A prospective, cohort study of 75 consecutive patients requiring management in the medical intensive care unit (MICU) of the Singapore General Hospital was carried out over a five-month period to determine thyroid and adrenocortical profiles and evaluate their use in predicting patient outcome. Up to 88% of patients had at least one abnormal thyroid function and 77% had abnormal adrenocortical function test results. There were significantly lower triiodothyronine, thyroxine and free thyroxine, but not thyrotropin levels, and higher cortisol levels in non-survivors compared to survivors (all P < 0.01). Of the endocrine parameters, triiodothyronine and cortisol concentrations were independent predictors of outcome. The overall predictive accuracy of combining these two variables on admission into the MICU was 74%. The APACHE II (acute physiology and chronic health evaluation II) score alone predicted outcome with 71% accuracy, and in combination with triiodothyronine and cortisol levels improved accuracy to 84%. The use of dopamine alone predicted outcome with 74% accuracy, and in combination with triiodothyronine and cortisol levels, improved accuracy to 84%. Measurements of total triiodothyronine and cortisol concentrations on admission to the MICU, and consideration of the use of dopamine improve on the APACHE II score in outcome prediction.