A novel approach to evaluation of adsorbents for sampling indoor volatile organic compounds associated with symptom reports†
This article addresses problems that complicate attempts to compare methods when several factors may be associated with an effect, but it is not known which factors are relevant. Chemicals that may contribute to ‘sick building syndrome’ (SBS), and thus should be sampled in investigations of SBS, are not currently known. A study was undertaken to compare the utility of three adsorbents (Carbopack B, Chromosorb 106 and Tenax TA) for detecting differences in personal chemical exposure to volatile organic compounds in indoor air, between persons with and without SBS symptoms (cases and controls). On the basis of office workers' responses to a questionnaire, 15 cases and 15 controls were chosen. They simultaneously carried diffusive samplers with adsorbents during a week at work, and the acquired samples were analysed by gas chromatography/mass spectrometry (GC/MS). The adsorbents were then compared in terms of their ability to separate cases and controls in partial least square discriminant analysis (PLS-DA) models. This method of comparison takes into account detected differences in chemical exposure between cases and controls measured with the different adsorbents. Tenax TA gave the best PLS-DA models for separating cases and controls, but a combination of measurements with Tenax TA and Carbopack B gave better PLS-DA models than models based on measurements from either adsorbent alone. Adding measurements from Chromosorb 106 did not improve the results.