Severe nonthyroidal illness has been claimed to cause secondary hypothyroidism. We reevaluated this concept measuring serum free T4 and free T3 by an ultrafiltration method and serum TSH by an ultrasensitive technique (detection limit, and serum TSH by an ultrasensitive technique (detection limit, 0.05 mU/L). Forty-five critically ill patients suffering from hepatic coma (n = 10), terminal cancer (n = 9), stroke (n = 8), and respiratory insufficiency not treated (n = 7) and treated (n = 11) with dopamine were studied. The mortality rate was 80%. No patients received glucocorticoids, and only patients in the last group received dopamine. Serum total as well as free thyroid hormone index values were grossly reduced in the majority of the patients. The 34 patients not receiving dopamine in general had normal values of serum free T4 (32 of 34) and free T3 (31 of 34), measurable TSH (33 of 34), and detectable TSH responses to iv TRH (33 of 34). In contrast, the dopamine-treated patients had reduced serum free T4 and TSH levels compared to normal subjects (P less than 0.05), as well as reduced TSH responses to TRH (P less than 0.01). Serum free T4 and free T3 were below the normal range in 3 patients and 1 patient, respectively, and serum TSH was below the detection limit in 2 patients. We conclude that critically ill patients with nonthyroidal illness not receiving dopamine have normal pituitary-thyroid function, whereas dopamine induces some degree of secondary hypothyroidism.