Improved practicality of insulin measurements by using urine as a source: relationship between serum and urine measurements as a function of dietary intake (Preprint)
BACKGROUND Obesity, insulin resistance and diabetes are taking epidemic proportions and novel approaches to addressing this world-wide public health challenge are required. The only molecular parameter that is currently routinely measured in this domain is blood glucose in patients with diabetes. However, measuring insulin concentrations would be more informative of the metabolic state of any person and applicable to people with obesity at varying levels of insulin resistance, including those with normal blood glucose levels. We recently demonstrated with 52 participants dieting the utility of determining insulin concentrations in urine as a molecular feedback mechanism. OBJECTIVE Our ultimate goal is to replace invasive blood insulin measurements with a more non-invasive approach. Towards this goal, we here demonstrate the use of a mobile health application to record diet and anthropometric data together with the measurement of insulin in urine and in blood, in controlled laboratory conditions and in the field. METHODS Five females aged 40-50 years were recruited and studied under laboratory conditions. Two of these were also studied in their work/home environment. The participants recorded events such as food intake, urine volume and exercise to the mobile health platform available as webinterface personalhealth.warwick.ac.uk and as mobile applications through the apple and google play stores. Urine samples were collected while varying dietary intake (low-carbohydrate, normal and ketogenic diets) and timing of food intake. Urine insulin values were measured by a highly sensitive immunosandwich electrochemiluminescence assay, which features 5 orders of magnitude in dynamic range and a fM detection limit. RESULTS We show that blood insulin and urine insulin values are linearly dependent, with urine concentrations being 3 times lower than the corresponding concentrations in serum. Characteristic urine insulin profiles were obtained by varying diet and participant and were found to be highly reproducible for the same diet/participant combination. As expected, concentrations of insulin were overall higher under normal diet conditions as compared to low-carbohydrate or ketogenic diet conditions. CONCLUSIONS This research demonstrates a practical and accurate approach to measure insulin in urine without the need for pre-processing of urine samples. The approach is applicable to a broad range of insulin concentrations as found under a variety of dieting conditions and inter-personal differences. Potential applications are improved diabetes care, and diet adherence monitoring, useful not only for clinicians but also for individuals who can thus obtain personalized metabolic feedback to food intake choices. This may empower individuals to visualize the metabolic effects of nutritional interventions with the quantitative biomarker insulin. CLINICALTRIAL NA