Blood serum analyses in population studies: parallel analyses and degree of agreement in a Norwegian-Russian co-project. The Svalbard Study.

In a survey of more than 300 russian coal workers living on Svalbard, parallel laboratory analyses of gamma-glutamyltransferase, serum total cholesterol, high density lipoprotein, and triglycerides were performed in Tromsö, Norway and in Archangelsk, Russia. Mean values for gamma-glutamyltransferase and serum total cholesterol were significantly higher in Tromsö than in Archangelsk while mean values of triglycerides and high density lipoproteins were significantly lower in Tromsö than in Archangelsk. Linear regression plots revealed both systematical and arbitrary differences between the two laboratories. Because no "true" standards were available, we were unable to assess the accuracy of laboratory results. We conclude that laboratory measurements from Norway and Russia are not directly comparable. Future research programs between the two laboratories should attempt to include calibrated standards of test serum from which to identify sources of variation. Alternatively, the serum tests should always only be analyzed at one site.