© 2002 American Thoracic Society
Studies of the Early Bactericidal Activity of New Drugs for TuberculosisA Help or a Hindrance to Antituberculosis Drug Development?Division of Tuberculosis Elimination Centers for Disease Control and Prevention Atlanta, Georgia During the past several years there has been renewed interest in the development of new antituberculosis drugs (1), and the newly established Global Alliance for Tuberculosis Drug Development (available online at www.tballiance.org) is providing needed leadership in this long-neglected area. One of the many impediments to progress has been the long and costly process of clinical development. Thus, methods to more efficiently assess drug activity and clinical efficacy, including studies of the early bactericidal activity (EBA) of new drugs, are receiving greater attention. In EBA studies, newly diagnosed patients with acid-fast bacillus (AFB)-positive, pulmonary tuberculosis are treated for periods ranging from 2 to 14 days with single drugs or drug combinations. During this period, quantitative counts of viable tubercle bacilli from carefully collected sputum specimens are made, with the EBA traditionally expressed as log-decrease in colony forming units/ml sputum/day over the first 48 hours. In this issue of the AJRCCM, Gillespie and colleagues (pp. 31–35) describe a novel method for the analysis of data from EBA studies (2). Gillespie proposes that daily measurements be made for the first 5 days and that the data be fitted to a regression curve with the results expressed as the time to achieve a 50 percent reduction in cultivatable bacilli. By incorporating a larger number of data points and providing for exclusion of outlier data that may arise from both intrinsic and extrinsic factors, Gillespie suggests that this method may both reduce the inherent variability commonly found in these studies and increase the precision of the measurement. Investigators in the United States have also suggested that EBA studies are optimally conducted over 5 days (3). The first EBA study was conducted in Kenya over 20 years ago, demonstrating the feasibility of the methodology (4). Subsequently, a series of studies have been reported, most commonly from countries in sub-Saharan Africa. The few EBA studies conducted elsewhere were often plagued with greatly variable results of the kind that Gillespie seeks to address. To explain these findings, it has been suggested that the success of an EBA study is largely determined by the type of patient evaluated (5). African patients with acute, far advanced pulmonary tuberculosis seem to have a more uniform and predictable EBA response, as contrasted to Asian or North American patients who may have more chronic forms of the disease. EBA studies can demonstrate that a drug is active early in treatment. However, failure to demonstrate an EBA effect does not mean that a drug has no role in treatment. Both the initial and subsequent studies found little or no EBA with aminoglycoside antibiotics (4), although their role in tuberculosis treatment is well established. A drug may have excellent sterilizing activity (i.e., the ability to kill organisms that persist after the initial bactericidal phase of treatment) but little bactericidal activity. This is certainly the case for pyrazinamide, a drug that is essential to modern short-course therapy (4). Thus, a drug with a low EBA might be abandoned for further development even though as a potent sterilizing agent it might permit shortening treatment or provide for more widely-spaced intermittent therapy, both priorities for antituberculosis drug development (1). One potentially useful piece of information that could come from an EBA study is the demonstration of a dose–response effect that might help in the selection of an appropriate dose for subsequent clinical trials. EBA studies have demonstrated a dose–response for isoniazid, rifampin, and ciprofloxacin (6). EBA studies might also be useful for the selection of one drug for further clinical evaluation from a class of potentially active compounds, as has been suggested for the fluoroquinolone antibiotics. Although EBA studies are thought to be important in advancing a new candidate drug into clinical development (7), the record of such studies to date has not been promising. Perhaps this is because EBA studies of new antituberculosis drugs have until now involved the longer-acting rifamycins, drugs for which sterilizing activity and not bactericidal activity is paramount. Moreover, because these drugs are intended for administration at widely spaced intervals, EBA studies are more problematic. A recently reported study of rifalazil showed no additional EBA when given once weekly with daily isoniazid over a 2-week period (8), although it has been well demonstrated that the potent EBA of isoniazid masks the effect of companion drugs. Nonetheless, following the results of this study, further clinical development ceased. Similarly, corporate interest in the further development of rifapentine appears to be stalled pending the outcome of a dose–response EBA study of that drug, although several Phase III clinical trials have demonstrated drug efficacy (9), and the drug is approved for use in the United States. Moreover, a recently completed Phase II dose escalation study has demonstrated the safety of higher doses that will likely improve the efficacy of once-weekly rifapentine-based therapy (10). It is difficult to see how the results of a yet-to-be completed EBA study would assist substantially in further development of this compound. Certainly, further evaluation of EBA methodology and standardization of study protocols are needed. Moreover, data should be developed for presentation to regulatory authority so that EBA studies are accepted as part of the drug development process, with the understanding that positive results may be helpful but are by no means a requirement for drug approval. It is hoped that refinements in study methodology such as those proposed by Gillespie are further assessed and found to improve the utility of these studies. However, markers of sterilizing activity that are assessed early in treatment are most needed to speed clinical development of new antituberculosis drugs. An early marker of sterilizing activity and treatment response that could serve as a surrogate for nonrelapsing cure would be invaluable, potentially shortening the time of clinical trials and leading to earlier drug approval. A recent suggestion that sterilizing activity may be accurately measured by serial sputum colony forming unit counts over the first 28 to 56 days of treatment certainly deserves further evaluation (11). Related studies of molecular markers in the sputum assessed during the first month of therapy also suggest a method for accurate prediction of treatment response (12). Further work in this area should be of highest priority to all those interested in improving tuberculosis treatment. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
