Comparison between radiometry and spectrophotometry for the determination of angiotensin-converting enzyme activity in cerebrospinal fluid

Determination of angiotensin-converting enzyme (ACE) acti-vity in cerebrospinal fluid (CSF) can help for establishing the diagnosis of neurosarcoidosis. We investigated the performance characteristics of two assays for ACE determination in 57 CSF, radiometry with [glycine-1-14C] benzoyl-L-histidyl-L-leucine and spectrophotometry with furylacry-loyl-phenylalanyl-L-glycyl-L-glycine (FAPGG) as substrates. We compared both kinetic assays to an ELISA specific for human ACE. Within run and between run imprecisions were 14-17% for radiometry, 6-19% for spectro-photometry and 5-8% for ELISA. The limit of detection was 0.04 U/L for radiometry, 1.0 U/L for spectrophotometry and 0.156 & mu;g/L for ELISA. The limit of quantification was 0.06 U/L for radiometry, 1.5 U/L for spectro-photometry, but not known for ELISA. The domain for quantification was 0.06-4.0 U/L for radiometry, 1.5-24 U/L for spectrophotometry and 0.156-10 & mu;g/L for ELISA. Deming regression and Bland-Altman plots show good correlations between the three assays, but with high slopes, because both kinetic assays use different substrates and ELISA measures ACE molecule but not activity. Radiometry was more sensitive than spectrophotometry, which has a limit of detection above most pathological levels. ELISA could be an alternative to radiometry but only after complete evaluation, deter-mination of normal values and assessment of its clinical value. We claim for standardization of ACE determination as well as in serum as in other biological fluids, in particular CSF.