© 2003 American Thoracic Society
Tuberculosis ChemotherapyStill a Double-edged SwordCenter for Tuberculosis Research Johns Hopkins University School of Medicine Baltimore, Maryland
Now, with the kindest motives in the world, I offer for the consideration of consumptive patients the variegated course of treatment I have lately gone through. Let them try it; if it don't cure, it can't more than kill them.—Mark Twain, Curing a Cold The conquest of tuberculosis with modern chemotherapy was one of the preeminent achievements of 20th century medicine. Between 1944 and 1976, effective drugs for tuberculosis went from none to more than 10, and tuberculosis went from a frequently fatal disease to one that could be cured with mostly oral drugs given for a matter of months. Presently, tuberculosis can be cured in the vast majority of patients with regimens received daily or two to three times weekly for 6 months at a cost of $10 to $200. To clinicians who treat patients with tuberculosis, the wonder of cheap and effective antituberculosis drugs is tempered by the counterweight of drug toxicity. Physicians have long dealt with the double-edged sword of medications: the historian Roy Porter wrote that for the ancient Greeks "the word pharmakos meant both remedy and poison—‘kill’ and ‘cure’ were apparently indistinguishable" (1). Although drug toxicity has been of intense interest in treating latent tuberculosis infection (2–4), the enormous benefits of antituberculosis agents in active disease have overshadowed their risks. But in this issue of AJRCCM (pp. 1472–1477), Yee and colleagues provide useful data on the rates of adverse reactions to antituberculosis drugs in routine clinical practice (5). These investigators analyzed rates of toxicity for specific antituberculosis agents in a retrospective cohort of 430 patients treated at the Montreal Chest Institute. They used a rigorous definition of toxicity and assigned blame in a fashion that greatly diminished bias. The results are noteworthy and sobering. Serious adverse events occurred in 9% of patients. The adverse events that occurred were by no means trivial; all required an alteration in therapy, and 5% of patients were hospitalized for a median of 16 days. Toxicity added substantially to the costs of care, and the duration of treatment was prolonged by an average of 5 months. The most unique aspect of the report by Yee and coworkers is the incidence rates of toxicity by specific drugs. By determining person-months of treatment with each agent, the authors calculated the incidence of serious toxicity for each drug. The striking observation is that pyrazinamide was associated with a rate of toxicity that was threefold higher than isoniazid and rifampin and 20-fold higher than ethambutol. Isoniazid has always stood out as the principal cause of drug toxicity during tuberculosis therapy, with pyrazinamide associated with frequent minor toxicity. Yet in this report, the rate of toxicity with pyrazinamide was 1.5 per 100 person-months compared with 0.5 per 100 person-months for isoniazid. But this finding must be interpreted with caution, as two important methodological issues influenced the outcome. First, reporting rates by person-months, although more quantitative than simple prevalence, may overestimate the toxicity of pyrazinamide compared with isoniazid and rifampin. The occurrence of toxicity during tuberculosis treatment is not linear over time, as was seen by Yee and coworkers. Ninety-seven percent of all toxicities occurred within 60 days of starting the treatment, with virtually no late events. Earlier studies suggest that pyrazinamide has little late toxicity (6), but here it was only given for the first 2 months of treatment. Using person-months of treatment underestimates the toxicity of isoniazid and rifampin by adding a large amount of risk-free time to the denominator without paying a penalty in the numerator, making the incidence of pyrazinamide toxicity seem relatively higher. When one compares the prevalence of toxicity of isoniazid and pyrazinamide (4 vs. 6% of patients, respectively), a slightly different impression is formed. The second limitation is the manner in which the attribution of toxicity was determined. Although it greatly reduces bias from the investigators and is praiseworthy, it does not eliminate bias on the part of the clinicians who actually stopped the medications. When more than one drug was stopped because of toxicity, the authors appropriately divided the blame evenly. For patients who were not rechallenged, however, it is not possible to determine the real culprit. Dividing the responsibility proportionally avoids bias in the analysis but may not provide a correct answer. Pyrazinamide may have been blamed for isoniazid hepatotoxicity (and vice versa), but this cannot be determined. In addition, rashes attributed to pyrazinamide may have led to inappropriate drug discontinuation; pyrazinamide-related rashes usually resolve spontaneously and are not considered a reason to stop therapy (7). Given the caveats about the denominators used to determine incidence for pyrazinamide and isoniazid, the real difference in toxicity becomes less dramatic. Nonetheless, the study demonstrates that pyrazinamide may have important toxicity that was previously underappreciated. Recent reports of severe hepatotoxicity in patients treated with rifampin and pyrazinamide for latent tuberculosis underscore the potential importance of this finding (8–11). What are the implications of the work of Yee and coworkers? First, toxicity during tuberculosis therapy is a real problem that can cause serious harm to patients. This is particularly important because, in most parts of the world, tuberculosis treatment is given with little or no clinical supervision. Second, a better understanding of the mechanisms of action and toxicity of pyrazinamide is necessary. We know remarkably little about this agent, and learning more about its physiologic activities could result in its safer use as well as contribute to the search for new drugs (12). Third, although modern tuberculosis therapy is indeed miraculous, it is not good enough. Some might argue that the currently available drugs are as good as it gets, given their remarkable efficacy and low cost. However, we can and must do better. Development of new tuberculosis drugs is critically important for reasons other than reducing toxicity, of course. New drugs are urgently needed for the treatment of drug-resistant tuberculosis, and agents that can simplify therapy further by reducing its duration and the number of doses required for cure are highly desirable. After decades of neglect and decline, the tuberculosis drug development field has been reignited over the past several years, and groups such as the Global Alliance for TB Drug Development, the TB Trials Consortium, and other university-based investigators are actively pursuing important new therapeutics. Tools for drug discovery have improved vastly since the last tuberculosis drug was developed more than 25 years ago, with extraordinary advances in genomics and molecular biology opening new avenues of research. Programs like the Grand Challenges in Global Health Initiative of the Bill and Melinda Gates Foundation may help draw more creative minds and fresh ideas into the effort. The prospects for finding new and better drugs for the treatment and prevention of tuberculosis are reasonably bright. The work of Yee and colleagues is yet another reminder of why this task is so important. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
