BackgroundDiagnosis of leprosy mainly relies on clinical examination due to the inconsistent sensitivity and poor reproducibility of the current laboratory tests. Utilisation of alternative methods to the standard Ziehl Neelsen (ZN), Fite-Faraco (FF) and Haematoxylin and Eosin (H&E) staining procedures may eventually improve leprosy diagnosis.Methodology/Principal findingsIn this comparative study, the performance of the fluorescent Auramine O (AO) staining and polymerase chain reaction (PCR) was assessed with different skin samples using a combination of ZN, FF and H&E staining as the gold standard. AO, ZN, FF, H&E and PCR tests were performed on slit skin smears (SSS) and/or punch biopsies collected from 141 clinically confirmed leprosy cases and 28 non-leprosy skin samples. DNA was extracted from punch biopsies using two different methods with or without mechanical lysis.Sensitivities were 87.6%, 59.3% and 77% for H&E, ZN and FF, respectively, whereas it reached 65.5% and 77.9% for AO in SSS and tissue sections and 91.1% for PCR in tissue samples. Morover, samples with low bacillary index, sensitivity of AO staining (61.8%) was similar to FF (60%, p>0.05) and lower than PCR (86.6%, p<0.05). Sensitivity of PCR also increased (96.8%, p<0.05) when mechanical lysis was used during DNA extraction compared to enzymatic treatment alone (84.6%).Conclusions/SignificanceOur results showed that for diagnostic purposes, analysis of skin section is more sensitive than SSS, especially for samples with low bacillary load. AO staining on SSS and tissue sections was not significantly better than other routine diagnostic tests but considerably more user friendly. The sensitivity of PCR was higher than current standard methods and increased when combined with more efficient DNA extraction using mechanical and chemical lysis. Therefore, we recommend AO staining for the diagnosis of leprosy in lower health facilities such as health centres and district hospitals and PCR diagnosis at referral level and research centres.