Therapeutic drug monitoring (TDM) is a standard clinical technique used for many disease states, including many infectious diseases. As for these other conditions, the use of TDM in the setting of tuberculosis (TB) allows the clinician to make informed decisions regarding the timely adjustment of drug therapy. Such adjustments may not be required for otherwise healthy individuals who are responding to the standard, four-drug TB regimens. However, some patients are slow to respond to treatment, have drug-resistant TB, are at risk of drug-drug interactions or have concurrent disease states that significantly complicate the clinical situation. Such patients may benefit from TDM and early interventions may preclude the development of further drug resistance. It is not possible to collect multiple blood samples in the clinical setting for logistical and financial reasons. Therefore, one typically is limited to one or two time points. When only one sample can be obtained, the 2-hour post-dose concentrations of isoniazid, rifampin, pyrazinamide and ethambutol are usually most informative. Unfortunately, low 2-hour values do not distinguish between delayed absorption (late peak, close to normal range) and malabsorption (low concentrations at all time points). A second sample, often collected at 6-hour post-dose, can differentiate between these two scenarios. The second time point can also provide some information about clearance and half-life, assuming that drug absorption was nearly completed by 2 hours. 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. Rifampin is stable for more than 6 hours under these conditions. During TB treatment, isoniazid causes the greatest early reduction in organisms and is considered to be one of the two most important TB drugs, along with rifampin. Although isoniazid is highly active against TB, low isoniazid concentrations were associated with poorer clinical and bacteriological outcomes in US Public Health Services (USPHS) TB Trial 22. Several earlier trials showed a clear dose-response for rifampin and pyrazinamide, so low concentrations for those two drugs also may correlate with poorer treatment outcomes. At least in USPHS TB Trial 22, the rifampin pharmacokinetic parameters were not predictive of the outcome variables. In contrast, low concentrations of unbound rifapentine may have been responsible, in part, for the worse-than-anticipated performance of this drug in clinical trials. The 'second-line' TB drugs, including p-aminosalicylic acid, cycloserine and ethionamide, are relatively weak TB drugs. Under the best conditions, treatment with these drugs takes over 2 years, as opposed to 6 to 9 months with isoniazid- and rifampin-containing regimens. Therefore, TB centres such as National Jewish Medical and Research Center in Denver, CO, USA, measure serum concentrations of the 'second-line' TB drugs early in the course of treatment. That way, poor drug absorption can be dealt with in a timely manner. This helps to minimise the time that patients are sputum smear- and culture-positive with multidrug-resistant TB, and may prevent the need for even longer treatment durations. Patients with HIV are at particular risk for drug-drug interactions. Because the published guidelines typically reflect interactions only between two drugs, these guidelines are of limited value when the patient is treated with three or more interacting drugs. Under such complicated circumstances, TDM often is the best available tool for sorting out these interactions and placing the patient the necessary doses that they require. TDM is only one part of the care of patients with TB. In isolation, it is of limited value. However, combined with clinical and bacteriological data, it can be a decisive tool, allowing the clinician to successfully treat even the most complicated TB patients.