Assessment of the Efficacy of New Anti-Tuberculosis Drugs
RESEARCH ARTICLE

Assessment of the Efficacy of New Anti-Tuberculosis Drugs

The Open Infectious Diseases Journal 29 Dec 2008 RESEARCH ARTICLE DOI: 10.2174/1874279300802010059

Abstract

The pathology of tuberculosis in humans starts with an initial Ghon focus in the lungs followed by transmission of bacilli though the blood and lymph to other regions in the lungs and to other organs. While these bacilli usually lie latent without causing further disease, some 10% start foci of adult type disease usually starting in the sub-apical regions of the lungs. Bacilli multiply, killing tissue by caseation and then forming colonies within the caseum. Cavities form connecting to the air in whose walls vigorous bacillary multiplication occurs. The history of the development of antituberculosis chemotherapy is described, starting with the use of multi-drug regimens to prevent the emergence of drug resistance and continuing with the shortening of the treatment period to 6 months by the incorporation in the regimens of rifampicin and pyrazinamide, which are the two drug responsible for bactericidal activity during treatment. Prospects for further shortening of treatment rest with the introduction of higher dosage with rifamycins and with new anti-tuberculosis drugs. These new drugs include the 8methoxyfluoroquinolones moxifloxacin and gatifloxacin which inhibit topoisomerases and protein formation, the diarylquinoline TM-207 which inhibits the mycobacterial ATP synthase and thus energy formation, the nitroimidazopyran PA-824 and the closely related OPC-676832 which are pro-drugs with uncertain modes of action and the pyrrole SQ-109, a cell wall inhibitor. Anti-tuberculosis drugs have widely variable pharmacokinetic characteristics but as they work efficiently together, it is unnecessary to match these when giving drug combinations. The effects of drug-drug interactions are usually small though the interactions with anti-retroviral drugs can pose problems. Dose sizes have usually been chosen to minimize side effects while retaining activity and thus tend to have low therapeutic margins, the exception being the margin of about 20 for isoniazid. The role of high plasma binding, important in limiting the efficacy of rifamycins, is uncertain for the newer drugs. Post antibiotic effects are vital to the prevention of drug resistance and need exploration for new drugs. The main aims of current drug development are (1) to shorten treatment, and (2) to make it more convenient, by for instance using widely intermittent regimens. The current techniques for measuring efficacy during drug development start with in vitro models, including the Hu/Coates models, which should contain bacterial populations resembling the bacterial persisters in lesions that are responsible for the long duration of treatment. The next stage is the mouse model of the chemotherapy of established tuberculosis, which has proved remarkably useful in assessing the value of the different drugs. The main problem in clinical assessment arises from the use of relapse after treatment as the main end-point, and the consequent need for very large numbers of patients required to provide measurable relapse rates in final phase III licensing studies. For this reason, surrogate studies are necessary in phase II which require much smaller numbers of patients. The first such investigations are phase IIA studies of early bactericidal activity which establish whether the drug given alone has bactericidal activity on cavitary bacilli and which can estimate the minimal effective dose of the drug, useful for decisions of dose size. The next step should be phase IIB studies which measure the rate of elimination of viable bacilli in sputum during the initial 8-weeks of treatment with various combinations of the new drug with established drugs. Measurement can be as (1) the proportion of patients with positive sputum at the end of the 8-weeks period, the easiest method but the least sensitive, or (2) as the speed with which sputum cultures become negative in a survival analysis, or (3) as the mean regression in modeling of serial sputum collections colony counts (SSCC). The relation between these surrogate estimates and the amoun of treatment shortening that can be obtained has still to be worked out.