Tuberculosis (TB) is the world's second leading infectious killer. Cases of multidrug-resistant (MDR-TB) and extremely drug-resistant (XDR-TB) have increased globally. Therapeutic drug monitoring (TDM) remains a standard clinical technique for using plasma drug concentrations to determine dose. For TB patients, TDM provides objective information for the clinician to make informed dosing decisions. Some patients are slow to respond to treatment, and TDM can shorten the time to response and to treatment completion. Normal plasma concentration ranges for the TB drugs have been defined. For practical reasons, only one or two samples are collected post-dose. A 2-h post-dose sample approximates the peak serum drug concentration (Cmax) for most TB drugs. Adding a 6-h sample allows the clinician to distinguish between delayed absorption and malabsorption. TDM requires that samples are promptly centrifuged, and that the serum is promptly harvested and frozen. Isoniazid and ethionamide, in particular, are not stable in human serum at room temperature. Rifampicin is stable for more than 6 h under these conditions. Since our 2002 review, several papers regarding TB drug pharmacokinetics, pharmacodynamics, and TDM have been published. Thus, we have better information regarding the concentrations required for effective TB therapy. In vitro and animal model data clearly show concentration responses for most TB drugs. Recent studies emphasize the importance of rifamycins and pyrazinamide as sterilizing agents. A strong argument can be made for maximizing patient exposure to these drugs, short of toxicity. Further, the very concept behind 'minimal inhibitory concentration' (MIC) implies that one should achieve concentrations above the minimum in order to maximize response. Some, but not all clinical data are consistent with the utility of this approach. The low ends of the TB drug normal ranges set reasonable 'floors' above which plasma concentrations should be maintained. Patients with diabetes and those infected with HIV have a particular risk for poor drug absorption, and for drug-drug interactions. Published guidelines typically describe interactions between two drugs, whereas the clinical situation often is considerably more complex. Under 'real-life' circumstances, TDM often is the best available tool for sorting out these multi-drug interactions, and for providing the patient safe and adequate doses. Plasma concentrations cannot explain all of the variability in patient responses to TB treatment, and cannot guarantee patient outcomes. However, combined with clinical and bacteriological data, TDM can be a decisive tool, allowing clinicians to successfully treat even the most complicated TB patients.