Editorial Commentary: Short‐Course Treatment of Community‐Acquired Pneumonia

One of the unfortunate but abiding truisms of much of infectious diseases clinical practice is that we usually do not know how long to treat most infections. In an era of evidence-based medicine, there are very few studies that are in fact based on appropriately designed clinical trials and that are meant to guide us in terms of length of treatment. With the exception of treatment of uncomplicated cystitis, tonsillopharyngitis acute otitis media, and acute exacerbations of chronic bronchitis, this holds true for most organ system infections, including meningitis, pneumonia, intra-abdominal abscesses, and diabetic foot infections, to name just a few. For pneumonia, in general, and community-acquired pneumonia (CAP), in particular, the time interval of 1–2 weeks has gained acceptance as a rule of thumb for length of treatment of most patients. Generally, though, physicians tend to overtreat patients, and 2 weeks is the usual length of treatment, particularly for patients who require admission to the hospital. During the past several years, however, concern regarding increasing antimicrobial resistance has led to attempts to shorten the duration of therapy. The argument has been put forward that, by reducing the total amount of an antibiotic used, antibiotic selection pressure is thereby reduced, and, in turn, resistance is less likely to occur. In one study that compared the impact of short-course, high-dose amoxicillin therapy (90 mg/kg q.d. for 5 days) with “standard-course” therapy (40 mg/kg q.d. for 10 days) for respiratory infections in children, Schrag et al. [1] observed a significantly lower rate of carriage of penicillin-resistant Streptococcus pneumoniae in children treated with the short-course, high-dose regimen. Other compelling arguments in favor of shorter courses of treatment included reduced costs and fewer adverse drug reactions, as well as enhanced patient compliance. The article by Dunbar et al. [2] addresses the efficacy of a shorter course of treatment of CAP. Unfortunately, a confounder interferes with the interpretation of the study results. The investigators used a multicenter, double-blind design in which patients were stratified according to severity of illness and then randomized to receive either an experimental or a control treatment regimen. The former was 750 mg of levofloxacin given intravenously or by mouth for 5 days, and the latter was 500 mg of levofloxacin given intravenously or by mouth for 10 days. By using the same drug for a shorter duration but at a higher dose, 2 variables were introduced. No statistically significant differences were found for any of the clinical and microbiological end points for both the intent-to-treat and clinically evaluable populations. It is important to note that 38.4% and 44.8% of patients in the experimental and control arms, respectively, had a Pneumonia Severity Index (PSI) class III, IV, or V. Also of note is the fact that, despite a substantial increase in the dose (albeit given for a shorter period of time), there were no differences in the incidence of adverse drug reactions between the 2 treatment arms. This is certainly not the first time that short-course treatment has been shown to be feasible for CAP. A number of studies comparing azithromycin with various control agents have shown that a shorter duration of treatment is possible using the azalide azithromycin [3, 4]. A course of 3–5 days of azithromycin is arguably the equivalent of a course of several days' duration for any other drug, because azithromycin has a uniquely prolonged half-life that distinguishes it from virtually all other antibiotics. Therefore, these data cannot be extrapolated to other agents or even to other macrolides. With regard to other drugs, however, Tellier et al. [5] reported the results of a multicenter, double-blind, randomized, controlled trial comparing telithromycin (800 mg q.d. for 5 or 7 days) with clarithromycin (500 b.i.d. for 10 days). The clinical and bacteriological response rates for telithromycin given for 5 days were comparable to those for clarithromycin given for 10 days. There is a sound scientific basis for the concept of short-course, higher-dose therapy, particularly for fluoroquinolone agents, and 2 arguments feature prominently in the design of the trial by Dunbar et al. [2] The first is that levofloxacin, like all fluoroquinolones, exhibits concentration-dependent bactericidal activity, and the second is that the therapeutic efficacy of fluoroquinolones is most closely aligned with ratio of the area under the concentration-time curve to the MIC [6]. The in vitro findings in the study by Lister [7], in which an increase in the dose of levofloxacin from 500 mg to 750 mg resulted in more rapid and complete bactericidal activity against the test organism, ciprofloxacin-resistant pneumococci, certainly sets the stage for a clinical trial comparing the 2 regimens. It has been suggested by some investigators that antibiotic levels in epithelial lining fluid may be better predictors of outcomes in the treatment of pneumonia than are drug levels in serum. Levofloxacin levels in epithelial lining fluid of 9.9 and 22.1 µg/mL have been found with daily doses of 500 mg and 750 mg, respectively [8]. The fact that CAP could be successfully treated for a period of time as short as 5 days should not be surprising if we consider other pulmonary infections, such as ventilator-associated pneumonia (VAP). VAP is a subset of nosocomial pneumonia and is associated with considerable morbidity and mortality. The standard length of treatment is usually a minimum of 10 days and often extends to ⩾14 days, depending on such variables as the pathogen, the presence or absence of bacteremia, and the severity of illness at the time of initial presentation. Two studies, however, have suggested that treatment for as short a period of time as 1 week may be all that is necessary for selected patients with VAP. In one study, it was shown that standard parameters used to gauge the presence of infection were significantly improved within 6 days after starting antibiotic therapy [9]. These parameters included temperature, WBC count, ratio of arterial oxygen tension to fractional inspired oxygen, and presence of etiologic pathogens in cultures of samples obtained by endotracheal aspiration. A significant finding was the fact that continuation of antibiotic therapy for a second week was associated with tracheal colonization with drug-resistant pathogens. A prospective observational cohort study by Luna et al. [10] used time-dependent analysis of clinical pulmonary infection score (CPIS) as a predictor of resolution of VAP. The CPIS was measured over time in survivors and nonsurvivors with VAP. A statistically significant decrease in the score over the first 5 days after onset of VAP in those who survived, compared with those who did not survive, was found. Both studies suggest that, even in infections as potentially lethal as VAP, the pathogens may be eradicated and the markers of infection may normalize in certain patients within 5–6 days after onset of VAP and initiation of antibiotic therapy. If such a short time frame for resolution of infection is seen with VAP, then certainly a 5-day regimen may be reasonable for many patients with CAP. The study by Dunbar et al. [2] is an interesting one, and the data are certainly compelling. The straightforward design with an appropriate control regimen lends credibility to their argument, as does the fact that a reasonable percentage of the patients were in PSI classes III and IV (only 3 patients in total were in class V). Some questions, however, still remain. Because the study actually tests 2 variables (shorter vs. longer time course and higher vs. lower dose) simultaneously, it is impossible to definitively know the effect of the shorter time course alone. It is certainly possible that 5 days of the 500-mg dose may have resulted in similar findings, but that is unfortunately conjecture at this point. Also, would a 5-day course be appropriate for patients in PSI class V and for those who are bacteremic? Is short-course treatment feasible only with a concentration-dependent bactericidal drug, such as a fluoroquinolone, or could other classes of drugs be used as well? There is, of course, a limit to how far one can push this question, because monotherapy for patients on a hospital ward with a PSI class of IV or V is limited to a fluoroquinolone. The other recommended treatment option for such patients is a β-lactam plus a macrolide. A shortened course of treatment is perhaps even more appealing if one uses combination therapy rather than monotherapy. If such studies are extended to bacteremic patients, those in PSI class V, or those treated with combination therapy, would it be possible to identify a subset of patients who are more likely to respond well to a short-course regimen? We strongly advocate more randomized, controlled trials specifically designed to evaluate the duration of therapy. Such studies should not only evaluate efficacy and safety but also the effect of therapy on the development of antimicrobial resistance.