxacin, clarithromycin, doxycycline, imipenem, and trimethoprim-sulfamethoxazole in each of four laboratories. At each site, isolates were tested three times on each of three separate days (nine testing events per isolate) using common lots of media and Etest strips. Interlaboratory agreement among MICs (i.e., mode +/- 1 twofold dilution) varied for the different drug-isolate combinations and overall was best for trimethoprim-sulfamethoxazole (75% for one isolate and 100% for all others), followed by doxycycline and ciprofloxacin. Interlaboratory agreement based on interpretive category also varied and overall was best for doxycycline (100% for all isolates), followed by trimethoprim-sulfamethoxazole and ciprofloxacin. Interlaboratory reproducibility among MICs was most variable for imipenem, and agreement by interpretive category was lowest for imipenem and amikacin. Modal Etest MICs agreed with those by broth microdilution only for doxycycline and the sulfonamides. For all other drugs, the modal MICs by the two methods differed by more than +/- 1 twofold dilution for one or more isolates. In all cases, the Etest MIC was higher and would have caused reports of false resistance. In summary, the Etest in this evaluation did not perform as well as broth microdilution for susceptibility testing of the rapidly growing mycobacteria. It was problematic for most species and drugs, primarily because of a trailing endpoint and/or high MICs compared to broth. Its use will necessitate further investigation, including determination of the optimal medium and incubation conditions and clarification of endpoint interpretation.

Identification of nocardia species by restriction endonuclease analysis of an amplified portion of the 16S rRNA gene.

Identification of clinical isolates of Nocardia to the species level is important for defining the spectrum of disease produced by each species and for predicting antimicrobial susceptibility. We evaluated the usefulness of PCR amplification of a portion of the Nocardia 16S rRNA gene and subsequent restriction endonuclease analysis (REA) for species identification. Unique restriction fragment length polymorphism (RFLP) patterns were found for Nocardia sp. type strains (except for the N. asteroides type strain) and representative isolates of the drug pattern types of Nocardia asteroides (except for N. asteroides drug pattern type IV, which gave inconsistent amplification). A variant RFLP pattern for Nocardia nova was also observed. Twenty-eight clinical isolates were evaluated both by traditional biochemical identification and by amplification and REA of portions of the 16S rRNA gene and the 65-kDa heat shock protein (HSP) gene. There was complete agreement among the three methods on identification of 24 of these isolates. One isolate gave a 16S rRNA RFLP pattern consistent with the biochemical identification but was not identifiable by its HSP gene RFLP patterns. Three isolates gave 16S rRNA RFLP patterns which were inconsistent with the identification obtained by both biochemical tests and HSP gene RFLP; sequence analysis suggested that two of these isolates may belong to undefined species. The PCR and REA technique described appears useful both for the identification of clinical isolates of Nocardia and for the detection of new or unusual species.

Multisite reproducibility of results obtained by the broth microdilution method for susceptibility testing of Mycobacterium abscessus, Mycobacterium chelonae, and Mycobacterium fortuitum.

A multicenter study was conducted to assess the interlaboratory reproducibility of broth microdilution testing of the more common rapidly growing pathogenic mycobacteria. Ten isolates (four Mycobacterium fortuitum group, three Mycobacterium abscessus, and three Mycobacterium chelonae isolates) were tested against amikacin, cefoxitin, ciprofloxacin, clarithromycin, doxycycline, imipenem, sulfamethoxazole, and tobramycin (M. chelonae only) in four laboratories. At each site, isolates were tested three times on each of three separate days (nine testing events per isolate) with a common lot of microdilution trays. Agreement among MICs (i.e., mode +/- 1 twofold dilution) varied considerably for the different drug-isolate combinations and overall was best for cefoxitin (91.7 and 97.2% for one isolate each and 100% for all others), followed by doxycycline, amikacin, and ciprofloxacin. Agreement based on the interpretive category, using currently suggested breakpoints, also varied and overall was best for doxycycline (97.2% for one isolate and 100% for the rest), followed by ciprofloxacin and clarithromycin. Reproducibility among MICs and agreement by interpretive category was most variable for imipenem. Based on results reported from the individual sites, it appears that inexperience contributed significantly to the wide range of MICs of several drugs, especially clarithromycin, ciprofloxacin, and sulfamethoxazole. New interpretive guidelines are presented for the testing of M. fortuitum against clarithromycin; M. abscessus and M. chelonae against the aminoglycosides; and all three species against cefoxitin, doxycycline, and imipenem.

Primary Mycobacterium avium complex infections correlate with lowered cellular immune reactivity in Matschie's tree kangaroos (Dendrolagus matschiei).

The National Zoological Park has maintained a breeding colony of Matschie's tree kangaroos (Dendrolagus matschiei) since 1975 with a documented history and continued prevalence of Mycobacterium avium complex (MAC) infections. No evidence of immunosuppressive retrovirus infections or loss of heterozygosity that may have led to an immune dysfunction in these animals was found. Isolates of MAC organisms from affected tree kangaroos and from their environment had no common restriction fragment DNA types. Cellular immune reactivity in apparently healthy tree kangaroos was 3- to 6-fold lower than in humans and other marsupial and eutherian mammals, as determined by lymphocyte proliferative assays. Thus, while MAC infections are typically opportunistic in humans and other mammals, tree kangaroos commonly develop primary progressive disease with MAC from random sources. Comparative information derived from this study should benefit both the endangered tree kangaroo and humans with immunosuppressive disorders that lead to mycobacterial infections.

Variables affecting results of sodium chloride tolerance test for identification of rapidly growing mycobacteria.

The sodium chloride tolerance test is often used in the identification of rapidly growing mycobacteria, particularly for distinguishing between Mycobacterium abscessus and Mycobacterium chelonae. This test, however, is frequently unreliable for the identification of some species. In this study we examined the following variables: medium manufacturer, inoculum concentration, and atmosphere and temperature of incubation. Results show that reliability is improved if the test and control slants are inoculated with an organism suspension spectrophotometrically equal to a 1 McFarland standard. Slants should be incubated at 35 degrees C in ambient air and checked weekly for 4 weeks. Growth on control slants should be critically evaluated to determine the adequacy of the inoculum; colonies should number greater than 50. Salt-containing media should be examined carefully to detect pinpoint or tiny colonies, and colonies should number greater than 50 for a positive reaction. Concurrent use of a citrate slant may be helpful for distinguishing between M. abscessus and M. chelonae. Molecular methodologies are probably the most reliable means for the identification of rapidly growing mycobacteria and should be used, if possible, when unequivocal species identification is of particular importance.

Clarithromycin lowers plasma zidovudine levels in persons with human immunodeficiency virus infection.

The use of antiretroviral agents and drugs for the treatment and prophylaxis of opportunistic infections has lengthened the survival of persons with AIDS. In the era of multidrug therapy, drug interactions are important considerations in designing effective and tolerable regimens. Clarithromycin has had a significant impact on the treatment of disseminated Mycobacterium avium complex infection, and zidovudine is the best-studied and one of the most widely used antiretroviral agents in this population. We conducted a study to determine the maximally tolerated dose of clarithromycin and the pharmacokinetics of clarithromycin and zidovudine individually and in combination. Mixing studies were conducted to simulate potential interaction in the gastric environment. The simultaneous administration of zidovudine and clarithromycin had little impact on the pharmacokinetics of clarithromycin or of its major metabolite. However, coadministration of zidovudine and clarithromycin at three doses (500 mg orally [p.o.] twice daily [b.i.d.], 1,000 mg p.o. b.i.d., and 2,000 mg p.o. b.i.d.) reduced the maximum concentration of zidovudine by 41% (P < 0.005) and the area under the concentration-time curve from 0 to 4 h for zidovudine by 25% (P < 0.05) and increased the time to maximum concentration of zidovudine by 84% (P < 0.05), compared with zidovudine administered alone. Mixing studies did not detect the formation of insoluble complexes due to chelation, suggesting that the decrease in zidovudine concentrations results from some other mechanism. Simultaneous administration of zidovudine and clarithromycin appears to decrease the levels of zidovudine in serum, and it may be advisable that these drugs not be given at the same time. Drug interactions should be carefully evaluated in persons with advanced human immunodeficiency virus infection who are receiving multiple pharmacologic agents.

Correlation of in vitro susceptibility results for amoxicillin-clavulanate and ampicillin-sulbactam tested against Escherichia coli.

The results of amoxicillin-clavulanate (AUG) and ampicillin-sulbactam (A/S) susceptibility testing by three different susceptibility testing methods, the MicroScan, Etest, and Kirby-Bauer methods, for 61 consecutive isolates of ampicillin-resistant Escherichia coli from different patients were compared. There was poor correlation of results for the two agents, the most and least marked discrepancies being observed by the MicroScan method (86.9% susceptible to AUG and 4.9% susceptible to A/S) and the Kirby-Bauer method (39.4% susceptible to AUG and 32.8% susceptible to A/S), respectively. More organisms were susceptible to AUG than A/S, regardless of the susceptibility testing methodology. The results from a College of American Pathologists survey with one E. coli isolate tested at different institutions also indicated greater susceptibility to AUG than to A/S. These agents are thought to be equally efficacious clinically. The discrepancies observed among methods for each antimicrobial inhibitor combination and the discrepancies observed between the two agents by each testing method suggest that the breakpoints for these agents need to be reevaluated.

Identification of mycobacteria by conventional methods.

The methodology and use of the conventional tests employed in the identification of the currently recognized human mycobacterial pathogens are reviewed. The common disease presentations of each species are briefly noted. Tabular summaries of the phenotypic characteristics of these organisms have also been provided. It should be re-emphasized that the use of conventional methods, unlike the rapid methods now available, is not recommended for the initial identification of the M. tuberculosis complex. We also urge caution in the identification of unfamiliar or atypical isolates. It is to be expected that additional species of human mycobacterial pathogens will be characterized in the future; many of these may be represented by isolates that differ phenotypically little, if at all, from species currently recognized.

Mycobacterial testing in clinical laboratories that participate in the College of American Pathologists Mycobacteriology Surveys. Changes in practices based on responses to 1992, 1993, and 1995 questionnaires.

OBJECTIVE: To determine whether the trend of increasing use of rapid methods of mycobacterial testing among participants in the College of American Pathologists (CAP) Mycobacteriology E Proficiency Testing Survey noted between 1992 and 1993 continued through 1995, and to collect information concerning mycobacterial staining and culture protocols from laboratories that do limited mycobacterial testing. METHODS: The 1993 CAP E Survey questionnaire addressing mycobacterial laboratory practices, test volumes, and rate of recovery of drug-resistant Mycobacterium tuberculosis was included with the CAP 1995 E-A Survey. A shortened list of these same questions, excluding those addressing mycobacterial identification and susceptibility test methods, was added to the CAP 1995 E1-A Survey, to which laboratories that do limited mycobacterial testing subscribe. RESULTS: A total of 802 and 1490 participants in the E and E1 surveys, respectively, returned responses to the CAP by the cutoff date for data analysis. For E Survey participants who answered questions concerning test methods in the years being compared, the percentage who used rapid techniques increased significantly over the study period. More participants used the fluorochrome stain (58% in 1992, 62% in 1993, and 72% in 1995), BACTEC TB plus a solid medium for culture (36% in 1992, 42% in 1993, and 50% in 1995), DNA probes for identification of M tuberculosis (68% in 1993, 79% in 1995), and BACTEC TB for susceptibility testing (65% in 1993, 71% in 1995). The percentages of E1 Survey participants who used a fluorochrome stain for detection of acid-fast bacilli and both a liquid and a solid medium for mycobacterial culture were lower than the percentages of E Survey participants who used these methods. Among participants who responded in all years being compared, the percentage processing respiratory specimens at least 7 times per week increased from 26% in 1992 to 30% in 1993 and 43% in 1995 (P < .001), and the percentages reporting an identification of M tuberculosis within 21 days and susceptibility test results within 28 days increased significantly over the study period (29% in 1992, 40% in 1993, and 56% in 1995 for identification; 13% in 1992, 19% in 1993, and 30% in 1995 for susceptibility testing). Turnaround times for E Survey participants were significantly shorter than those for E1 Survey participants. The number of specimens tested per month appeared to remain relatively stable between 1993 and 1995; however, the number of new patients with tuberculosis and the number of known tuberculosis patients with positive cultures declined significantly. CONCLUSIONS: The recent emphasis placed on utilization of rapid methods of mycobacterial testing appears to have influenced laboratories that subscribe to the CAP E Survey. Significantly more of these laboratories were following the Centers for Disease Control and Prevention's recommendations in 1995 than in 1993 and 1992. However, many laboratories that provide only limited mycobacterial testing still have not adopted the more rapid techniques. Because tuberculosis remains a public health problem, the efforts directed at its control must not wane if the recent downward trend in incidence is to be maintained.

Susceptibility testing of Mycobacterium avium complex in clinical laboratories. Results of a questionnaire and proficiency test performance by participants in the College of American Pathologists Mycobacteriology E Survey.

OBJECTIVES: To obtain information regarding the frequency and methodology of susceptibility testing of Mycobacterium avium complex (MAC) in clinical microbiology laboratories, and to assess interlaboratory reproducibility of MAC susceptibility testing. DESIGN: Questions addressing MAC susceptibility testing were added to the College of American Pathologists' 1994 Mycobacteriology E Proficiency Testing Survey, and participants were asked to complete the questionnaire. In addition, participants in the 1994 E Survey were asked to test susceptibility of a MAC isolate recovered from a proficiency testing specimen as an ungraded exercise if they offered such testing for patients. RESULTS: Of the 1003 participants enrolled in the 1994 Mycobacteriology E-A Survey, 806 responded to one or more supplemental questions. In regard to the demand for MAC susceptibility testing, 606 participants indicated that the test is requested by physicians in their institutions, and 188 said that they do the test routinely on at least one MAC isolate per patient. Eighty-two percent (630/765) of participants refer the test to an outside laboratory, most commonly a commercial reference laboratory or state health laboratory. Of the 70 participants who perform MAC susceptibility testing in-house and indicated the method on the questionnaire, 54 (77%) used a solid medium, whereas only 14 (20%) used BACTEC TB, which currently is the recommended method. The most frequently tested drugs were ethambutol, rifampin, isoniazid, and streptomycin; other commonly evaluated agents were ciprofloxacin, amikacin, and clarithromycin. Only eight participants modify the pH of the medium when testing a macrolide. In regard to reporting test results, 56% (45/80) report a qualitative result only, 35% (28/80) report a quantitative result with a qualitative interpretation, and 9% (7/80) report only a quantitative result. Participant performance on the MAC proficiency testing specimen showed lack of interlaboratory reproducibility; 80% or fewer participants reported the correct result for all drugs except amikacin, for which 92% (11/17) of laboratories responded correctly. CONCLUSIONS: Given the obvious interest in MAC susceptibility testing, standardized methodology that demonstrates interlaboratory reproducibility and, optimally, shows some correlation with clinical outcome is needed. Moreover, recommendations concerning indications for performing the test would be useful.

Effect of PANTA on growth of Mycobacterium kansasii in BACTEC 12B medium.

Mycobacterium kansasii isolates from two patients showed relatively slow growth in BACTEC 12B medium (12B) (Becton Dickinson Diagnostic Instrument Systems, Sparks, Md.) compared with the more rapid growth of these isolates on solid media. This finding prompted an evaluation of the effect of PANTA (Becton Dickinson) on the growth rate of these isolates. Suspensions of one isolate from each of these two patients (A and B), six additional isolates from six other patients (C through H), and one M. kansasii American Type Culture Collection isolate were inoculated into 12B with PANTA, 12B with reconstituting fluid only, and Middlebrook 7H11 agar plates (Remel, Lenexa, Kans.). For the isolates from patients A and B, the average times to detection for 12B with PANTA, 12B with reconstituting fluid, and Middlebrook 7H11 agar were 12.3, 7.4, and 9.0 days, respectively. For the remaining six patient isolates and the American Type Culture Collection strain, the average times to detection for these media were 9.2, 8.1, and 9.6 days. Susceptibility tests performed with the isolates from patients A and B with the individual component antibiotics of PANTA and testing of four of the other isolates with nalidixic acid alone suggested that nalidixic acid exerts some degree of inhibition on the growth of M. kansasii. The eight patient isolates were also inoculated onto Lowenstein Jensen medium (Remel) and onto a variety of selective mycobacterial media containing nalidixic acid and other antimicrobial agents. All isolates showed some degree of inhibition on at least one of these selective media.

Mycobacterial testing in clinical laboratories that participate in the College of American Pathologists' Mycobacteriology E survey: results of a 1993 questionnaire.

Participants in the College of American Pathologists' Mycobacteriology E proficiency testing survey in 1993 were asked to complete a questionnaire addressing mycobacterial test methods, test volume, and frequency of detection of drug-resistant Mycobacterium tuberculosis (MTB). A similar questionnaire had been distributed in 1992. The population responding to the 1993 questionnaire changed, because of a shift of small hospitals to the limited Mycobacteriology E1 survey, and the format of some questions was altered, so a direct comparison of 1992 and 1993 responses was not always possible. Among participants who answered the questions in both years, there was a significant increase in the use of the fluorochrome stain (57% in 1992, 61% in 1993), BACTEC TB for culture (34% in 1992, 38% in 1993) and susceptibility testing (51% in 1992, 61% in 1993), and DNA probes for identification (30% in 1992, 51% in 1993). The percentage of participants who processed respiratory specimens at least seven times per week increased from 9% in 1992 to 13% in 1993, and the percentage processing five times per week increased from 68 to 72%. The percentage of respondents who reported an identification of MTB within 21 days of specimen receipts and susceptibility test results within 28 days in 1992 and 1993 increased from 30 to 41% and from 12 to 19%, respectively. In regard to resistant MTB, 177 institutions in 1991 and 291 in 1992 reported resistance to isoniazid, and 114 in 1991 and 187 in 1992 reported resistance to both isoniazid and rifampin. Laboratorians are to be applauded for using the more rapid mycobacterial testing methods; however, given that tuberculosis remains a problem, this trend must continue.

College of American Pathologists Mycobacteriology E Proficiency Testing Survey. Summary of participant performance, 1979-1992.

The College of American Pathologists first offered a program of proficiency testing in mycobacteriology in 1969 to laboratories that offered any extent of diagnostic service. This program was intended to provide a mechanism for evaluation of methods of staining, culture, identification, and susceptibility testing. From 1979 to 1992, the period covered by this review, participation in the Mycobacteriology E Survey increased almost sixfold. On graded smears to be stained for detection of acid-fast bacilli, more than 85% of Extent 4 and Extent 3 laboratories and more than 80% of Extent 2 laboratories responded correctly to all specimens except one. Performance on specimens that contained Mycobacterium tuberculosis was similar for Extent 4 and Extent 3 laboratories. For all specimens containing M tuberculosis, a mean of more than 90% of Extent 4 and Extent 3 laboratories provided a correct identification each year except 1979, when a mean of 83% of Extent 3 laboratories responded correctly. Only Extent 4 laboratories were required to identify isolates other than M tuberculosis to the species level. For specimens that contained nontuberculous mycobacteria, the means of the yearly averages of correct responses for Extent 4 laboratories were 90% or greater for M kansasii, M marinum, M avium complex, and M fortuitum-chelonae complex and less than 85% for M bovis, M simiae, M scrofulaceum, M szulgai, M flavescens, M xenopi, M terrae complex, and M gastri. In general, on these same specimens, a slightly higher percentage of Extent 3 laboratories (which were required to identify only M tuberculosis to the species or complex level) gave correct or acceptable responses, and the performance of Extent 2 laboratories (which were only required to report whether or not a mycobacterium was present) was the best of all extents. The data suggest that laboratory performance improved somewhat after initial experience with uncommonly encountered organisms. For the most part, however, performance with a given species changed minimally from year to year.

Antimicrobial susceptibilities of mycobacteria as determined by differential light scattering and correlation with results from multiple reference laboratories.

The DAWN Model B laser light scattering instrument (Wyatt Technology Corporation, Santa Barbara, Calif.) was evaluated to assess its potential to provide rapid mycobacterial antimicrobial susceptibility test results. For Mycobacterium tuberculosis there was a clear separation between susceptible and resistant results with the isolates tested, and there was excellent correlation with reference laboratory results. For Mycobacterium avium there was no obvious breakpoint between susceptible and resistant results with the isolates tested, and correlation with reference laboratory results was less good than for M. tuberculosis. However, for M. avium there was also less agreement among reference laboratory results than for M. tuberculosis. Significant instrument design and software program changes would be required for the instrument to become a useful tool for mycobacterial susceptibility testing in the diagnostic laboratory.

Orally administered clarithromycin for the treatment of systemic Mycobacterium avium complex infection in children with acquired immunodeficiency syndrome.

OBJECTIVE: To determine the safety, tolerance, pharmacokinetics, and antimycobacterial activity of orally administered clarithromycin in children with acquired immunodeficiency syndrome and disseminated Mycobacterium avium complex (MAC) infection. DESIGN: Phase I study with a 10-day pharmacokinetic phase followed by a 12-week continuation therapy phase. PATIENTS: Twenty-five patients with a median age of 8.3 years were enrolled. Ten were receiving zidovudine and 13 were receiving didanosine at the time of enrollment. INTERVENTION: Clarithromycin suspension was administered to each patient at one of three dose levels: 3.75, 7.5, and 15 mg/kg per dose every 12 hours. Clarithromycin and antiretroviral pharmacokinetics were measured during single-drug and concurrent-drug administration. Clinical and laboratory monitoring was performed biweekly. MEASUREMENTS AND MAIN RESULTS: Clarithromycin was well tolerated at all dose levels. Plasma clarithromycin concentrations increased proportionately with increasing doses, and significant pharmacokinetic interactions were not observed during concurrent administration with zidovudine or didanosine. Decreases in mycobacterial load in blood were observed only at the highest clarithromycin dose level. Decreased susceptibility to clarithromycin developed rapidly (within 12 to 16 weeks) in the majority of MAC strains isolated from study patients.