Neonatal screening for G6PD deficiency has long been established in many countries. The aim of the study was to determine whether the routine semiquantitative fluorescent spot test could detect all cases of G6PD deficiency, including those cases with partial deficiency (residual red cell G6PD activity between 20-60% of normal). We compared the results of G6PD screening by the semiquantitative fluorescent spot test and quantitative G6PD activity assay on a group of 976 neonates and 67 known female heterozygotes. The values for mean G6PD activity of G6PD-normal neonates and 293 healthy adult females were determined. There was no significant difference in the mean normal G6PD activity between the two racial groups in the neonates (669 Malays, 307 Chinese) and in the 293 healthy adult females (150 Malays, 143 Chinese) group. The values for the upper limits of total deficiency (20% of normal residual activity) for neonates and adult females were 2.92 U/gHb and 1.54 U/gHb, respectively. The upper limits of partial deficiency (60% of normal residual activity) were 8.7 U/gHb and 4.6 U/gHb respectively. The prevalence of G6PD deficiency among the male neonates was 5.1%(26) by both the fluorescent spot test and the enzyme assay method. The G6PD activity levels of all 26 cases of G6PD-deficient male neonates were< 20% normal (severe enzyme deficiency). In the female neonate group, the frequency of G6PD deficiency was 1.3%(6 of 472) by the fluorescent spot test and 9.35%(44 of 472) by enzyme assay. The 6 cases diagnosed as deficient by the fluorescent spot test showed severe enzyme deficiency (< 2.92 U/gHb). The remaining 38 female neonates had partial enzyme deficiency and all were misdiagnosed as normal by the fluorescent spot test. In the female heterozygote group, G6PD deficiency was diagnosed in 53%(35 of 67) by enzyme assay and in 7.5%(4 of 67) of cases by the fluorescent spot test. The 4 cases detected by fluorescent spot test had severe enzyme deficiency (< 1.6 U/gHb). The remaining 31 (46.3%) cases, diagnosed as normal by fluorescent spot test, showed partial G6PD deficiency. In conclusion, we found that the semiquantitative fluorescent spot test could only diagnose cases of total G6PD deficiency and misclassified the partially-deficient cases as normal. In this study, the overall prevalence of G6PD deficiency was 3.28% by the semiquantitative fluorescent spot test and 7.17% by enzyme assay. This means that 3.9% of G6PD-deficient neonates were missed by the routine fluorescent spot test and they were found to be exclusively females. This study demonstrates a need to use a method that can correctly classify female heterozygotes with partial G6PD deficiency. The clinical implication is that these individuals may be at risk of the hemolytic complication of G6PD deficiency. duction of NADPH required for detoxification of toxic products of oxidative stress. G6PD deficiency is the commonest enzymopathy in humans, estimated to affect 400 million individuals worldwide. It has a wide geographical distribution, reaching high frequencies in certain parts of Africa, the Mediterranean and Asia. Although most individuals are asymptomatic, the clinical importance is that exposure of G6PD-deficient individu-