Continuous glucose monitoring reveals thyroid feedback-related residual glycemic variability and hypoglycemia during inpatient glucose optimization in euthyroid adults with type 2 diabetes: a CGM-based observational study

Continuous glucose monitoring (CGM) captures short-term glucose excursions and hypoglycemia beyond what HbA1c reflects, but endocrine traits that shape inpatient glycemic variability (GV) during therapy intensification are poorly defined. We evaluated whether thyroid feedback indices that proxy central thyroid hormone sensitivity are associated with CGM-derived residual GV and hypoglycemia in euthyroid inpatients with type 2 diabetes. We conducted a retrospective cohort study of 224 euthyroid adults with type 2 diabetes admitted for structured glucose optimization who wore blinded CGM for at least 6 consecutive days. Baseline FT3, FT4 and TSH were used to derive thyroid feedback indices (TFQI, TSHi and TT4RI). Daily CGM metrics were computed; Day 6 was prespecified as the primary later-inpatient analysis day at the cohort level. High GV was defined as Day 6 MAGE ≥ 3.9 mmol/L, and hypoglycemia as ≥ 2 consecutive readings < 3.9 mmol/L. Associations were tested using linear regression (including day-specific models) and multivariable logistic regression with sequential adjustment, including mean glucose and clinical covariates; post hoc sensitivity analyses additionally adjusted for available discharge medication classes. Overall CGM profiles improved from Day 2 to Day 6 (TIR: 62 ± 25% to 84 ± 15%; MAGE: 6.34 ± 2.73 to 4.78 ± 2.04 mmol/L; both p < 0.001), while hypoglycemia still occurred in a subset. The TFQI–MAGE association increased in magnitude across inpatient days and was significant on Day 6 (β = 1.38, p < 0.001). In fully adjusted models, higher TFQI was associated with high GV (OR 3.11, 95% CI 1.27–7.60) and hypoglycemia (OR 4.78, 95% CI 1.56–14.63). In euthyroid adults with type 2 diabetes undergoing inpatient titration, higher thyroid feedback indices, particularly TFQI, were linked to greater residual CGM variability and a higher risk of hypoglycemia on the later inpatient day. These readily available indices may augment CGM-based inpatient risk stratification for monitoring intensity and titration pace; prospective studies incorporating treatment dose, timing, dietary intake and stress physiology are needed. Not applicable.