The aim of this study was to compare the in-house polymerase chain reaction (PCR) with Löwenstein-Jensen (LJ) culture method in the diagnosis of tuberculosis. In the Group of Florence Nightingale Hospitals totally 333 clinical specimens [200 (66.66%) from pulmonary tuberculosis (PT) and 100 (33.33%) from extrapulmonary tuberculosis (EPT) (totally 300 cases)] were included in the study. The specimens were stained with Ziehl-Neelsen (ZN) stain for direct microscopic examinationon and decontaminated with NALC + 4% NaOH for LJ culture. Additionally, the in-house PCR was performed for each sample. The primary pairs specific to the IS 6110 region of Mycobacterium tuberculosis complex were used. In 200 cases diagnosed as PT clinically, 22 (11%) were ZN positive, 51 (25.50%) were culture positive and 55 (27.50%) were PCR positive; 53 (26.5%) of these cases were finally diagnosed as tuberculosis with the help of parameters like response to therapy. In 100 cases diagnosed as EPT clinically at the time of admission, two (2%) were positive in ZN, 23 (23%) in LJ, 26 (26%) in PCR; however, 38 (38%) cases were finally diagnosed as tuberculosis with the help of culture and parameters except PCR (clinical, radiological, histopathological findings and response to antituberculosis therapy). The concordance of PCR with LJ in PT was found statistically at high ratios (Kappa 94% > 65%). However, in EPT the concordance ratio was found just below the statistically acceptable concordance ratios (Kappa 59% < 65%). As a result, the study showed that in PT cases the in-house PCR results and culture positiveness had a great concordance whereas in EPT the concordance of the PCR results and culture positiveness was found just below the accepted ratios.