Background: There is no gold standard on how national tuberculosis programs should design retreatment regimens. Often drug susceptibility testing (DST) is not available for all patients, and representative DST patterns in patient populations are used to guide therapy.
Objectives: To examine DST patterns in different patient populations based on previous treatment and to estimate the number of effective anti-tuberculosis agents in several retreatment regimens.
Methods: We reviewed DST results from patients treated with individualized regimens in Peru between January 1998 and July 2004. We stratified patients into four groups based on previous treatment exposure from Group 1 who had failed only one regimen to Group 4 who had failed three regimens. We compared resistance frequencies across the four groups. In Groups 1 and 3, the number of likely effective agents under six possible retreatment regimen scenarios was estimated.
Results: Resistance to second-line drugs was significantly higher in groups with more previous courses of treatment. A few retreatment regimens could be identified that would allow at least 80% of patients to receive at least four likely effective drugs.
Conclusion: Because it is associated with resistance frequencies, previous treatment exposure can serve to guide the design of non-individualized MDR-TB regimens.