Addition Quinolones Hospital Formularies Changes Pneumonia Treatment

To determine patterns and predictors of antimicrobial drug use for outpatients with community-acquired pneumonia, we examined office visit and pharmacy claims data of 4 large third-party payer organizations from 2000 to 2002. After patients with coexisting conditions were excluded, 4,538 patients were studied. Despite lack of coexisting conditions, fluoroquinolone use was commonly observed and increased significantly (p < 0.001) from 2000 to 2002 (24%–39%), while macrolide use decreased (55%–44%). Increased age correlated with increased fluoroquinolone use: 18–44 years (22%), 45–64 years (33%), and >65 years (40%) (p < 0.001). Increased use of fluoroquinolones occurred in healthy young and old patients alike, which suggests a lack of selectivity in reserving fluoroquinolones for higher risk patients. Clear and consistent guidelines are needed to address the role of fluoroquinolones in treatment of outpatient community-acquired pneumonia.

Community-acquired pneumonia

(CAP) is a leading cause of death due to infection in the United States and a primary indication for antimicrobial drug use in inpatient and outpatient settings. The fluoroquinolone class of antimicrobial agents has become increasingly popular for the management of CAP because of coverage of common CAP pathogens, toleration by patients, and excellent oral absorption (1). “Respiratory” fluoroquinolones, such as levofloxacin, gatifloxacin, moxifloxacin, and gemifloxacin, have activity against most strains of drugresistant Streptococcus pneumoniae (DRSP) (2). Concerns for infection due to DRSP may drive fluoroquinolone use because providers fear that traditional CAP regimens will fail (3,4). However, fluoroquinolone resistance, while generally low, appears to be increasing in S. pneumoniae (5–8). Therapeutic failures have been reported in patients infected with fluoroquinolone-resistant organisms treated with levofloxacin (9,10). Although data from community

settings are lacking, resistance to fluoroquinolones is also increasing among gram-negative organisms in hospital settings (11,12). Increased use of fluoroquinolones for outpatient respiratory tract infections may lead to increased resistance rates among community-acquired gram-negative organisms. Fluoroquinolones may promote colonization and infection with methicillin-resistant Staphylococcus aureus (MRSA) (13–15). Communityacquired MRSA infections, once rare, have increased in frequency (16,17). Clinicians are faced with the dilemma of attempting to limit broad-spectrum antimicrobial drug use on a population level while trying to maximize therapeutic success in individual patients (18). Routine prescribing of fluoroquinolones for CAP may limit the possibility of therapeutic failure due to drug-resistant organisms but may compromise the future effectiveness of this class of drugs. Practice guidelines available for management of CAP from various professional societies provide mixed messages on the use of fluoroquinolones, particularly for patients eligible for outpatient treatment.

The American Thoracic Society recommends reserving fluoroquinolones for outpatients with cardiopulmonary disease or other modifying factors, advocating a macrolide or doxycycline for patients without such coexisting conditions (19). The DrugResistant Streptococcus pneumoniae Working Group recommends reserving fluoroquinolones for patients whose treatment has failed on other regimens or those with documented infections due to DRSP (20). Previous guidelines of the Infectious Diseases Society of America considered macrolides, doxycycline, or fluoroquinolones as equivalent options for treating outpatients, with the suggestion that older patients and those with underlying disease have a stronger indication for fluoroquinolone therapy (21). A recent update of these guidelines categorizes patients according to whether they recently received antimicrobial drugs and presence of underlying conditions: patients without underlying illnesses and no recent antimicrobial drug therapy should receive a macrolide or doxycycline, whereas fluoroquinolones are an option for the other groups (22).

Such extensive subclassification and conflict among guidelines may pose difficulties for clinicians practicing in busy outpatient settings (23). Without clear and consistent guidelines, clinicians may base their therapeutic decisions on tradition, the practice of colleagues, or advice from pharmaceutical sales representatives, rather than the best evidence. We examined a database of office visit and pharmacy claims from 4 large managed-care organizations in Colorado. Our objective was to determine the patterns of antimicrobial drug prescribing, especially fluoroquinolone use, in a group of outpatients with CAP without serious underlying conditions.


We used administrative claims data from 4 healthcare organizations in Colorado. Identifiable patient information was removed before the information was provided, and patients were assigned a unique identification number for the purpose of data manipulation. Information contained in the database included the patient’s date of birth, sex, visit date, health plan, provider identification and specialty, up to 3 International Classification of Diseases, 9th edition, Clinical Modification (ICD-9-CM) diagnostic codes, and drugs prescribed during the visit. Data were available from March 1, 2000, to March 1, 2003. The study received approval from the institutional review board of the University of California, San Francisco. Our criteria for inclusion in the study were age >18 years, primary diagnosis of CAP (based on ICD-9-CM codes 481, 482, 483, 485, and 486), and prescription of an antimicrobial agent associated with the visit. As serious coexisting conditions may justify the use of fluoroquinolones according to some guidelines, we excluded those patients with coexisting conditions to examine prescribing patterns in an otherwise healthy population.

Specifically, we excluded patients who had a second or third diagnosis of chronic obstructive pulmonary disease, congestive heart failure, diabetes, lung cancer, renal failure, atrial fibrillation, respiratory failure, pleural effusion, Parkinson disease, multiple sclerosis, and asphyxia. We excluded patients who had sought treatment for an acute respiratory tract infection (bronchitis, pharyngitis, otitis media, sinusitis, and upper respiratory tract infection) or urinary tract infection during the 4 weeks before the visit for pneumonia. Consequently, we excluded results for the first month of the study (March 2000) since data regarding prior visits were unavailable for that group. Finally, we limited the dataset to 1 pneumonia visit per patient to reduce the likelihood of including patients whose previous therapy had been unsuccessful. We categorized patients by age into 3 strata: 18–44 years, 45–64 years, and >65 years.

We also categorized patients according to health plan (1 through 4). For categorization by year, we used March 1 as the start and end date (e.g., year 2001 was March 1, 2001, to March 1, 2002). Antimicrobial agents were identified by using National Drug Codes. Antimicrobial drugs were assigned to one of the following categories: tetracyclines (doxycycline, tetracycline), macrolides (azithromycin, clarithromycin, erythromycin), fluoroquinolones (ciprofloxacin, levofloxacin, ofloxacin, moxifloxacin, gatifloxacin), aminopenicillins (amoxicillin, amoxicillin/clavulanate), cephalosporins (primarily cefuroxime and cefprozil), sulfonamides (trimethoprim-sulfamethoxazole), and other. Only 1 antimicrobial agent was recorded per patient visit. We used the visit date to analyze data by year of prescription (2000, 2001, or 2002). Comparisons between proportions in groups were performed by using the Mantel-Haenszel χ2 test for trend, with a significance level of 0.05. Logistic regression analysis was performed on the group of patients without coexisting conditions. The outcome was prescription of a fluoroquinolone. Variables included in the model were year of treatment, age (by category), patient sex, and health plan; and interactions between year of treatment with age, patient sex, and health plan were also tested. All analyses were performed by using SAS, version 8.2 (SAS, Cary, NC, USA).