RESUMEN
Multidrug-resistant (MDR) isolates of Mycobacterium tuberculosis complex (MTBC) are defined by resistance to at least rifampin (RMP) and isoniazid (INH). Rapid and accurate detection of multidrug resistance is essential for effective treatment and interruption of disease transmission of tuberculosis (TB). Overdiagnosis of MDR TB may result in treatment with second-line drugs that are more costly, less effective, and more poorly tolerated than first-line drugs. CDC offers rapid confirmation of MDR TB by the molecular detection of drug resistance (MDDR) for mutations associated with resistance to RMP and INH along with analysis for resistance to other first-line and second-line drugs. Simultaneously, CDC does growth-based phenotypic drug susceptibility testing (DST) by the indirect agar proportion method for a panel of first-line and second-line antituberculosis drugs. We reviewed discordance between molecular and phenotypic DST for INH and RMP for 285 isolates submitted as MTBC to CDC from September 2009 to February 2011. We compared CDC's results with those from the submitting public health laboratories (PHL). Concordances between molecular and phenotypic testing at CDC were 97.4% for RMP and 92.5% for INH resistance. Concordances between CDC's molecular testing and PHL DST results were 93.9% for RMP and 90.0% for INH. Overall concordance between CDC molecular and PHL DST results was 91.7% for RMP and INH collectively. Discordance was primarily attributable to the absence of known INH resistance mutations in isolates found to be INH resistant by DST and detection of mutations associated with low-level RMP resistance in isolates that were RMP susceptible by phenotypic DST. Both molecular and phenotypic test results should be considered for the diagnosis of MDR TB.