The diagnostic accuracy of urine dipsticks was investigated using two different brands in five endemic communities of south-western Nigeria. The BM-5L test was used in 1992 to screen 566 subjects in two communities in Ogun State, while 1457 subjects in three other communities in Osun State were screened with the Combur-9 test in 2006. Haematuria gave a higher prevalence of infection than proteinuria irrespective of which strip brand was used (e.g. BM-5L test: 58.3 and 36.2%; Combur-9 test: 46.5 and 41.9%, respectively). Compared with egg microscopy (gold standard), haematuria identified over 90% of egg-positive samples using either the BM-5L test in 1992 or the Combur-9 test in 2006. The corresponding values for proteinuria were 58% using the BM-5L test and 82% using the Combur-9 test. Sensitivity of haematuria to infection was higher using the BM-5L test (92.4-93.5%) than Combur-9 (58.6-73.3%), while sensitivity of proteinuria to infection was higher using Combur-9 (55.5-80.4%) than BM-5L test (26.0-58.3%). However, both strip brands have comparable specificity for haematuria (BM-5L test, 88.3-99.5%; Combur-9, 88.9-100%) and proteinuria (BM-5L test, 94.4-100%; Combur-9, 98.7-100%) to infection. Based on these results we conclude that neither brand nor manufacturer has a significance effect on the performance of chemical reagent strips. However, the diagnostic value of both haematuria and proteinuria varied according to the positivity level adopted, intensity of infection and age, but was not affected by sex and village of residence.