To evaluate the regulation of TSH secretion in nonthyroidal illness (NTI) we studied the nocturnal TSH surge in 11 healthy controls and 26 NTI patients; none of the patients was on medication known to interfere with TSH secretion. The presence of a nocturnal TSH surge was defined as a mean nighttime TSH (the mean of 5 samples taken hourly from 0000-0400 h) significantly greater than the mean daytime TSH (the mean of 5 samples taken from 1500-1900 h). A nocturnal TSH surge was present in 11 of 26 NTI patients and in 11 of 11 controls (P less than 0.01). Both the absolute (0.3 +/- 0.1 vs. 1.0 +/- 0.2 mU/L; P less than 0.01) and relative (11 +/- 6% vs. 71 +/- 12%; P less than 0.001) nocturnal TSH surges were lower in NTI patients than in controls. NTI patients had lower plasma T3 (1.11 +/- 0.08 vs. 1.84 +/- 0.11 nmol/L; P less than 0.001) and higher plasma rT3 (0.81 +/- 0.24 vs. 0.23 +/- 0.01 nmol/L; P less than 0.001) concentrations than controls, but T4, FT4, and TSH values were similar in both groups. No differences were found between the 15 NTI patients without nocturnal TSH surge and the 11 patients with a nocturnal TSH surge in sex distribution, age, caloric intake, or plasma T4 and T3, but hospital mortality was slightly, although not significantly, higher in those with an absent nocturnal TSH surge. An absent nocturnal TSH surge occurred in 2 of 2 patients with a low TSH (less than 0.4 mU/L), in 11 of 20 patients with a normal TSH (0.4-4.0 mU/L), and in 2 of 4 patients with a high TSH (greater than 4.0 mU/L). Pituitary TSH responsiveness to TRH was similar in patients with or without a nocturnal TSH surge. We conclude that NTI is frequently associated with a decreased nocturnal TSH surge. This phenomenon is not related to ambient plasma T4, T3, or TSH concentrations or pituitary TSH responsiveness to TRH. A decreased nocturnal TSH surge appears to be one of the features of the sick euthyroid syndrome and is probably related to hypothalamic dysregulation.