Thyrotropin reference values in the first trimester of pregnancy

Sir,

A recent consensus document¹ and the clinical guidelines of the Spanish Society of Endocrinology and Nutrition² recommend that each center should establish its own normal reference value for each trimester of pregnancy in their population using adequate laboratory techniques, but as stated by the American Thyroid Association (ATA),³ in the absence of such values, it is recommended that a value of 2.5 mcU/mL be used as the cut-off point for thyroid releasing hormone (TSH) in the first trimester of pregnancy.

In Spain, normal TSH levels adequate for the first trimester of pregnancy have only been reported in four populations: Aragon, with reference values ranging from 0.41 to 2.63 mcU/L⁴; Catalonia, 0.12–4.75 mcU/mL⁵; Cartagena, 0.13–3.71 mcU/mL⁶; and Jaén, 0.23–4.18 mcU/mL.⁷ A simple calculation of the non-weighted mean of these populations would give an upper limit close to 4 mcU/mL, very far from that proposed by the ATA (Table 1).

Table 1.

Thyroid releasing hormone levels in the first trimester of pregnancy in women with negative anti-thyroid peroxidase antibodies.

	Median	2.5th p	97.5th p	Gestational weeks	Method	Laboratory	No.	Normal values
Bocos-Terraz et al.^4	1.0	0.41	2.63	<14	IMA	Abbot Architect	481	<4.94
Vila et al.^5	1.36	0.12	4.75	9	IMA	Advia-Bayer	178	0.4–4.0
García de Guadiana Romualdo et al.^6	1.44	0.13	3.71	11–13	IMA	Roche	400	0.27–4.2
Santiago et al.^7	1.52	0.23 (p3)	4.18 (p97)	7–10	IMA	Beckman	305	0.26–5.6
Oviedo (current)	1.76	0.17	4.15	6–12	IMA	Roche	264	0.45–5

IMA: Immunometric assay; 2.5th p: 2.5th percentile; 97.5th p: 97.5th percentile.

TSH data in mcIU/mL.

At our center, plasma TSH and anti-thyroid peroxidase antibody (anti-TPO Ab) levels were measured in 309 women in the first trimester of pregnancy (6–12 weeks) using a chemiluminescence immunoassay from Roche Diagnostics. Thirty-nine women (11.7%) had positive anti-TPO Ab (>35 U/mL). In this group, the 2.5th and 97.5th percentiles were 0.45 and 7.89 mcIU/mL respectively. Excluding these women with positive anti-TPO and those with TSH levels higher than 5 mcIU/mL, the 2.5th and 97.5th percentiles of TSH in the remaining 264 women were 0.17 and 4.15 mcIU/mL respectively, which are considered the reference levels in our population using this test method.

Using these normal reference values for thyroid function screening performed at our health area during 2010 and 2011 on 4461 women in their first trimester of pregnancy, the prevalence of hypothyroidism with elevated TSH (>4.17 mcIU/mL) was 7.2%, higher than that found in other populations.⁸ If we had not had our own reference values and had used instead the level recommended by the ATA of 2.5 mcIU/mL, the result would have been a prevalence of 28.2%, a very high rate, and the number of false diagnoses would therefore have been totally unacceptable.

Based on the foregoing, and before any universal or selective screening program is considered, we believe that each center should have its own reference values appropriate for its method and population as an indispensable precondition for adequate thyroid function assessment in pregnancy. In Spain, the use of the cut-off values recommended by scientific societies from other countries could lead to overdiagnosis, with significant healthcare and financial implications.