Analysis of fat-soluble vitamins. XIX. Collaborative studies on the chemical assay for vitamin D concentrates

In 1971, a chemical method for the assay of vitamin D in concentrates containing only vitamin D was collaboratively studied by 14 laboratories, using 6 different samples from 2 European manufacturers. On the basis of these results, the laboratories were divided into 2 groups: 5 with significant laboratory biases of greater than or equal to 2%, and 9 laboratories with nonsignificant bias. The 9 laboratories were subdivided into 2 groups which differed significantly as to reproducibility within laboratories. The reproducibility between laboratories, expressed as a standard deviation in per cent with 95% confidence limits, was 1.2% (confidence range 0.6-7.3) and 4.7% (confidence range 2.4-29.3) for 3 and 6 laboratories, respectively. A second collaborative test was performed in 1974, using 12 vitamin D resin samples in oil from 3 United States manufacturers, to compare 2 chemical vitamin D assay methods (with and without maleic anhydride) and to compare results from the chemical and biological methods; 9 laboratories participated in the chemical method study and 3 in the rat bioassay study. The correlation of results of the chemical method including maleic anhydride treatment and the rat bioassays was satisfactory. The reproducibility of the chemical method was about the same as that in the first collaborative test.