Epidemiology and outcome of invasive fungal infection in adult hematopoietic stem cell transplant recipients: analysis of Multicenter Prospective Antifungal Therapy (PATH) Alliance registry

BACKGROUND

With use of data from the Prospective Antifungal Therapy (PATH) Alliance registry, we performed this multicenter, prospective, observational study to assess the epidemiologic characters and outcomes of invasive fungal infection (IFI) in hematopoietic stem cell transplant (HSCT) recipients.

METHODS

Sixteen medical centers from North America reported data on adult HSCT recipients with proven or probable IFI during the period July 2004 through September 2007. The distribution of IFIs and rates of survival at 6 and 12 weeks after diagnosis were studied. We used logistic regression models to determine risk factors associated with 6-week mortality for allogeneic HSCT recipients with invasive aspergillosis (IA).

RESULTS

Two hundred thirty-four adult HSCT recipients with a total of 250 IFIs were included in this study. IA (59.2%) was the most frequent IFI, followed by invasive candidiasis (24.8%), zygomycosis (7.2%), and IFI due to other molds (6.8%). Voriconazole was the most frequently administered agent (68.4%); amphotericin B deoxycholate was administered to a few patients (2.1%). Ninety-three (46.7%) of 199 HSCT recipients with known outcome had died by week 12. The 6-week survival rate was significantly greater for patients with IA than for those with invasive candidiasis and for those with IFI due to the Zygomycetes or other molds (P < .07). The 6-week mortality rate for HSCT recipients with IA was 21.5%. At 6 weeks, there was a trend toward a worse outcome among allogeneic HSCT recipients with IA who received myeloablative conditioning (P = .07); absence of mechanical ventilation or/and hemodialysis (P = .01) were associated with improved survival.

CONCLUSIONS

IA remains the most commonly identified IFI among HSCT recipients, but rates of survival in persons with IA appear to have improved, compared with previously reported data. Invasive candidiasis and IFI due to molds other than Aspergillus species remain a significant problem in HSCT recipients.

Use of the PATH Alliance database to measure adherence to IDSA guidelines for the therapy of candidemia

Candidemia is an increasing complication of the care of complex patients. Adherence to Infectious Diseases Society of America (IDSA) guidelines for the treatment of candidemia was investigated to assess the impact of compliance on outcomes of therapy. Data on the epidemiology, diagnosis, and treatment of patients with invasive fungal infections (IFIs) was extracted from the PATH Alliance registry, a prospective, multicenter, observational database of invasive fungal infections. Patients with proven candidemia were evaluated for adherence to the IDSA guidelines in terms of choice, dosage, and duration of antifungal therapy. Removal of central venous catheters and time to treatment initiation were assessed. Four-week survival data were compared. Of the 418 patients with candidemia who were included in the study, 361 patients with the necessary data sets were identified, of whom 262 (72.6%) were treated within the IDSA guidelines for the treatment of candidemia (IDSA group); the remaining 99 (27.4%) patients received treatment different from the guidelines (non-IDSA group). Kaplan-Meier (KM) survival analysis for patients in the IDSA and non-IDSA group showed mortality rates of 21.9 and 13.6%, respectively; the difference between the two groups was not statistically significant (P = 0.093). Following the exclusion of patients requiring mechanical ventilation or acute cardiac support, the modified survival KM curves were similar between the two groups. Significantly more patients in the IDSA group required mechanical ventilation and tunneled central catheters, had a concomitant IFI, and received caspofungin. The duration of treatment and inappropriate dosing did not appear to have had a significant impact on survival. Most of the deviations from IDSA guidelines were due to the inappropriate duration and dosing of therapy. No significant difference in mortality was noted between the IDSA and non-IDSA groups. The basis of these differences merit further study.

Quality control and reference guidelines for CLSI broth microdilution susceptibility method (M 38-A document) for amphotericin B, itraconazole, posaconazole, and voriconazole

Although standard conditions are available for testing the susceptibilities of filamentous fungi to antifungal agents by the Clinical and Laboratory Standards Institute (CLSI; formerly National Committee for Clinical Laboratory Standards) broth microdilution assay, quality control (QC) MIC limits have not been established for any mold-agent combination. This multicenter (eight-center) study documented the reproducibility of tests for one isolate of Paecilomyces variotii ATCC MYA-3630 and 11 other mold isolates (three isolates of Aspergillus fumigatus; two isolates of A. terreus; one isolate each of A. flavus, A. nidulans, Fusarium moniliforme, and F. solani; and two isolates of Scedosporium apiospermum) by the CLSI reference broth microdilution method (M 38-A document). Control limits (amphotericin B, 1 to 4 microg/ml; itraconazole, 0.06 to 0.5 microg/ml; posaconazole, 0.03 to 0.25 microg/ml; voriconazole, 0.015 to 0.12 microg/ml) for the selected QC P. variotii ATCC MYA-3630 were established by the analysis of replicate MIC results. Reference isolates and corresponding MIC ranges were also established for 6 of the 12 molds evaluated. MIC limits were not proposed for the other five molds tested due to low testing reproducibility for these isolates.

Multicenter comparison of the Sensititre YeastOne colorimetric antifungal panel with the NCCLS M27-A2 reference method for testing new antifungal agents against clinical isolates of Candida spp

A multicenter (three centers) study compared MICs obtained by the Sensititre YeastOne Colorimetric Antifungal plate to reference microdilution broth (NCCLS M27-A2 document) MICs of three new triazoles (posaconazole, ravuconazole, and voriconazole) and the echinocandin caspofungin acetate for 100 isolates of Candida spp. In addition, amphotericin B and fluconazole were tested as control drugs. Colorimetric MICs of caspofungin and amphotericin B corresponded to the first blue well (no growth), and MICs of the other agents corresponded to the first slightly purple or blue well. Two comparisons of MIC pairs by the two methods were evaluated: 24-h colorimetric MICs were compared to NCCLS MICs at 24 and at 48 h. The interlaboratory reproducibility of YeastOne and reference MICs was also examined. The best performance of the YeastOne plate was with 24-h MICs (overall, 95 to 99% agreement) for all the species and antifungal agents. These results suggest the potential value of the YeastOne plate for use in the clinical laboratory for the four new antifungal agents evaluated.

Quality control limits for fluconazole disk susceptibility tests on Mueller-Hinton agar with glucose and methylene blue

An international collaborative study was performed in order to propose quality control limits for fluconazole disk diffusion susceptibility tests on Mueller-Hinton agar with 2% glucose and 0.5 micro g of methylene blue per ml. The supplements may be added before autoclaving the agar, or Mueller-Hinton agar plates may be flooded with a glucose-methylene blue solution. Replicate tests on both types of agar plates generated data to propose zone size limits for tests of Candida parapsilosis ATCC 22019 (22 to 33 mm), C. tropicalis ATCC 750 (26 to 37 mm), and C. albicans ATCC 90028 (28 to 39 mm). C. krusei ATCC 6258 was not useful for this purpose.

Comparison of porphyrin-based, growth factor-based, and biochemical-based testing methods for identification of Haemophilus influenzae

The accurate identification of Haemophilus spp. is essential for optimizing the role of the clinical microbiology laboratory in the diagnosis and management of Haemophilus infections. One laboratory-prepared medium and eight commercially available test systems were examined in parallel as a means of identifying 378 clinical isolates of Haemophilus spp. as either Haemophilus influenzae or non- Haemophilus influenzae spp. At least one discordant result was noted with 187 (49.5%) of the isolates tested. Discordant results were resolved either by majority rule for isolates with less than three discordant test results or by confirming the identity using conventional biochemical tests for isolates with three or more discordant test results ( n=20). Among these 20 isolates, 2 were judged not to belong to the Haemophilus genus. Comparisons of three porphyrin-based methods, three growth factor-based methods (1 of which also incorporates a porphyrin testing component), and three biochemical-based methods revealed varying discrepancy rates within each testing method. In general, porphyrin-based methods, with overall discrepancy rates of 1.3% or less, outperformed other testing methods. One important exception was the performance of the porphyrin testing component of the Haemophilus Identification Test Kit (Remel, USA), which produced an overall discrepancy rate of 28.5% and a false-negative rate of 52.2% with non- Haemophilus influenzae isolates. Growth factor-based methods yielded overall discrepancy rates ranging from 1.6% ( Haemophilus Identification Agar Quad; Remel) to 10.4% (hemin and nicotinamide adenine dinucleotide disk component of the Haemophilus Identification Test Kit). Biochemical-based assays produced overall discrepancy rates ranging from 4.5% (API NH; bioMérieux Vitek, USA) to 10.1% ( Neisseria Haemophilus Identification Card; bioMérieux Vitek). Collectively, these results suggest that porphyrin-based testing methods represent the most reliable means for identifying Haemophilus spp.

Risk factors for Candida species colonization of neonatal intensive care unit patients

BACKGROUND

Candida spp. are increasingly important pathogens in neonatal intensive care units (NICU). Prior colonization is a major risk factor for candidemia, but few studies have focused on risk factors for colonization, particularly in NICU patients.

METHODS

A prospective, multicenter cohort study was performed in six NICUs to determine risk factors for Candida colonization. Infant gastrointestinal tracts were cultured on admission and weekly until NICU discharge and health care worker hands were cultured monthly for Candida spp.

RESULTS

The prevalence of Candida spp. colonization was 23% (486 of 2157 infants); 299 (14%), 151 (7%) and 74 (3%) were colonized with Candida albicans, Candida parapsilosis and other Candida spp., respectively. Multiple logistic regression analysis adjusting for length of stay, birth weight < or = 1000 g and gestational age < 32 weeks revealed that use of third generation cephalosporins was associated with either C. albicans (155 incident cases) or C. parapsilosis (104 incident cases) colonization. Use of central venous catheters or intravenous lipids were risk factors for C. albicans, whereas delivery by cesarean section was protective. Use of H2 blockers was an independent risk factor for C. parapsilosis. Of 2989 cultures from health care workers' hands, 150 (5%) were positive for C. albicans and 575 (19%) for C. parapsilosis, but carriage rates did not correlate with NICU site-specific rates for infant colonization.

CONCLUSIONS

We speculate that NICU patients acquire Candida spp., particularly C. parapsilosis, from the hands of health care workers. H2 blockers, third generation cephalosporins and delayed enteral feedings alter gastrointestinal tract ecology, thereby facilitating colonization.

Cloning and characterization of a complex DNA fingerprinting probe for Candida parapsilosis

Candida parapsilosis accounts for a significant number of nosocomial fungemias, but in fact, no effective and verified genetic fingerprinting method has emerged for assessing the relatedness of independent isolates for epidemiological studies. A complex 15-kb DNA fingerprinting probe, Cp3-13, was therefore isolated from a library of C. parapsilosis genomic DNA fragments. The efficacy of Cp3-13 for DNA fingerprinting was verified by a comparison of its clustering capacity with those of randomly amplified polymorphic DNA analysis and internally transcribed spacer region sequencing, by testing species specificity, and by assessing its capacity to identify microevolutionary changes both in vitro and in vivo. Southern blot hybridization of EcoRI/SalI-digested DNA with Cp3-13 provides a fingerprinting system that (i) identifies the same strain in independent isolates, (ii) discriminates between unrelated isolates, (iii) separates independent isolates into valid groups in a dendrogram, (iv) identifies microevolution in infecting populations, and (v) is amenable to automatic computer-assisted DNA fingerprint analysis. This probe is now available for epidemiological studies.

Elevated phenotypic switching and drug resistance of Candida albicans from human immunodeficiency virus-positive individuals prior to first thrush episode

Strains of Candida albicans obtained from human immunodeficiency virus (HIV)-positive individuals prior to their first episode of oral thrush were already in a high-frequency mode of switching and were far more resistant to a number of antifungal drugs than commensal isolates from healthy individuals. Switching in these isolates also had profound effects both on susceptibility to antifungal drugs and on the levels of secreted proteinase activity. These results suggest that commensal strains colonizing HIV-positive individuals either undergo phenotypic alterations or are replaced prior to the first episode of oral thrush. They also support the suggestion that high-frequency phenotypic switching functions as a higher-order virulence trait, spontaneously generating in colonizing populations variants with alterations in a variety of specific virulence traits.

Risk factors for candidemia in Neonatal Intensive Care Unit patients. The National Epidemiology of Mycosis Survey study group

BACKGROUND

Candida species are important nosocomial pathogens in neonatal intensive care unit (NICU) patients.

METHODS

A prospective cohort study was performed in six geographically diverse NICUs from 1993 to 1995 to determine the incidence of and risk factors for candidemia, including the role of gastrointestinal (GI) tract colonization. Study procedures included rectal swabs to detect fungal colonization and active surveillance to identify risk factors for candidemia. Candida strains obtained from the GI tract and blood were analyzed by pulsed field gel electrophoresis to determine whether colonizing strains caused candidemia.

RESULTS

In all, 2,847 infants were enrolled and 35 (1.2%) developed candidemia (12.3 cases per 1,000 patient discharges or 0.63 case per 1,000 catheter days) including 23 of 421 (5.5%) babies < or =1,000 g. After adjusting for birth weight and abdominal surgery, forward multivariate logistic regression analysis demonstrated significant risk factors, including gestational age <32 weeks, 5-min Apgar <5; shock, disseminated intravascular coagulopathy, prior use of intralipid, parenteral nutrition, central venous catheters, H2 blockers, intubation or length of stay > 7 days before candidemia (P < 0.05). Catheters, steroids and GI tract colonization were not independent risk factors, but GI tract colonization preceded candidemia in 15 of 35 (43%) case patients.

CONCLUSIONS

Candida spp. are an important cause of late onset sepsis in NICU patients. The incidence of candidemia might be decreased by the judicious use of treatments identified as risk factors and avoiding H2 blockers.

A head-on comparison of the in vitro antifungal activity of conventional and lipid-based amphotericin B: a multicenter study

A comparative study of conventional amphotericin B, Abelcet and AmBisome was performed using a microdilution format of the NCCLS M27-A methodology for susceptibility testing against 300 fungal isolates (152 yeasts, 148 filamentous fungi) in both RPMI-1640 and antibiotic medium #3 (AB3). The clinical isolates included Candida albicans (n=54), Candida glabrata (n=25), Candida parapsilosis (n=23), Candida krusei (n=19), Candida lusitaniae (n=14), Cryptococcus neoformans (n=5), Candida tropicalis (n=12), Aspergillus flavus (n=34), Aspergillus fumigatus (n=46) and 68 other filamentous fungi encompassing 22 different genera. The minimal inhibitory concentrations (MIC) for all drugs were defined as the lowest concentrations in which there was no visible growth. MICs were determined after 48 h for yeasts and 72 h for filamentous fungi. The mean MICs +/- standard error (microg/ml) for yeasts and filamentous fungi, respectively, were: Abelcet, 0.51+/-0.21, 4.34+/-0.61; AmBisome, 1.28+/-0.24, 5.68+/-0.57; amphotericin B, 0.29+/-0.11, 1.12+/-0.19, respectively. Overall, against both yeasts and filamentous fungi Abelcet proved to have more potent antifungal activity than AmBisome. Using AB3 as opposed to RPMI-1640 generally produced lower MIC values but did not have any effect on the order of relative activity with all of the antifungal agents tested. In conclusion, our data shows that Abelcet is more active than AmBisome against pathogenic yeast and filamentous fungi when assayed in AB3 in vitro. Comparison of the activities of these antifungals in experimental animal models is necessary to determine whether these in vitro findings are correlated with in vivo efficacy.

National epidemiology of mycoses survey (NEMIS): variations in rates of bloodstream infections due to Candida species in seven surgical intensive care units and six neonatal intensive care units

Candida species are the fourth most frequent cause of nosocomial bloodstream infections, and 25%-50% occur in critical care units. During an 18-month prospective study period, all patients admitted for > or = 72 hours to the surgical (SICUs) or neonatal intensive care units (NICUs) at each of the participant institutions were followed daily. Among 4,276 patients admitted to the seven SICUs in six centers, there were 42 nosocomial bloodstream infections due to Candida species (9.8/1,000 admissions; 0.99/1,000 patient-days). Of 2,847 babies admitted to the six NICUs, 35 acquired a nosocomial bloodstream infection due to Candida species (12.3/1,000 admissions; 0.64/1,000 patient-days). The following were the most commonly isolated Candida species causing bloodstream infections in the SICU: Candida albicans, 48%; Candida glabrata, 24%; Candida tropicalis, 19%; Candida parapsilosis, 7%; Candida species not otherwise specified, 2%. In the NICU the distribution was as follows: C. albicans, 63%; C. glabrata, 6%; C. parapsilosis, 29%; other, 3%. Of the patients, 30%-50% developed incidental stool colonization, 23% of SICU patients developed incidental urine colonization, and one-third of SICU health care workers' hands were positive for Candida species.

Multicenter comparison of the sensititre YeastOne Colorimetric Antifungal Panel with the National Committee for Clinical Laboratory standards M27-A reference method for testing clinical isolates of common and emerging Candida spp., Cryptococcus spp., and other yeasts and yeast-like organisms

National Committee for Clinical Laboratory Standards (NCCLS) standard guidelines are available for the antifungal susceptibility testing of common Candida spp. and Cryptococcus neoformans, but NCCLS methods may not be the most efficient and convenient procedures for use in the clinical laboratory. MICs of amphotericin B, fluconazole, flucytosine, itraconazole, and ketoconazole were determined by the commercially prepared Sensititre YeastOne Colorimetric Antifungal Panel and by the NCCLS M27-A broth microdilution method for 1,176 clinical isolates of yeasts and yeast-like organisms, including Blastoschizomyces capitatus, Cryptococcus spp., 14 common and emerging species of Candida, Hansenula anomala, Rhodotorula spp., Saccharomyces cerevisiae, Sporobolomyces salmonicolor, and Trichosporon beigelii. Colorimetric MICs of amphotericin B corresponded to the first blue well (no growth), and MICs of the other agents corresponded to the first purple or blue well. Three comparisons of MIC pairs by the two methods were evaluated to obtain percentages of agreement: 24- and 48-h MICs and 24-h colorimetric versus 48-h reference MICs. The best performance of the YeastOne panel was with 24-h MICs (92 to 100%) with the azoles and flucytosine for all the species tested, with the exception of C. albicans (87 to 90%). For amphotericin B, the best agreement between the methods was with 48-h MIC pairs (92 to 99%) for most of the species tested. The exception was for isolates of C. neoformans (76%). These data suggest the potential value of the YeastOne panel for use in the clinical laboratory.

Molecular epidemiology and antifungal susceptibility of Cryptococcus neoformans isolates from Ugandan AIDS patients

Little is known of the antifungal susceptibility patterns and molecular epidemiology of Cryptococcus neoformans from tropical regions. We studied 164 clinical isolates of C. neofomans from 120 Ugandan AIDS patients with cryptococcal meningitis by analyzing their electrophoretic karyotypes and antifungal susceptibility profiles. Computer-assisted analysis of karyotype patterns was performed to generate dendrograms. MICs of fluconazole and flucytosine were determined by reference methods. A total of 43 distinguishable DNA types were identified among the 164 isolates. Only 30 patients (25%) were infected with their own unique strain of c. neoformans, whereas 75% of the patients shared their infecting strain with at least one other patient. Among 17 patients with more than one CSF isolate of C. neoformans, sequential isolates were identical or highly related in 12 (71%) and were different in five patients (29%). The isolates were susceptible to both fluconazole and flucytosine and there were no instances in which a stepwise increase in either fluconazole or flucytosine MICs was observed among serial isolates. These findings suggest that the epidemiology of cryptococcal disease in AIDS patients from tropical regions may be somewhat different from that observed in more temperate climates.

Effects of sucralfate vs antacids on gastric pathogens: results of a double-blind clinical trial

BACKGROUND

Unblinded studies suggested that sucralfate prophylaxis for stress ulcers is associated with a lower rate of nosocomial pneumonia than acid-reducing approaches. We performed a randomized, double-blind, double-sham clinical trial comparing the exact microbial effects of each treatment.

METHODS

One hundred forty patients entered this study before major elective surgery, allowing baseline cultures of gastric and pulmonary secretions to be obtained intraoperatively. Postoperatively, the patients were treated with standard doses of either sucralfate or antacids, plus a sham of the other drug. Cultures were repeated twice daily for 3 days. Molecular epidemiological typing was used to track the appearance of specific microbes and their transmission from site to site, and clinical end points were compared. The number of patients chosen was for sufficient statistical power to detect differences in the microbial measures, as detecting differences in clinical measures would have required increasing the sample size by an order of magnitude.

RESULTS

Gastric pH was affected by the form of stress ulcer prophylaxis throughout the study, and this pH effect affected the number of new gastric organisms appearing in the 2 different groups. Colonization of the airway with new gastric organisms occurred more frequently in the antacid than in the sucralfate group, and colonization of the airway with organisms of gastric origin was associated with occurrence of postoperative pneumonia.

CONCLUSIONS

Both sucralfate and antacids offered safe and effective stress ulcer prophylaxis in this double-blind clinical trial of postoperative patients in an intensive care unit. In association with the drug's effects on gastric pH, more new pathogens appeared in the gastric contents of antacid-treated than sucralfate-treated patients.

Multicenter evaluation of broth microdilution method for susceptibility testing of Cryptococcus neoformans against fluconazole

We have developed a microdilution method for measuring the susceptibility of Cryptococcus neoformans to fluconazole. The present study evaluated the interlaboratory agreement of the results for the microdilution method obtained at three different sites and compared this method with the National Committee for Clinical Laboratory Standards M27-P reference method. Excellent interlaboratory agreement among the results obtained at the three sites was achieved with this method (83 and 96% agreement within 1 and 2 log2 dilutions, respectively). An overall agreement of 90% between the microdilution method and the M27-P method was observed, demonstrating the comparability of the two methods. However, there are inherent problems with the M27-P method in relation to measuring C. neoformans susceptibility, including suboptimal growth of the organism in RPMI 1640, a longer incubation period, and a narrow range of MICs. On the basis of these data, the microdilution method tested in this study is recommended for inclusion in the National Committee for Laboratory Standards method for testing the antifungal susceptibility of C. neoformans.

Interlaboratory evaluation of Etest method for testing antifungal susceptibilities of pathogenic yeasts to five antifungal agents by using Casitone agar and solidified RPMI 1640 medium with 2% glucose

An interlaboratory evaluation (two centers) of the Etest method was conducted for testing the antifungal susceptibilities of yeasts. The MICs of amphotericin B, fluconazole, flucytosine, itraconazole, and ketoconazole were determined for 83 isolates of Candida spp., Cryptococcus neoformans, and Torulopsis glabrata. Two buffered (phosphate buffer) culture media were evaluated: solidified RPMI 1640 medium with 2% glucose and Casitone agar. MIC endpoints were determined after both 24 and 48 h of incubation at 35 degrees C. Analysis of 3,420 MICs demonstrated higher interlaboratory agreement (percentage of MIC pairs within a 2-dilution range) with Casitone medium than with RPMI 1640 medium when testing amphotericin B (84 to 90% versus 1 to 4%), itraconazole (87% versus 63 to 74%), and ketoconazole (94 to 96% versus 88 to 90%). In contrast, better interlaboratory reproducibility was determined between fluconazole MIC pairs when RPMI 1640 medium rather than Casitone medium was used (96 to 98% versus 77 to 90%). Comparison of the flucytosine MICs obtained with RPMI 1640 medium revealed greater than 80% reproducibility. The study suggests the potential value of the Etest as a convenient alternative method for testing the susceptibilities of yeasts. It also indicates the need for further optimization of medium formulations and MIC endpoint criteria to improve interlaboratory agreement.

The use and abuse of routine stool microbiology: a College of American Pathologists Q-probes study of 601 institutions

OBJECTIVE

To examine the efficiency with which physicians use routine stool microbiology tests.

DESIGN

Questionnaire and structured review of 100 consecutive stool bacteriology and parasitology examinations at each participating institution.

SETTING

Six hundred one institutions enrolled in the College of American Pathologists Q-probes Program.

RESULTS

Of 59500 bacteriology specimens, 3808 (6.4%) contained a pathogen. The vast majority (99%) of bacterial pathogens were detected in either the first or second specimen submitted. Almost 40% of inpatient specimens were collected after the third day of hospitalization, but only 0.6% of these specimens were positive for enteric pathogens that had not been previously recovered. More than half of the laboratories reported having no limits on the number of bacteriology specimens per patient that could be submitted for testing, and fewer than 8% of laboratories rejected specimens from inpatients after a certain number of days in the hospital. The frequency with which laboratories performed tests for Clostridium difficile varied widely. Of 58500 parasitology specimens, 1463 (2.5%) contained a pathogen; 97.6% of pathogens were detected by the second stool specimen, and 99.8% were detected by the third specimen. Only 0.7% of specimens from inpatients hospitalized more than 4 days contained a new pathogen.

CONCLUSIONS

We recommend that no more than two bacteriology specimens and no more that two or three parasitology specimens be processed per patient without consultation. Standard stool examination for a bacterial pathogens has a low yield and should not be performed after 3 days of hospitalization. Likewise, parasitology examinations should not be performed after 4 days of hospitalization.

Comparative antimicrobial activities of the penem WY-49605 (SUN5555) against recent clinical isolates from five U.S. medical centers

The in vitro activity of WY-49605 (SUN5555) (WY) was compared with those of cefaclor, cefixime, and amoxicillin-clavulanic acid against 2,958 consecutive clinical isolates from five medical centers and 402 respiratory pathogens from 18 other facilities. Most members of the family Enterobacteriaceae were inhibited by WY (MIC at which 50% of the isolates are inhibited [MIC50], < or = 2.0 micrograms/ml). MIC90s of > or = 8.0 micrograms/ml were observed for Enterobacter cloacae, Serratia spp., and Proteus mirabilis. WY was the most active drug against methicillin-susceptible Staphylococcus aureus (MIC90, 0.12 microgram/ml) and other coagulase-negative staphylococci (MIC90, 4.0 micrograms/ml). The four drugs were not active against nonenteric gram-negative bacilli, methicillin-resistant Staphylococcus aureus, and Staphylococcus haemolyticus. At 2.0 micrograms/ml, WY inhibited 82% of Enterococcus faecalis strains and was equal to or superior to the other drugs against streptococci, Haemophilus influenzae, and Moraxella catarrhalis.

Comparative and collaborative evaluation of standardization of antifungal susceptibility testing for filamentous fungi

The purpose of the study was to evaluate the interlaboratory agreement of broth dilution susceptibility methods for five species of conidium-forming (size range, 2 to 7 microns) filamentous fungi. The methods used included both macro- and microdilution methods that were adaptations of the proposed reference method of the National Committee for Clinical Laboratory Standards for yeasts (m27-P). The MICs of amphotericin B, fluconazole, itraconazole, miconazole, and ketoconazole were determined in six centers by both macro- and microdilution tests for 25 isolates of Aspergillus flavus, Aspergillus fumigatus, Pseudallescheria boydii, Rhizopus arrhizus, and Sporothrix schenckii. All isolates produced clearly detectable growth within 1 to 4 days at 35 degrees C in the RPMI 1640 medium. Colony counts of 0.4 x 10(6) to 3.3 x 10(6) CFU/ml (mean, 1.4 x 10(6) CFU/ml) were demonstrated in 90% of the 148 inoculum preparations. Overall, good intralaboratory agreement was demonstrated with amphotericin B, fluconazole, and ketoconazole MICs (90 to 97%). The agreement was lower with itraconazole MICs (59 to 79% median). Interlaboratory reproducibility demonstrated similar results: 90 to 100% agreement with amphotericin B, fluconazole, miconazole, and ketoconazole MICs and 59 to 91% with itraconazole MICs. Among the species tested, the MICs for S. schenckii showed the highest variability. The results of the study imply that it may be possible to develop a reference method for antifungal susceptibility testing of filamentous fungi.

Efficacy of microbial identification system for epidemiologic typing of coagulase-negative staphylococci

The lack of an adequate typing system hampers our understanding of the epidemiology of infections caused by coagulase-negative staphylococci (CoNS). CoNS have become recognized as important nosocomial pathogens and the principal cause of infections associated with invasive devices. Sensitive, specific, and convenient methods are needed to evaluate whether implementing infection control guidelines reduces the risk of nosocomial infections from CoNS and other pathogens. The Microbial Identification System (MIDI) (Microbial ID Inc., Newark, Del.), a semiautomated system for fatty acid methyl ester analysis, shows considerable promise for clinical and epidemiologic applications. Its predictive accuracy and reliability were tested by using epidemiologically related and replicated CoNS isolates as well as CoNS from epidemiologically unrelated clinical infections, which were obtained from five established hospital culture collections in diverse geographic locations. Two hundred isolates were fully characterized in 5 days by one person using MIDI, and the results were similar to those produced by more expensive and time-consuming conventional typing methods. MIDI appears to be a useful screening tool that could be used before more expensive and labor-intensive molecular methods. It offers important advantages to hospital epidemiologists and clinical microbiologists who must identify and type CoNS isolates.

Resistance of Candida albicans to fluconazole during treatment of oropharyngeal candidiasis in a patient with AIDS: documentation by in vitro susceptibility testing and DNA subtype analysis

We describe a patient with recurrent episodes of oropharyngeal candidiasis who required progressively higher doses of fluconazole to control and infection. The patient was treated for 14 infections over a 2-year period with doses of fluconazole that ranged from 100 to 800 mg per day. Clinical response, two methods of in vitro susceptibility testing, and molecular epidemiologic techniques were evaluated for 12 of the 14 episodes. Ultimately, the patient became unresponsive clinically to a dose of 800 mg of fluconazole per day. In vitro susceptibility testing of isolates obtained during these successive episodes of infection revealed the development of resistance to fluconazole, and molecular epidemiologic techniques confirmed the persistence of the same Candida albicans strain throughout all 12 episodes.

Evaluation of two commercial systems for identification of coagulase-negative staphylococci to species level

Bloodstream (224) and urine (nine) isolates of coagulase-negative staphylococci (CNS) were evaluated using MicroScan Pos ID and Rapid Pos ID panels. A modification of the conventional method of Kloos and Schleifer served as the reference method. The isolates were selected to include a broad range of CNS species, including 44 S. epidermidis, 50 S. hominis, 39 S. warneri, 33 S. capitis, 21 S. haemolyticus, 12 S. simulans, 11 S. saprophyticus, six S. cohnii, five S. lugdunensis, three S. xylosus, four S. auricularis, two S. schleiferi, two S. intermedius, and one S. sciuri. The Pos ID panel had an overall rate of agreement (correct plus probably correct) with the reference method of 79%, including 95% for S. epidermidis, 95% for S. haemolyticus, 64% for S. hominis, 67% for S. simulans, 79% for S. warneri, and 100% for S. saprophyticus. The Rapid Pos ID panel had an overall rate of agreement (correct plus probably correct) of 76%, including 91% for S. epidermidis, 90% for S. haemolyticus, 64% for S. hominis, 58% for S. simulans, 77% for S. warneri, and 100% for S. saprophyticus. Both systems are acceptable for the identification of the clinically significant species S. haemolyticus, S. saprophyticus, and S. epidermidis, but are less reliable for the infrequently isolated species of CNS.

Postoperative pneumonia: a prospective study of risk factors and morbidity

BACKGROUND

Postoperative pneumonia (PP) is a major complication that has been linked to microaspiration of pathogens originating in the gastrointestinal tract. This prospective study was performed to define the role of gastric bacterial aspiration in the development of PP.

METHODS

Informed consent was obtained before operations from 140 veterans scheduled for major elective procedures requiring nasogastric tubes, and cultures were obtained of the gastric contents and sputum twice daily after operation.

RESULTS

PP developed in 26 (18.6%) of 140 patients. The patients who had PP did not differ from those with no pneumonia after operation except for a history of chronic obstructive pulmonary disease, which was found in 38.5% of those with PP compared with 20% of patients with no pneumonia (p < 0.05). Morbidity was increased in association with PP, because length of stay in the surgical intensive care unit (6.2 vs 2.6 days), days intubated (2.7 vs 0.6), total postoperative days (15.3 vs 8.4), and mortality rates (19.2% vs 1.7%) were greater than in patients with no pneumonia. Gastric pathogens were present on entry in 38% of patients, and 32% of these had PP compared with 13% whose initial gastric cultures were sterile (p = 0.01). Colonization of sputum for greater than 24 hours with gastric pathogens occurred in 28% of patients. These patients had a 40% incidence of PP compared with 12% in patients without such evidence of microaspiration (p < 0.01).

CONCLUSIONS

PP is a morbid postoperative complication associated with not only chronic obstructive pulmonary disease but also the presence of gastric bacteria during operation and transmission of gastric bacteria to the pulmonary tree after operation.

Ampicillin-sulbactam susceptibility testing criteria

In vitro studies in five different medical centers documented the susceptibility of 2,440 consecutive isolates of the Enterobacteriaceae against ampicillin-sulbactam disks of different potencies. For determination of MICs, both 2:1 or 1:1 ratios were used as long as the concentrations of sulbactam at the breakpoints remained the same, i.e. MIC < or = 16/8.0 micrograms/ml or < or = 8.0/8.0 micrograms/ml for the susceptible category. Disks containing 10 micrograms of ampicillin and 10 micrograms of sulbactam are still to be preferred with interpretive criteria of > or = 15 mm for susceptible and < or = 11 mm for resistant (MIC > or = 64/32 micrograms/ml or > or = 32/32 micrograms/ml). The reliability of the disk test actually diminished when the amount of sulbactam in the disk was increased.

Relative efficacy of tazobactam, sulbactam and clavulanic acid in enhancing the potency of ampicillin against clinical isolates of Enterobacteriaceae

Three beta-lactamase inhibitors were combined with ampicillin in a fixed 2:1 ratio. The activity of ampicillin was enhanced by tazobactam and by clavulanic acid, and to a lesser extent by sulbactam when tested against fresh clinical isolates of Enterobacteriaceae. At a concentration of 8 micrograms/ml, ampicillin alone inhibited 49.6% of 2,434 consecutive isolates of enteric bacilli compared to 81% inhibited by ampicillin combined with tazobactam or clavulanic acid and 69.3% inhibited by the sulbactam/ampicillin combination. A four-fold or greater reduction in ampicillin MICs was observed in comparable numbers of isolates with all three combinations, but the most marked effects were seen with strains that were highly resistant to ampicillin.

Impact of the changing epidemiology of fungal infections in the 1990s

The increase in fungal infections over the past decade is striking. This is particularly true for hospitalized patients where the rate of candidal bloodstream infection has increased by as much as 487% over the decade of the 1980s. This increase in fungal infections is accompanied by a significant excess mortality and excess length of stay in hospital. The emergence of "new" fungal pathogens such as Candida krusei, Torulopsis glabrata, Fusarium and Trichosporon beigelii is now recognized as a significant problem in many patient populations. The documentation of nosocomial transmission of fungal pathogens and the recognition of resistance to both new and established anti-fungal agents poses a significant problem entering the 1990s. Continued effort is needed to develop new and better therapeutic agents and more effective strategies for prophylaxis of endogenous infections and prevention of transmission within the hospital setting.

Methods for estimating days of hospitalization due to nosocomial infections

While there is little agreement at the individual patient level of analysis, estimates of mean NI-attributed days of stay for the two methods were essentially the same. The lack of agreement at the individual patient level may reflect fundamental differences in the methods used to derive these estimates: incorporation of noninfected patient data versus exclusive reliance on data from infected patients; and, focus on length of stay rather than the actual care being received. The potential advantages of the AEP-based method include the following: 1) all patients with NI can be included in developing estimates; 2) estimates are based on the care provided rather than simple length of stay differences; 3) data on which to form the NI-day estimates are readily available in the medical record; 4) the AEP is a validated and commonly used utilization review instrument; 5) the AEP-based method has acceptable reliability; 6) this method is designed to provide individual and group estimates of NI-attributed days; 7) because every day of stay is reviewed, additional information is available, which results in greater precision of study of the development, diagnosis, and treatment of the NI relative to the other care that originally brought the patient into the hospital. The AEP-based method for estimating NI-days is a promising alternative to the historical cohort approach. Additional applications of this approach are encouraged to further assess its reliability,validity, and additional information yield.

Effects of fluconazole on the sterol and carbohydrate composition of four species of Candida

The effects of fluconazole, a bis-triazole antifungal agent, on the sterol and carbohydrate composition of Candida albicans, Candida tropicalis, Candida krusei and Candida parapsilosis were investigated. Exposure of Candida species to fluconazole resulted in a profound depletion of ergosterol with a corresponding increase in lanosterol content versus control cells. Carbohydrate analysis revealed a significant increase in chitin and either a decrease (Candida albicans, Candida tropicalis and Candida parapsilosis) or an increase (Candida krusei) in glucan content in fluconazole-treated cells. The decreased ergosterol and increased lanosterol content is consistent with 14-alpha-demethylase inhibition by fluconazole. The increase in cell wall chitin is most likely due to deregulation of chitin synthesis secondary to ergosterol depletion in the cell membrane. Because chitin, glucan and ergosterol are critical components of the fungal cell, perturbation of the production and localization of these components by fluconazole is likely to contribute to the selective toxicity of this compound to Candida species and other fungi.

In vitro selection of resistance to vancomycin in bloodstream isolates of Staphylococcus haemolyticus and Staphylococcus epidermidis

The purpose of the study was to determine whether vancomycin-resistant strains of Staphylococcus haemolyticus could be selected regardless of the initial MIC of vancomycin. Twenty-one bloodstream isolates of Staphylococcus haemolyticus were studied by broth and agar selection methods. The broth method selected strains for which MICs of vancomycin ranged from 4 to 32 micrograms/ml and MBCs from 16 to greater than 128 micrograms/ml. The agar method selected strains for which MICs ranged from 8 to 32 micrograms/ml and MBCs from 8 to greater than 128 micrograms/ml. For comparison, seven strains of Staphylococcus epidermidis were evaluated by the agar selection method. Final MICs of vancomycin ranged from 8 to 16 micrograms/ml; MBCs ranged from 16 to 64 micrograms/ml. Clearly, in vitro exposure to vancomycin can select strains of Staphylococcus haemolyticus and Staphylococcus epidermidis for which MIC values are beyond the susceptible breakpoint.

Assigning responsibility: using feedback to achieve sustained control of methicillin-resistant Staphylococcus aureus

Sustained control of endemic methicillin-resistant Staphylococcus aureus (MRSA) originating from multiple sources has not been reported. We describe a simple, inexpensive program based on feedback to physicians that resulted in significant reduction of nosocomial MRSA. When nosocomial cases were identified, the epidemiologist contacted the team resident to encourage increased emphasis on hand washing. Handouts, periodic hand cultures of house staff, and monthly presentations at morning report were also employed. In the first 15 months, nosocomial MRSA decreased from 1.025 to 0.508 cases per 1,000 patient days (p less than 0.01). Monthly rates were significantly decreased for 9 months of 1989 and the first 2 months of 1990. Feedback and assignment of responsibility resulted in a 50% reduction in nosocomial MRSA that has been sustained for 15 months.

Application of DNA probes to antimicrobial susceptibility testing of Legionella pneumophila

The commercial availability of a DNA probe assay for the detection of Legionella (Gen-Probe Incorporated, San Diego, CA) provides a unique opportunity to investigate the application of this technology to antibiotic susceptibility testing of L. pneumophila. We examined the ability of erythromycin, rifampin, and ciprofloxacin to kill L. pneumophila in buffered ACES-yeast extract broth (YEB). The test organism was incubated for a total of 96 hr in the presence of 10 micrograms/ml erythromycin, 1 micrograms/ml rifampin, or 1 micrograms/ml ciprofloxacin. Growth was monitored at 24-hr intervals by quantitative plating and the DNA probe assay. The correlation between organism concentration [colony-forming units (CFU) per ml] and DNA probe activity (counts per min) was excellent (r = 0.97). The percent decrease in CFU/ml at 96 hr relative to control counts was greater than 99% for erythromycin, rifampin, and ciprofloxacin. The percent decrease in CPM at 96 hr versus control was 87% for erythromycin, 89% for rifampin, and 93% for ciprofloxacin. This data documents a novel application of DNA probe technology, which may be useful in future studies of in vitro susceptibility of Legionella to various agents.

Staphylococcus aureus colonization and infection in patients on continuous ambulatory peritoneal dialysis

Staphylococcus aureus is the most common cause of peritonitis in patients undergoing peritoneal dialysis in Brazil. Using restriction endonuclease analysis of plasmid DNA, we investigated the importance of chronic carriage of S. aureus in the development of peritonitis in patients on continuous ambulatory peritoneal dialysis at the Division of Nephrology, Escola Paulista de Medicina, Sao Paulo, Brazil. A total of 117 isolates (30 patients) of S. aureus were available for typing, including 51 isolates (22 patients) from the nares, 58 isolates (27 patients) from pericatheter skin, and 8 isolates (6 patients) from peritoneal fluid, from patients with peritonitis. Restriction endonuclease subtyping showed that although most patients harbored more than one subtype of S. aureus, in the majority of patients nasal and/or pericatheter skin isolates with identical restriction endonuclease digest patterns were recovered on more than one occasion. Furthermore, 95% of patients with both nasal and pericatheter colonization were colonized with the same subtypes at both sites. All of the patients with peritonitis were infected with a subtype which colonized the nares, pericatheter skin, or both. These results demonstrate the importance of an endogenous source of S. aureus in the development of continuous ambulatory peritoneal dialysis-associated peritonitis.

Effects of cilofungin (LY121019) on carbohydrate and sterol composition of Candida albicans

Cilofungin (LY 121019) is a novel analogue of echinocandin B with potent activity against Candida albicans. The effects of cilofungin on the sterol and cell wall carbohydrate composition of Candida albicans were investigated. Exposure of Candida albicans to cilofungin resulted in a 55-60% decrease in ergosterol and a 4-13% decrease in lanosterol content relative to controls. Carbohydrate analysis revealed a 72-79% decrease in glucan content and no significant decrease in mannan or chitin content relative to controls. These studies suggest that cilofungin specifically inhibits glucan synthesis in Candida albicans and thus may be less toxic to mammalian cells. The effects of cilofungin on sterol composition may be non-specific and will require additional investigation.

Value of the Cand-Tec Candida antigen assay in the diagnosis and therapy of systemic candidiasis in high-risk patients

A total of 911 sera from 171 patients at risk for systemic candidiasis and 24 sera from 24 non-hospitalized control subjects were analyzed for the presence of candida antigen using a commercially available latex agglutination test (Cand-Tec). Thirty-seven (22%) patients had systemic candidiasis documented by positive blood cultures, deep biopsy culture and histopathology or autopsy. Six patients had transient candidemia, 20 patients had candiduria, 53 patients had mucous membrane colonization, 21 patients were not colonized but received empiric amphotericin B, and 34 patients were not colonized and not treated with amphotericin B. The intraobserver reproducibility was 90% for the exact titer and 100% for a deviation of one dilution. The sensitivity and specificity of the candida antigen test in detection of systemic candidiasis was 95% and 50% (greater than or equal to 1:2), 73% and 72% (greater than or equal to 1:4), and 46% and 80% (greater than or equal to 1:8) respectively. Despite the poor specificity, serial antigen determinations in patients with documented systemic candidiasis demonstrated both an early diagnostic and prognostic role for the candida antigen test. Seventy-one percent of patients whose antigen titer increased during the course of amphotericin B therapy of documented infection died versus only 13% of those whose titer decreased while on therapy (p = 0.01). The candida antigen test has a limited yet potentially useful role in the diagnosis and management of systemic candidiasis in high-risk patients.

In vitro comparison of cefpirome and four other beta-lactam antibiotics alone and in combination with tobramycin against clinical isolates of Pseudomonas aeruginosa

In vitro susceptibility studies of cefpirome versus cefotaxime, ceftazidime, imipenem, and piperacillin alone and in combination with tobramycin were performed against 153 clinical isolates of Pseudomonas aeruginosa from four medical centers. The minimal inhibitory concentration (MIC) for each antibiotic alone was determined by a standardized dilution method. Antibiotic combination studies were performed using a modified checkerboard technique. Cefpirome alone was more active (MIC90 64 micrograms/ml) than piperacillin (MIC90 128 micrograms/ml) or cefotaxime (MIC90 256 micrograms/ml) but less active than imipenem (MIC90 2 micrograms/ml) or ceftazidime (MIC90 32 micrograms/ml). The addition of tobramycin reduced the MICs of all of the beta-lactam antibiotics except for imipenem. The MIC90 for cefpirome when combined with tobramycin was 8 micrograms/ml compared to 16 micrograms/ml for cefotaxime and piperacillin, 8 micrograms/ml for ceftazidime, and 4 micrograms/ml for imipenem. The combination of tobramycin and cefpirome proved to be additive or synergistic for 82% of the isolates (highest rate) compared to 31% with imipenem (lowest rate). The potent in vitro antipseudomonal activity of cefpirome alone and in combination with an aminoglycoside (tobramycin) suggests that this agent may play a useful role in the therapy of infections due to P. aeruginosa.

The application of molecular and immunologic techniques to study the epidemiology of Legionella pneumophila serogroup 1

We applied monoclonal antibody typing and restriction endonuclease analysis of plasmid DNA to study 28 clinical and 35 environmental (potable water) isolates of Legionella pneumophila serogroup 1 from three hospitals in Iowa between 1981 and 1986. Monoclonal antibody typing employed a panel of seven antibodies and delineated eight different subtypes. Plasmids were present in 57% of the isolates including 12 of 28 (43%) clinical and 25 of 35 (69%) potable water isolates. The plasmids ranged in size from 28 to 98 kilobase pairs and comprised eight distinct subtypes by restriction endonuclease analysis with Eco RI. Combination of monoclonal antibody and restriction endonuclease subtyping (composite subtyping) revealed 19 different composite subtypes of Legionella pneumophila serogroup 1. The most common composite subtype, 09:04, comprised 29% (18 of 63) of the isolates and was only found in clinical and potable water samples from a single pavilion in hospital A during an outbreak of Legionella pneumophila serogroup 1 pneumonia. Aside from this cluster the diversity of composite subtypes of Legionella pneumophila serogroup 1 observed in clinical and potable water sources over the 5-year period was striking. The combination of monoclonal antibody and restriction endonuclease typing resulted in improved strain delineation and a more useful use of epidemiologic markers for Legionella pneumophila serogroup 1.

Fungicidal activity of cilofungin (LY121019) alone and in combination with anticapsin or other antifungal agents

Cilofungin (LY121019) was shown to have potent fungicidal activity against clinical isolates of Candida albicans and Candida tropicalis but not Candida parapsilosis. Fungicidal activity was evident against both replicating and non-replicating Candida albicans and was progressive over the first 12 h of incubation. The combination of cilofungin (LY121019) with anticapsin but not with amphotericin B, ketoconazole or 5-fluorocytosine resulted in synergistic fungicidal activity. This compound warrants further investigation of its safety and efficacy in the treatment of Candida infections.

Cost of nosocomial infection: relative contributions of laboratory, antibiotic, and per diem costs in serious Staphylococcus aureus infections

This study reports an analysis of the relative importance of laboratory antibiotic, and per diem costs of caring for 58 patients with serious Staphylococcus aureus nosocomial infections. Laboratory costs accounted for 2%, antibiotics for 21%, and per diem costs for 77% of total infection-related costs. Only 45% of patients were hospitalized for additional days specifically because of infection, but these patients stayed an average of 18 extra days. Nosocomial infections with S. aureus resistant to penicillinase-resistant penicillins (PRP) were more frequently associated with additional infection-related days of hospitalization than were PRP-susceptible infections. The cost of PRP-resistant infections was also significantly greater than PRP-susceptible infections, primarily because of the costs of additional days of hospitalization. Rational strategies to control costs of nosocomial infection should focus on two approaches: (1) prevention and (2) reduction of acute hospital days attributable to infections.

Multiple Candida strains in the course of a single systemic infection

Species and strain variabilities have been monitored during the history of a prolonged Candida infection in a single compromised bone marrow transplant patient by analyzing sugar assimilation patterns, high-frequency switching repertoires, and Southern blot hybridization patterns with two cloned mid-repeat sequences (Ca3 and Ca7) which are species specific for Candida albicans and one cloned mid-repeat sequence (Ct13-8) which is species specific for Candida tropicalis. Evidence is presented that during the course of this infection (i) two strains of C. albicans and three strains of C. tropicalis were distinguished by their switching repertoires, Southern blot hybridization patterns, and sugar assimilation patterns; (ii) the three C. tropicalis strains were in a high-frequency mode of switching; (iii) two C. tropicalis strains coexisted in the blood and three C. tropicalis strains coexisted in the throat at different times during the history of the infection; (iv) amphotericin B treatment selectively removed one of two C. tropicalis strains coexisting in the blood and this strain exhibited greater susceptibility to amphotericin B in vitro (the remaining strain was subsequently removed from the blood by flucytosine treatment); and (v) both the strain removed from the blood by amphotericin B and the strain removed from the blood by flucytosine reappeared several days later at another site of infection. It is demonstrated for the first time that C. tropicalis is capable of high-frequency switching of colony morphology just as C. albicans is, that there is more than one strain-specific switching repertoire in C. tropicalis, and that a C. tropicalis mid-repeat sequence can be used for discriminating species and assessing strain relatedness, as previously demonstrated for C. albicans mid-repeat sequences.

Development of the quantitative micro-test for slime production by coagulase-negative staphylococci

The macro-test for slime production by coagulase-negative staphylococci was adapted to a spectrophotometric micro-test assay to obtain more objective and quantitative information on slime production. A total of 135 isolates of coagulase-negative staphylococci (70 macro-test-positive and 65 macro-test-negative) were tested by both methods. The quantitative micro-test optical density readings were (mean +/- SD) 1.176 +/- 0.294 and 0.130 +/- 0.095 for the macro-test-positive and -negative groups, respectively. The micro-test was reproducible and demonstrated quantitative differences in slime production among the different species of coagulase-negative staphylococci. Aside from Staphylococcus epidermidis, the majority of the coagulase-negative staphylococci had very low optical density readings, indicating little or no slime production under the conditions employed in this assay. This test allows one to study the relative production of slime by different strains and species of coagulase-negative staphylococci, and may be useful in studying the effects of different conditions, such as antibiotic exposure, on slime production.

Antimicrobial activity of U-76,252 (CS-807), a new orally administered cephalosporin ester, including recommendations for MIC quality control

U-76,253A (R-3746), the active metabolite of the new cephalosporin ester, U-76,252 (CS-807), was tested against 4,742 fresh clinical isolates from four large medical centers. U-76,253A was very active against nearly all species of Enterobacteriaceae (87.7% inhibited at less than or equal to 4.0 micrograms/ml). Staphylococcus spp., and the streptococci. The U-76,253A spectrum was superior to the comparison orally administered cephalosporins (cephalexin, cephradine, cefaclor). Pseudomonas spp., Acinetobacter spp., and the enterococci were resistant to U-76,253A and the other tested drugs. Broth microdilution MIC quality control (QC) limits were established for U-76,253A in a multilaboratory investigation using a minimum of 125 MIC determinations per organism. The following MIC QC ranges were recommended; Escherichia coli (ATCC 25922) = 0.25-1.0 micrograms/ml, Staphylococcus aureus (ATCC 29213) = 2.0-4.0 micrograms/ml and Pseudomonas aeruginosa (ATCC 27853) = greater than 32 micrograms/ml.

Multicenter comparison of MicroScan and BACTEC blood culture systems

Recently, MicroScan (Baxter MicroScan Div., W. Sacramento, Calif.) introduced a radiolabeled-blood-culture system that is compatible with the BACTEC 460 (Johnston Laboratories, Inc., Towson, Md.). A multicenter blood culture study was initiated to evaluate this new system. Approximately 20 ml of blood was obtained from each patient and divided equally between BACTEC and MicroScan bottles which were incubated and processed identically. Aerobic bottles were examined twice on days 1 and 2 and once on days 3, 4, 5, 6, and 7. Anaerobic bottles were examined once a day for 7 days. There were 3,451 cultures evaluated, and 414 of these subsequently grew microorganisms. Of these positive cultures, 64 were judged to be contaminated. Of the remaining 350 positive cultures, 253 grew in both systems, 54 grew in BACTEC bottles only, and 43 grew in MicroScan bottles only. The average times to detect positive cultures were 1.8 and 2.1 days by the BACTEC and the MicroScan systems, respectively. No significant difference in the number or kind of organisms recovered or in the detection times for positive cultures was observed between the two blood-culturing systems.

Predictive value of surveillance cultures for systemic infection due to Candida species

Weekly fungal surveillance cultures (1,542 cultures) of urine (475), stool (520) and oropharyngeal (547) specimens from 111 patients on the bone marrow transplant and hematologic malignancy services were analyzed. Forty-three percent of the patients were colonized by Candida albicans and 10.8% by Candida tropicalis. There were 22 proven systemic fungal infections, ten due to Candida albicans, eight to Candida tropicalis, one each to Candida pseudotropicalis and Torulopsis glabrata, and two to Aspergillus species. Positive surveillance cultures for Candida tropicalis were highly predictive of systemic infection. The finding of two or more positive cultures yielded high positive predictive values (100%) as a function of body site. Positive surveillance cultures for Candida albicans were not predictive of disease but negative cultures for Candida albicans and Candida tropicalis had a high negative predictive value (95-99%). Surveillance culture data for specific Candida species may aid in diagnostic and therapeutic decision making.