ORIGINAL INVESTIGATION. Azithromycin Monotherapy for Patients Hospitalized With Community-Acquired Pneumonia

ORIGINAL INVESTIGATION Azithromycin Monotherapy for Patients Hospitalized With Community-Acquired Pneumonia A3 1 2-Year Experience From a Veterans Affairs Hospital Randy B. Feldman, MD; David C. Rhew, MD; John Y. Wong, BS; Robert Antoine Charles, PharmD, MS; Matthew Bidwell Goetz, MD Background: Current American Thoracic Society (ATS) community-acquired pneumonia treatment guidelines recommend azithromycin monotherapy for a limited subset of hospitalized patients. We evaluated the effectiveness of azithromycin monotherapy in a more generalized population of patients hospitalized with mild-tomoderate community-acquired pneumonia. Methods: We reviewed medical records from a Veterans Affairs facility for patients admitted with communityacquired pneumonia between December 1, 1997, and June 30, 2001, comparing those receiving azithromycin monotherapy, other ATS-recommended antibiotics, and non ATS-recommended antibiotics. We excluded patients with immunosuppression, metastatic cancer, or hospitalacquired pneumonia. Outcome measures included times to stability, meeting criteria for change to oral therapy, and eligibility for discharge; length of stay; intensive care unit transfer; and mortality. Outcomes were adjusted for pneumonia severity, skilled nursing facility status, and processes of care. Results: A total of 442 patients were eligible for the study (221 in the azithromycin monotherapy group, 129 in the ATS group, and 92 in the non-ats group). Times to clinical stability and to fulfilling early switch criteria were not statistically significantly different among the 3 groups. Mean time to fulfilling early discharge criteria was 2.48 days for patients receiving azithromycin monotherapy vs 2.84 days for those receiving ATS antibiotics (P=.008) and 2.58 days for those receiving non-ats antibiotics (P=.64). Overall mean length of stay was shorter in the azithromycin monotherapy group (4.35 days) vs the ATS (5.73 days) (P=.002) and non-ats (6.21 days) (P.001) groups. Mortality, intensive care unit transfer, and readmission rates were similar across the groups. Conclusion: Azithromycin monotherapy is equally as efficacious as other ATS-recommended regimens for treating hospitalized patients with mild-to-moderate community-acquired pneumonia. Arch Intern Med. 2003;163:1718-1726 From the Division of Infectious Diseases, the VA Greater Los Angeles Healthcare System, Los Angeles, Calif (Drs Feldman, Rhew, and Goetz); Zynx Health Inc, Cedars-Sinai Health System Department of Health Services Research, Beverly Hills, Calif (Dr Rhew and Messrs Wong and Charles); and Department of Medicine, The David Geffen School of Medicine at UCLA, Los Angeles (Drs Feldman, Rhew, and Goetz). The authors have no relevant financial interest in this article. COMMUNITY-ACQUIRED pneumonia (CAP) remains a significant cause of morbidity and mortality and is the leading cause of death due to infection in the United States. 1 An estimated 5 1 2 million cases occur annually in the United States, of which 20% require admission to the hospital. 2 In recent years, several organizations, including the British Thoracic Society (2001), 3 the American Thoracic Society (ATS) (2001), 4 the Infectious Diseases Society of America (2000), 5 the Centers for Disease Control and Prevention (2000), 6 and the Canadian Infectious Diseases Society in collaboration with the Canadian Thoracic Society (2000), 7 have published guidelines that address the empiric choice of antibiotics for hospitalized patients with nonsevere CAP. Each of these organizations recommends a -lactam plus a macrolide or an antipneumococcal ( respiratory ) fluoroquinolone alone as an appropriate empiric antibiotic drug choice for individuals from the community admitted to the general medical ward. The ATS also recommends monotherapy with a macrolide, specifically, azithromycin, for a subset of patients requiring intravenous therapy who are admitted to the hospital ward. According to ATS guidelines, to be eligible for azithromycin monotherapy, patients should not have cardiopulmonary disease or any of the modifying risk factors outlined in Table 1. Data from comparative (vs intravenous cefuroxime sodium with or without intravenous erythromycin lactobionate) and noncomparative phase 3 trials have demonstrated that parenteral azithromycin monotherapy is an effective treatment for hospitalized patients with CAP. 8 Several patients in these trials had risk factors for complicated episodes. For example, approximately 50% were older than 1718

65 years, a third were smokers, 20% had diabetes mellitus, and many experienced chronic obstructive airway disease. A randomized, multicenter study 9 (mean patient age, 68 years; 30% of patients were from a nursing home) has shown that monotherapy with intravenous azithromycin has the same efficacy as and fewer adverse effects than the regimen of cefuroxime with or without erythromycin. These studies have also demonstrated shorter treatment durations and possible hospital stay advantages with azithromycin monotherapy compared with combination therapy. However, few data are available regarding the utility of parenteral azithromycin monotherapy for the treatment of CAP in clinical practice as opposed to the setting of a research study. 10,11 Furthermore, some experts have expressed concerns that increasing rates of macrolide resistance among Streptococcus pneumoniae might decrease the clinical effectiveness of azithromycin monotherapy. 5 For the past 5 1 2 years, the VA Greater Los Angeles Healthcare System (VAGLAHS) has used guidelines that recommend the use of intravenous azithromycin monotherapy for the empiric treatment of patients hospitalized with nonsevere CAP (ie, not requiring care in the intensive care unit [ICU]). The objectives of this study are to evaluate the clinical and bacteriologic outcomes in patients treated with azithromycin monotherapy and to determine the effectiveness of this regimen in a realworld setting. METHODS STUDY DESIGN We performed a retrospective cohort study to assess clinical and bacteriologic outcomes in hospitalized patients with nonsevere CAP. We reviewed the medical records of patients hospitalized at the VAGLAHS between December 1, 1997 (when azithromycin monotherapy was first recommended as empiric treatment for hospitalized patients with nonsevere CAP), and June 30, 2001. This study was approved by the institutional review board at the VAGLAHS. STUDY POPULATION Patients were eligible for the study if they were 18 years or older, were admitted to the general medical ward, and had a documented International Classification of Diseases, Ninth Revision, code for CAP. 12 To ensure accuracy of the CAP diagnosis, a physician trained in infectious diseases (R.B.F.) reviewed every potentially eligible patient medical record to verify the presence of (1) radiographic evidence of a new pulmonary infiltrate within 48 hours of admission not attributable to another cause and (2) clinical evidence suggestive of pneumonia (ie, the presence of 2 of the following: temperature 38 C, new or worsening cough, production of purulent sputum, pleuritic chest pain, dyspnea, rales/rhonchi, egophony, white blood cell count 10.0/µL, or leukocyte band forms 15%). Patients were excluded from the study if they were known to be human immunodeficiency virus infected, neutropenic, or immunosuppressed for other reasons (eg, receiving corticosteroids, chemotherapy, or other immunosuppressive agents or the presence of metastatic malignancy, leukemia, or lymphoma) or if they had documented mycobacterial infection. Patients with presumed hospital-acquired pneumonia were also excluded; this group included patients who were admitted as Table 1. Modifying Risk Factors for Pneumonia* Risk of pneumonia owing to penicillin- and drug-resistant pneumococci Multiple medical comorbidities Immunosuppressive illness (including therapy with corticosteroids) -Lactam therapy within the past 3 mo Alcoholism Age 65 y Exposure to a child in a day care center Risk of pneumonia owing to Pseudomonas aeruginosa Structural lung disease (bronchiectasis) Corticosteroid therapy ( 10 mg of prednisone per day) Broad spectrum antibiotic therapy for 7 d in the past month Malnutrition Risk of pneumonia owing to enteric gram-negative organisms Multiple medical comorbidities Underlying cardiopulmonary disease Recent antibiotic therapy Residence in a nursing home *Adapted from Niederman et al. 4 Copyright 2001, American Thoracic Society. inpatient transfers from an outside hospital or who were discharged from any hospital within 14 days preceding their admission. Patients who were admitted from an outside emergency department (ED) or urgent care facility were included in the study. Finally, patients who were transferred to the ICU within 24 hours of hospital admission were excluded from the primary analysis. It was assumed that the most likely reason these patients required ICU admission within 24 hours was that they were inappropriately admitted to the ward in the first place (ie, they should have been admitted to the ICU initially) as opposed to having failed their initial antibiotic drug regimen. INSTITUTION The VAGLAHS comprises the combined resources of the Veterans Affairs Sepulveda Ambulatory Care Center (Sepulveda, Calif), the Veterans Affairs Los Angeles Ambulatory Care Center, and the West Los Angeles Healthcare Center (West Los Angeles, Calif). In addition, the VAGLAHS includes 2 large satellite clinics (in Bakersfield and Santa Barbara, Calif). The catchment area of the VAGLAHS includes more than 750000 veterans, and it is the largest and most complex health care facility in the Department of Veterans Affairs. Comprehensive ambulatory services and all inpatient and tertiary care services for VAGLAHS are provided at the West Los Angeles Healthcare Center. This facility has 283 active acute care beds. Almost all surgical services (excluding transplantation), all subspecialties of internal medicine, and comprehensive neurologic and psychiatric services are offered. In the 12 months ending September 30, 2000, the VAGLAHS provided medical care to more than 75000 veterans. More than 26% of these patients were African American, more than 12% were Hispanic, and approximately 95% were male. DATA COLLECTION The following data were collected for all eligible patients: (1) choice of initial antibiotic agent, (2) patient demographics, (3) Pneumonia Severity Index (PSI) score, 13 (4) adherence to key process-of-care indicators, (5) clinical outcomes, and (6) microbiologic data. For this study, the components of the pa- 1719

1026 Patients Initially Identified 968 Patients 793 Patients 778 Patients 766 Patients 602 Patients 514 Patients No Evidence of CAP (n = 58) Metastatic Malignancy (n = 175) Immunosuppression (n = 15) Mycobacterial Disease (n = 12) Admitted to ICU (n = 142) or Transferred Within 24 h of Admission (n = 22) Transfer From Outside Hospital (n = 88) Discharged Within Preceding 14 d (n = 72) site in a patient admitted to the study was presumed to be the cause of the pneumonia. Such sites included blood, pleural fluid, and bronchoalveolar lavage fluid; 10000 colony-forming units per milliliter was used as the cutoff value to define a positive result. 18 Also, organisms recovered from suitable sputum specimens (ie, containing 25 white blood cells and 10 squamous epithelial cells per high-power field) 5 were considered true pathogens. Positive IgM titers or a demonstrated 4-fold rise in IgG titers to known pulmonary pathogens (Mycoplasma pneumoniae and Chlamydia pneumoniae) or a positive Legionella urinary antigen test result was also considered diagnostic for the presence of the corresponding atypical organism. All laboratory testing was performed at the discretion of the treating physician. In vitro susceptibility testing was performed using agarbased Kirby-Bauer disk diffusion techniques, and minimum inhibitory concentrations (MICs) were determined for antibiotics using an antibiotic gradient strip (Etest; AB-BIODISK North America Inc, Piscataway, NJ). STATISTICAL ANALYSIS 442 Patients Included in the Study Flowchart of excluded patients. CAP indicates community-acquired pneumonia; ICU, intensive care unit. tient s initial antibiotic regimen were defined as those antibiotic drugs that were administered consistently after the first dose. Most patients received their first dose of antibiotic(s) in the ED; however, this did not always reflect the antibiotic course received during the remainder of the hospitalization. We reasoned that by the second dose of antibiotic, the admitting physician had evaluated the patient and started therapy for CAP. Any addition to or discontinuation of this initial regimen was noted. Demographic data included age, race, sex, smoking history, and any previous history of pneumonia. Severity of illness was measured using the PSI. Key process-of-care indicators included the timing of antibiotic drug administration relative to hospital admission, 14 administration of antibiotics before drawing of blood cultures, 14 and measurement of oxygenation within 24 hours of admission. 15 Clinical outcomes included transfer to the ICU after 24 hours of admission, time to clinical stability (defined as systolic blood pressure 90 mm Hg, pulse rate 100 beats/min, temperature 38.3 C, respiratory rate 24 breaths/min, arterial oxygen saturation 90% or no further need for supplemental oxygen, able to eat, and no altered mental status), 16 time to fulfilling early switch to oral antibiotic drug treatment criteria (improved cough, tolerating oral intake, defervescence, and normalized white blood cell count) and early discharge criteria (no need to treat comorbid conditions, no need for further diagnostic workup[s], and no social needs), 17 number of days hospitalized, mortality (in the hospital and within 30 days of discharge), return to the ED within 30 days of discharge, and readmission to the hospital within 30 days of discharge. Microbiologic data consisted of identification of any organisms obtained from bacterial cultures and their corresponding antibiotic sensitivities and results of serologic testing (Mycoplasma pneumoniae and Chlamydia pneumoniae) or urine Legionella antigen testing. MICROBIOLOGIC DATA Any microbiologic organism(s) with the potential for respiratory pathogenicity that was recovered from a usually sterile body Univariate analyses were performed to assess differences in patient characteristics. Differences in the means of continuous variables were evaluated using unpaired t tests, and differences in proportions were evaluated using 2 tests. Cochran-Mantel- Haenszel statistics were used for severity-stratified analyses. Least squared means was used to severity adjust continuous variables (age). Cox proportional hazards were used to adjust for severity of illness (PSI score), nursing home residence, and antibiotic drug therapy within 8 hours of hospital admission. Time to clinical stability was censored by inpatient death and discharge before achieving clinical stability. Time to early switch was censored by inpatient death and eligibility to switch to oral medication. Time to early discharge was censored by inpatient death and eligibility for early discharge. Length of stay was censored by inpatient death. Logistic regression was used to adjust simultaneously for potential confounding variables such as PSI score (classes I and II, III, IV, and V), skilled nursing facility status, receiving antibiotic agents within the first 8 hours of treatment, and having blood cultures drawn within 24 hours of hospital admission. Odds ratios and 95% confidence intervals were computed. RESULTS PATIENT GROUPS One thousand twenty-six patients were initially identified for the study by International Classification of Diseases, Ninth Revision, coding. After reviewing the medical records, we excluded 584 patients from the analysis (Figure). The remaining 442 patients were divided into 3 comparison groups: (1) the azithromycin monotherapy group consisted of 221 patients receiving azithromycin alone as the initial antibiotic regimen for ward treatment of pneumonia, (2) the ATS group consisted of 129 patients receiving a 2001 ATS-recommended antibiotic regimen other than azithromycin monotherapy as the initial treatment of mild-to-moderate CAP, and (3) the non- ATS group consisted of 92 patients receiving an initial antibiotic regimen that was not part of the 2001 ATS recommendations for treatment of mild-to-moderate CAP. 1720

Table 2. Demographic and Clinical Characteristics of the 3 Patient Groups Characteristic Azithromycin Monotherapy ATS Regimen (n = 129) vs ATS) Non-ATS Regimen (n = 92) vs Non-ATS) Race, % White 57.5 56.6.95 63.0.47 African American 30.3 29.5.95 26.1.47 Hispanic 10.4 12.4.95 8.7.47 Asian 1.8 1.6.95 1.1.47 Other 0 0.95 1.1.47 Sex, % male 94.6 95.3.75 98.9.08 Age, mean, y 64.6 66.9.57 68.9.09 Current smoker, % 48.0 42.6.33 38.0.11 Previous pneumonia history 18.1 13.3.25 18.7.90 during study, % Admitted from SNF, % 4.1 7.8.14 26.1.001 Addition or change of antibiotics 1.8 1.6.86 6.5.03 within 48 h, % PSI score, mean (95% CI) 89.2 (85.5-92.8) 95.1 (89.9-100.2).07 98.7 (92.7-104.6).14 PSI class, % I or II 24.9 22.5 19.6 III 31.2 20.2 22.8 IV 35.3 45.7 43.5 V 8.6 11.6 14.1 Abbreviations: ATS, American Thoracic Society; CI, confidence interval; PSI, Pneumonia Severity Index; SNF, skilled nursing facility. Table 3. Process-of-Care Measures in the 3 Patient Groups Measure Azithromycin Monotherapy ATS Regimen (n = 129) vs ATS) Non-ATS Regimen (n = 92) vs Non-ATS) s received 94.1 99.2.02 91.3.36 within 8 h, % Blood cultures drawn within 24 h, % 75.6 92.2.001 90.2.003 Of blood cultures drawn, those drawn 98.8 (167/169) 100 (119/119).23 98.8 (83/84).45 within 24 h, % (No./total No.) Oxygen assessment within 24 h, % 93.7 99.2.003 88.0.09 Abbreviation: ATS, American Thoracic Society. PATIENT DEMOGRAPHICS The 3 groups did not differ significantly in terms of average age, racial composition, smoking history, or previous pneumonia history. A small but statistically significantly greater percentage of male patients received non ATS-recommended antibiotic regimens than received azithromycin alone (Table 2). There were no significant differences in the individual PSI risk classes among the 3 groups. Also, there were no significant differences in the individual comorbid conditions composing the PSI among the 3 groups except for a higher percentage of cerebrovascular disease in the non-ats group (data not shown). Patients admitted from skilled nursing facilities were more likely to receive non ATS-recommended antibiotic agents than either azithromycin monotherapy or ATS-recommended antibiotics. We also evaluated whether the initial antibiotic regimens were changed as a potential surrogate marker for increased severity of illness. Individuals in the non-ats antibiotic group were more likely to have their regimens completely changed or to have other antibiotics added within 48 hours of hospital admission than were patients undergoing azithromycin monotherapy. PROCESS-OF-CARE MEASURES In terms of adherence to process-of-care measures that have previously been shown to be associated with improved clinical outcomes, 13 patients receiving either an ATS or a non-ats antibiotic regimen were significantly more likely to have blood cultures drawn within 24 hours of admission than were patients receiving azithromycin monotherapy, and patients receiving ATS-recommended antibiotics had a higher probability of receiving oxygen assessment within 24 hours of hospital admission than patients receiving azithromycin monotherapy (Table 3). A higher percentage of patients in the ATS group than in the azithromycin monotherapy group received antibiotics within 8 hours of admission. There were no significant differences among the 3 groups in terms of receiving antibiotics within 4 hours of admission (data not shown) or in having blood cultures drawn within 24 hours of admission (for those pa- 1721

Table 4. Clinical Outcomes in the 3 Patient Groups Outcome Azithromycin Monotherapy ATS Regimen (n = 129) vs ATS) Non-ATS Regimen (n = 92) vs Non-ATS) Time to fulfilling clinical stability 1.72 (1.56-1.89) 2.18 (1.85-2.51).02 2.42 (1.72-3.13).20 criteria, mean (95% CI), d (n = 200) (n = 111) (n = 78) Time to fulfilling early switch 2.56 (2.30-2.82) 3.20 (2.71-3.69).02 3.85 (2.96-4.76).007 criteria, mean (95% CI), d (n = 168) (n = 82) (n = 60) Time to fulfilling early discharge criteria, 3.53 (3.22-3.85) 4.83 (4.03-5.62).003 5.45 (4.46-6.44).004 mean (95% CI), d (n = 152) (n = 73) (n = 41) Length of stay, mean (95% CI), d 4.30 (3.88-4.73) 5.78 (4.91-6.64).003 6.32 (5.27-7.36).006 (n = 129) (n = 92) Transferred to ICU, No. (%) 3 (1.4) 3 (2.3).50 3 (3.3).26 Died in the hospital, No. (%) 1 (0.5) 0.44 2 (2.2).15 ED visit within 30 d of discharge, No. (%) 30 (13.6) 18 (14.0).92 12 (13.0).90 ED visit within 30 d if discharge for pulmonary 7 (3.2) 10 (7.8).054 5 (5.4).34 infection (COPD, AECB, or pneumonia), No. (%) Readmission within 30 d of discharge, No. (%) 20 (9.0) 16 (12.4).32 10 (10.9).62 Readmission within 30 d of discharge for 0 8 (6.2).02 5 (5.4).31 pulmonary infection, No. (%) Died within 30 d of discharge (all causes), No. (%) 1 (0.5) 1 (0.8).70 3 (3.3).04 Abbreviations: AECB, acute exacerbation of chronic bronchitis; ATS, American Thoracic Society; CI, confidence interval; COPD, chronic obstructive pulmonary disease; ED, emergency department; ICU, intensive care unit. Table 5. Multivariate Analysis: Adjusted End Points for the 3 Patient Groups* End Point Azithromycin Monotherapy ATS Regimen (n = 129) vs ATS) Non-ATS Regimen (n = 92) vs Non-ATS) Time to fulfilling clinical stability criteria 1.42 (1.32-1.55) (n = 200) Time to fulfilling early switch criteria 1.82 (1.70-1.94) (n = 168) Time to fulfilling early discharge criteria 2.48 (2.33-2.65) (n = 152) Time to actual discharge 4.35 (3.83-4.86) 1.61 (1.45-1.76) (n = 111) 1.81 (1.64-1.99) (n = 82) 2.84 (2.62-3.08) (n = 73) 5.73 (5.03-6.43) (n = 129).09 1.62 (1.44-1.79) (n = 78).92 2.02 (1.82-2.22) (n = 60).008 2.58 (2.26-2.89) (n = 41).002 6.21 (5.40-7.03) (n = 92).08.10.64.001 Abbreviation: ATS, Amercian Thoracic Society. *Data are given as mean (95% confidence interval) days. The end points were adjusted for Pheumonia Severity Index class, skilled nursing facility status, receiving the first antibiotic drug within 8 hours of admission, and having blood cultures drawn within 24 hours of admission. tients who had blood cultures drawn). Influenza and pneumococcal vaccination (0.9% and 4.1% of all patients, respectively) were rarely administered but were more often given in the azithromycin monotherapy group (data not shown). CLINICAL OUTCOMES There were no significant differences among the 3 groups in terms of the mean number of days needed to reach clinical stability or to fulfill eligibility criteria for early switch from intravenous to oral antibiotic drug therapy (Table 4). However, patients receiving either non ATS- or ATSrecommended antibiotic regimens required a longer time to fulfill early discharge criteria than those in the azithromycin monotherapy group. In addition, overall mean length of stay was significantly longer in the ATS and non-ats groups than in the azithromycin monotherapy group. Except for the mean number of days to reach eligibility for early discharge in the non-ats antibiotic group, these same relationships held true when end points were adjusted in a multivariate analysis using the following variables: PSI score, nursing home residence status, blood cultures drawn within 24 hours of admission, and receiving antibiotic drugs within 8 hours of admission (Table 5). All 3 groups had similar rates of transfer to the ICU (after 24 hours), return to the ED within 30 days of discharge, and 30-day readmission. Individuals receiving non ATS-recommend regimens had a higher 30-day mortality rate (P=.04) than those who received azithromycin monotherapy. Also, individuals who received ATSrecommended regimens had a higher rate of readmission to the hospital for pulmonary infections than patients who received azithromycin monotherapy. One in-hospital death occurred in the azithromycin monotherapy group (cardiac arrest on hospital day 7 in a patient with cardiomyopathy and congestive heart failure), 2 in the non- ATS antibiotic group (1 patient had bowel perforation with Escherichia coli and Bacteroides fragilis sepsis and the other had an acute cerebrovascular accident, car- 1722

Table 6. Results in Patients in Whom Streptococcus pneumoniae Was Isolated Variable Patients, No. (n = 45) PSI Score, Mean Time to Clinical Stability, Mean, d Time to Early Switch, Mean, d Time to Early Dischange, Mean, d Erythromycin susceptible (n = 36) Azithromycin monotherapy 17 85.3 2.4 3.1 4.0 ATS therapy 17 92.5 2.4 3.3 5.2 Non-ATS therapy 2 74.5 2.0 2.0 6.0 Erythromycin resistant (n = 9) Azithromycin monotherapy 6 101.0 2.4 2.5 3.7 ATS therapy 1 115.0 3.0 4.0 4.0 Non-ATS therapy 2 80.0 1.5 2.0 4.5 Abbreviations: ATS, Amercian Thoracic Society; ICU, intensive care unit; PSI, Pneumonia Severity Index. diac arrest, and anoxic encephalopathy on hospital day 3), and none in the ATS group. All the patients who died were severely ill on hospital admission (ie, had PSI class IV or V). The number of deaths was not statistically significant among the 3 groups. MICROBIOLOGIC FINDINGS Presumed etiologic organisms were recovered in 112 cases. Consistent with findings from other studies, 19 the most common pathogens recovered were S pneumoniae (41% of all isolates recovered), Haemophilus influenzae (20%), Staphylococcus aureus (13%), and Moraxella catarrhalis (7%). 4 Other isolates included Klebsiella (5%), Pseudomonas (4%), other gram-negative rods (6%), and group B Streptococcus species (4%). Eighty-two percent of all isolates were recovered from sputum, 12% from blood, and 5% from other sites (eg, pleural fluid and bronchoalveolar lavage fluid). Eighty percent (36/45) of the S pneumoniae isolates were sensitive to erythromycin therapy. These isolates were obtained in essentially equal numbers across each year of the study (data not shown). There were no ICU transfers or deaths in any patient from whom S pneumoniae was recovered. A subset analysis demonstrated similar outcomes in patients with erythromycinsusceptible S pneumoniae isolates vs those with erythromycin-resistant isolates (Table 6). COMMENT Azithromycin monotherapy has been recommended by the ATS for empiric treatment of CAP in hospitalized (ward) patients in limited circumstances. 4,5 Hesitation in broadening this recommendation may be due to the relative paucity of comparative clinical trials using azithromycin alone, concerns about increasing rates of S pneumoniae resistance to macrolides, 3,7,20 and lack of published clinical experience regarding the effectiveness of parenteral azithromycin in the treatment of CAP in routine clinical practice. 21 This study sought to provide a real-world comparison of azithromycin monotherapy and other treatment options for hospitalized patients with CAP and to potentially determine whether azithromycin monotherapy could safely and effectively be applied in a broader population of patients than defined by the ATS. 5 In this study, azithromycin monotherapy resulted in equivalent times to fulfilling clinical stability and early switch criteria. In addition, there were no significant differences in rates of ICU transfer, death during hospitalization, or utilization of the ED. Patients receiving azithromycin monotherapy reached the criteria for early discharge sooner than those in the 2 comparison groups and had an overall shorter length of stay when adjusted for severity indicators. Also, patients receiving azithromycin monotherapy had lower 30-day mortality rates than patients receiving non ATS-recommended regimens. No patients receiving azithromycin monotherapy were readmitted to the hospital within 30 days of discharge for new or recurrent pulmonary infection, whereas 6% and 5% of patients in the ATS and non-ats groups, respectively, were readmitted for such conditions. We determined postdischarge outcomes from the medical record with reasonable certainty because most Veterans Affairs patients receive postdischarge follow-up in Veterans Affairs clinics, and the Department of Veterans Affairs has a computerized medical record system that is fully integrated with all of its clinics and hospitals. Furthermore, there is no a priori reason to expect bias in the rate of reporting of outpatient mortality in persons who did, or did not, receive monotherapy with azithromycin. Reports of increasing pneumococcal macrolide resistance have led to recommendations for azithromycin monotherapy to be limited 4,21 or altogether avoided. 3,7,20,22,23 However, some authors have noted that these recommendations are primarily based on data from in vitro studies using standardized MICs, 24 which may not have ideally reflected the in vivo activity or the effectiveness of the drug. 25,26 In an attempt to address these issues, we performed a subset analysis of patients in whom S pneumoniae was isolated. The results showed that nearly 20% of S pneumoniae isolates were resistant to erythromycin, all but one of which were also resistant to penicillin. However, none of the 10 patients in our study with erythromycin-resistant pneumococci died or were transferred to the ICU, including the 6 who received azithromycin monotherapy. Furthermore, the initial choice of empiric antibiotic(s) did not have an effect on the times to reaching clinical stability or eligibility for switch from parenteral to oral antibiotics in patients infected with erythromycin-resistant pneumococci. This lack of correlation among erythromycin resistance, use of macrolides, and clinical outcomes has also been demonstrated in 2 prospective clinical trials. 10,27 Although we do not test the azithromycin susceptibility of pneumococcal iso- 1723

lates at our facility, previous studies have shown an excellent correlation between the susceptibility of S pneumoniae isolates to erythromycin and azithromycin. 21,28,29 In North America, 60% to 70% of macrolide resistance in pneumococci is due to the M-phenotype, an efflux pump associated with the mefe gene. 30-32 These organisms exhibit moderate resistance to erythromycin and azithromycin with MICs of 1 to 32 µg/ml. 31 In contrast, high-level resistance with MICs greater than 64 µg/ml due to ribosomal methylation associated with the ermam gene (MLSB phenotype) is more common in Europe. 33 Other mechanisms for macrolide resistance are rare. susceptibility data from the 9 erythromycin-resistant S pneumoniae isolates in this study showed that 67% (6/9) of the isolates were resistant to clindamycin (data not shown), which suggested that the MLSB phenotype (as determined by the expression of macrolide and clindamycin resistance) predominated in most of our erythromycin-resistant S pneumoniae isolates. Despite our findings, recent reports of increasing erythromycin MICs among S pneumoniae isolates expressing the M-phenotype 30,31 and failures of azithromycin monotherapy in the treatment of persons infected by pneumococci with erythromycin MICs in the 8- to 16-µg/mL range 23 indicate the need for continued surveillance of the effectiveness of azithromycin monotherapy for CAP. Other studies 34,35 suggest that serotype may play a more important role than antibiotic susceptibility in mortality due to pneumococcus. To our knowledge, this study is the first to compare azithromycin monotherapy with other pneumonia treatments based on whether the treatment regimens were consistent with the 2001 ATS CAP treatment guidelines. This study corroborates the efficacy of azithromycin monotherapy for patients with CAP, as has been documented in other studies, 9,10 but does so outside the artificial conditions of a clinical trial. A recently published retrospective study by Lentino and Krasnicka 11 compared intravenous azithromycin with other parenteral antibiotics for the treatment of Veterans Affairs patients hospitalized with CAP. However, there are some primary differences between the present study and that by Lentino and Krasnicka: (1) the present study enrolled a larger number of patients (442 vs 92) over a longer study period (3 1 2 years vs 2 1 2 years); (2) the present study assessed the appropriateness of the initial antibiotic regimen by classifying it according to the 2001 ATS CAP treatment guidelines, whereas the study by Lentino and Krasnicka compared patients receiving azithromycin with those receiving any other parenteral antibiotic(s); and (3) the present study evaluated several measures to assess efficacy of treatment, including time to clinical stability, time to fulfilling early switch criteria, various clinical outcomes, and length of stay, whereas the study by Lentino and Krasnicka evaluated length of hospital stay only. This study has several limitations. First, it is a retrospective analysis, not a randomized controlled trial. Therefore, it is possible that the 3 groups may have had differing severity of illnesses or may have been managed in varying ways (beyond having received different initial empiric antibiotic choices). The azithromycin monotherapy group did not have as many individuals admitted from a skilled nursing facility or with a history of stroke as the non-ats group, and they did not have additions or changes to their antibiotic regimens made as frequently as in the non-ats group. These facts suggest that patients in the azithromycin monotherapy group may have been less ill compared with the other groups despite the fact that PSI scores were similar among the 3 groups. On the other hand, it does not seem that patients receiving azithromycin monotherapy received other clinical interventions that may have resulted in improved outcomes. For example, adherence to process-of-care measures associated with improved outcomes such as early initiation of antibiotic drug therapy and blood cultures drawn before administering antibiotics 14 was not higher in the azithromycin monotherapy group than in the other groups. Furthermore, decreased adherence to some process-of-care measures, including the drawing of blood cultures and the assessment of oxygenation, may potentially have also been a surrogate marker for decreased severity of illness. In fact, the azithromycin monotherapy group had blood cultures drawn (within 24 hours of admission) less frequently than the other 2 groups and oxygen assessment less frequently than the ATS group. Second, we excluded from analysis 22 patients who were transferred to the ICU within 24 hours of admission because we believe that it was not feasible to define the exact role (if any) that the initial empiric choice of antibiotics played in the rapid progression of illness and that these individuals were likely to have been misclassified as having mild-to-moderate pneumonia when they should have been categorized as cases of severe CAP. Using intention-to-treat analysis, these 22 patients would have been equally distributed among the 3 groups (7 in the azithromycin monotherapy group, 7 in the ATS group, and 8 in the non-ats group) had they not been excluded owing to transfer to the ICU within 24 hours of hospital admission. Furthermore, there were no deaths in the 7 patients who would otherwise have been classified in the azithromycin monotherapy group, whereas there were 2 deaths in the 7 patients who would have been classified in the ATS group and 4 deaths in the 8 potential non-ats group patients. However, patients in the latter groups were sicker (ATS group: mean age, 73.9 years; mean length of stay, 23.1 days; and mean PSI score, 128.3; non-ats group: mean age, 65.5 years; mean length of stay, 9.4 days; and mean PSI score, 140.5) than those who would have been classified in the azithromycin group (mean age, 67.9 years; mean length of stay, 12.9 days; and mean PSI score, 95.4). Another large retrospective study 36 that evaluated the relationship between the initial choice of antibiotics and clinical outcomes in patients hospitalized with CAP, similar to our study, classified patients who were admitted to or transferred to the ICU within 24 hours of admission as having received initial ICU care. Third, our definition as to what encompassed the initial antibiotic regimen may have been oversimplified and may have allowed for some variability in the overall antibiotic agents a patient may have received. For example, if a patient received a dose of a non ATSrecommended antibiotic in the ED, a dose of azithromycin on the ward, and then an ATS-recommended antibiotic 1724

for the remainder of the hospitalization, then this patient would have been classified as having initially received azithromycin monotherapy (with a change in therapy). However, most patients were either continued on the antibiotic regimen started in the ED or received one dose of antibiotics in the ED that was modified on arrival to the ward and subsequently continued on the modified regimen throughout the remainder of the hospitalization. There were 13 cases in which this did not occur (ie, patients received several antibiotic drug changes and may have been misclassified in the study). Reclassification of these patients would not have affected the death rates or the outcome measures among the 3 groups in analysis (data not shown). Nonetheless, some patients received one dose of intravenous antibiotic in the ED (or possibly an oral medication in the ambulatory setting) that was different from the antibiotic regimen on which the patient was maintained during the hospitalization, and this may have confounded our results. Finally, our data are reflective of the VAGLAHS and other similar institutions, and this study may not be applicable to all patients with CAP as a whole. During the study, patients meeting the inclusion criteria had relatively low rates of bacteremia overall (12% in patients from whom an organism was isolated), had rates of macrolide resistance that may have been lower than those seen elsewhere, 37 were infrequently admitted from skilled nursing facilities, and were mainly men. These data should be considered when exploring therapeutic options for CAP in the community. Despite these limitations, the results of this study suggest that azithromycin monotherapy is safe and effective for patients hospitalized with mild-to-moderate CAP, including many patients who have cardiopulmonary disease and those with risk factors (eg, age 65 years, resident of a nursing home, and medical comorbidities) for infection with gram-negative rods or resistant pneumococcus. Decreasing the use of third-generation cephalosporins could also potentially help decrease the rates of vancomycin-resistant enterococcus 38 and extendedspectrum -lactamases 39 in outbreak settings or prevent the emergence of these types of infections in the hospital. In summary, these data suggest that azithromycin monotherapy is efficacious in treating patients hospitalized with mild-to-moderate CAP even in the presence of ATS modifying factors such as underlying cardiopulmonary disease and age greater than 65 years. It is possible that some of the patients with CAP who were excluded from the study may also have potentially benefited from the use of azithromycin alone (such as those with metastatic malignancy, human immunodeficiency virus, and immunosuppression). Future randomized controlled trials are needed to better define which groups of patients hospitalized with CAP are most suitable for azithromycin monotherapy. Accepted for publication October 21, 2002. This study was funded by an unrestricted grant from Pfizer, Inc, Global Outcomes Research. We thank Kevin Knight, MD, PhD, for his assistance with data analysis for this study. Corresponding author and reprints: David C. Rhew, MD, Zynx Health Inc, 9100 Wilshire Blvd, Suite 655E, Beverly Hills, CA 90212 (e-mail: drhew@cerner.com). REFERENCES 1. Anderson RN. Deaths: leading causes for 2000. Natl Vital Stat Rep. 2002;50: 1-86. 2. Niederman MS, McCombs JS, Unger AN, Kumar A, Popovian R. The cost of treating community-acquired pneumonia. Clin Ther. 1998;20:820-837. 3. British Thoracic Society Standards of Care Committee. BTS Guidelines for the management of community acquired pneumonia in adults. Thorax. 2001;56 (suppl 4):IV1-IV64. 4. NiedermanMS,MandellLA,AnzuetoA,etal.Guidelinesforthemanagementofadults with community-acquired pneumonia: diagnosis, assessment of severity, antimicrobial therapy, and prevention. Am J Respir Crit Care Med. 2001;163:1730-1754. 5. Bartlett JG, Dowell SF, Mandell LA, File TM Jr, Musher DM, Fine MJ. Practice guidelines for the management of community-acquired pneumonia in adults: Infectious Diseases Society of America. Clin Infect Dis. 2000;31:347-382. 6. Heffelfinger JD, Dowell SF, Jorgensen JH, et al. Management of communityacquired pneumonia in the era of pneumococcal resistance: a report from the Drug-Resistant Streptococcus pneumoniae Therapeutic Working Group. Arch Intern Med. 2000;160:1399-1408. 7. Mandell LA, Marrie TJ, Grossman RF, Chow AW, Hyland RH. Summary of Canadian guidelines for the initial management of community-acquired pneumonia: an evidence-based update by the Canadian Infectious Disease Society and the Canadian Thoracic Society. Can Respir J. 2000;7:371-382. 8. Plouffe J, Schwartz DB, Kolokathis A, et al. Clinical efficacy of intravenous followed by oral azithromycin monotherapy in hospitalized patients with communityacquired pneumonia: the Azithromycin Intravenous Clinical Trials Group. Antimicrob Agents Chemother. 2000;44:1796-1802. 9. Vergis EN, Indorf A, File TM Jr, et al. Azithromycin vs cefuroxime plus erythromycin for empirical treatment of community-acquired pneumonia in hospitalized patients: a prospective, randomized, multicenter trial. Arch Intern Med. 2000; 160:1294-1300. 10. Lentino JR, Krasnicka B. Association between initial empirical therapy and decreased length of stay among veteran patients hospitalized with community acquired pneumonia. Int J Antimicrob Agents. 2002;19:61-66. 11. Gotfried MH, Neuhauser MM, Garey KW, Saubolle MA, Danziger LH. In vitro Streptococcus pneumoniae resistance: correlation with outcomes in patients with respiratory infections. In: Program and abstracts of the 5th International Conference on Macrolides, Azalides, Streptogramins, Ketolides, and Oxazolidinones (ICMAS/KO 5); January 28-29, 2000; Seville, Spain. Abstract 4.05. 12. Joint Commission s Advisory Council on Performance Measurement. ORYX Core Measure Profile for Community-Acquired Pneumonia. Washington, DC: Joint Commission on Accreditation of Healthcare Organizations; 1999. 13. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med. 1997;336:243-250. 14. Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA. 1997;278:2080-2084. 15. Levin KP, Hanusa BH, Rotondi A, et al. Arterial blood gas and pulse oximetry in initial management of patients with community-acquired pneumonia. J Gen Intern Med. 2001;16:590-598. 16. 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;279:1452-1457. 17. Ramirez JA, Vargas S, Ritter GW, et al. Early switch from intravenous to oral antibiotics and early hospital discharge: a prospective observational study of 200 consecutive patients with community-acquired pneumonia. Arch Intern Med. 1999; 159:2449-2454. 18. Torres A, El-Ebiary M. Invasive diagnostic techniques for pneumonia: protected specimen brush, bronchoalveolar lavage and lung biopsy methods. Infect Dis Clin North Am. 1998;12:701-722. 19. Gotfried MH. Epidemiology of clinically diagnosed community-acquired pneumonia in the primary care setting: results from the 1999-2000 Respiratory Surveillance Program. Am J Med. 2001;111(suppl 9A):25S-29S. 20. Bartlett JG. Empirical therapy of community-acquired pneumonia: macrolides are not ideal choices. Semin Respir Infect. 1997;12:329-333. 21. Lynch JP III, Martinez FJ. Clinical relevance of macrolide-resistant Streptococcus pneumoniae for community-acquired pneumonia. Clin Infect Dis. 2002;34 (suppl 1):S27-S46. 22. Doern GV. Antimicrobial use and the emergence of antimicrobial resistance with Streptococcus pneumoniae in the United States. Clin Infect Dis. 2001;33(suppl 3):S187-S192. 1725

23. Kelley MA, Weber DJ, Gilligan P, Cohen MS. Breakthrough pneumococcal bacteremia in patients being treated with azithromycin and clarithromycin. Clin Infect Dis. 2000;31:1008-1011. 24. National Committee for Clinical Laboratory Standards. Performance Standards for Antimicrobial Susceptibility Testing: Twelfth Informational Supplement M100- S12. Villanova, Pa: National Committee for Clinical Laboratory Standards; 2001. 25. Amsden GW, Duran JM. Interpretation of antibacterial susceptibility reports: in vitro versus clinical break-points. Drugs. 2001;61:163-166. 26. Amsden GW. Pneumococcal macrolide resistance: myth or reality? J Antimicrob Chemother. 1999;44:1-6. 27. Moreno S, Garcia-Leoni ME, Cercenado E, Diaz MD, Bernaldo de Ouiros JC, Bouza E. Infection caused by erythromycin-resistant Streptococcus pneumoniae: incidence, risk factors, and response to therapy in a prospective study. Clin Infect Dis. 1995;20:1195-1200. 28. Klugman KP, Capper T, Widdowson CA, Koornhof HJ, Moser W. Increased activity of 16-membered lactone ring macrolides against erythromycin-resistant Streptococcus pyogenes and Streptococcus pneumoniae: characterization of South African isolates. J Antimicrob Chemother. 1998;42:729-734. 29. Visalli MA, Jacobs MR, Appelbaum PC. Susceptibility of penicillin-susceptible and -resistant pneumococci to dirithromycin compared with susceptibilities to erythromycin, azithromycin, clarithromycin, roxithromycin, and clindamycin. Antimicrob Agents Chemother. 1997;41:1867-1870. 30. Gay K, Baughman W, Miller Y, et al. The emergence of Streptococcus pneumoniae resistant to macrolide antimicrobial agents: a 6-year population-based assessment. J Infect Dis. 2000;182:1417-1424. 31. Hyde TB, Gay K, Stephens DS, et al. Macrolide resistance among invasive Streptococcus pneumoniae isolates. JAMA. 2001;286:1857-1862. 32. Shortridge VD, Doern GV, Brueggemann AB, Beyer JM, Flamm RK. Prevalence of macrolide resistance mechanisms in Streptococcus pneumoniae isolates from a multicenter antibiotic resistance surveillance study conducted in the United States in 1994-1995. Clin Infect Dis. 1999;29:1186-1188. 33. Lagrou K, Peetermans WE, Verhaegen J, Van Lierde S, Verbist L, Van Eldere J. Macrolide resistance in Belgian Streptococcus pneumoniae. J Antimicrob Chemother. 2000;45:119-121. 34. Henriques B, Kalin M, Ortqvist A, et al. Molecular epidemiology of Streptococcus pneumoniae causing invasive disease in 5 countries. J Infect Dis. 2000;182: 833-839. 35. Plouffe JF, Breiman RF, Facklam RR, Franklin County Pneumonia Study Group. Bacteremia with Streptococcus pneumoniae: implications for therapy and prevention. JAMA. 1996;275:194-198. 36. Gleason PP, Meehan TP, Fine JM, Galusha DH, Fine MJ. Associations between initial antimicrobial therapy and medical outcomes for hospitalized elderly patients with pneumonia. Arch Intern Med. 1999;159:2562-2572. 37. Blondeau JM, Tillotson GS. Antimicrobial susceptibility patterns of respiratory pathogens: a global perspective. Semin Respir Infect. 2000;15:195-207. 38. Smith DW. Decreased antimicrobial resistance after changes in antibiotic use. Pharmacotherapy. 1999;19:129S-132S. 39. Pena C, Pujol M, Ardanuy C, et al. Epidemiology and successful control of a large outbreak due to Klebsiella pneumoniae producing extended-spectrum -lactamases. Antimicrob Agents Chemother. 1998;42:53-58. Call for Photographs Archives W of Internal Medicine Covers ith the January 13, 2003 issue, the Archives of Internal Medicine introduced photographs as cover art for the journal. Do you have a scenic photograph you have taken that you think would make a great cover shot? Submissions should be from our readers, reviewers, and authors, and must be formatted horizontally. They should be in color and at least 3.5 5 in but no larger than 8 10 in. Due to legal concerns, no recognizable people should appear in the picture. Please include your name and address and where the picture was taken. Send submissions to Archives of Internal Medicine, 1840 E River Rd, Suite 207, Tucson, AZ 85718. Cover photos will be chosen at the discretion of the ARCHIVES editorial staff. We look forward to seeing your photo on the cover of a future issue of the ARCHIVES! James E. Dalen, MD, MPH Editor 1726