Metabolic trajectories during treatment of diabetic ketoacidosis described by breath analysis

Mo Awchi; Singh Kapil Dev; Sara Bachmann-Brenner; Marie-Anne Burckhardt; Melanie Hess; Urs Zumsteg; Jiafa Zeng; Datta Alexandre N.; Urs Frey; Gabor Szinnai; Pablo Sinues

Objective: This feasibility study aimed to investigate the anabolic effect of insulin on metabolites captured in exhaled breath during acute diabetic ketoacidosis (DKA) for a better pathophysiological understanding.

Research Design and Methods: Children and adolescents with type 1 diabetes (T1D) with DKA (n=5) and without DKA (n=7) and children with epilepsy without ketogenic diet (n=18) were recruited into an observational breath analysis study. Study participants repeatedly inflated Nalophan bags, and its metabolic content was subsequently interrogated by secondary electrospray ionization high-resolution mass spectrometer (SESI-HRMS). Routine laboratory parameters, such as glucose, pH, base excess, bicarbonate, and ketone bodies, were measured for participants with DKA admitted to the intensive care unit (ICU) and were used to identify associated metabolic markers in breath.

Results: In 30 participants, the total amount of longitudinal individual breath measurements was 107 (88 in 12 diabetic participants and 19 in 18 non-diabetic participants). In 12 diabetic participants, 24/88 measurements were conducted during DKA in 5 participants and 56/88 in 5 participants after and in 7 participants without DKA. SESI-HRMS analysis showed that acetone, pyruvate and acetoacetate that are well known to be altered in DKA, were readily detectable in breath in participants with DKA (base excess (BE) < -2) compared to non-diabetic participants (acetone P= 3x10^-7, pyruvate P= 5x10^-3 and acetoacetate P= 4x10^-7 respectively). Further, they followed the expected dynamics during the switch from catabolic to anabolic state in DKA participants. In addition, a total of 665 mass spectral features were found to significantly correlate (p < 0.05; q < 0.18) with BE (219/446 features showed significant positive/negative correlation). These correlating features with BE prompt metabolic trajectories towards in-control state as they progress towards homeostasis.

Conclusion: This study provides proof-of-principle for using exhaled breath analysis in an ICU setting for DKA research. We found that, apart from the well-known association between acetone and DKA state, a completely new panel of exhaled metabolites (e.g. acetoacetate), describe the transition between catabolic to anabolic state in DKA patients. This non-invasive new technology provides new insights and a more comprehensive overview on the effect of insulin during DKA treatment.

ESPE Abstracts

P1-31