Com o objetivo de avaliar se a retirada de maior número de fragmentos pleurais em um único sítio (A), bem como se a biópsia em um segundo sítio (B) da sua superfície, através do mesmo orifício cutâneo, acrescenta em rendimento diagnóstico aos três fragmentos habitualmente retirados, conduzimos um estudo prospectivo no Hospital Octávio Mangabeira, envolvendo 115 pacientes com indicação de biópsia pleural. Os fragmentos pleurais foram colocados em três frascos distintos: FRA1-3 = 3 primeiros fragmentos, FRA4-8 = 5 fragmentos subseqüentes e FRB1-5 = 5 fragmentos em segundo sítio da superfície pleural. Era retirada a agulha até o subcutâneo, feitas inclinação lateral e reintrodução no espaço pleural. A análise dos fragmentos provindos do FRA1-3 evidenciou granuloma em 72/92 casos de tuberculose (78,3%) e no FRA4-8 e FRB1-5, em 74/92 casos (80,5%). Em 6/92 pacientes foram encontrados pleurite inespecífica no FRA1-3 e granuloma no FRA4-8, resultando em acréscimo diagnóstico de 8,3% (6 aos 72 pacientes). Em 7/92 pacientes (7,6%) a HP foi inespecífica no sítio A e mostrou granuloma no sítio B, resultando em acréscimo diagnóstico de 9,0% (7 aos 78 pacientes). A concordância histológica intra-observador teve um índice kappa de 0,62 (0,54 a 0,70) e interobservador de 0,64 (0,56 a 0,72). Concluímos que a realização de múltiplas biópsias em dois sítios diferentes da superfície pleural diagnosticou 13 pacientes que não obteriam diagnóstico definitivo se fossem retirados apenas três fragmentos pleurais, resultando em incremento diagnóstico de 18,0% (13 aos 72 diagnosticados no FRA1-3), sendo considerado significativo do ponto de vista médico e social.For diagnostic purposes of pleural tuberculosis, three pleural fragments are routinely obtained from a single site by needle biopsy. The objective of this study is to compare the diagnostic yield of the routine procedure with two different approaches: harvesting five fragments on a single site (A), and collection of fragments at an additional site (B); from the same cutaneous orifice, the needle was partially removed up to the subcutaneous tissue and redirected at a 45o angle. The authors performed a prospective study in 115 patients submitted to pleural biopsy. The first three pleural fragments were placed in a vial (VA1-3), the 5 subsequent fragments obtained at the same site were placed in a second vial (VA4-8), and the five fragments from the second site were placed in a third vial (VB1-5). Analysis of fragments from VA1-3 revealed granuloma in 72/92 tuberculosis cases (78.3%), whereas in fragments VA5-8 and VB1-5, granuloma was observed in 74/92 cases (80.5%). In 6/92 patients the diagnosis at VA1-3 was non-specific pleuritis, but granuloma was found in VA4-8, resulting in a diagnosis increment of 8.3% (adding six to the 72 patients). In 7/92 patients (7.6%) the histopathological examination at site A led to a non-specific result but granuloma was found at site B, leading to a diagnostic increment of 9.0% (7 to 78 patients). Histological concordance of two examinations by the same pathologist had a kappa index of 0.62 (0.54 to 0.7) whereas concordance between two different pathologists had a kappa index of 0.64 (0.56 to 0.72). The performance of multiple biopsies at two different sites at the pleural surface led to an additional diagnosis of 13 patients, i.e., an increase of 18% (13 to 72) relative to the routine procedure. Considering the low cost increase of obtaining the fragments, adoption of this method is recommended in routine pleural tuberculosis diagnosis.