Subclinical Thyroid Dysfunction in the First Trimester of Pregnancy

Disease' Versus Physiological (Pulsatile) Variation in TSH Concentrations

Krzysztof C. Lewandowski; Karolina Garnysz; Wojciech Horzelski; Joanna Kawalec; Karolina Budzen; Mariusz Grzesiak; Andrzej Lewinski


Clin Endocrinol. 2020;93(6):739-745. 

In This Article


The study population included 110 healthy pregnant women, age 30.21 ± 5.7 years, BMI 24.9 ± 5.4 kg/m2, at 9.9 ± 2.6 weeks of gestation, and 19 healthy female controls, age 28.9 ± 10.7 years, BMI 28.3 ± 8.3 kg/m2. There were no significant differences in age and BMI between the pregnant and control group.

Mean TSH concentration (based on data from five measurements [n = 550]) was 1.62 mIU/L ± 1.26 mIU/L in the pregnant group and 1.67 ± 0.84 mIU/L in controls (n = 95), P = ns. Variation of TSH concentrations at different time-points is presented in Table 1. Notably, in the control group, TSH concentrations were significantly higher at 7 AM than at 9 AM (P = .038, Mann-Whitney U test).

Eleven patients in the pregnant group (10%) and a single patient in the control had raised titre of anti-TPO antibodies. Raised titres of anti-thyroglobulin and anti-TSH receptor antibodies were found only in the pregnant group in nine and one patient, respectively. Patients with raised titre of anti-TPO antibodies had higher TSH concentrations (mean TSH 3.17 mIU/L [range 0.01–8.98] vs 1.62 mIU/L [range 0.01–6.6], P = .0108, Mann-Whitney U test).

Analysis of TSH concentrations above 2.5 and 4.0 mIU/L is presented in Table 2, while respective lower cut-off points, that is 0.27 mIU/L—nonpregnant assay reference range—and 0.1 mIU/L—low first-trimester reference range—as suggested by the Endocrine Society[2] and the American Thyroid Association 2011 guidelines,[1] are presented in Table 3. The number of women with free T4 or free T3 concentrations below or above assay reference range is presented in Table 4. TSH concentrations were assessed at 30-minute intervals between 7 AM and 9 AM (ie five consecutive measurements), while free T4 and free T3 concentrations were assessed at 7 AM only. Notably, the percentage of pregnant women with TSH >2.5 mIU/L varied between 20.9% and 12.7%, and between 9.1% and 3.6% for TSH above 4.0 mIU/L. This discrepancy was even more pronounced for women with normal titre of aTPO antibodies, that is 21.2% vs 8.1% for TSH >2.5 mIU/L and 6.06% vs 1.01% for TSH above 4.0 mIU/L. This implies that for aTPO-negative women, as much as 13.1% would be misclassified as 'healthy' or 'subclinically hypothyroid' for TSH cut-off of 2.5 mIU/L. Interestingly, the discrepancy was even greater for TSH cut-off point of 4.0 mIU/L (6.06% vs 1.01%).

Analysis of TSH variability based on data from five consecutive measurements (pregnant n = 550, control n = 95) is presented in Table 5. In pregnant women, a maximum TSH increment above the mean was 0.387 mIU/L (23.74%); however, high standard deviation (0.514 mIU/L) reflected large interindividual difference variability of TSH concentrations. We therefore demonstrate examples of four individual cases, where individual TSH concentrations at consecutive time-points fell either below or above the respective cut-off points of 2.5 and 4.0 mIU/L, respectively (Figure 1A-D). If an average decline below the mean (15.81%) was taken into account, then total TSH variation was 39.55% (42.47% in controls). There was no difference in TSH variability between pregnant and control group (P = .139).

Figure 1.

A-D, Patients 1, 2 and 4 demonstrate variation in TSH concentrations of 217%, 177% and 193%, respectively, based on comparison between the lowest and the highest TSH value. As consecutive TSH concentrations fall either below or above a 2.5 or 4.0 mIU/L 'cut-off' point, then this fits the description: 'give me the guideline and I will give you a TSH to fit it'. Patient 2 was found to have genuine subclinical hypothyroidism (she also had high titre of aTPO antibodies). Note, however, a fast decline in TSH within 90 min (from 8.98 mIU/L to 4.2 mIU/L, ie by 53.3%), possibly caused by a previous TSH pulse.