Investigation of faecal volatile organic compounds as biomarkers for the diagnosis of necrotising enterocolitis

Mayor, Arnaud
(2015) Investigation of faecal volatile organic compounds as biomarkers for the diagnosis of necrotising enterocolitis. Master of Philosophy thesis, University of Liverpool.

[img] Text
Available under License Creative Commons Attribution.

Download (5MB)


Necrotising enterocolitis (NEC) is the most prevalent and harmful illness in the neonatal intensive care unit. It primarily affects premature babies, with higher incidence in low birthweight patients. NEC usually occurs in the early days of life and may develop rapidly. Its diagnosis is based on medical observation because no early diagnostic tool is currently available. Over the last two decades, developments in analytical chemistry have allowed scientists to perform untargeted investigation of biological samples. Biomarkers are being sought, for a range of disorders, by investigating volatile organic compounds (VOCs) in bio-fluids. For the analysis of gas samples, solid phase micro-extraction (SPME) is a suitable technique to preconcentrate volatile compounds prior to analysis using devices such as gas chromatography – mass spectrometry (GC–MS). Recently, much attention has been directed towards the development of sensors and electronic noses to be used as diagnostic tools in hospitals, as they are generally more compact than a GC-MS and do not require specialised personnel. The work presented here is based on the hypothesis that faeces from patients suffering from necrotising enterocolitis show a specific pattern of volatile organic compounds in the days prior to diagnosis when compared with faeces from healthy patients. The objectives of this work were to develop two methods for the analysis of premature faeces using headspace–SPME–GC–MS and headspace – gas chromatography – sensor (HS–GC–Sensor), to analyse samples from healthy premature infants and premature infants affected by NEC using both instruments and to analyse the data collected. Methods were developed individually for each analytical technique. Two pipelines were applied for mass spectrometric data analysis while classification models were exclusively built using sensor data. Results obtained from HS–SPME–GC–MS data showed that the age at sampling had an influence on the number of compounds identified and on their intensities or relative abundance. Heptanal, 2-E-pentenal, hexanal and 2-methylbutanoic acid were identified as relevant compounds. Classifiers were built at days 1 to 6 prior to diagnosis. Accuracy, sensitivity and specificity of up to 74%, 62% and 79%, respectively, were obtained one day prior to diagnosis based on mass spectrometric data, while accuracy, sensitivity and specificity of 100% were obtained based on sensor data at two days prior to diagnosis. Therefore, classification of samples based on headspace analysis of faeces might have potential for the early diagnosis of necrotising enterocolitis.

Item Type: Thesis (Master of Philosophy)
Subjects: Q Science > QD Chemistry
Q Science > QP Physiology
Divisions: Fac of Health & Life Sciences > The School of Tropical Medicine
Depositing User: Symplectic Admin
Date Deposited: 15 Aug 2016 10:12
Last Modified: 15 Aug 2016 10:12
Repository Staff Access