Rapid pre-symptomatic diagnosis of sepsis by vibrational spectroscopy

  • Lila Lovergne

Student thesis: Doctoral Thesis


Sepsis is a dysregulated host response to an infection that causes life-threatening organ dysfunction. Each year, over 30 million cases and 5 million deaths are estimated worldwide. Diagnosis of sepsis is based on non-specific clinical signs and time consuming positive identification of the causative pathogen(s). Rapid presymptomatic detection of sepsis would enable the early administration of therapeutics, maximising their effects, reducing mortality and healthcare costs. Vibrational spectroscopy can provide a "molecular fingerprint" of biological materials and presents many advantageous aspects. Indeed, this technique is label-free, nondestructive, non-contact, rapid, cost-effective, simple to operate, and requires only simple sample preparation. The objective of this study was to develop and evaluate the potential of vibrational spectroscopy applied on human serum with the aim to improve diagnosis of patients with sepsis. Challenges of serum spectroscopy inherent to the sample nature and to the technique have been assessed. Different infrared and Raman spectroscopy modalities as well as different serum sample preparations have been compared to determine the most suitable methodology approach with an overall clinical application purpose. Then, some aspects of the pre-analytical phase have been addressed in order to standardise protocols in sample handling and preparation for spectral acquisitions to ensure quality and reproducibility of spectral data collected.;Finally, based upon the developed methodology, patient serum samples (n=380) collected before surgery, up to 3 days before sepsis diagnosis, and on the day of sepsis diagnosis have been analysed. Control serum samples (n=353) from age/ sex/ procedure-matched patients who did not go on to develop sepsis have been also analysed over similar timeframes post-surgery as well as samples (n=190) from patients with systemic inflammatory response syndrome. Spectral data acquired have been interrogated by chemometric methods to discriminate spectral zones reflecting differences in molecular composition.
Date of Award23 Mar 2018
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SponsorsUniversity of Strathclyde & Defence Science and Technology Laboratory DSTL MoD
SupervisorMatthew Baker (Supervisor) & Lynn Dennany (Supervisor)

Cite this