RESUMO
BACKGROUND: NT-proBNP may be useful for ruling out heart failure in primary health care. In this study we examined the analytical quality of NT-proBNP in primary health care on the Cobas h 232 point-of-care instrument compared with measurements performed in a hospital laboratory. MATERIALS AND METHODS: Blood samples requested for NT-proBNP were collected in primary health care (n = 95) and in a hospital laboratory (n = 107). NT-proBNP was measured on-site on Cobas h 232 instruments both in primary health care centres and at the hospital laboratory and all samples were also analyzed with a comparison method at the hospital. Precision, trueness, accuracy, and lot-variation were determined at different concentration levels and evaluated according to acceptance criteria. Furthermore user-friendliness was assessed by questionnaires. RESULTS: For Cobas h 232 repeatability CV was 8.5-10.7% in the hospital setting and 5.3-10.0% in the primary health care and within the analytical quality specifications, but higher than with the comparison method (< 4%). NT-proBNP results obtained in primary health care were significantly higher than by the hospital comparison method (bias ranged from 14.3-23.7%), whereas there was no significant bias when Cobas h 232 was used in the hospital setting (bias ranged from - 4.9 to 7.0%). User-friendliness of Cobas h 232 was overall acceptable. CONCLUSION: Cobas h 232 point-of-care instrument for measurement of NT-proBNP performed satisfactorily with regard to precision, user-friendliness, and lot-variation. A decrease in NT-proBNP levels observed in samples transported to a central laboratory needs further attention and investigation.