Ethanol contamination of cerebrospinal fluid during standardized sampling and its effect on (1)H-NMR metabolomics

Standardization of body fluid sampling, processing and storage procedures is pivotal to ensure data quality in metabolomics studies. Yet, despite strict adherence to standard sampling guidelines, we detected variable levels of ethanol in the (1)H-NMR spectra of human cerebrospinal fluid (CSF) samples (range 9.2 × 10(−3)–10.0 mM). The presence of ethanol in all samples and the wide range of concentrations clearly indicated contamination of the samples of some sort, which affected the (1)H-NMR spectra quality and the interpretation. To determine where in the sampling protocol the ethanol contamination occurs, we performed a CSF sampling protocol simulation with 0.9 % NaCl (saline) instead of CSF and detected ethanol in all simulation samples. Ethanol diffusion through air during sampling and preparation stages appeared the only logical explanation. With a bench study, we showed that ethanol easily diffuses into ex vivo CSF samples via air transmission. Ethanol originated from routinely used skin disinfectants containing ethanol and from laboratory procedures. Ethanol affected the CSF sample matrix at concentrations above ~9.4 mM and obscured a significant part of the (1)H-NMR spectrum. CSF sample preparation for (1)H-NMR-based metabolomics analyses should therefore be carried out in a well-ventilated atmosphere with laminar flow, and use of ethanol should be avoided. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s00216-015-8663-9) contains supplementary material, which is available to authorized users.