Volatile organic compound markers of psychological stress in skin: a pilot study. (21st November 2016)

• Record Type: Journal Article
• Authors: Martin, H J; Turner, M A; Bandelow, S; Edwards, L; Riazanskaia, Svetlana; Thomas, C L P
• Is Part Of: Journal of breath research. Volume 10:Number 4(2016:Dec.)
• Publication Date: 2016-11-21
• Subjects: skin -- polydimethylsilicone sampling -- stress -- benzoic acid -- n-decanoic acid -- thermal desorption GC-MS; Volatile organic compounds -- Analysis -- Periodicals; Clinical chemistry -- Periodicals; Bad breath -- Periodicals; Bad breath -- Treatment -- Periodicals; Bad breath -- Diagnosis -- Periodicals; 616.0756
• DOI: 10.1088/1752-7155/10/4/046012
• Languages: English
• ISSNs: 1752-7155

Abstract

Abstract: The forehead was studied as a possible sampling site for capturing changes in volatile organic compound (VOC) profiles associated with psychological-stress. Skin-VOCs were sampled with a polydimethylsilicone (PDMS)-coupon and the resulting VOCs were recovered and analysed with two-stage thermal desorption gas chromatography-mass spectrometry. Fifteen young adult volunteers (19 years–26 years) participated in two interventions run in a randomised crossover design. One intervention, termed 'Neutral', required the participants to listen to peaceful music, the other, termed a 'paced audio serial addition task', required the participants to undertake a series of rapid mental arithmetic calculations in a challenging environment that induced a stress response. Skin-VOC samples were taken during each intervention. The resultant data were processed with dynamic background compensation, deconvolved, and registered to a common retention index scale. The importance of freezing skin patch samplers to −80 °C was determined during the method development phase of this study. The cumulative distribution function of the GC-MS data indicates the possibility that PDMS-coupons are selective towards the lower volatility VOC components in skin. The frequency distribution of the GC-MS data was observed to be approximately log-normal, and on the basis of this study, a further two-orders of magnitude reduction in sensitivity may be required before the complete skin-VOC profile may be characterised. Multi-variate analysis involving Pareto-scaling prior to partial least squares discriminant analysis identified four VOCs with the highest probability of contributing to the variance between the two states, and the responses to these VOCs were modelled with principle components analysis (PCA). Two VOCs, benzoic acid and n -decanoic acid were upregulated (14 and 8 fold respectively) and appear to be PASAT sensitive, with areas under (AUC) their receiver operator characteristic (ROC) curves of 0.813 and 0.852 respectively. A xylene isomer and 3-carene were down regulated 75% and 97% respectively, and found to be predictive of the neutral intervention (ROC AUC values of 0.898 and 0.929 respectively). VOC profiles in skin appear to change with stress either due to increased elimination, elevated bacterial activity, or perhaps increased oxidative pathways.