Evaluation of a radiometrically-determined regrowth model for the study of anti-tuberculosis drugs

Abstract

Background: A post-exposure regrowth model utilizing the well-tried Bactec radiometric system, which would simulate in vivo situations at the site of invasive disease, was developed to measure drug activity against multiplying Mycobacterium tuberculosis. Aims: The aims of this dissertation were to (a) construct a radiometric model simulating drug efficacy relating to the combined bactericidal activity and delays in regrowth due to the action of anti¬tuberculosis (TB) agents, (b) compare the killing kinetics of drugs singly and in combinations by the time-kill curve method, with the radiometrically-determined regrowth model, and (c) assess whether the Bactec radiometric regrowth model could predict likely bactericidal activities of drugs. Design and methods: Drug concentrations in the time-kill curve method were in a range of achievable drug concentrations at the site of infection and in multiples of the minimal inhibitory concentration (MIC), (1x, 2x or 3x, and 8x). Exposure times of 6h, 24h, 48h and 72h were used and these were based on pharmacokinetic data reflecting likely periods of in vivo exposure in TB lesions. Standardized inocula of approximately 106CFU/mi of actively multiplying M. tuberculosis strains were used. The same concentrations, exposure times and bacterial concentrations were used for the assessment of radiometrically-determined post-exposure regrowth times of M. tuberculosis. Growth times were recorded as the number of days required to reach a predetermined growth index (GI) level in the Bactec system, and were expressed as T400 readings in days. Simple linear regression and a mathematical logistic model were used to assess whether the radiometric post-exposure regrowth model could predict the bactericidal activity of the drugs. For drug combination studies, 1MIC of isoniazid (INH) and rifampicin (RMP) were used singly and in combination while 2MIC of ethambutol (EMB), streptomycin (SM), ofloxacin (OFL) and amikacin (AMK) were used in combinations studies. Colony counts at Oh and following 24h exposures were performed and regrowth patterns were determined using the T400 method. M. tuberculosis H37Rv was tested and subsequently resistant strains. Results: INH and RMP were highly bactericidal while EMS showed moderate activity in the time-kill curve method. The three drugs produced the best curves, showing longer regrowth times and markedly depressed rates of regrowth in the Sactec post-exposure regrowth model. Using simple linear regression, a linear relationship between bacterial survivors and the radiometric regrowth times, T400, was achieved for all drugs tested. Even better agreement was found when control-related regrowth times, (T-C) 400, were used in the analysis. Conditions compromising the linear relationsbip in the radiometric regrowth model, for OFL and less markedly EMS and AMK, were likely postantibiotic effects (PAEs) brought on by the short exposure time (6h), and drug carry-over effects due to concentrations ≥ 8 MIC for INH, RMP and 8M (10x and 20x MICs). The mathematical logistic model showed good correlation between bactericidal activity and regrowth for INH and RMP but not for EMB, SM, OFL and AMK. Drug combination effects in the two techniques depended on the criteria used to describe synergy. Generally, it was found in drug combination experiments that the drugs did not influence each other to a meaningful extent. Discussion and conclusions: For prediction of bactericidal activity, interpretation of the radiometrically-determined regrowth model needs to accommodate PAEs and the effect of subinhibitory concentrations. The validity of the mathematical logistic model is not clear. Technical aspects of future studies such as better organism dispersal, need to be improved to achieve a more reliable evaluation based on the logistic model. For drug combination studies, the radiometric regrowth model yielded findings that were difficult to interpret in relation to published data, reinforcing the need for the use of internationally standardized techniques which would give statistically reliable data. The radiometrically-determined regrowth model showed good discrimination between the standard activities of anti- TB agents, correlating with clinical efficacy. It is simple to perform and could prove to be useful for the screening of candidate anti- TB drugs. Improved technical stringency and the evaluation of poorly active control drugs, are however needed before proof of validity of the model can be established.