Agreement between histopathological results in clinically diagnosed cases of indeterminate leprosy in São Paulo, Brazil.

Histopathological slides from skin biopsies of fifty-seven self-reporting patients diagnosed as indeterminate leprosy by the Leprosy Control Programme in São Paulo, were sent to three independent histopathologists. Agreement between the reports were based on the following diagnosis: "indeterminate leprosy", "suggestive leprosy" or "no leprosy". A great variation was observed in the interpretation of the histopathological examination. The three pathologists reported "indeterminate leprosy" respectively in 7.0%, 54.4% and 84.2%, of the cases studied. A kappa index of agreement between any two pathologists ranged from 0.08 to 0.32, showing poor agreement between observers. Agreement improved by pooling together the reports "suggestive leprosy" and "indeterminate leprosy". The three pathologists agreed in the results of 24 biopsies of the 27 classified as leprosy by any one of the three observers. Eight cases were considered as "no leprosy" by all pathologists. Higher agreement indices were obtained for positive and negative proportionate concordance between any two examiners. The implications of the variation in the diagnosis of indeterminate leprosy and early leprosy are discussed in the context of public health and case-management.