Two automated methods for measuring plasma thiocyanate compared

In an interlaboratory comparison of two continuous-flow analytical procedures for measuring thiocyanate, we used ferric nitrate (y) and p-phenylenediamine/pyridine (x) as colorimetric reagents to measure its concentrations in plasma of 100 consecutive patients attending a peripheral vascular disease clinic. The results correlated well (r = 0.987, p less than 0.001; y = 0.938x + 1.2 mumol/L). However, there were small, systematic, positive differences between the phenylenediamine values and the corresponding ferric nitrate values (paired t = 5.4, p less than 0.001). These differences were linearly related to the means of the pairs of results (r = 0.42, p less than 0.001; y = 0.0739x - 2). Nevertheless, when we used previously determined cutoff points the two sets of SCN concentrations concurred completely in classifying the 100 patients as smokers or nonsmokers. On the basis of self-classification by 71 of these subjects, the measurement techniques had a sensitivity of 91% and a specificity of 75%; when five patients claiming to be nonsmokers but found to have abnormally high values for carboxyhemoglobin (2.7 - 6.9%) were reclassified as smokers, specificity increased to 89%.