Diabetes biomarkers in breath gas, findings, setbacks and perspectives.

Currently, cardiovascular, oncologic, neurodegenerative and chronic respiratory diseases as well as diabetes form a severe burden to the countries worldwide. Breath gas analysis (BGA) is supposed to build an efficient, non-invasive diagnostic tool for the detection of these diseases in their early stages. Volatile organic components (VOCs) in the exhaled breath reflect the volatiles composition of the bloodstream and airways and therefore the metabolic status of the whole organism. However, in most cases, there is no one single volatile biomarker indicating a distinct disease. Therefore, an approach to identify a disease (diabetes mellitus) by the response of the organism to a metabolic challenge was used. First results were quite promising but some setbacks are forcing us to rethink the approach. Introduction It has been shown, that the analysis of exhaled breath and exhaled breath condensates (EBC) are useful, non-invasive tools for the detection of biomarkers of respiratory diseases like asthma and lung cancer [1]. Due to the intense exchange with the blood stream, exhaled breath does not only reflect the conditions within the airways, but contains also volatile organic compounds (VOCs) from the internal metabolism. Thus, exhaled breath analysis has a strong potential to monitor biomarkers of diseases like diabetes mellitus (DM)[2]. Due to the rapidly increasing prevalence in both, adult and childhood obesity, DM is a worldwide growing problem. Gestational diabetes mellitus (GDM), defined as a glucose intolerance, first detected during pregnancy, is the most frequent metabolic problem during gestation and its prevalence is also increasing across most racial/ethnic groups studied so far [3]. GDM is associated with an increase in perinatal morbidity and mortality. The risk for the development of diabetes postpartum in these women as well as for the development of obesity and diabetes for their offspring in later life is significantly increased [4]. Therefore, screening of GDM is important and, based on an oral glucose tolerance test (oGTT), currently practiced in women considered at risk in many Western countries. The diagnosis of the oGTT test is based on the blood glucose concentration measured just before and repeatedly after the ingestion of 75 mg glucose. A non-invasive test, based on breath gas analysis, could increase the numbers of women screened and subsequently treated, leading to a reduction in perinatal morbidity. Experimental Methods In close connection with a GDM screening (75 g glucose oGTT) we performed breath gas analysis by means of a standard proton transfer reaction quadrupole mass spectrometry (PTR-MS) device (Ionicon GmbH, Innsbruck, Austria). Tested persons breathed at 6 to 10 min. intervals in a buffered end-tidal sampler (Ionicon GmbH) connected to the PTR-MS by a heated flexible PTFE tube. This set-up allowed the controlled measurement of the end-tidal breath fraction.  From a broad spectrum of initially 142 signals (m/z 18…174) the responsive signals were selected for further analysis. Two distinct response patterns were observed between 0 and 120 min of the oGTT, a “kinetic” response (19 signals), resulting in an impulsive peak response and a “linear” response (eight signals), with either decreasing or increasing trend. For each type of response, signals were fitted to mathematical equations with a time evolution parameter µ. These parameters as well as the basal and peak levels of the measured signals were used to distinguish between the different patient classes. Since PTR-MS is not able to identify individual components from given m/z values, additional measurements using needle trap micro extraction (NTME) in combination with thermal desorption and comprehensive two dimensional gas chromatography time of flight mass spectrometry to identify individual VOCs in the exhaled breath were carried out. Using short time sampling of about one to two minutes, we were able to identify several isomers or isobars belonging to one m/z value and, due to the time dependent concentration profile, to assign one isomer or isobar as responsible for the direct measured response within the PTR-MS. To asses confounding factors, further PTR-MS measurements were performed in conjunction with oral glucose tolerance test. These tests were carried out under a particular oral hygiene. For this purpose, the bacterial coating of the tongue was mechanically scraped and subsequently the oral cavity rinsed for about one minute with an antibacterial agent (chlorhexidine). Further, for some tests, the commonly used oGTT juice was replaced by an aqueous glucose solution (75 g). Results With the approach on evaluating the time dependent evolution of the different parameters (m/z) it was possible, in a first study [5] to differentiate the four groups (gestational diabetes (GDM), impaired glucose tolerance (IGT), marginal and the healthy control group (for details see [5]) Unfortunately, in a follow up study on women with previously diagnosed GDM [6], this result could not be repeated. A closer look to the dataset, the identification of individual components representing the m/z values as well as further investigations on confounding factors with healthy (male) volunteers showed that the results are strongly influenced by oral bacteria. The second study on women was performed during autumn and winter time. Besides the unknown hormone status, medication and tobacco smoking, a great number of the patients were influenced by infections of the upper airways. The combination of all these unknowns leads to strong variations in the breath profile, which make a clear classification impossible. Summary Since two studies for breath gas biomarkers of diabetes mellitus show different results, we have to critically re-think the analytical approach and the boundary conditions. Although breath analysis seems to be an excellent tool for the detection of diseases in their early states, oral bacteria play an important role. This has to be taken into account in further discussions about biomarkers of diseases as well as for the planning of related studies and the breath sampling