The quality of variable number of tandem repeats (VNTR) typing of Mycobacterium tuberculosis was first investigated in 2009 in 37 laboratories worldwide. The results revealed an inter- and intra-laboratory reproducibility of respectively 60% and 72%. These data spurred an improvement in laboratory-specific assays and global standardisation of VNTR typing.To measure the effects of the technical improvements and increased standardisation, a test panel consisting of 30 M. tuberculosis complex DNA samples was distributed for VNTR typing in 41 participating laboratories from 36 countries.The inter- and intra-laboratory reproducibility increased overall to respectively 78% and 88%. The 33 laboratories that participated in both the first and second proficiency studies improved their inter- and intra-laboratory reproducibility from 62% and 72% to respectively 79% and 88%. The largest improvement in reproducibility was detected in 10 laboratories that use an in-house polymerase chain reaction technique and perform amplicon sizing using gel electrophoresis. Detailed error analysis revealed a reduction in the number of systematic errors, sample exchange events and non-amplifiable loci.This second worldwide proficiency study indicates a substantial increase in the reproducibility of VNTR typing of M. tuberculosis. This will contribute to a more meaningful interpretation of molecular epidemiological and phylogenetic studies on the M. tuberculosis complex.