On the use of different measures of exposure-experiences from a case-control study on testicular cancer and PVC exposure.

Associations between exposure to PVC plastics and testicular cancer have been reported. To improve the exposure-response analysis in a matched case-control study on testicular cancer and occupational exposures, a self-administered exposure questionnaire and expert assessment was applied and different exposure measures were developed. The questionnaires regarding work histories and employment in PVC production, manufacturing, and handling of PVC products were completed by 1582 subjects (90%). By expert assessment, 360 subjects were considered exposed, and the exposure intensity to PVC plastics for different working periods was determined. Different exposure measures to PVC plastics were then developed, such as ever/never exposed, duration, maximum intensity, median intensity, and cumulative median intensity. The correlation between the different measures of exposure was high for exposure duration and the cumulative median exposure intensity (Spearman rank coefficient r(s) = 0.94), as was the correlation between the maximum intensity and the median intensity (r(s) = 0.94). The agreement between the answers in the questionnaire and the expert assessments was moderate, Kappa value 0.56. The odds ratio for "ever" exposed based on the exposure as reported in the questionnaire was 1.1 (95%, CI 0.82-1.56), and as determined by expert assessment 1.3 (CI 1.05-1.69). The odds ratios for all four different categories of exposure measures varied between 0.86 and 2.6 but decreased by increasing exposure. An overall excess of testicular cancer for the PVC exposed vs. the unexposed was not supported by the pattern seen in a standard exposure-response analysis based on several exposure measures. The findings stress the importance of using several exposure measures as dose surrogates when the underlying toxic mechanisms are unknown and when there are indications of an overall effect.