We assessed the diagnostic performances of homeostasis model assessment indices (HOMA) of β-cell function (HOMA-%β) and of insulin resistance (HOMA-IR) for cystic fibrosis related diabetes (CFRD) screening.

Data were collected from a prospective cohort of 228 patients with CF (117 adults and 111 children). Fasting insulin and glucose levels were measured to calculate HOMA-%β and HOMA-IR. HOMA-%β <100 indicated insulin secretion deficiency and HOMA-IR >1 insulin resistance. Both were used to calculate sensitivity, specificity, and positive and negative predictive values (PPV and NPV). Two-hour oral glucose tolerance tests (2h-OGTT) defined CFRD. Analyses were conducted separately for children and adults. Performances of HOMA-%β and HOMA-IR were calculated at inclusion, for each year of follow-up and for pooled data over the follow-up period.

Sensitivity, specificity, NPV and PPV were respectively: 88%, 45%, 98% and 11% for HOMA-%β and 42%, 48%, 91% and 6% for HOMA-IR in the pooled data of children; and 83%, 18%, 90% and 10% for HOMA-%β, and 39%, 80%, 92% and 18% for HOMA-IR in the pooled data of adults. Combining HOMA-%β and HOMA-IR did not improve performances.

Within both age groups, HOMA-%β <100 provided good sensitivity and NPV. HOMA-IR >1 had low sensitivity. Calculation of the HOMA-%β could be an interesting first-line screening approach to exclude CFRD and thus avoid unnecessary OGTT in patients for whom value is ≥100. However, HOMA-%β<100 does not support the diagnosis of CFRD and should be complemented by OGTT.