Relationship between nocturnal serum thyrotropin peak and metabolic control in diabetic patients

Circadian variations in serum TSH, especially its nocturnal rise, are often blunted in nonthyroidal illness. We analyzed TSH secretion in 15 diabetic patients (7 with type I and 8 with type II diabetes mellitus). Patients were evaluated when diabetes was poorly controlled (fasting blood glucose ranging from 13.7-19.2 mmol/L with absence of ketoacidosis) and after achieving glycemic control. Before correction of hyperglycemia, the nocturnal serum TSH peak (2230-0200 h) was abolished in 11 of 15 patients (73%); the mean (+/- SE) night TSH/morning TSH x 100 was 109.0 +/- 9.5 (range, 66.7-166.7) vs. a mean of 216.5 +/- 27.0 (range, 139.8-462.5) in normal controls. The mean morning TSH value in diabetics (1.9 +/- 0.4 mU/L) did not differ from that in normal age- and sex-matched controls. The mean TSH increase after iv administration of TRH was only slightly reduced (8.4 +/- 1.2 mU/L pretreatment vs. 10.8 +/- 1.6 mU/L posttreatment), with the TRH test blunted in 3 cases. No differences were found between type I and type II patients. Correction of hyperglycemia was associated with the reappearance of a nocturnal TSH peak in all but 1 patient (mean TSH peak, 198.2 +/- 13.0; P = NS vs. controls). This change paralleled the normalization of serum total T3 and rT3, which were reduced and increased, respectively, when diabetes was poorly controlled. An inverse relationship was found between serum fructosamine levels and the nocturnal TSH peak, suggesting that metabolic decompensation accounts for the abolishment of the latter.