[13C]glucose breath testing provides a noninvasive measure of insulin resistance: calibration analyses against clamp studies.

BACKGROUND: Exhaled (13)CO2 following ingestion of [(13)C]glucose with a standard oral glucose tolerance load correlates with blood glucose values but is determined by tissue glucose uptake. Therefore exhaled (13)CO2 may also be a surrogate measure of the whole-body glucose disposal rate (GDR) measured by the gold standard hyperinsulinemic euglycemic clamp. SUBJECTS AND METHODS: Subjects from across the glycemia range were studied on 2 consecutive days under fasting conditions. On Day 1, a 75-g oral glucose load spiked with [(13)C]glucose was administered. On Day 2, a hyperinsulinemic euglycemic clamp was performed. Correlations between breath parameters and clamp-derived GDR were evaluated, and calibration analyses were performed to evaluate the precision of breath parameter predictions of clamp measures. RESULTS: Correlations of breath parameters with GDR and GDR per kilogram of fat-free mass (GDRffm) ranged from 0.54 to 0.61 and 0.54 to 0.66, respectively (all P<0.001). In calibration analyses the root mean square error for breath parameters predicting GDR and GDRffm ranged from 2.32 to 2.46 and from 3.23 to 3.51, respectively. Cross-validation prediction error (CVPE) estimates were 2.35-2.51 (GDR) and 3.29-3.57 (GDRffm). Prediction precision of breath enrichment at 180 min predicting GDR (CVPE=2.35) was superior to that for inverse insulin (2.68) and the Matsuda Index (2.51) but inferior to that for the log of homeostasis model assessment (2.21) and Quantitative Insulin Sensitivity Check Index (2.29) (all P<10(-5)). Similar patterns were seen for predictions of GDRffm. CONCLUSIONS: (13)CO2 appearance in exhaled breath following a standard oral glucose load with added [(13)C]glucose provides a valid surrogate index of clamp-derived measures of whole-body insulin resistance, with good accuracy and precision. This noninvasive breath test-based approach can provide a useful measure of whole-body insulin resistance in physiologic and epidemiologic studies.

Intra-individual variability of CO2 breath isotope enrichment compared to blood glucose in the oral glucose tolerance test.

BACKGROUND: Glucose tolerance can be assessed noninvasively using (13)C-labeled glucose added to a standard oral glucose load, by measuring isotope-enriched CO(2) in exhaled air. In addition to the clear advantage of the noninvasive measurements, this approach may be of value in overcoming the high variability in blood glucose determination. METHODS: We compared within-individual variability of breath CO(2) isotope enrichment with that for blood glucose in a 75-g oral glucose tolerance test (OGTT) by adding 150 mg of d-[(13)C]glucose ((13)C 99%) to a standard 75-g glucose load. Measurements of whole blood glucose (by glucose oxidase) and breath isotope enrichment (by isotope ratio mass spectrometry) were made every 30 min for 3 h. Subjects underwent three repeat tests over a 3-week period. Values for variability of breath isotope enrichment at 3 h (∂‰180) and of area under the curve for enrichment to 180 min (AUC180) were compared with variability of the 2-h OGTT blood glucose. RESULTS: Breath test-derived measures exhibited lower within-subject variability than the 2-h OGTT glucose. The coefficient of variation for ∂‰180 was 7.4 ± 3.9% (mean ± SD), for AUC180 was 9.4 ± 6.3%, and for 2-h OGTT blood glucose was 13 ± 7.1% (P = 0.005 comparing ∂‰180 versus 2-h blood glucose; P = 0.061 comparing AUC180 versus 2-h blood glucose; P = 0.03 comparing ∂‰180 versus AUC180). CONCLUSIONS: Breath test-derived measurements of glucose handling had lower within-subject variability versus the standard 2-h blood glucose reading used in clinical practice. These findings support further development of this noninvasive method for evaluating glucose tolerance.

Effect of triglyceride structure on fecal excretion of 13C-labeled triglycerides.

OBJECTIVE: The aim of this work was to determine the effects of specific changes in the structure of (13)C-labeled triglyceride (TG*) on its fecal excretion relative to total stool fat excretion determined simultaneously in patients with reduced exocrine pancreatic function. METHODS: A series of 47 studies were conducted in 26 young cystic fibrosis (CF) patients and 11 adult patients with chronic pancreatitis over a five year period. Each test consisted of ingesting a single high fat test meal containing both (13)C-labeled triglyceride (TG*) and dysprosium chloride (DyCl(3)) a nonabsorbable marker of intestinal transit; in most studies the food colorant brilliant blue (FD&C blue #1) was administered along with the DyCl(3). The TG*s tested were: P*P*P* = TRIPALMITIN-1,1,1-(13)C(3); SO*S = 2-OCTANOYL-1,3-DISTEARIN-2-octanoyl-1,2-(13)C(2); and P*LP* = 2-LAURYL-1,3-DIPALMITIN-dipalmitoyl-1,1,2,2-(13)C(4). Ingestion of the test meal was followed by collection of individual stools for at least 72 hours. Stools were analyzed for (13)C-Excess ((13)C*), total fat, and Dy. RESULTS: Excretion of P*LP* showed a high degree of linear correlation with stool fat (r(2) = 0.924) over a wide-range of fecal fat values. Excretion of SO*S was also significantly correlated with stool fat, but its excretion was less than 10% at all levels of steatorrhea and the slope of the regression line relating TG* excretion to stool fat was some four to five times smaller than observed for P*LP*. Fecal excretion of P*P*P* was highly correlated with stool fat (r(2) = 0.941) in patients with moderate steatorrhea (<25 g fat/24 hours) and the slope of the regression line (3.20) was considerably greater than for P*LP*. Only results from those studies in which stool collections were complete (Dy excretion >90%) were utilized in the statistical comparisons (36 of 47 studies). CONCLUSIONS: The observed highly significant linear correlation between P*LP* and stool fat over the entire range of steatorrhea suggests that P*LP* excretion may be a suitable surrogate for fecal fat in patients with reduced exocrine pancreatic function. Because fecal excretion of TG* administered as described can be accurately determined by sampling only two visually marked stools, development of a noninvasive test to replace the current 72-hour stool fat test using this approach is possible. Use of other engineered TG*s and/or labeled fatty acids, may provide a method for non-invasive in vivo assessment of the specific defect(s) leading to steatorrhea in other patient groups.

Dysprosium chloride as a nonabsorbable gastrointestinal marker for studies of stable isotope-labeled triglyceride excretion in man.

OBJECTIVE: The aim of this work was to determine if dysprosium chloride (DyCl(3)) is a suitable nonabsorbable marker for studies of labeled-triglyceride excretion in cystic fibrosis patients allowing excretion to be determined accurately after analysis of one or two stools. METHODS: A series of 66 absorption studies were conducted in 36 young cystic fibrosis patients over a five year period. All tests consisted of ingesting a single test meal containing both (13)C-labeled triglyceride (TG*) and DyCl(3); in most studies the food colorant brilliant blue (FD&C blue #1) was administered along with the DyCl(3). Ingestion of the test meal was followed by collection of individual stools for 72 to 96 hours. Stools were analyzed for (13)C-Excess ((13)C*) and Dy. RESULTS: Excretion of Dy in cystic fibrosis patients who exhibited a wide-range of steatorrhea was quantitative. Fractional excretion of Dy and (13)C* in individual stools showed a high linear correlation (r(2) = 0.969) with a slope and y-intercept close to unity and zero, respectively. As a result, estimates of TG* excretion based on analysis of only two stools (partial pool method, PPM) were not different from those based on the analysis of all stools or stool composites. This was true both when Dy content and when stool color due to ingested brilliant blue was used to determine which stools to analyze for the PPM. CONCLUSIONS: Combining the use of Dy and brilliant blue permits reasonably accurate estimates of fecal TG* excretion after analysis of samples from two easily identified stools. This practical method can be used to address many important clinical and experimental questions regarding triglyceride digestion and absorption that may otherwise go unanswered.