Correlation Between Breath Acetone and Ketone Bodies in Blood and Urine Among Individuals with Different...

Background/Objectives: Rapid and accurate ketone detection is clinically valuable for both diabetic and non-diabetic individuals. Breath acetone represents an attractive non-invasive and real-time alternative for monitoring ketones. This study aimed to explore the correlation between breath acetone measured by proton transfer reaction time-of-flight mass spectrometry (PTR-TOF-MS) and ketone bodies in blood and urine in individuals with normoglycemia, prediabetes and diabetes. Methods: A total of 4000 participants were recruited. Urinary ketone levels, blood β-hydroxybutyrate (βHB) concentrations, and breath acetone (measured by PTR-TOF-MS) were assessed. Participants were classified into urine ketone-negative and urine ketone-positive groups. They were further divided according to the semiquantitative degree of urinary ketone (-, +, ++, +++ and ++++). Spearman rank correlation analysis was performed to evaluate the relationship between breath acetone and blood βHB. Receiver operating characteristic (ROC) curve analysis was conducted to determine the preliminary thresholds of breath acetone for detecting blood βHB positivity. Multiple linear regression, including a quadratic term for blood βHB, was performed, and the inflection point of the curve was calculated. Results: Blood βHB and breath acetone were significantly higher in the urine ketone-positive group than in the urine ketone-negative group (p < 0.001). A progressive elevation in blood βHB and exhaled breath acetone concentrations was observed with increasing urinary ketone levels (p < 0.001). The Spearman rank correlation coefficient between breath acetone and blood βHB was 0.493, 0.522, 0.379, and 0.636 in the general population, normoglycemia, prediabetes, and diabetes groups, respectively (all p < 0.001). The preliminary thresholds of breath acetone concentration for detecting blood βHB positivity were 2.37 ppmv in diabetes, 1.10 ppmv in prediabetes, and 2.05 ppmv in normoglycemia, respectively. Blood βHB exhibited a significant quadratic association with breath acetone in the general population, normoglycemia, and diabetes (B = -1.698, -1.409, and -16.070, respectively; p < 0.001), with inflection points of the quadratic curves at 2.73 mmol/L, 3.28 mmol/L and 1.61 mmol/L, respectively, whereas in prediabetes, the relationship did not significantly deviate from linearity. Conclusions: In normoglycemia and diabetes, breath acetone based on PTR-TOF-MS exhibited an inverted U-shaped relationship with blood βHB, while in prediabetes, breath acetone showed a linear positive correlation with blood βHB. In populations where blood βHB is below the quadratic inflection point, breath acetone measured by PTR-TOF-MS holds promise as a non-invasive ketone-monitoring tool.