Multiple-marker risk parameters predictive of conversion to diabetes
Methods, systems and circuits evaluate a subject's risk of developing type 2 diabetes using defined mathematical models of short term risk (STR) and longer term risk of progression. The evaluations can stratify risk for patients having the same glucose measurement, particularly those with intermediate or low (normal) fasting plasma glucose (FPG) values. The STR or IR (insulin resistance) model(s) may include an inflammatory biomarker such as an NMR derived measurements of GlycA and a plurality of selected lipoprotein components of at least one biosample of the subject. Embodiments of the invention also provide methods, systems and circuits that generate STR scores as a marker of beta-cell dysfunction or impairment.
1. A method of evaluating a subject's risk of developing type 2 diabetes and/or having pancreatic beta cell impairment and/or dysfunction, comprising:
performing Nuclear Magnetic Resonance spectroscopy on a biosample from the subject to measure a concentration of a defined high density lipoprotein particle (HDL-P) subpopulation and a concentration of GlycA;
programmatically calculating a short term diabetes risk factor (SDRF) score of a subject using a defined mathematical model; and
comparing the calculated SDRF score to a SDRF score of a study population,
wherein the defined mathematical model includes a concentration measurement of a defined high density lipoprotein particle (HDL-P) subpopulation, a concentration measurement of inflammatory marker GlycA to generate the calculated SDRF score.
2. The method of claim 1 , wherein the defined HDL-P subpopulation is medium HDL-P.
3. The method of claim 2 , further comprising an interaction parameter of GlycA multiplied by the concentration of medium HDL-P, wherein the subject is at risk of beta cell dysfunction when medium HDL-P and GlycA values are in a third tertile values of the population norm.
4. The method of any of claim 1 , further comprising programmatically calculating an insulin resistance (IR) score of the subject using a defined mathematical model of insulin resistance.
5. The method of claim 4 , wherein the defined mathematical model of insulin resistance for the IR score includes a plurality of components including a concentration measurement of a defined HDL-P subpopulation, a concentration measurement of GlycA, a concentration measurement of a defined subpopulation of VLDL-P (very low density lipoprotein/chylomicron particle subclasses), an IR index with a range of between 0-100, the range representing from low to high, insulin sensitivity to insulin resistance, and a concentration measurement of a branched chain amino acid (BCAA), obtained from the biosample of the subject.
6. The method of claim 5 , wherein the components of the defined mathematical model of insulin resistance are mathematically combined to generate the IR score.
7. The method of claim 5 , wherein the measurement of the defined sub-population of VLDL-P is a concentration of medium VLDL-P, and wherein the BCAA is Valine.
8. The method of claim 4 , further comprising programmatically calculating a diabetes risk score by combining the SDRF score and the IR score.
9. The method of claim 8 , wherein coefficients of the SDRF and IR scores are derived from a logistic regression model that includes SDRF and IR to predict actual 5 year conversion to diabetes using at least one defined population study to thereby generate a DRI score with a risk of conversion within 5 years irrespective of a glucose value of the subject.
10. The method of any of claim 9 , further comprising evaluating a measured glucose and/or HbA1c value of the subject, and electronically providing a report with a 5 year risk of conversion to Type 2 diabetes risk estimate based on the glucose measurement and the DRI score.
11. The method of claim 4 , wherein the IR score is an eLP-IR score and is calculated using the following equation:
eLP-IR=(A) (LP-IR)+(B)(Valine)−(C) (VLDL-P MED )−(D)(HDL-P MED )+(E)(GlycA),
wherein A, B, C, D and E are defined beta coefficients from a linear regression model for insulin resistance, HDL-P mED is the concentration of a medium HDL-P subpopulation, VLDL-P MED is a concentration of a medium VLDL-P subpopulation, valine is a branched chain amino acid, and LP-IR is a lipoprotein insulin resistance index calculated using six defined lipoprotein subclasses and has a numerical value in a range of 0-100 representing insulin sensitive to insulin resistance.
12. The method of any of claim 4 , further comprising generating a DRI score using the following equation:
DRI score=X(IR score)+Y(SDRF),
wherein X and Y are coefficients defined by a logistic regression model for 5-year diabetes conversion in people with glucose <110 mg/dL using a defined study population, and wherein the DRI score is mathematically altered into a range of between 0-10 or 1-10 or 0-100 or 1-100 using a plurality of equal subparts over a range of possible DRI raw scores.
13. The method of any of claim 1 , wherein the SDRF score is calculated using the following equation:
SDRF score=−(A) (HDL-P MED )−(B) (GlycA)+(C) (GlycA×HDL-P MED ),
wherein A, B and C are defined beta coefficients from a logistic regression model for short term conversion to diabetes as the defined mathematical model of risk of developing type 2 diabetes, and HDL-P MED is a medium size HDL-P subpopulation, and GlycA χHDL-P MED an interaction parameter.
14. The method of claim 1 , further comprising at least one of: evaluating a drug therapy, evaluating a clinical trial, or evaluating candidates for drug discovery, using the SDRF score.
15. The method of claim 1 , further comprising calculating a plurality of the SDRF scores over time from respective biosamples to thereby evaluate a change in SDRF score to identify a change in β-cell dysfunction.
16. The method of claim 1 , wherein raw scores associated with the SDRF score are between −6.4 and −1.6, wherein −4.1 is associated with about a 25th percentile and >−3.8 is associated with about >75th percentile of the study population, and wherein >−3.8 values indicate an increased risk of beta cell dysfunction and/or early conversion to type 2 diabetes independent of glycemic value that can stratify risk of conversion to type 2 diabetes for subjects having a common glycemic measurement with different SDRF scores.
17. The method of claim 1 , wherein the SDRF score is provided in a report in a defined numerical score range, with scores associated with a fourth quartile (4Q), fifth quintile (SQ) or 10th decile of the study population reflecting an increased risk of developing type 2 diabetes within 2 years and/or beta cell dysfunction and/or impairment relative to scores in a first quartile, quintile or decile of the population.
18. The method of claim 1 , wherein the HDL-P subpopulation comprises only medium HDL particle subclasses with diameters between 8.3 nm and 9.4 nm, between 8.3 nm and 10.0 nm, or between 8.3 nm and 10.2 nm.
19. A method of identifying at-risk patients that may benefit from therapies for improving or stabilizing beta-cell function and/or improve a patient's ability to produce insulin, comprising:
performing Nuclear Magnetic Resonance spectroscopy on a biosample from a subject to measure concentrations of GlycA and defined lipoprotein and metabolite components; and
electronically generating a short term risk (STR) score by combining measurements of defined lipoprotein and metabolite components of a biosample of a subject, wherein the components include a high density lipoprotein particle (HDL-P) subpopulation and an interaction parameter using the concentration measurements of the HDL-P subpopulation and GlycA.
20. A method of identifying subjects that are likely to benefit from a drug therapy such as reconstituted HDL infusion for improving pancreatic beta cell function and/or to inhibit Type 2 diabetes mellitus (T2DM), comprising:
performing Nuclear Magnetic Resonance on a biosample from a subject to measure concentrations of GlycA and defined lipoprotein and metabolite components;
generating a defined short term diabetes risk factor (SDRF) score using measurements of defined lipoprotein and metabolite components of a biosample of a subject, wherein the components include a high density lipoprotein particle (HDL-P) subpopulation and GlycA; and
comparing the generated SDRF score to the SDRF score of a defined population norm to identify if a subject is likely to benefit from therapy to improve pancreatic beta cell function and/or inhibit T2DM.