Very little is known about the growth and mutation rates of Mycobacterium tuberculosis during latent infection in humans. However, studies in rhesus macaques have suggested that latent infections have mutation rates that are higher than that observed during active tuberculosis disease. Elevated mutation rates are presumed risk factors for the development of drug resistance. Therefore, the investigation of mutation rates during human latency is of high importance. We performed whole genome mutation analysis of M. tuberculosis isolates from a multi-decade tuberculosis outbreak of the New Zealand Rangipo strain. We used epidemiological and phylogenetic analysis to identify four cases of tuberculosis acquired from the same index case. Two of the tuberculosis cases occurred within two years of exposure and were classified as recently transmitted tuberculosis. Two other cases occurred more than 20 years after exposure and were classified as reactivation of latent M. tuberculosis infections. Mutation rates were compared between the two recently transmitted pairs versus the two latent pairs. Mean mutation rates assuming 20 hour generation times were 5.5 X 10(-10) mutations/bp/generation for recently transmitted tuberculosis and 7.3 X 10(-11) mutations/bp/generation for latent tuberculosis. Generation time versus mutation rate curves were also significantly higher for recently transmitted tuberculosis across all replication rates (p = 0.006). Assuming identical replication and mutation rates among all isolates in the final two years before disease reactivation, the u 20 hr mutation rate attributable to the remaining latent period was 1.6 × 10(-11) mutations/bp/generation, or approximately 30 fold less than that calculated during the two years immediately before disease. Mutations attributable to oxidative stress as might be caused by bacterial exposure to the host immune system were not increased in latent infections. In conclusion, we did not find any evidence to suggest elevated mutation rates during tuberculosis latency in humans, unlike the situation in rhesus macaques.