Clinical Outcomes for Hospitalized Patients with Legionella Pneumonia in the Antigenuria Era: The Influence of Levofloxacin Therapy

Transcription

1 MAJOR ARTICLE Clinical Outcomes for Hospitalized Patients with Legionella Pneumonia in the Antigenuria Era: The Influence of Levofloxacin Therapy Analía Mykietiuk, 1 Jordi Carratalà, 1 Núria Fernández-Sabé, 1 Jordi Dorca, 2 Ricard Verdaguer, 3 Frederic Manresa, 2 and Francesc Gudiol 1 1 Infectious Disease, 2 Respiratory Medicine, and 3 Microbiology Services, Institut d Investigació Biomèdica de Bellvitge, Hospital Universitari de Bellvitge, University of Barcelona, Barcelona, Spain Background. Although the reduction in case-fatality rate recently observed among patients with Legionella pneumonia has been largely attributed to the progressive utilization of urine antigen testing, other factors, such as changes in empirical antibiotic therapy, may also have contributed. We have analyzed more-recent outcomes of Legionella pneumonia in an institution where urine antigen testing was reflexly performed in cases of communityacquired pneumonia without an etiological diagnosis. Methods. From a prospective series of 1934 consecutive cases of community-acquired pneumonia in nonimmunocompromised adults, 139 cases of Legionella pneumophila pneumonia were selected for observational review. Legionella cases were analyzed for outcome with respect to antibiotic, mortality, complications, length of stay, time to defervescence, and stability. Results. The early case-fatality rate was 2.9% (4 of 139 patients), and the overall case-fatality rate was 5% (7 of 139 patients). One hundred twenty patients (86.3%) received an appropriate initial therapy, which included macrolides (i.e., erythromycin or clarithromycin) in 80 patients and levofloxacin in 40. Levofloxacin progressively replaced macrolides as the initial therapy during the study period. Compared with patients who received macrolides, patients who received levofloxacin had a faster time to defervescence (2.0 vs. 4.5 days; P!.001) and to clinical stability (3 vs. 5 days; P p.002). No differences were found regarding the development of complications (25% vs. 25%; P p.906) and case-fatality rate (2.5% vs. 5%; P p.518 ). The median length of hospital stay was 8 days in patients treated with levofloxacin and 10 days in those who received macrolides ( P p.014 ). Conclusions. Legionella pneumonia is still associated with significant complications in hospitalized patients, but recent mortality is substantially lower than that found in earlier series. Levofloxacin may produce a faster clinical response than older macrolides, allowing for shorter hospital stay. Legionella pneumophila is a leading cause of community-acquired pneumonia [1 5]. It has been estimated that ,000 people are hospitalized with legionellosis annually in the United States [6]. Moreover, Legionella is particularly frequent among patients with community-acquired pneumonia who require admission to an intensive care unit [7, 8]. Classically, L. pneumophila has been classified as a high-risk pathogen in community-acquired pneumo- Received 20 September 2004; accepted 28 October 2004; electronically published 17 February Reprints or correspondence: Dr. Jordi Carratalà, Infectious Disease Service, Hospital Universitari de Bellvitge, Feixa Llarga s/n, L Hospitalet, Barcelona, Spain (jcarratala@csub.scs.es). Clinical Infectious Diseases 2005; 40: by the Infectious Diseases Society of America. All rights reserved /2005/ $15.00 nia, accounting for substantial morbidity and with casefatality rates of 5% 25% among immunocompetent hosts [1, 9 12]. However, a recent surveillance study [13] showed a remarkable reduction in case-fatality rates during the past decade, suggesting that new models of management may have improved outcomes in patients with Legionella pneumonia. Although it is reasonable to suppose that early recognition of Legionella pneumonia by means of urine antigen testing may have contributed to decreasing mortality, the impact of other factors, including new antibiotic options, is less clear. Furthermore, to assess more adequately the role of these factors, outcome measures other than mortality should be considered [14]. In this study, we analyzed more recent clinical outcomes of Legionella pneumonia, including response to antibiotic therapy, time to clinical stability, develop- 794 CID 2005:40 (15 March) Mykietiuk et al.

2 ment of in-hospital complications, length of stay, and mortality in an institution where urine antigen testing was consistently performed during the study period. METHODS Setting and study design. The study was conducted at a 900- bed university hospital for adults in Barcelona, Spain (Hospital Universitari de Bellvitge). The hospital serves an urban area of 1,100,000 inhabitants and has 24,000 admissions annually. All 1934 non severely immunosuppressed patients with community-acquired pneumonia admitted to the hospital between 13 February 1995 and 31 December 2003 were prospectively recruited and followed up. Patients with neutropenia (i.e., those with a granulocyte count of! cells/l), those with AIDS, and transplant recipients were not included. The definition of pneumonia and hospitalization criteria have been described elsewhere [5]. For the purposes of the present observational study, we analyzed data on 139 cases of Legionella pneumonia, diagnosed with use of 1 of the following methods: urine antigen test; isolation of Legionella in sputum, transthoracic needle aspiration specimen, or pleural fluid; and/or a 4-fold increase in the antibody titer. We compared clinical outcomes for patients treated with levofloxacin with outcomes for those who received macrolides. Clinical evaluation, antibiotic therapy, and follow-up. At the initial visit to the emergency department, patients gave a complete clinical history, underwent physical examination, and underwent chemical and hematological testing. Microbiological studies included blood cultures and sputum Gram stain and culture, when available. A urine antigen test for Legionella species was performed at hospital admission, if indicated by the attending physician, but was performed 48 h after admission for all patients for whom nonetiological diagnosis had been made (reflex testing). Paired serum samples were obtained during the acute and convalescent phases of infection (separated by a 3 8 week interval) for serological studies. Antibiotic therapy was initiated in the emergency department in accordance with hospital guidelines, which recommended the administration of a b-lactam (either ceftriaxone or amoxicillin-clavulanate, administered intravenously) with or without a macrolide (either erythromycin 1000 mg iv q.i.d. or clarithromycin 500 mg iv b.i.d.); from 1998 onward, levofloxacin (500 mg iv q.d.) was also allowed. Patients with a urine antigen test result positive for Legionella at admission were treated with a macrolide (with or without rifampin, on the basis of physician preference) or levofloxacin. Those patients initially treated with other antibiotics were switched to appropriate therapy. Recommended duration of therapy was days. Patients were seen daily during their hospital stay by 1 of the investigators, who provided medical advice when requested and stored clinical data in a computer-assisted protocol. All patients were evaluated for adverse events. Assessments included evaluation of changes in vital signs, clinical laboratory results, and physical examination findings. A long-term followup visit took place 1 month after discharge. Definitions. To stratify patients into risk classes, we used the validated prediction rule, calculated according to the pneumonia severity index (PSI) [15]. The time to overall clinical stability was defined as the time required for 5 vital signs to became stable (temperature, 37.2 C; heart rate, 100 beats/ min; respiratory rate, 24 breaths/min; systolic blood pressure, 90 mm Hg; oxygen saturation on room air, 190%) with normalization of mental status and the ability to eat [16]. A variable was considered stable if all measurements in a 24-h period met stability criteria. Complications were defined as any untoward circumstance occurring during hospitalization, with the exception of the side effects of the. Length of hospital stay was calculated as the time from the admission date to the date of discharge from the hospital. Early case-fatality rate was defined as death (as proportion of cases) due to any cause 48 h after hospitalization. Overall case-fatality rate was defined as death (as proportion of cases) due to any cause 30 days after hospitalization. Microbiologic studies. Investigation of pathogens in blood, normally sterile fluids, sputum, and other samples was performed using standard microbiological procedures. L. pneumophila serogroup 1 antigen in urine was detected by an immunoenzymetric commercial method (Legionella Urinary Antigen; Binax). Isolation of Legionella species was attempted in sputum samples and other respiratory samples by the selective medium buffered charcoal yeast extract-a. EIA was used to detect antibodies against L. pneumophila serogroups 1 6. Statistical analysis. To detect significant differences between antibiotic groups, we used the x 2 test with continuity correction for categorical variables and the Student s t test or the Mann-Whitney U test for continuous variables, when appropriate. In all analyses, P values!.05 were considered to be statistically significant. RESULTS During the study period, 1934 non severely immunosuppressed adults with community-acquired pneumonia were admitted to our institution. L. pneumophila was the causative organism in 139 (7.2%) of the cases. The diagnosis was established with use of 1 of the following methods: urine antigen test (in 120 cases), seroconversion (in 80), and/or positive culture (in 43). Culture results were positive for 32 of 59 cases in which sputum samples were available, in 9 of 14 cases in which transthoracic needle aspiration specimens were available, and in 2 of 10 cases in which pleural fluid Outcomes of Legionella Pneumonia CID 2005:40 (15 March) 795

3 samples were available. All cases identified by culture corresponded to L. peumophila serogroup 1. Only 19 cases were diagnosed by serological methods alone. Two cases of mixed infection were detected (1 case of Chlamydia pneumoniae infection and 1 case of Chlamydia psittaci infection). Cases were uniformly distributed during the study period. Figure 1 shows the total number of cases distributed every 3 years and the number of patients who underwent urine antigen testing in that period, which was also constant. One hundred twenty (86.3%) of the 139 patients with Legionella pneumonia were men, with a mean age ( SD) of years. Seventy-one (51.1%) of the patients were smokers, and 59 (42.4%) had a history of alcohol abuse. Twenty-five patients (17.9%) had been immunized with influenza vaccine for the influenza season, and 8 (5.8%) had received 23-valent pneumococcal capsular polysaccharide vaccine in the 5 years before admission to the hospital. Seventy-four patients (53.2%) had underlying diseases, mainly chronic heart disease (in 25 patients), diabetes mellitus (in 20), chronic obstructive pulmonary disease (in 15), chronic liver disease (in 6), and cancer (in 5). Five patients (3.6%) had a history of long-term steroid use. Sixty-three patients (45.3%) had received antibiotic therapy before hospitalization (mainly b-lactam antibiotics [in 56 cases]). Sixty-four patients (46%) were classified into high-risk pneumonia classes (risk classes IV V) according to PSI score. One hundred twenty (86.3%) of the patients received an appropriate initial therapy, which included levofloxacin for 40 patients and macrolides for the remaining 80 patients (erythromycin for 53 patients and clarithromycin for 27 patients). Forty-eight patients who were treated with macrolides were also given rifampin. Ten (25%) of the patients in the levofloxacin group and 21 (26.2%) in the macrolide group received steroids for the current episode of pneumonia ( P p.883). Levofloxacin progressively replaced macrolides as the initial therapy during the study (figure 1). Eighteen patients (13%) received an inappropriate empirical therapy with b-lactam antibiotics, which was promptly modified to cover Legionella when the result of the urine antigen test was available; all of these patients remained febrile at the time they were switched to macrolides (13 patients) or levofloxacin (5 patients). One additional patient received initial therapy with both levofloxacin and clarithromycin. As shown in table 1, demographic and clinical characteristics were similar when comparing patients treated initially with levofloxacin or macrolides. There were no differences between groups regarding the presence at baseline of significant prognostic factors, such as age 170 years, presence of comorbid conditions, hypoalbuminemia, renal insufficiency, and multilobar pneumonia. The mean partial pressure of oxygen breathing room air at admission to the hospital was similar in the 2 groups (59.72 mm Hg vs mm Hg; P p.950). The percentage of patients classified into high-risk pneumonia classes according to PSI score (i.e., those with 190 points) did not differ between groups (42.5% vs. 47.5%; P p.688). The mean duration of intravenous therapy ( SD) was days for patients treated with levofloxacin and days for those treated with macrolides ( P p.043). Complete duration of antibiotic therapy ( SD) was days and days, respectively ( P p.004). Table 2 shows the main clinical outcomes for patients treated with either levofloxacin or macrolides. As can be seen, patients receiving levofloxacin had a faster time to defervescence and to resolution of cough. No difference was observed in time to resolution of chest pain between the 2 groups. Patients treated with levofloxacin achieved overall clinical stability significantly earlier than did patients treated with macrolides (3.0 vs. 5.0 days). Figure 1. Data for patients hospitalized with Legionella pneumonia in Spain, Levofloxacin, cases treated with levofloxacin; LP, cases of Legionella pneumonia; macrolides, cases treated with macrolides; mortality, mortality associated with Legionella pneumonia; urinary antigen test, cases of Legionella pneumonia confirmed with use of urine antigen testing. 796 CID 2005:40 (15 March) Mykietiuk et al.

4 Table 1. Baseline characteristics of 120 patients with community-acquired Legionella pneumonia treated with either levofloxacin or macrolides. Variable Lvfx group (n p 40) Macrolide group a (n p 80) P Age, mean years SD Age 170 years 7 (17.5) 13 (16.2).862 Male sex 36 (90) 69 (86.2).558 Alcohol consumption 180 g per day 17 (42.5) 37 (46.2).697 Smoking 20 (50) 41 (51.2).845 Previous antibiotic therapy 2 (5) 9 (11.2).263 Comorbid condition 14 (35) 27 (33.7).892 Hypoalbuminemia b 14 (35) 40 (50).127 Renal insufficiency c 7 (17.5) 9 (11.2).341 Partial pressure of oxygen breathing room air, mean mm Hg SD Multilobar pneumonia 18 (45) 35 (43.7).804 Pleural effusion 5 (12.5) 14 (17.5).477 Pneumonia severity index, mean score SD High-risk pneumonia d All 17 (42.5) 38 (47.5).688 Risk class IV 13 (32.5) 26 (32.5).723 Risk class V 4 (10 ) 12 (15).723 Admission to the ICU 5 (12.5) 2 (2.5).783 NOTE. Data are no. (%) of patients, unless otherwise indicated. ICU, intensive care unit; Lvfx, levofloxacin. a Forty-eight patients who were treated with macrolides also received rifampin. b Albumin level,!3 g/dl. c Serum creatinine level, 11.5 mg/dl. d Pneumonia severity index, 190 points (i.e., risk classes IV and V). As shown in table 2, there were no significant differences between groups regarding the development of complications, which occurred in 25% of both antibiotic groups. The most frequent complications were respiratory failure and a worsening of comorbid conditions. The median length of hospital stay was significantly shorter for patients treated with levofloxacin. The length of stay for other patients with pneumonia did not change significantly over the study period, ranging from a mean of days in 1995 to a mean of in No differences in outcome between patients who received either erythromycin or clarithromycin or patients who received regimens that included rifampin were observed. The incidence of -related adverse events was similar for the 2 groups, occurring in 8 (20%) of 40 patients in the levofloxacin group, compared with 24 (30%) of 80 patients in the macrolide group ( P p.243). All of the drugrelated adverse events were minor, the most common events being phlebitis (occurring in 12.5% of patients in the levofloxacin group, compared with 10% of patients in the macrolide group), gastrointestinal disorders (5% vs. 5%), liver function test abnormalities (2.5% vs. 3.7%), and rash (2.5% vs. 3.7%). When considering the total number of patients with Legionella pneumonia, the early case-fatality rate was 2.9% (4 of 139 patients), and the overall case-fatality rate was 5% (7 of 139 patients). No patients died during the last 3 years of the study (0 of 46 patients). As shown in table 2, early and overall casefatality rates did not differ significantly between patients receiving levofloxacin and those receiving macrolides. Although there was a trend towards a lower case-fatality rate in the levofloxacin group, small numbers of fatal cases of pneumonia may produce type II error. Among patients who received an initial inappropriate empirical therapy, the early case-fatality rate was 5.5% (1 of 18 patients), and the overall case-fatality rate was 11.1% (2 of 18 patients). DISCUSSION The present study offers a detailed analysis of clinical outcomes for a large series of patients with community-acquired Legionella pneumonia documented in recent years in a single institution. Clinical management and diagnostic evaluation of patients with community-acquired pneumonia, including urine antigen testing, did not undergo substantial changes during the study period, although levofloxacin has progressively replaced Outcomes of Legionella Pneumonia CID 2005:40 (15 March) 797

5 Table 2. Clinical outcomes for 120 patients with community-acquired Legionella pneumonia treated with either levofloxacin or macrolides. Clinical outcome Lvfx group (n p 40) Macrolide group a (n p 80) P Time to defervescence, days b!.001 Mean SD Median (IQR) 2.0 ( ) 4.50 ( ) Time to resolution of cough, days!.001 Mean SD Median (IQR) 3.0 ( ) 5.0 ( ) Time to resolution of chest pain, days.893 Mean SD Median (IQR) 3.0 ( ) 3.0 ( ) Time to achieve clinical stability, days.002 Mean SD Median (IQR) 3.0 ( ) 5.0 ( ) In-hospital complications c Any 10 (25) 20 (25).906 Respiratory failure 4 (10) 12 (15).507 Worsening of comorbid conditions 5 (12.5) 6 (7.5).371 Acute renal failure 3 (7.5) 6 (7.5) Length of hospital stay, days.014 Mean SD Median (IQR) 8.0 ( ) 10.0 ( ) Early case-fatality rate d 0 3 (3.7).550 Overall case-fatality rate e 1 (2.5) 4 (5).518 NOTE. Data are no. (%) of patients, unless otherwise indicated. IQR, interquartile range; Lvfx, levofloxacin. a Forty-eight patients who were treated with macrolides also received rifampin. b Temperature, 37.0 C. c Some patients had 11 complication. d Death!48 h after admission. e Death!30 days after admission. macrolides as the initial therapy for Legionella pneumonia since Benin et al. [13] recently reported data from a large-scale study by the Centers for Disease Control and Prevention showing a decrease in the case-fatality rate for community-acquired Legionella pneumonia from 26% to 10% for the period These investigators observed a relationship between the increasing use of urine antigen testing as a useful diagnostic tool for Legionella pneumonia and the sustained decrease in mortality. However, they stated that improvement in empirical antibiotic s, promoted by the implementation of community-acquired pneumonia guidelines, could be another reason for the reduction in mortality. We observed a low casefatality rate throughout the entire study period; this low rate was more evident during the last 3 years of the study period, when levofloxacin accounted for most initial s (figure 1). This finding is in accordance with some recent case studies of patients with Legionella pneumonia who received diagnoses by means of urine antigen testing; these case studies have reported case-fatality rates of 0% 5% [9, 17, 18]. Although there have been no prospective controlled trials addressing the subject, some authorities have recommended therapy with fluoroquinolones or azithromycin, rather than with older macrolides, for the of Legionella pneumonia [1]. Indeed, these agents have been shown to be better than erythromycin in inhibiting the intracellular growth of L. pneumophila, both in vitro and in animal models [19 22]. In this regard, Yu et al. [18] have reported a successful experience with levofloxacin therapy in 75 hospitalized and ambulatory patients with Legionella pneumonia who were identified through a database of patients enrolled in 6 trials involving community-acquired pneumonia. Similarly, Plouffe et al. [17] reported similar results using azithromycin in an open, noncomparative trial involving 25 hospitalized patients with Legionella pneumonia who had significant comorbidities. Both studies indicate that use of these highly active agents may 798 CID 2005:40 (15 March) Mykietiuk et al.

6 shorten the duration of antibiotic, although neither study analyzed outcomes other than clinical response (i.e., cure, improvement, or failure) and safety. In our study, we observed a nonsignificant trend toward a lower case-fatality rate in patients who received levofloxacin therapy. However, there were significant differences favoring levofloxacin in important outcomes other than mortality, such as time to defervescence, time to overall clinical stability, duration of antibiotic, and length of hospital stay. These data provide evidence about the safety of an early switch from intravenous to oral levofloxacin for of Legionella pneumonia and suggest that the use of agents with enhanced activity against Legionella species may result in a faster resolution of clinical signs and symptoms. Moreover, they provide valuable information in terms of better understanding the continuous improvement in Legionella pneumonia outcomes following the generalized use of urine antigen testing. Because our study is not randomized, the possible superiority of levofloxacin therapy over therapy with the older macrolides should be interpreted with caution. It is possible that concomitant early recognition of Legionella and a trend toward levofloxacin in the past few years coincided to produce less morbidity and fewer fatalities and that levofloxacin is not a better than macrolides. A prospective, randomized trial comparing therapy with the more effective macrolide, azithromycin, with levofloxacin therapy is warranted. However, it seems unlikely that any randomized trials involving the antibiotic of Legionella pneumonia will be possible in the near future [23, 24]. In summary, our data show that, although Legionella pneumonia continues to be associated with significant morbidity in hospitalized patients, the current mortality rate is substantially lower than that traditionally reported. Furthermore, our findings suggest that levofloxacin therapy produces a faster clinical response than does therapy with the older macrolides, allowing a shorter length of hospital stay. Acknowledgments Financial support. Fondo de Investigación Sanitaria de la Seguridad Social (grants 98/0783 and 00/0438); Institut d Investigació Biomèdica de Bellvitge (IDIBELL), Hospital Universitari de Bellvitge ( ; to A.M.); University of Barcelona, Beca de Formació en la Recerca i la Docència (2000; to N.F.-S). Potential conflicts of interest. All authors: no conflicts. References 1. Mandell LA, Bartlett JG, Dowell SF, File TM Jr, Musher DM, Whitney C. Update of practice guidelines for the management of communityacquired pneumonia in immunocompetent adults. Clin Infect Dis 2003; 37: Stout JE, Yu VL. Legionellosis. N Engl J Med 1997; 337: File TM Jr. Community-acquired pneumonia. Lancet 2003; 362: Fang GD, Fine M, Orloff J, et al. New and emerging etiologies for community-acquired pneumonia with implications for therapy: a prospective multicenter study of 359 cases. Medicine (Baltimore) 1990; 69: Rosón B, Carratalà J, Dorca J, Casanova A, Manresa F, Gudiol F. Etiology, reasons for hospitalization, risk classes, and outcomes of communityacquired pneumonia in patients hospitalized on the basis of conventional admission criteria. Clin Infect Dis 2001; 33: Marston BJ, Plouffe JF, File TM Jr, et al. Incidence of community-acquired pneumonia requiring hospitalizations: results of a population-basedactive surveillance study in Ohio. The Community-Based Pneumonia Incidence Study Group. Arch Intern Med 1997; 157: Rello J, Bodi M, Mariscal D, et al. Microbiological testing and outcome of patients with severe community-acquired pneumonia. Chest 2003; 123: Vergis EN, Yu VL. New directions for future studies of communityacquired pneumonia: optimizing impact on patient care. Eur J Clin Microbiol Infect Dis 1999; 18: Lettinga KD, Verbon A, Weverling GJ, et al. Legionnaires disease at Dutch flower show: prognostic factors and impact of therapy. Emerg Infect Dis 2002; 8: Lieberman D, Porath A, Schlaeffer F, Lieberman D, Boldur I. Legionella species community acquired pneumonia: a review of 56 hospitalizedadult patients. Chest 1996; 109: Yu VL, Kroboth FJ, ShonnardJ, Brown A, McDearman S, Magnussen M. Legionaires disease: new clinical perspective from a prospective pneumonia study. Am J Med 1982; 73: Fernandez-Sabé N, Rosón B, Carratalà J, Dorca J, Manresa F, Gudiol F. Clinical diagnosis of Legionella pneumonia revisited: evaluation of the Community-Based Pneumonia Incidence Study Group Scoring System. Clin Infect Dis 2003; 37: Benin AL, Benson RF, Besser RE. Trends in legionnaires disease, : declining mortality and new patterns of diagnosis. Clin Infect Dis 2002; 35: Barlow GD, Lamping DL, Davey PG, Nathwani D. Evaluation of outcomes in community-acquired pneumonia: a guide for patients, physicians and policy-makers. Lancet Infect Dis 2003; 3: Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify lowrisk patients with community-acquired pneumonia. N Engl J Med 1997; 336: Halm EA, Fine MJ, Marrie TJ, et al Time to clinical stability in patients hospitalized with community-acquired pneumonia: implications for practice guidelines. JAMA 1998; 279: Plouffe JF, Breiman RF, Fields BS, et al. Azithromycin in the of Legionella pneumonia requiring hospitalization. Clin Infect Dis 2003; 37: Yu VL, Greenberg RN, Zadeikis N, et al. Levofloxacin efficacy in the of community-acquired Llegionellosis. Chest 2004; 125: Meyer RD. Role of the quinolones in the of legionellosis. J Antimicrob Chemother 1991; 28: Jonas D, Engels I, Friedhoff C, Spitzmüller B, Daschner FD, Frank U. Efficacy of moxifloxacin, trovafloxacin, clinafloxacin and levofloxacin against intracellular Legionella pneumophila. J Antimicrob Chemother 2001; 47: Baltch AL, Smith RP, Franke MA, Michelsen PB. Antibacterial effects of levofloxacin, erythromycin and rifampin in a human monocyte system against Legionella pneumophila. Antimicrob Agents Chemother 1998; 42: Edelstein PH, Edelstein MA, Lehr KH, Ren J. In-vitro activity of levofloxacin against clinical isolates of Legionella spp, its pharmacokinetics in guinea pigs, and use in experimental Legionella pneumophila pneumonia. J Antimicrob Chemother 1996; 37: Edelstein PH. Antimicrobial chemotherapy for legionnaires disease: time for a change. Ann Intern Med 1998; 129: Fields BS, Benson RF, Besser RE. Legionella and legionnaire s disease: 25 years of investigation. Clin Microbiol Rev 2002; 15: Outcomes of Legionella Pneumonia CID 2005:40 (15 March) 799