© 2004 American Thoracic Society
Blood Cultures for Community-Acquired PneumoniaPiecing Together a Mosaic for Doing LessDepartment of Infectious Diseases Lahey Clinic Medical Center Tufts University School of Medicine Burlington, Massassuchetts
The full truth in science has been likened to a mosaic constructed by piecing together the partial truths of many studies. Each year in the United States, over 1 million patients are hospitalized with community-acquired pneumonia. Guidelines from the Infectious Diseases Society of America and the American Thoracic Society recommend two blood cultures for patients hospitalized with pneumonia (2, 3). Low rates of secondary bacteremia (4 to 18%) have led many investigators to question the clinical value and cost-effectiveness of routine blood cultures (4–8). In this issue of the Journal (pp. 342–347), Metersky and coworkers, using elegant epidemiological techniques and a retrospectively analyzed database of 13,043 Medicare patients hospitalized with pneumonia, present new data on risk factors for secondary bacteremia (9). Eight independent predictors of bacteremia (recent antibiotic therapy, liver disease, systolic blood pressure, pulse and temperature, serum blood urea, nitrogen, and sodium, and white blood cell count) were used in combination to formulate a "decision support tool" for obtaining blood cultures in patients at low (2%), moderate (5%), and high (11%) risk of bacteremia. No blood cultures were recommended for low-risk patients, one culture for moderate risk, and two cultures for high-risk patients. The clinical tool, validated in a cohort of 12,771 Medicare patients, resulted in 38% fewer blood cultures and identification of bacteremia in nearly 90% of patients. A simplified version of this tool identified 87% of the bacteremias with 44% fewer cultures. Based on a direct charge of $70/blood culture, an estimated $340,000 to $393,000 would be saved in this study, and millions of dollars annually in the United States. Most clinicians agree that blood cultures are needed for clinical management of severely ill patients, but delays in initial therapy and the 24 to 48 hours needed to obtain culture results limit clinical usefulness (4–9). Data from two groups of investigators, who studied more than 1,100 hospitalized patients with pneumonia, concluded that blood cultures have limited clinical utility and questionable cost-effectiveness (4, 5). Waterer and coworkers confirmed that blood culture results often did not result in de-escalation or narrowing the spectrum of antibiotic therapy for pneumococcal bacteremia, and suggested that blood cultures be limited to patients with severe pneumonia (pulmonary severity index grades IV and V) (6, 7). Fine and coworkers identified bacteremia as an independent predictor of mortality, which was highest in patients with pulmonary severity index grades IV and V (10, 11). Metersky and coworkers, and other investigators, reported no correlation between bacteremia and age or the pulmonary severity index grade (4, 9). These differences may be related to variations in the study population, design, definitions, and primary endpoints. The rapid emergence of Streptococcus pneumoniae strains resistant to penicillin and other antibiotics is another concern (2, 3). Yu and coworkers, however, demonstrated that discordant antibiotic therapy for pneumococcal pneumonia and bacteremia, based on in vitro antibiotic resistance testing, did not correlate with poorer outcomes or increased mortality (12). Others have reported that patients with bacteremia do not require longer courses of therapy (13). Updated guidelines have endorsed higher antibiotic sensitivity breakpoints for pneumococci, and the urine pneumococcal capsular antigen test to increase diagnostic sensitivity (2). The critical question remains, "In how many patients do blood cultures alter therapy or improve outcome, and is it worth the cost?" The cost–benefit and effectiveness of routine diagnostic tests, such as blood cultures, should be carefully examined, especially with rising healthcare costs, and absent health insurance for more than 42 million Americans. Blood cultures are currently a national "quality indicator" based on data from Meehan and coworkers indicating that performance of blood cultures was associated with an improved 30-day mortality (14). This finding, however, probably reflects differences in overall quality of medical care rather than a causal relationship. The clinical tools for obtaining blood cultures proposed by Metersky and coworkers are simple, sensitive, and if appropriately implemented, would significantly reduce the number of wasted blood cultures, with substantial savings (9). Maintaining the status quo would require 10,000 more blood cultures to identify 100 secondary bacteremias, most of which would be effectively treated without blood cultures. Metersky and coworkers reported a 4.9% rate of contaminated blood cultures, which was probably underestimated but similar to other reports (5, 9). Several studies have reported that rates of blood culture contaminants, such as coagulase-negative staphylococci, are similar to rates of pathogen isolation. Contaminated blood cultures often create clinical confusion and are a significant economic burden. Bates and coworkers compared 94 patients with false-positive blood cultures to 1,097 patients with negative blood cultures and demonstrated that contaminated blood cultures increased median length of stay by 4.5 days at a total cost of approximately $6,000/patient (15). Increased costs, length of stay, and inappropriate use of antibiotics, such as vancomycin, may be more important than failure to diagnose all secondary bacteremias. Metersky and coworkers recommend one blood culture for patients in the moderate risk group. The volume of blood cultured is an important determinant for diagnostic yield that should increase sensitivity (16). Recovery rates from two blood culture samples with different volumes of blood (6 ml versus 2 ml), collected simultaneously from different sites in 300 children, were significantly increased in the high volume group (72% versus 35%, p < 0.01). Maximizing the volume of blood into each blood culture bottle is critical to increase sensitivity. Pneumonia is a dynamic disease with new pathogens, rapidly emerging antibiotic-resistance, and the added threat of bioterrorism (2, 3). Because the cause of pneumonia is not identified in nearly 50% of patients, it is imperative that all patients have ongoing assessment of their clinical response to initial antibiotics. Blood cultures are clearly needed in nonresponders. The data of Metersky and coworkers shed new light on the controversy over the utility and cost-effectiveness of initial blood cultures with some "partial truths" that include: limitations of retrospective databases, the definitions used for bacteremia and contaminants, and lack of data on clinical outcomes from undiagnosed bacteremia. The current mosaic of "partial truths" supports targeted blood cultures for higher-risk patients, which should reduce healthcare expenditures and preserve resources without compromising clinical outcomes. We now have ample evidence to alter recommendations for routine use of blood cultures in pneumonia management guidelines and as a "quality indicator" for hospitals (2, 3, 13). FOOTNOTES Conflict of Interest Statement: D.E.C. has no declared conflict of interest. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
