Piperacillin/tazobactam (YTR 830) combination. Comparative antimicrobial activity against 5889 recent aerobic clinical isolates and 60 Bacteroides fragilis group strains

Piperacillin combined with tazobactam (formerly YTR 830) was tested at a ratio of 8:1 against 5889 aerobic isolates and 50 strains from the Bacteroides fragilis group. Imipenem was the most active agent tested against Enterobacteriaceae (99.3% at less than or equal to 4 micrograms/ml), ceftazidime was most effective against nonenteric Gram-negative bacilli (80.7% at less than or equal to 8 micrograms/ml), and piperacillin/tazobactam possessed a superior spectrum against Gram-positive cocci (92.2% at less than or equal to 16/2 micrograms/ml). Against all aerobic strains, piperacillin/tazobactam had a spectrum (90.3% at less than or equal to 16/2 micrograms/ml) comparable to imipenem (93.6% at less than or equal to 4 micrograms/ml) and was distinctly greater than that of ticarcillin/clavulanic acid (73.3% at less than or equal to 16/2 micrograms/ml) and ceftazidime (75.5% at less than or equal to 8 micrograms/ml). Against the B. fragilis group isolates, all piperacillin/tazobactam MICs were less than or equal to 64/8 micrograms/ml. This activity was superior to piperacillin alone (MIC:50, 8-64 micrograms/ml) and cefoxitin (MIC50, 4-64 micrograms/ml). Piperacillin/tazobactam appears to be a promising parenteral antimicrobial combination, with a spectrum effective against a wide variety of clinical pathogens.

Susceptibility of clinical isolates of Candida spp. to terconazole and other azole antifungal agents

Terconazole is a triazole ketal derivative with potent, broad-spectrum antifungal activity. We investigated the in vitro activity of terconazole, miconazole, and clotrimazole, against 94 clinical isolates of Candida spp.: C. albicans (n = 68), C. tropicalis (n = 18), and C. parapsilosis (n = 8). In vitro susceptibility testing was performed using a broth microdilution method. The minimal inhibitory concentrations of terconazole were less than those of miconazole against C. albicans and C. parapsilosis but higher against C. tropicalis. Terconazole was more active than clotrimazole against C. parapsilosis and less active against C. albicans and C. tropicalis. Terconazole inhibited the uptake of 14C-labeled glucose, leucine, and hypoxanthine into C. albicans and caused the rapid release of intracellular K+. Based on these studies, terconazole has promising anticandidal activity and warrants further in vitro and in vivo investigation.

Nosocomial gram-negative bloodstream isolates: a comparison of in vitro antibiotic potency

Nosocomial bloodstream infections add to the morbidity, mortality and length of hospitalization that is attributed to the underlying diseases alone. We have compared the in vitro potency of fifteen antibiotics against 136 isolates from clinically significant nosocomial gram-negative bacteraemias. Ciprofloxacin was the most potent antibiotic and had the broadest spectrum of activity (98% of isolates susceptible, MIC90 range: 0.06-0.5 micrograms ml-1). We subjected all isolates to beta-lactamase induction but antibiotic susceptibility was unaffected by this procedure.

Plasma tumor necrosis factor levels in patients with presumed sepsis. Results in those treated with antilipid A antibody vs placebo

Using an enzyme-linked immunosorbent assay, we measured plasma levels of tumor necrosis factor (TNF) in 38 patients who were treated with either antilipid A antibody or a placebo for presumed gram-negative bacteremia. Sixteen of the 38 patients had positive blood cultures: 14 with gram-negative rods and 2 with Streptococcus pneumoniae. Initial serum samples for TNF determinations were obtained within 2 to 72 hours (mean, 18.8 hours) after the onset of clinical signs of sepsis. Six (16%) of 38 patients had detectable TNF levels: 4 of 14 with positive blood cultures for gram-negative rods but only 2 of 22 with negative blood cultures (odds ratio, 4; 95% confidence limits, 0.5 and 24.3). Of the 6 patients, 4 had received the placebo and 2 had received the antibody. Tumor necrosis factor levels did not predict adult respiratory distress syndrome, shock, disseminated intravascular coagulation, renal failure, or mortality. The highest TNF levels (500 and 250 pg/mL) were observed in 2 patients with Enterobacter cloacae bacteremia who had received the placebo and antilipid A antibody, respectively. The other 2 patients with bacteremia and detectable TNF levels had positive blood cultures for Haemophilus influenzae (50 pg/mL) and Bacteroides fragilis (120 pg/mL), respectively. Despite negative blood cultures, the remaining 2 patients repeatedly had detectable TNF levels and a clinical picture consistent with gram-negative sepsis.

Use of cellular hydrophobicity, slime production, and species identification markers for the clinical significance of coagulase-negative staphylococcal isolates

Determining the clinical relevance of coagulase-negative staphylococci isolated from cultures of clinical specimens remains a common dilemma. One hundred eighteen strains of coagulase-negative staphylococci isolated from patients with and without indwelling foreign bodies were characterized with regard to cell-surface hydrophobicity, slime production, and species to determine the predictive value of these phenotypic markers in distinguishing clinically significant from insignificant isolates. The single test with the highest positive predictive value was hydrophobicity (79%). Hydrophobicity and speciation had the greatest combined predictive value of any two tests (89%), and this increased to only 90% when determination of slime production was added. These tests provide additional clinical information when coagulase-negative staphylococci are isolated in culture.

Infection control: opportunistic fungal infections--the increasing importance of Candida species

Candida spp are clearly among the most important of the opportunistic fungal pathogens that cause nosocomial infections. Given the increasing incidence, high attributable mortality and excess length of stay associated with nosocomial candidemia, it is imperative that more effective preventive, diagnostic and therapeutic measures be developed to deal with this emerging group of pathogens. Implicit in these issues is the need for an increased understanding of the epidemiology of nosocomial candidiasis. The development of new epidemiologic typing methods provides useful tools that can assist in clinical and epidemiologic investigation. These studies are essential to the rational development of effective measures to prevent and to control nosocomial infections caused by Candida spp.

Multicenter in vitro evaluation of SM-7338, a new carbapenem

A new carbapenem, SM-7338, was compared with imipenem, cefotaxime, and ceftazidime at five medical centers. Nearly 6,000 strains were tested by reference methods of the National Committee for Clinical Laboratory Standards, and SM-7338 inhibited the largest percentage of gram-negative bacilli. Its spectrum included all members of the family Enterobacteriaceae (99.7% were susceptible to less than or equal to 4 micrograms/ml), Pseudomonas spp. (but not Xanthomonas maltophilia), and Acinetobacter spp. The potency and spectrum of SM-7338 against the gram-positive organisms were less than those of imipenem and superior to those of ceftazidime. Only the enterococci and some oxacillin-resistant staphylococci were less susceptible to SM-7338 (MICs for 90% of isolates, greater than or equal to 8 micrograms/ml). Organisms resistant to ceftazidime were generally susceptible to SM-7338 and imipenem (76%). However, for one-third of the imipenem-resistant gram-negative bacilli (MICs, greater than 8 micrograms/ml), SM-7338 MICs were less than or equal to 4 micrograms/ml. Some endemic differences in patterns of SM-7338 activity against selected gram-negative species were found among some medical centers.

Clinical evaluation of a direct fluorescent monoclonal antibody test for detection of Pseudomonas aeruginosa in blood cultures

A direct fluorescent monoclonal antibody test (DFA; Genetic Systems Corp., Seattle, Wash.) was evaluated for the detection of Pseudomonas aeruginosa in 178 blood culture broths obtained from 128 patients. The DFA identified 44 (98%) of 45 blood cultures positive for P. aeruginosa and was negative in 131 (98%) of 133 blood cultures which grew gram-negative rods other than P. aeruginosa. Upon further investigation, saline suspensions of the organism from the false-negative blood culture were strongly (4+) DFA positive. The false-positive reactions were not due to cross-reactivity, as shown by lack of DFA staining of the non-P. aeruginosa isolates following subculture to agar media. The specificity of the reagent was further demonstrated by directly staining culture isolates including 10 serotypes of P. aeruginosa (all positive) and 57 selected gram-negative bacilli including eight species of Pseudomonas that were not P. aeruginosa (all negative). DFA staining of blood culture broths was easy to perform and read with minimal background fluorescence. The DFA method can be performed in 50 min and appears promising as a rapid method for the identification of P. aeruginosa bacteremia.

In-vitro susceptibility of nosocomial gram-negative bloodstream pathogens to quinolones and other antibiotics--a statistical approach

We examined the in-vitro activity of 12 antibiotics against Gram-negative bacillary isolates from 141 distinct episodes of nosocomial bloodstream infection occurring from July 1984 through November 1986. At least ten strains of each of the seven most frequently encountered species were tested. Relative potency was carefully assessed by extending the concentrations from 0.004 to 64 mg/l in microdilution tests performed in duplicate. We estimated MIC50 and MIC90 and, importantly, calculated 95% confidence intervals (CI95) for MIC90. Against all isolates, ciprofloxacin, the most potent antibiotic, had an MIC50 of 0.03 and an MIC90 of 0.13 (CI95 0.11 to 0.16 mg/l). Norfloxacin and enoxacin had MIC90 of 0.50 and 0.71 mg/l, respectively. Imipenem, ceftazidime, aztreonam, and cefoperazone had MIC90 of 1.0, 2.0, 5.3, and 5.7 mg/l, respectively. Both cefotaxime and ceftriaxone had MIC90 of 16 mg/l (CI95 13-19.7 mg/l). For tobramycin, gentamicin and amikacin, the MIC90 values were 1.4, 5.7, and 8 mg/l, respectively. Against Pseudomonas aeruginosa (n = 26), Serratia marcescens (n = 19), and Klebsiella pneumoniae (n = 26), the CI95s about the MIC90 for ciprofloxacin were 0.31-0.81, 0.07-0.23, and 0.04-0.10 mg/l, respectively. For optimal comparison of antibiotics used to treat hospital-acquired bacteraemias, only clinically significant nosocomial bloodstream isolates should be studied with regard to their antibiotic susceptibilities; the isolates should be unique (only one isolate per episode of bacteraemia occurring over a defined period of time); an adequate number of isolates of a particular species should be studied; MICs should be determined over a wide range of concentrations; and both the MIC90 and the CI95 should be reported.

Susceptibility of nosocomial isolates of Candida species to LY121019 and other antifungal agents

LY121019 is a novel analog of the polypeptide antifungal antibiotic echinocandin B. We investigated the in vitro activity of LY121019, amphotericin B, ketoconazole and 5-fluorocytosine against 131 nosocomial isolates of Candida species: C. albicans (n = 50), C. tropicalis (n = 30), C. rugosa (n = 12), C. parapsilosis (n = 11), C. lusitaniae (n = 10), C. guillermondii (n = 9), and C. krusei (n = 9). In vitro susceptibility testing was performed using a broth microdilution method. The minimal inhibitory concentrations (MIC) of LY121019 were less than or equal to that of the other antifungal agents against C. albicans and C. tropicalis but were generally higher for the other species of Candida. Paradoxical growth at high concentrations, but not at low concentrations, of LY121019 was observed with isolates of C. albicans and C. tropicalis.

Coagulase-negative staphylococcal bacteremia. Mortality and hospital stay

OBJECTIVE

To determine the attributable mortality and the excess length of hospital stay resulting from coagulase-negative staphylococcal bacteremia.

DESIGN

Matched historical cohort study.

SETTING

Large university-based tertiary care center.

PATIENTS

Of 171 patients with hospital-acquired coagulase-negative staphylococcal bacteremia identified by prospective surveillance of nosocomial infections from 1 July 1984 to 30 June 1987, 118 met criteria for the study and were matched to a control patient by age, sex, primary diagnoses, operative procedures, and date of admission.

MEASUREMENTS AND RESULTS

Success was achieved in 621 of 650 (96%) variables used for matching. Staphylococcus epidermidis accounted for 92% of the bacteremias. Twenty cases (17%) had evidence of septic shock, and 10 had disseminated intravascular coagulation. The mortality rate in cases was 36 of 118 (30.5%) compared with 20 of 118 (16.9%) in controls. The attributable mortality was 13.6% (95% CI, 4.2 to 22.9) and the risk ratio for dying was 1.8 (95% CI, 1.2 to 2.7; P = 0.006). The median length of stay was 46 days for cases and 37.5 for controls (P = 0.0002).

CONCLUSIONS

Coagulase-negative staphylococci, the leading organisms causing hospital-acquired bacteremias, are associated with mortality in excess of that due to the underlying diseases alone. Moreover, they significantly prolong the length of hospital stay. These findings show the importance of coagulase-negative staphylococcal bacteremia in hospitalized patients.

Hospital-acquired candidemia. The attributable mortality and excess length of stay

Between 1977 and 1984, estimates of hospital-acquired bloodstream infections caused by Candida species increased in the United States from 0.5 to 1.5 per 10,000 admissions (National Nosocomial Infection Study data). We examined crude and attributable mortality rates and excess length of stay in 88 closely matched pairs of cases and controls with illnesses occurring between July 1983 and December 1986. The crude mortality rates for cases and controls were 57% and 19%, respectively; thus the attributable mortality rate was 38% with a 95% confidence interval of 26% to 49%. The risk ratio was 2.94 with a 95% confidence interval of 1.95 to 4.43. The median length of stay was 48 days for all cases and 40 days for all controls. An analysis of the length of stay for the 34 matched pairs that survived showed a median of 70 days for cases and 40 days for controls. Candida bloodstream infections represented 10% of all nosocomial bloodstream infections in the period studied at our University Hospital; they are associated with a significant medical and economic burden well above that expected of the underlying diseases alone.

In vitro activity of FCE 22101, imipenem, and ceftazidime against over 6,000 bacterial isolates and MIC quality control limits of FCE 22101

Six geographically separate laboratories within the USA tested 6,198 bacterial isolates against FCE 22101 (a penem), imipenem (a carbapenem) and ceftazidime (a third-generation cephalosporin). Ninety-three percent of 2,749 Enterobacteriaceae were inhibited by FCE 22101, while 95% were susceptible to ceftazidime and 99% were susceptible to imipenem. FCE 22101 had little activity against Pseudomonas spp. but was active against most gram-positive pathogens, including enterococci. FCE 22101 MICs for standard quality control strains were defined as 0.5-2.0 micrograms/ml for Escherichia coli ATCC 25922, 2-8 micrograms/ml for Enterococcus faecalis ATCC 29212 and 0.06-0.25 microgram/ml for Staphylococcus aureus ATCC 29213.

Prospective evaluation of the Gen-Probe assay for detection of legionellae in respiratory specimens

A prospective evaluation of a DNA probe assay for detection of Legionella species was performed on 427 consecutive respiratory specimens submitted over an 18-month period. The Gen-Probe assay utilizing both low (greater than or equal to 4.0) and high (greater than 7.0) ratio threshold values was compared to direct fluorescent antibody staining (DFA) as a predictor of isolation of Legionella on culture. The highest sensitivity (63%) was obtained with the lower threshold ratio, but was not significantly different from the result obtained with a threshold ratio of greater than 7.0 (50%, p = 0.722) or DFA results (44%, p = 0.479). The specificity of the DNA probe assay was improved with the high threshold (99%) compared either to the low threshold ratio (95%, p = 0.002) or DFA (97%, p = 0.055). When the DNA probe was compared to DFA and/or Legionella isolation on culture, a significantly lower specificity (97% versus 99%, p = 0.0006) and higher sensitivity (74% versus 37%, p = 0.013) was obtained with a threshold value of greater than or equal to 4.0 than greater than 7.0. Ten of 20 specimens with a DNA probe ratio between 4.0 and 7.0 were DFA positive, although only two were isolated on culture. The DFA assay and both probe threshold ratios have a high negative predictive value when compared to culture. However, only the threshold ratio of greater than 7.0 has a sufficiently high positive predictive value to be useful alone.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparison of the Quantum II, API Yeast Ident, and AutoMicrobic systems for identification of clinical yeast isolates

The Quantum II Yeast Identification System (Abbott Laboratories) is a microprocessor-based spectrophotometric system for identification of clinical yeast isolates within 24 h. We compared the Quantum II system with the API Yeast Ident (Analytab Products) and the AutoMicrobic System Yeast Biochemical Card (AMS-YBC; Vitek Systems, Inc.) for the identification of 221 clinical yeast isolates, including 120 common clinical isolates (Candida albicans, C. tropicalis, C. parapsilosis, Torulopsis glabrata, and Cryptococcus neoformans) and 101 relatively uncommon clinical isolates. The API 20C (Analytab) was used as the reference system. The Quantum II and AMS-YBC systems correctly identified 181 (82%) and 184 (83%) isolates, respectively, whereas the Yeast Ident system correctly identified 132 (60%) isolates. Of the 120 common clinical isolates, 113 (94%) were correctly identified by Quantum II, 103 (86%) were correctly identified by AMS-YBC, and 83 (69%) were correctly identified by Yeast Ident. Of the 101 uncommon clinical isolates tested, 68 (67%) were correctly identified by Quantum II, 81 (80%) were correctly identified by AMS-YBC, and 49 (49%) were correctly identified by Yeast Ident. The overall accuracy of the Quantum II, AMS-YBC, and API Yeast Ident was not sufficient to recommend any of these systems for routine use in the clinical microbiology laboratory without substantial expansion of the respective data bases.

Implications of acquired oxacillin resistance in the management and control of Staphylococcus aureus infections

Refinements in testing for resistance to penicillinase-resistant penicillins (PRP) in Staphylococcus aureus have resulted in confusion in classifying isolates as PRP susceptible or resistant. Specifically, a group of organisms has been identified that produce large amounts of beta-lactamase and appear borderline resistant. These organisms have been called "occult resistant" or "acquired oxacillin-resistant" S. aureus (AORSA). A retrospective study was conducted to evaluate the implication of this in vitro phenomenon in managing patients with AORSA infections. Among 134 patients with S. aureus infections, 89 were infected with oxacillin-susceptible S. aureus (OSSA), 26 with AORSA, and 19 with oxacillin-resistant S. aureus (ORSA). There were no significant differences in outcomes when OSSA and AORSA infections were treated with PRP (chi 2MH = .990; P = .32). These results do not suppor the contention that AORSA infections should be managed differently from OSSA infections. Identifying AORSA may not be helpful in guiding antimicrobial therapy or predicting the outcome of infections with AORSA.

Removal of nosocomial pathogens from the contaminated glove. Implications for glove reuse and handwashing

STUDY OBJECTIVE

To evaluate the effectiveness of three different types of handcleansing agents in decontaminating gloved hands that were inoculated with a series of four nosocomial pathogens.

DESIGN

A controlled, experimental trial.

SETTING

Tertiary care referral center.

PATIENTS OR OTHER PARTICIPANTS

Five healthy volunteers participated in all portions of the study.

INTERVENTIONS

A standard concentration of one of four representative nosocomial pathogens was placed on the gloved hand, spread, and allowed to dry. One of three different handcleansing agents--a nonmedicated soap, a 60% isopropyl alcohol preparation, or 4% chlorhexidine gluconate--was used to cleanse the gloves, which were cultured using a broth-bag technique. The gloves were then removed and the hands were cultured in a similar manner.

MEASUREMENTS AND MAIN RESULTS

The handwashing agents reduced the median log10 counts of organisms to 2.1 to 3.9 after an inoculation of 10(7) colony forming units. The proportion of positive glove cultures for Staphylococcus aureus, 8% to 100%; Serratia marcescens, 16% to 100%; and Candida albicans, 4% to 60% varied greatly after use of the different handcleansers (P less than 0.001), and varied considerably for Pseudomonas aeruginosa, 20% to 48% (P = 0.085). After the gloves were removed, the differences among the observed proportions of hands contaminated with the test organisms varied from 5% to 50%, depending on the handcleansing agent used (P less than 0.001).

CONCLUSIONS

In the era of universal precautions these data suggest that it may not be prudent to wash and reuse gloves between patients. Further, handwashing is strongly encouraged after removal of gloves.

Influence of in vitro susceptibility testing conditions on the anti-candidal activity of LY121019

LY121019 is a new antifungal antimicrobic that is structurally similar to the lipopeptide agents echinocandin B and aculeacin A. Because of the importance of in vitro test conditions on the activity of other antifungal agents, we studied the effects of inoculum size, time and temperature of incubation, pH, and medium composition on the in vitro activity of LY121019 against Candida albicans, Candida tropicalis, and Candida parapsilosis. LY121019 was highly active against Candida albicans and Candida tropicalis and inactive against Candida parapsilosis. The in vitro activity of LY121019 is marked by a paradoxical dose-response with isolates of Candida albicans and Candida tropicalis and is influenced by choice of inoculum size, time and temperature of incubation, medium composition, and pH. We recommend the use of an inoculum size of less than 10(5) organisms/ml, a defined medium buffered to a pH of 7.0, and incubation at 30 degrees C for 24 hr for future in vitro studies of LY121019.

Multicenter evaluation of four methods of yeast inoculum preparation

We initiated a comparative study of four methods of yeast inoculum preparation: a spectrophotometric method, the Wickerham card method, a hemacytometer method, and the Prompt inoculation system. The variability in inoculum size obtained when each method was applied to two strains each of Candida albicans, Candida tropicalis, Candida parapsilosis, Torulopsis glabrata, Cryptococcus neoformans, and Saccharomyces cerevisiae was analyzed in a single laboratory. Each method was performed in triplicate on the same day and on three separate days to provide estimates of within-day and between-day variations. Inoculum size was determined by viable colony counts. The greatest range of inoculum sizes was seen with the Wickerham card method. Viable counts ranged from 1.1 X 10(6) to 24.2 X 10(6) CFU/ml among the 12 yeast isolates. The greatest variation was observed with the Prompt system. Within-day coefficients of variation averaged 19% (range, 4 to 45%), and between-day coefficients of variation averaged 22% (range, 3 to 51%). Variation between laboratories was evaluated by comparing inoculum values obtained by each method in three different laboratories for two strains of C. albicans. The spectrophotometric method was the least variable and the Wickerham card and hemacytometer methods were the most variable methods between laboratories. The spectrophotometric method is recommended as the method of choice for preparation of a standardized inoculum suspension for susceptibility testing of yeasts.

Multicenter in vitro evaluation of lomefloxacin (NY-198, SC-47111), including tests against nearly 7,000 bacterial isolates and preliminary recommendations for susceptibility testing

Lomefloxacin (NY-198 or SC-47111) is a difluoro-quinolone derivative having a C-methyl at the 3-position of the piperazine ring, thus minimizing its metabolic alteration in vivo. In our research, its antimicrobial activity was most similar to that of difloxacin, enoxacin, fleroxacin, and norfloxacin but usually less than that of ciprofloxacin and ofloxacin against most species. Lomefloxacin shared cross-resistance with other 4-quinolones but remained very active against ceftazidime-resistant organisms, including stably derepressed beta-lactamase producing Gram-negative bacilli. Lower pH increased the lomefloxacin MICs. MBCs were usually identical to the measured MIC, and the lomefloxacin MICs were not significantly increased by high inoculum concentrations. The Enterobacteriaceae were found to have a very low rate of spontaneous mutation to lomefloxacin resistance (10(-8)-10(-9). In vitro tests by 5-micrograms and 10-micrograms lomefloxacin disks and dilution methods were correlated, and the 10-micrograms disk was recommended for clinical trials using a less than or equal to 4 micrograms/ml susceptible breakpoint. The quality assurance guidelines for dilution tests were determined by a multilaboratory study.

Laboratory, clinical, and epidemiological aspects of coagulase-negative staphylococci

Coagulase-negative staphylococci, particularly Staphylococcus epidermidis, are increasingly important causes of nosocomial infection. Microbiologists and clinicians no longer can afford to disregard clinical isolates of coagulase-negative staphylococci as contaminants. Accurate species identification and antimicrobial susceptibility testing, in a clinically relevant time frame, are important aids in the diagnosis and management of serious coagulase-negative staphylococcal infections. Emphasis in the clinical laboratory should be placed on the routine identification of S. epidermidis and Staphylococcus saprophyticus, with identification of other species of coagulase-negative staphylococci as clinically indicated. The application of newer techniques, such as plasmid analysis and tests for slime production and adherence, contribute to our understanding of the epidemiology and pathogenesis of coagulase-negative staphylococci and may also be helpful in establishing the diagnosis of infection.

Comparison of the lysis centrifugation and radiometric blood culture systems for recovery of yeast

The performance of the Isolator (lysis centrifugation) and Bactec (radiometric) detection systems for the recovery of fungi from blood was studied prospectively by comparison of 2,188 paired cultures obtained at two geographically separated teaching hospitals. Eight-three yeast isolates were recovered from 78 (3.6%) cultures that were obtained from 43 patients. Seventy-three (88%) yeast strains were recovered using the Isolator system, and 60 (72%) were recovered in the Bactec system. The average time for recovery of yeast was 2.3 days for the Isolator system and 3.1 days for the Bactec system. Optimal recovery can be accomplished through the use of both systems.

Legionnaires' disease associated with a hospital water system. A five-year progress report on continuous hyperchlorination

In 1981, sixteen cases of nosocomial legionellosis occurred among 456 patients admitted to a new hematology-oncology unit (35 per 1000 admissions). Monoclonal antibody typing and restriction endonuclease plasmid analysis identified a unique strain (09,04) of Legionella pneumophila serogroup 1 isolated from both patients and water outlets. Continuous hyperchlorination of the hot and cold water began in January 1982, and chlorine levels of 3 to 5 mg/L have been maintained most recently. Water samples have been consistently negative for Legionella for more than five years. Four sporadic cases of nosocomial legionellosis have occurred in the hematology-oncology unit during the same period (one per 1000 admissions) associated with a different strain of L pneumophila serogroup 1 (09,00). The environmental reservoir(s) of L pneumophila serogroup 1 in these cases has not been identified. Levels of trihalomethanes (potential carcinogens) were high (greater than 100 micrograms/L) when chlorine levels of hot water exceeded 4 mg/L. Some corrosion damage to the water distribution system has occurred: the average number of leaks per month increased steadily from zero in 1982 to 5.2 in 1986. The chlorinator installation costs were +75,800, and annual operation expenses were +12,500. Continuous hyperchlorination is a promising but still experimental technique for control of nosocomial legionellosis. In our experience, epidemic disease has been controlled, but sporadic cases have continued to occur.