806-P: Quantifying Carbohydrate Intake: A Reliable, Valid, Sensitive Alternative to 24-Hour Dietary Recall

Carbohydrate (carb) ingestion strongly influences postprandial glucose (PPG). With recent interest in the role PPG plays in A1c and diabetic complications, there are research and clinical needs to quantify routine carb consumption. The 24-hour dietary recall is the gold standard for estimating carbs eaten, but it is tedious and narrowly focused. We developed a simple self-report to measure carbs routinely consumed (CRC). We hypothesized that the CRC would: 1) be reliable, 2) correlate with the 24-hour dietary recall, 3) be sensitive to a PPG-lowering intervention, and 4) reflect changes in A1c. Adults with type 2 diabetes (n = 116; mean age: 56 ± 11.6 years; mean disease duration: 5.3 ± 3 years; mean A1c: 8.2 ± 1.0) participated in a clinical trial comparing weight reduction to PPG reduction. At baseline and 3-month follow-up, participants completed the CRC, 3 days of ASA24 dietary recall, and A1c. The CRC score was the sum of the number of servings of 16 different high-carb food classes eaten (e.g., dried fruits) in an average week. The CRC mean was 35.9 ± 16.1. One-week test-retest reliability (r) was .96 (p = .02). Validity: At baseline, carbs eaten by CRC and ASA24 dietary recall were correlated (r = .43, p < .001). Change in CRC from baseline to follow-up was significantly lower for the PPG group (36.8 to 20.3, p <.001) but not for the weight-lowering group (33.0 to 34.4, p = .7). Correlation of pre-post change in A1c with CRC and ASA24 were r = .39 (p = .002) and r = -.03 (p = .8) respectively. Regression analysis indicated only the change in CRC predicted change in A1c. These data indicate the CRC is reliable and shows concurrent, discriminant and predictive validity - only changes in CRC reflected changes in A1c. The CRC may be useful to researchers in evaluating programs designed to lower carb intake. Clinicians might use it to evaluate a patient’s routine consumption of carbs and responsiveness to lifestyle interventions. Limitation: the CRC is not a comprehensive nutritional analysis. Disclosure D. Cox: None. M.A. Moncrief: None. T. Banton: None. V.M. Ngo: None. H. Singh: None. A.M. Diamond: None. A.L. McCall: None. Funding National Institutes of Health (R01DK108957); Dexcom, Inc. (IIS-2016-047)