Validation of the BETA-2 Score: An Improved Tool to Estimate Beta Cell Function After Clinical Islet Transplantation Using a Single Fasting Blood Sample

Abstract

The beta score, a composite measure of beta cell function after islet transplantation, has limited sensitivity because of its categorical nature and requires a mixed-meal tolerance test (MMTT). We developed a novel score based on a single fasting blood sample. The BETA-2 score used stepwise forward linear regression incorporating glucose (in millimoles per liter), C-peptide (in nanomoles per liter), hemoglobin A1c (as a percentage) and insulin dose (U/kg per day) as continuous variables from the original beta score data set (n = 183 MMTTs). Primary and secondary analyses assessed the score's ability to detect glucose intolerance (90-min MMTT glucose ≥8 mmol/L) and insulin independence, respectively. A validation cohort of islet transplant recipients (n = 114 MMTTs) examined 12 mo after transplantation was used to compare the score's ability to detect these outcomes. The BETA-2 score was expressed as follows (range 0-42): [Formula: see text] A score <20 and ≥15 detected glucose intolerance and insulin independence, respectively, with >82% sensitivity and specificity. The BETA-2 score demonstrated greater discrimination than the beta score for these outcomes (p < 0.05). Using a fasting blood sample, the BETA-2 score estimates graft function as a continuous variable and shows greater discrimination of glucose intolerance and insulin independence after transplantation versus the beta score, allowing frequent assessments of graft function. Studies examining its utility to track long-term graft function are required.

Keywords: clinical research/practice; diabetes: type 1; endocrinology/diabetology; immunosuppressant; immunosuppressive regimens; islet transplantation; islets of Langerhans; translational research/science.

© 2016 The Authors. American Journal of Transplantation published by Wiley Periodicals, Inc. on behalf of the American Society of Transplantation and the American Society of Transplant Surgeons.

Figures

Figure 1

Relationship between BETA‐2 score and indices of graft function. Relationship between BETA‐2 score and indices of graft function: (A) stimulated glucose, (B) beta score and acute insulin response (C) to glucose and (D) to arginine were highly statistically significant (all p < 0.01). AIR, acute insulin response.

Figure 2

Relationship between beta score intervals and the BETA‐2 score in the derivation cohort. The relationship between the beta score intervals and the BETA‐2 score were closely associated in the derivation cohort as demonstrated (r = 0.99; p < 0.01), with similar values in the validation cohort (data not shown). Data are shown as median (interquartile range).

Figure 3

BETA‐2 score's detection of 90‐min MMTTglucose level ≥8 mmol/L and insulin independence following islet transplantation in derivation and validation cohorts. (A) Participants in the original Edmonton derivation cohort (n = 58, 183 MMTTs), with 90‐min MMTT glucose levels ≥8 mmol/L (69% of cohort). The AUROC curve was 86 ± 2%, reflecting excellent discrimination (*p < 0.001) between the glycemic groups. A BETA‐2 score <20 detected a 90‐min MMTT glucose level ≥8 mmol/L with sensitivity of 82% (95% CI 74–88%) and specificity of 82% (95% CI 69–91%). (B) BETA‐2 and beta scores for participants (n = 114) in the validation cohort with 90‐min MMTT glucose levels ≥8 mmol/L (66% of cohort) versus <8 mmol/L confirmed the greater accuracy of the BETA‐2 score for detecting 90‐min MMTT glucose levels ≥8 mmol/L (ROC comparison; ‡p = 0.04). (C) Participants in the derivation cohort were segregated by insulin independence with HbA1c <6.5% (56% of cohort) versus insulin dependence. The AUROC curve was 96.4 ± 1.1%, reflecting excellent discrimination (*p < 0.001) between the groups. A BETA‐2 score ≥15 predicted insulin independence with sensitivity of 89% (95% CI 82–94%) and specificity of 88% (95% CI 78–94%). (D) BETA‐2 and beta scores for participants (n = 114) in the validation cohort with insulin independence (HbA1c <6.5%; 37% of cohort) versus insulin‐dependent participants confirmed greater accuracy of the BETA‐2 score for detecting insulin independence (ROC comparison; ‡p = 0.03). AUROC curve, area under the receiver operator characteristic curve; CI, confidence interval; MMTT, mixed‐meal tolerance test; ROC, receiver operator characteristic.

Figure 4

Frequency of beta and BETA‐2 scores according to 90‐min MMTTglucose level ≥8 mmol/L and insulin independence. Frequency of beta and BETA‐2 scores in islet transplant recipients is demonstrated according to 90‐min MMTT glucose level ≥8 or <8 mmol/L (A and B) and insulin dependence versus independence (C and D) in the validation cohort. The BETA‐2 scores demonstrated greater resolution than the beta scores (p < 0.001). MMTT, mixed‐meal tolerance test.

Figure 5

Example of beta versus BETA‐2 scores in a participant after islet transplantation. The beta score and the BETA‐2 score were demonstrated for a participant with type 1 diabetes receiving two islet infusions 11 mo apart. MMTTs were performed 1 mo following each transplant. Fasting blood samples were measured frequently, allowing frequent calculation of the BETA‐2 score, whereas the participant could attend MMTTs only infrequently; therefore, the beta score was calculated for the first time at 4 weeks after transplant. MMTT, mixed meal tolerance test.