Biological, analytical and experimental components of variance in a long-term study of plasma constituents in rat.

Statistical analysis of variance was applied to data from determinations of 14 plasma constituents in 25 rats in order to evaluate the analytical, experimental and biological (inter-and intraindividual) component of variance. Blood was taken seven times in intervals of 8-10 days, the last one by catheter technique and the other by heart puncture. The analytical portion of variance was determined by the concurrent analysis of a pool plasma standard. The experimental component of variance was evaluated by the comparison of the variation of the catheter values with that of the pooled data from heart puncture. The coefficient of variation for the latter may be grouped into three categories: less than 10% for protein, Na+, K+, Ca2+; 10-20% for urea, phosphate and the enzymes as alanine aminotransferase, choline esterase, alkaline phosphatase and leucine arylamidase and 20-65% for the other enzymes lactate dehydrogenase, malate dehydrogenase, aspartate aminotransferase and creatine kinase. The results from the samples taken by catheter technique generally revealed the lower values for the mean as well as for the variance. It became evident that the procedure of heart puncture is afflicted with the most aggravating interference factors, thus accounting for most of the experimental component of variance. The observed differences between the single blood drawings, the non-Gaussian distribution for several constituents, and the interactions between the components of variance do not always fit for the statistical concept of additivity of the single components.(ABSTRACT TRUNCATED AT 250 WORDS)