Inaccuracy detected in acid phosphatase kit method.

Output peak distortion and noise are a problem in the AutoAnalyzer II (Technicon Instruments Corp., Tarrytown, N. Y. 10591) direct cholesterol procedure no. SE4-0026 FC4. We originally believed this difficulty to be inherent in the two-step sampling procedure used in this method, but we later observed sample diluent back-up into the bubbler tube of the original injector block, apparently because of the relatively large cross-sectional area and the resulting dead-space volume of the bubbler tube. This effect may have been increased by the relatively large sample volume (0.4 ml) introduced. The fluid level in the bubbler tube varied continually. Hence sample slug length and concentration varied, producing considerable noise and abnormally shaped output curves. The following modifications considerably reduce noise and improve the output curve shape in this procedure. (a) Replace the sample injector fitting no. 116-0490-01 with injector fitting no. 177-B004-02, which has a considerably smaller cross-sectional area and interacts little with stream flow, except during bubbling. (b) Replace the original sample diluent (distilled water) with a 2 g/liter solution of the surfactant Brij-35, which improved bubble pattern. (c) Replace the pump tube feeding the sample injector bubbler so as to decrease the flow rate from 0.42 ml/min (no. 116-0549-08 tubing) to 0.32 ml/min (no. 116-0549-07 tubing), to limit surging. (d) Rotate the AlO (no. 116-B034-01) sample fitting from parallel to upright position. These modifications have significantly decreased both the coefficients of variation as well as the frequency of need to repeat assays. Three controls, a borderline low, borderline high, and normal were used for two months before the modification and two months afterwards. The coefficients of variation before the modification were 2.8, 3.7, and 3.1%, respectively, as compared with 1.6, 2.5, and 1.1% afterward. The probability (P) that these improvements in precision occurred by chance are less than 0.01, 0.05, and 0.001, respectively (F test). Student’s t-test showed no significant difference in the means at any of the three analyte concentrations before and after modification.