Controversy remains about the value of combined treatment with levothyroxine (LT4) and liothyronine (LT3), compared with LT4 alone in primary hypothyroidism. We compared combined treatment with LT4 and LT3 in a ratio of 5:1 or 10:1 with LT4 monotherapy. We conducted a double-blind, randomized, controlled trial in 141 patients (18-70 yr old) with primary autoimmune hypothyroidism, recruited via general practitioners. Inclusion criteria included: LT4 treatment for 6 months or more, a stable dose for 6 wk or more, and serum TSH levels between 0.11 and 4.0 microU/ml (mU/liter). Randomization groups were: 1) continuation of LT4 (n = 48); 2) LT4/LT3, ratio 10:1 (n = 46); and 3) LT4/LT3, ratio 5:1 (n = 47). Subjective preference of study medication after 15 wk, compared with usual LT4, was the primary outcome measure. Secondary outcomes included scores on questionnaires on mood, fatigue, psychological symptoms, and a substantial set of neurocognitive tests. Study medication was preferred to usual treatment by 29.2, 41.3, and 52.2% in the LT4, 10:1 ratio, and 5:1 ratio groups, respectively (chi2 test for trend, P = 0.024). This linear trend was not substantiated by results on any of the secondary outcome measures: scores on questionnaires and neurocognitive tests consistently ameliorated, but the amelioration was not different among the treatment groups. Median end point serum TSH was 0.64 microU/ml (mU/liter), 0.35 microU/ml (mU/liter), and 0.07 microU/ml (mU/liter), respectively [ANOVA on ln(TSH) for linear trend, P < 0.01]. Mean body weight change was +0.1, -0.5, and -1.7 kg, respectively (ANOVA for trend, P = 0.01). Decrease in weight, but not decrease in serum TSH was correlated with increased satisfaction with study medication. Of the patients who preferred combined LT4/LT3 therapy, 44% had serum TSH less than 0.11 microU/ml (mU/liter). Patients preferred combined LT4/LT3 therapy to usual LT4 therapy, but changes in mood, fatigue, well-being, and neurocognitive functions could not satisfactorily explain why the primary outcome was in favor of LT4/LT3 combination therapy. Decrease in body weight was associated with satisfaction with study medication.