Design: Diagnostic accuracy analysis within a multicentre cohort
Population: 422 US youth with type 1 diabetes, stimulated C-peptide within 12 months of diagnosis
Primary outcome: Accuracy of IDAA1c ≤9 for C-peptide-defined partial remission
Effect: Sensitivity 61%, specificity 79% overall; sensitivity 45% in overweight or obesity

Insulin dose-adjusted HbA1c (IDAA1c: HbA1c in percent plus four times the daily insulin dose in units per kg) of 9 or below is the usual definition of partial remission in children with new type 1 diabetes, and is used to select patients for disease-modifying trials. The SEARCH for Diabetes in Youth study tested it against mixed-meal stimulated C-peptide in 422 young people within a year of diagnosis.

Overall, IDAA1c of 9 or below had 61% sensitivity and 79% specificity for preserved C-peptide (area under the curve 0.77). In children with overweight or obesity, sensitivity fell to 45%. These children had higher C-peptide than leaner peers (1.88 vs 1.25 ng/mL) yet slightly higher IDAA1c, because insulin resistance pushes the dose up. A cut-off of 11.6 raised sensitivity to 93% in this group, at the cost of specificity dropping to 54%.

The practical point is that a heavier child whose IDAA1c sits above 9 may still have meaningful beta-cell function. That matters when counselling families about the honeymoon phase and when considering referral to trials of teplizumab or other agents that target preserved beta cells.

Do not use IDAA1c above 9 alone to conclude a heavier child has lost residual beta-cell function.
Where trial eligibility or a disease-modifying treatment depends on it, consider a stimulated C-peptide instead.
Expect higher insulin doses in children with overweight or obesity even while C-peptide is preserved.
Explain to families that the honeymoon phase is judged imperfectly by dose and HbA1c alone.

Why it matters

Trials of beta-cell-preserving drugs select patients with this score, so heavier children may be wrongly excluded.

The statistics, in plain English

Sensitivity of 45% means that of every 100 heavier children who truly still made enough insulin, the score labelled only 45 as being in remission. Moving the cut-off to 11.6 catches 93 of them but also mislabels nearly half of those not in remission. No single threshold does both jobs well, which is the usual trade-off with a screening score.