Reference intervals in the diagnosis of thyroid dysfunction: treating patients not numbers.

Although assigning a diagnosis of thyroid dysfunction appears quite simple, this is often not the case. Issues that make it unclear whether thyroid function is normal include transient changes in thyroid parameters, inter-individual and intra-individual differences in thyroid parameters, age-related differences, and ethnic variations. In addition, a statistically calculated distribution of thyroid analytes does not necessarily coincide with intervals or cutoffs that have predictive value for beneficial or adverse health outcomes. Based on current clincial trial data, it is unclear which individuals with mild thyroid-stimulating hormone elevations will benefit from levothyroxine treatment. For example, only a small number of patients with thyroid-stimulating hormone values of more than 10 mIU/L have been studied in a randomised manner. Even if therapy is initiated for abnormal thyroid function, not all treated individuals are maintained at the desired treatment target, and therefore might still be at risk. The consequence of this is that each patient's thyroid function needs to be assessed on an individual basis with the entire clinical picture in mind. Monitoring also needs to be vigilant, and the targets for treatment reassessed continually.