Neonatal plasma TSH – estimated upper reference intervals for diagnosis and follow up of congenital hypothyroidism

Abstract

Age- and method-dependent plasma TSH reference intervals are essential for the diagnosis and management of congenital hypothyroidism. However, accurate reference intervals for plasma TSH have not been adequately defined due to the difficulties in obtaining samples from a healthy paediatric population. To overcome the difficulties in generating such intervals we estimated method-dependent plasma TSH upper-reference intervals by determining the blood spot TSH upper-reference interval from newborn blood spot TSH screening data (N = 10,697) and then derived method-dependent conversion factors for blood spot TSH to plasma TSH concentration from paired-blood spot and plasma TSH measurements. The upper reference interval for blood spot TSH of 3.04 mU/L was obtained from the 97.5th centile of the selected data. Using experimentally-derived conversion factors, estimates of plasma TSH upper reference intervals of 7.6, 6.3, 7.3, 8.3 and 6.5 mU/L were obtained for the Siemens Centaur, Abbott Architect, Roche Elecsys E170, Siemens Immulite 2000 and Beckman access HYPERsensitive TSH assays respectively. These estimated method-dependent plasma TSH upper reference intervals will be of great practical use to clinicians to diagnose and to follow up infants found to have increased blood spot TSH concentrations identified by Newborn Screening programmes.

Adult and neonate TSH blood spot reference intervals

“…The first study by Hofman et al 2003 found DBS TSH to be approximately 40% lower than serum TSH when measured using the Abbott AxSYM instrument. 20 The study by Evans et al 2011 showed that DBS TSH is approximately 50% lower than plasma TSH measured using the same Abbott Architect assay we used in our study. 21 This difference in results may be explained by a number of factors and highlights the challenges when trying to establish a DBS reference interval for TSH.…”

Intervalos de referencia para tirotropina sérica durante los primeros 14 días de vida

“…El estadístico usado por nosotros tiene menor probabilidad de error pues se calculó sobre una población con distribución normal (posconversión logarítmica). 26 La metodología estadística de la transformación logarítmica aplicada por nosotros no es novedad, en la bibliografía sobre intervalos de referencia es defendida y recomendada por distintos autores. Kawahara y cols.…”

“…Los límites de decisión separan una población, los individuos por fuera de ellos (arriba o debajo) son tratados en forma diferente y estadísticamente las curvas de respuesta relativa [curvas ROC (Receiver-operating characteristic)] y los valores predictivos son las técnicas más utilizadas. 26 Si este concepto lo aplicamos en la interpretación de la pesquisa neonatal se podría decir que los intervalos de referencia de TSH serían de gran ayuda para una evaluación apropiada de los niños en riesgo de disfunción tiroidea; es decir, aquellos que fueron positivos para el valor de corte (límite de decisión) de pesquisa neonatal. En un esfuerzo para incrementar la especificidad del programa de pesquisa, nuestro objetivo fue definir intervalos de referencia para TSH sérica en RN de entre 2 y 14 días de vida.…”

Newborn Screening for Congenital Hypothyroidism

“…In recent years, more sensitive and automated methods (chemiluminescence, fluoroimmunoassay, etc.) for determining both TSH and T4 in dried blood spots have been introduced (16,17,18,19,20,21). These new methods have increased sensitivity and specificity in the detection of CH.…”

“…However, despite the development of more accurate test programs, approximately 5% of CH cases may still be missed in any screening program. The reasons could be failure of sample collection, unsatisfactory samples, misinterpretation of samples and unsatisfactory recalls, the condition being subclinical or, as is true for programs which measure only TSH, failure to detect infants with central CH (14,15,16,17). …”