Rational approach to the primary evaluation of thyroid disease in paediatrics. Full thyroid profile vs. thyroid-stimulating hormone and free thyroxine only

Abstract

Healthcare budgets worldwide are facing increasing pressureto reduce costs and improve efficiency, while maintainingquality. Laboratory testing has not escaped thispressure.The clinical laboratory control of demand is oftenconfused with demand management. Control of demandrefers to the reduction of costs, while demand managementfocusses on ensuring appropriate requesting. Hence,the latter has an inbuilt quality aspect and may result inincreased as well as decreased testing (i.e. to reduce overordering,underordering and misordering of tests) [1, 2].It has been widely accepted that the major changes inthyroid function in adult subjects may initially be studiedby assessing of thyroid-stimulating hormone (TSH) [3].Free thyroxine (fT4) should be measured in the setting ofan abnormal TSH, and free triiodothyronine (fT3) only inspecific circumstances, such as cases of suspected hyperthyroidismwith a normal fT4 and suppressed TSH [4].The recommendations by Wisely et al. [5] from theAmerican Society for Clinical Pathology (ASCP) adviseagainst ordering multiple tests in the initial investigationof patients with suspected non-neoplastic thyroid disease.In the paediatric population the impact of thyroidhormone deficiency on neuro development and growth,warrant evaluation of an aetiology of central origin. Herewe propose the use of serum TSH and fT4 to evaluateprimary and central thyroid dysfunction in this population[6, 7]. However, to our knowledge, currently noevidence exists on whether serum TSH and fT4 providesufficient information for the screening of thyroid functionin children.The aim of the study was to retrospectively assessserum full thyroid profile (TSH, fT4, T4 and T3) vs. TSHand fT4 only for the primary evaluation of thyroid diseasein children describing the results obtained and subsequentmedical management.We retrospectively analysed consecutive patients(n = 5999) in whom thyroid function was studied with afull thyroid test seen between November 2014 and September2015. Tests were performed on the day of sampling. Weexcluded those subjects who were previously diagnosedwith thyroid disease, those that were not seen by anyphysician after the studies, those who had been treatedwith levothyroxine (LT), methimazole, or antiepilepticdrugs, and severely ill patients. The remaining 5738 subjectsincluded were categorised into 12 groups accordingto age from 1 week to 12 years. TSH, T3 and fT4 weremeasured with Architect i4000 (Abbott) and we used ourown reference interval (RI) [8] for total T4 IMMULITE 2000(Siemens) and was used the RI by Elmlinger et al. [9]. Allwith percentiles between 2.5 and 97.5.Subjects with serum TSH and fT4 within the (RI)according to age were selected (n = 4007, 69.8%).From this group, children who had T3 and/or T4outside the RI were selected (n = 390, 6.79%) (Figure 1)(Supplementary Material 1 and 2). In the clinical records (CR) of 381 patients (6.63%) nocomment was made on thyroid function or the thyroidprofile was reported to be normal, in spite of T3 and/or T4outside the RI.In nine patients (0.15%) a comment was made inthe CR; five patients were sent home without further interventions, in one patient with high T3 and T4, highserum levels of thyroxine-binding globulin were found,and the three remaining were finally treated with LT(0.05% from de selected subjects). All of them had fT4below the 10th percentile of the RI and one additionallyhad TSH above the 90th percentile of the RI.Of the study population n = 5738, 69.8% presentedwith normal serum TSH and fT4 levels. Overall patients,6.79% had T3 and/or T4 outside the RI, which is statisticallyacceptable for a reference population. Accordingto the definition of RI using the central 95% values fromreference population in both analytes, 5% of healthy subjectsis statistically expected to be found outside the RI, ifwe used two analytes is possible to found almost 9.75% ofhealthy subjects outside the RI ([1?(0.95)2] *100 = 9.75%).In the CR of 381 patients (6.63%) no comment wasmade, this may have happened because in all cases T3and/or T4 were within the reference change value (RCV)on the limits of the RI (Supplementary Material 2).As to the nine patients that had some medicalcomment in their CR regarding thyroid test results, onlythree were put on treatment, and all had fT4 below the10th percentile and one had TSH above the 90th percentileas well. Moreover, the three patients who duringfollow-up showed a drop in serum thyroid hormones,and showed clinical signs of hypothyroidism and werestarted on supplementary treatment with LT had medulloblastoma,neurofibromatosis and Prader-Willi syndrome,diseases that may be associated with centralhypothyroidism. In adults the use of only serum TSH by primary careproviders, may be sufficient for screening of thyroiddisease. However, in children functional disorders of thethyroid, such as hypothyroidism and hyperthyroidism dueto chronic lymphocytic thyroiditis and Graves? Basedowdisease, are frequent. The combination of TSH and fT4allows detection of these two disorders as well as the centralabnormalities of the hypothalamus pituitary thyroid axis;the findings in our series support this concept.When considering cost-effectiveness, costs could bereduced by 46% in each patient studied, with a positiveimpact on the management of demand without effectingclinical outcome.In conclusion, we found that the use of TSH and fT4is useful for primary evaluation of thyroid disease in paediatrics.However, TSH and fT4 should be assessed with astricter RI between the 10th and 90th percentile. In caseof persistent clinical signs suggestive of a disorder of thethyroid gland in a child, assessment of a complete thyroidprofile, including antithyroid antibodies, total or free T3,and total T4, is recommended.This study suggests that questionable testing patternscontribute to inappropriate thyroid test utilisation. A prospectivestudy in a paediatric population would be necessaryto clinically confirm these results and subsequentlydefine medical management.