It is controversial whether the administration of thyroid hormone to patients with nonthyroidal illness has any beneficial effect. Two groups of patients undergoing abdominal hysterectomy under the same general anesthesia were studied. Group A consisted of 15 women taking chronically l-thyroxine therapy (1.8 mug/kg daily), and group B (control) consisted of 16 apparently healthy euthyroid women taking placebo. Thyroid hormones, cortisol, and interleukin (IL)-6 were measured 1 day before and 1, 2, 3, and 6 days after surgery. Total triiodothyronine (TT(3)) decreased to a significantly greater degree (P <.05) and for a longer period of time in group B than in group A. The significant increase of reverse T(3) (rT(3)) noted early in group B failed to reach the baseline levels until the end of the study, whereas in group A, rT(3) returned to the preoperative values by day 2. Both groups had similar free thyroxine (FT(4)) at baseline. FT(4) increased significantly at day 1 and remained significantly elevated throughout the postoperative period in group B only. Serum TT(4), thyroid-stimulating hormone (TSH), and cortisol did not change significantly in either group. In all patients, IL-6 increased significantly to a peak value at day 1, showing a slow decrease thereafter. A stronger negative correlation was found between T(3) and IL-6 in group B than in group A (r = -.66, P <.0001 v r = -.38, P <.001, respectively) and a strong positive correlation was observed between rT(3) and IL-6 in group B only (r =.57, P <.001). The long-term treatment with T(4) seems to attenuate the decrease of serum T(3), which occurs during the development of nonthyroidal illness postoperatively. The elevation of IL-6 accounted for a greater proportion of the variations of the T(3) and rT(3) in the control group B than in the T(4)-treated group A.