Enzymatic assays for urea nitrogen and uric acid on the Cobas Bio centrifugal analyzer.

1992 CLINICAL CHEMISTRY, Vol. 28, No. 9, 1982 data were collected before their communication appeared, the plan of data collection was not influenced by their hypothesis of systematic biases affecting the last two samples of the SMAC batch. Our choice of locations for control samples within the batch was in part determined by the possibility that control results might be influenced by carryover from the preceding samples. The sequentialpair of controls at the end of the batch was intended to provide in each batch a control specimen that was protected from carryover effects. The possibility raised by Walmsley et al. that systematic biases affect these two final specimens complicates interpretation of the data from these positions. Sample-interaction effects would increase the variance of the first control of the pair, while the end-of-batch effects postulated by Walmsley et al. would increase the variance of the second control more than that of the first. Actually, however,for four of the predilution channels the first control had a greater variance, while for two the second control had the greater variance. Their interpretation of our data assumes that preceding patients’ samples increased the control results by positive carryover effects. The carryover factors in force when our data were collected are not known at this time, nor is the exact mean value of the patients’ specimens run during the study known. Review of patients’ data approximately 18 months later indicates that the mean of the non-control samples was lower than the controls for albumin, but significantly higherthan the controls for cholesterol, triglycerides, and alkaline phosphatase. It may not be possible to make a definite choicebetweenthe two alternative interpretations on the basis of these data alone. We have, however, reviewed additional data and found that they favor the interpretation of Abernethy et al. These data may be summarized as follows: For six months a Dade abnormal control serum (lot no. XPT-549; American Hospital Supply Corp., Miami, FL 33152) was run in positions 8 and 18, unprotected from sample interaction effects. None of the six predilution channels studied (albumin, cholesterol, direct bilirubin, inorganic phosphorus, total bilirubin, and total protein) showed a statistically significant drift between positions 8 and 18 (p >0.05 in each case).