Multidrug-resistant tuberculosis (MDR-TB) is defined as resistance to at least isoniazid (INH) and rifampicin (RIF), and it complicates the implementation of tuberculosis control programmes. The rapid detection of MDR-TB is crucial to reduce the transmission of disease. The nitrate reductase assay (NRA) is one of the colorimetric susceptibility test methods for rapid detection of MDR-TB and based on the ability of reduction of nitrate to nitrite by Mycobacterium tuberculosis. The aim of this study was to evaluate the performance of the NRA for the rapid detection of MDR-TB. A total of 237 M.tuberculosis complex (MTC) isolates that were identified by the same method (BD MGIT (TM) TBc Identification Test, USA) from nine different medical centers in Turkey were included in the study. The susceptibility results of the isolates against INH and RIF obtained by reference test (Bactec MGIT (TM) 960, BD, USA) were then compared with NRA. In order to ensure consistency between centers, Lowenstein-Jensen (LJ) medium with antibiotics and without antibiotics (growth control) and Griess reagent solution were prepared in a single center (Ondokuz Mayis University School of Medicine, Medical Microbiology Department) and sent to all participant centers with the standardized test procedure. After the inoculation of bacteria into the test tubes, the tubes were incubated at 37 degrees C, and after seven days of incubation, 500 mu l Griess reagent was added to the LJ medium without antibiotics. If a color change was observed, an equal volume of Griess reagent was added to test LJ media with antibiotics. When a color change was observed in LJ media with antibiotics, it was considered that the isolate was resistant to tested antibiotics. Among 237 MTC isolates, 16 were resistant only to INH and nine were resistant only to RIF; 93 isolates (39.2%) were resistant (MDR) and 119 isolates (50.2%) were susceptible to both of the drugs determined with the reference susceptibility test. In the study, five INH-resistant isolates determined with reference method were found susceptible with NRT and eight INH-susceptible isolates determined with reference method were found resistant with NRT. In contrast, one RIF-resistant isolate determined with reference method was found susceptible with NRT and three RIF-susceptible determined isolates were found resistant with NRT. Accordingly, the concordance rate between the reference method and NRA were estimated as 94.5% for INH and 98.3% for RIF. The sensitivity, specificity, positive and negative predictive values of NRA were detected as 95.4%, 93.7%, 92.8% and 96% for INH, and 99%, 97.8%, 97.1% and 99.2% for RIF, respectively. The results of the 111 isolates were obtained on the seventh day, while the rest of the results were obtained between 10-14 days. In conclusion, the data of this multicenter study showed that NRA is a reliable, relatively inexpensive and practical method to perform for the rapid detection of MDR-TB.