Both T₄ and TSH can be measured accurately in blood spotted on filter paper, the standard method for sample collection from newborn infants for screening tests. T₄ refers to total T₄, because methods to measure free T₄ on blood spots are not widely available. In the early days of screening, the greater accuracy of T₄ measurements around the cut-off values relative to that of TSH led to T₄ being initially measured by a number of programs, with TSH being measured only on the samples with the lowest T₄. This T₄-first approach misses some infants who have functional disorders or ectopic thyroid tissue and who, under TSH stimulation, produce enough T₄ so that the blood T₄ concentration is above the cut-off value (9). The primary T₄ approach can theoretically detect both central (secondary/tertiary) and primary hypothyroidism (10). However, the former is at least 10-fold less frequent than the latter and can be diagnosed on the basis of the clinical signs and symptoms of associated pituitary hormone deficiencies; furthermore, central hypothyroidism is generally milder than primary hypothyroidism and a T₄ below the cut-off was only found in 19% of patients in one study (11). The primary TSH approach has the additional advantage that trends in neonatal TSH distribution can be used to assess changes in iodine status at the population level (see chapter 11D) and has now become the most widely used. Some programs measure T₄ as a second step to guide decision-making when the blood-spot TSH value is only slightly elevated; this is the case of the Québec screening program, whose algorithm is shown on Fig. 54A.1.

It is important to remember that when a disease is subjected to population-based biochemical screening, its prevalence usually increases, sometimes markedly. For congenital hypothyroidism, the birth prevalence has almost tripled, from about 1 in 6,700 newborns by clinical diagnosis to about 1 in 2,500 in the screening era. Further increases in prevalence likely reflect changes in screening algorithms rather than an increase in the true prevalence of the disease (12).

While most screening programs measure T₄ or TSH in blood eluted from filter paper and collected a few days after birth, a few use umbilical cord serum (9). One limitation of cord-serum screening is that the likelihood of a falsely normal serum TSH value in the hypothyroid twin of a discordant pair, due to fetal blood mixing, is even higher than on blood spots obtained postnatally (13).

As noted above, screening for congenital hypothyroidism started after screening for phenylketonuria, and this led to the use of the same filter-paper blood spots, which were usually obtained several days after birth. As postnatal hospital stays have shortened, blood sampling has been done sooner after birth. TSH secretion rises abruptly immediately after delivery (see Chapter 53), and, although the surge usually lasts only hours, it results in blood-spot TSH values of 15 mU/L or higher in 2% to 3% of normal infants sampled within 24 hours of birth (9). Therefore, in Québec, infants whose first sample is
obtained within 24 hours after birth are systematically recalled for a second filter paper sample. However, a sample is accepted before 24 hours for infants who die or who are to be transferred to another hospital.