Introduction A major and largely underestimated [1] concern of today’s society is metabolic health (e.g. obesity) that is related to several diseases (e.g., diabetes, cardiovascular illnesses). Various diet strategies (e.g., ketogenic [2], intermittent fasting [3]) as well as exercising are explored but the assessment of treatment effectiveness on an individual level remains difficult. Desired are simple and accurate integrated devices that communicate wirelessly [4] and monitor metabolic changes conveniently at the point-of-care. This is possible through non-invasive acetone detection, a metabolic breath marker of lipolysis [5]. In specific, acetone is formed during hepatic β-oxidation of fatty acids that further divide into acetoacetate [6] that undergoes decarboxylation and enzymatic degradation to acetone [6], which is highly volatile and measureable in exhaled breath [7]. This has led to the development of several acetone sensors in the last decade.Most, however, suffer from cross-interference to endogenous (e.g., isoprene) and background interferants (e.g., alcohol from disinfectants) [8], and their evaluation in the application has been hardly considered, a general challenge in sensor science [9]. Catalytic filters [10] offer a simple and most effective approach to enhance selectivity and continuously mitigate the interference of confounders. Here, we present a low-cost and compact detector based on a flame-made Pt-loaded Al2O3 catalytic filter [11] coupled to a Si/WO3 sensor [8] for rapid and highly selective acetone detection. We apply it for breath analysis to monitor exercising and fasting. Method Nine volunteers attended two separate appointments. In the first, they performed an exhaustive spiroergometry test for determination of the cardiorespiratory fitness, maximum oxygen uptake (VO2) and second ventilatory threshold (VT2). During the second appointment, all volunteers performed a cardiorespiratory fitness-adapted submaximal aerobic exercise protocol [12]. The protocol started at 20% of the individual VT2 (determined in the first appointment) and increased by 10% every 5 min. Breath was sampled every 5 min during exercise, as well as every 30 min during 3 hours of post-exercise rest with an end-tidal breath sampler [13] and analyzed by PTR-ToF-MS and with and without a Pt/Al2O3 filter-enhanced Si/WO3 sensor. Results and Conclusions Figure 1 shows the normalized (i.e., to the baseline at t = 0) end-tidal acetone concentrations during exercising (0 ≤ t < 60 min) and fasting (60 < t ≤ 240 min) for all nine volunteers, as determined with bench-top PTR-ToF-MS (circles) and the Pt/Al2O3 filter-enhanced Si/WO3 sensor (triangles). The error bars indicate the standard error of the mean (SEM) for all volunteers. Baseline acetone concentrations (0.4 - 1.7 ppm) increase marginally during the exercise phase (i.e., < 5%). In contrast, a strong increase (up to 124%) takes place during subsequent fasting. This indicates that the exercise stimulated the body fat metabolism.The impact of the filter is shown best when monitoring breath acetone in situ during exercising. In fact, the sensor with filter follows closely the bench-top PTR-ToF-MS, while the sensor alone deviates (Figure 1a, squares). This might be related to endogenous isoprene (Figure 1b, diamonds), that spikes immediately in breath during muscle activity (up to 656 ppb within the first 5 min [14]) and then declines. While the sensor responds to isoprene [8], this is mitigated by the filter (triangles). Also varying background ethanol (stars) from hand disinfection [15] may affect the sensor. Thus, with filter, this detector is promising for robust and in situ metabolic monitoring to guide personalized dieting and exercising.
Read full abstract