Introduction
Maternal thyroid dysfunction has been associated with a variety of adverse fetal and
maternal effects, including increased risk of pre-term birth, placental abruption,
fetal demise, future impaired neurological development, and maternal post-partum
thyroid disease [1 ]
[2 ]. Thyroid hormones regulate gene expression
during fetal developmental stages and appropriate levels of these hormones appear to
be critical for fetal brain development through control of gene transcription in
neural cells [3 ]. Maternal laboratory
measurement of thyroid function plays an important role in both low and high-risk
pregnancy follow up [4 ] and is critical in
order to achieve optimum fetal and neonatal outcomes. However, reference ranges for
thyroid function in pregnant women differ significantly from those of non-pregnant
women [4 ] due to the physiological changes
mediated primarily by human chorionic gonadotropin (hCG) [5 ]. Therefore, utilization of non-pregnant
reference intervals to interpret thyroid function in pregnant women may potentially
cause suboptimal patient care [4 ]
[6 ] through misclassification of normal results
as abnormal and vice versa.
The interpretation of gestational thyroid function tests depends on the stage of
pregnancy [7 ]. Reference intervals for thyroid
hormones in pregnancy are mostly trimester-specific and predominantly refer to
thyroid stimulating hormone (TSH) alone [6 ]
[8 ]. The American Thyroid
Association (ATA) and the United States National Academy of Clinical Biochemistry
(NACB) both recommend using population-based, trimester-specific reference ranges
for TSH and serum free Thyroxine (FT4) [6 ]
[8 ]. TSH is generally considered
the primary test for evaluating thyroid status during pregnancy. FT4, which also
changes with gestational age advancement, particularly between the first and second
trimester [9 ]
[10 ] is considered a complementary test as well as free triiodothyronine
(FT3), which is less commonly tested. In this respect it is noteworthy that
triiodothyronine (T3) has been shown to regulate gene transcription in neural cells
during fetal developmental stages and appropriate levels of this hormone appear to
be critical for fetal brain development [3 ].
When population-based references are unavailable, the ATA recommends using a fixed
upper threshold of 2.5 mIU/l for TSH during pregnancy for thyroid
peroxidase antibodies (TPOAb) positive women. It is not clear what the
recommendations are for the 2.5–4 mIU/l range of TSH levels.
For the range of TSH between 4 and 10 mIU/l for TPOAb negative women
no clear recommendation is given ("may be considered…"). The
ATA recommends "consideration" (as opposed to "maybe
considered") of levothyroxine therapy for thyroid peroxidase antibodies
(TPOAb) negative women when their TSH level is greater than
10.0 mIU/l [8 ]. Clearly, this
prestigious group cannot make clear recommendations because the evidence base is
insufficient. Similarly, in the UpToDate physicians’ reference, guidelines
for screening and therapy of hypothyroidism in pregnancy are based on expert
opinions with different recommendations by different experts in the same algorithm!
[11 ].
Furthermore, the different TSH cutoffs, used for deciding whether to repeat thyroid
function measurements or to initiate thyroid hormone treatment, are based on the
assumption that between the low and high TSH cutoffs the levels of FT4 and FT3
levels are "normal". However, there are insufficient data regarding
the normal range of thyroid hormones during pregnancy and their relation to the TSH
cut-off levels.
Our objective was to analyze FT4 and FT3 levels in relation to recommended pregnancy
TSH cut-off levels in order to evaluate the relevance of TSH cut-off levels to
thyroid function in pregnancy. We did not intend to evaluate whether subclinical
hypothyroidism is detrimental, and this issue may require further study.
Subjects and Methods
We performed a retrospective study utilizing the Clalit health service, Jerusalem
district, database between January 2006 and February 2017. Data were collected from
electronic medical files of all pregnant women tested by request of physicians in
community clinics for TSH (mIU/L) and in some patients, FT4 (pmol/l)
and FT3 (pmol/L) as well. If multiple tests were done during pregnancy only
the first test was used for calculating gestation-duration appropriate levels (see
results in [Table 1 ] and [Fig. 1 ]).
Fig. 1 Flow chart for application of exclusion criteria.
Table 1 Population based thyroid function tests in pregnant
women by trimester of pregnancy.
TSH (mIU/l)
FT4 (pmol/l)
FT3 (pmol/l)
1st Trimester
n
19374
3076
2865
Mean±SD (median)
1.93±1.72 (1.61)
14.53±2.39 (14.40)
4.92±0.78 (4.80)
2.5 percentile
0.19
10.40
3.70
97.5 percentile
5.42
19.90
6.90
2nd trimester
n
7281
1716
1257
Mean±SD (median)
2.21±1.77 (1.86)
12.99±1.9 (12.89)
4.21±0.61 (4.17)
2.5 percentile
0.39
9.63
3.21
97.5 percentile
5.95
16.91
5.78
3rd trimester
n
2567
849
620
Mean±SD (median)
2.38±1.69 (2.09)
12.49±2.05 (12.30)
3.98±0.64 (3.90)
2.5 percentile
0.41
9.14
3.08
97.5 percentile
5.84
16.88
5.35
TSH: Thyrotropin; FT3: Free triiodothyronine; FT4: Free thyroxine.
Inclusion criteria were women with a TSH measurement during pregnancy and a
consequent live birth. Exclusion criteria were gestational trophoblastic disease,
diabetes mellitus, chronic hypertension or any chronic medical therapy for at least
12 months prior to the current pregnancy. Women receiving anti-epileptic drugs or
lithium prior to or during pregnancy were also excluded. Similarly, women with a
known diagnosis of hyperthyroidism, hypothyroidism, positive titers of anti-thyroid
peroxidase or anti-thyroglobulin and women who had been or were currently treated
with methimazole, propylthiouracil, levothyroxine, recombinant thyrotropin, or
glucocorticoids were excluded from the study.
Gestational age was calculated using the date of last menstrual period in combination
with ultrasound measurements of crown-rump length (CRL) in the first trimester, as
CRL is considered to be the most reliable sonographic parameter for evaluation of
gestational age (especially in the first trimester) [12 ]. For thyroid function tests in the second and third trimesters,
gestational age was established based on last menstrual period and ultrasound dating
performed in the first trimester if available or, otherwise, a later sonographic
dating of gestational age.
TSH, FT3, and FT4 were measured using a commercial kit (Cobas kits used on modular
analytics E-170 analyzer, Roche Diagnostics, Mannheim, Germany). TSH, FT4 and FT3
tests were ordered by either gynecologists or community physician's as they
saw fit, and even when no thyroid abnormality was suspected.
We evaluated FT4 and FT3 mean, median, standard deviation (SD) and 2.5 and 97.5
percentiles in each trimester and compared those of our cohort with widely used
gestational TSH cut-off levels [8 ]
[13 ] as follows: (1) Below
0.1 mIU/l (first trimester), 0.2 (second trimester) and 0.3 (third
trimester) mIU/l; (2) 0.1–2.5 mIU/l (first
trimester), 0.2–3.0 mIU/l (second trimesters) and
0.3–3 mIU/l (third trimester); (3)
2.6–4.0 mIU/l (first trimester) and
3.1–4.0 mIU/l (second and third trimester); (4).
4.1–10.0 mIU/l; (5) TSH above
10.0 mIU/l.
We also calculated mean, median, standard deviation (SD) and 2.5 and 97.5 percentiles
according to bi-weekly gestational age (5 weeks or less, 6–7, 8–9,
10–11, 12–13, 14–15, 16–17, 18–19,
20–21, 22–23, 24–25, 26–27, 28–29,
30–31, 32–33, 34–35, 36–37, 38–39) in our
population. This enabled us to create population specific reference ranges and
assess trends in these parameters throughout gestation. We defined trimesters of
pregnancy as follows: First trimester, i. e., ≤ 13 weeks gestation;
second trimester: 14–27 weeks and the third trimester: beyond 28 weeks of
gestation.
Ethics
All data retrieval and analyses were computerized and anonymous. Clalit Health
Service institutional review board approved the study (IRB ID number
015/2015, received in March 2015).
Statistical analysis
We calculated frequency of every TSH level category through the trimesters of
pregnancy. We presented continuous data as mean ±standard
deviation (SD) and assessed significance of associations between TSH and FT4 and
FT3 with the student’s t-test and Pearson correlation coefficient
("R"). All statistical tests were two-tailed and a p-value
of<0.05 indicated a statistically significant difference. All analyses
were conducted using statistical package for the social sciences (SPSS) (for
Windows software, version 23; IBM Corp) and graphs were prepared with
"excel" 2013 software.
Results
Study population
Overall, 32 430 TSH measurements from pregnant women whose pregnancy resulted
with a live birth were evaluated. We excluded women for evidence of being
unhealthy and outliers with extreme thyroid hormone results (see details in
[Fig. 1 ]). FT4 levels were available
for 5641 of these women and FT3 measurements were available, as well as FT3, for
4742 women.
In the entire study population, median maternal age was 29.44 years (range
16–49). In the first trimester, 19 374 women were tested, with a mean
maternal age for women tested during this trimester of 29.6 years and median
gestational age of 8 weeks. In the second trimester, 7281 women were tested. The
mean maternal age and median gestational age for these women were 29.62 years
and 18 weeks respectively. For the third trimester, 2567 women were tested.
Their mean maternal age and median gestational age were 30.04 years and 32
weeks, respectively. The mean, median standard deviation and 2.5th and 97.5th
percentiles of TSH, FT4 and FT3 for each trimester and are shown in [Table 1 ]
TSH, FT4, and FT3 levels throughout pregnancy
The progress of TSH, FT4, and FT3 levels through pregnancy is presented in [Fig. 2 ] (detailed data in [Table 2 ]). Note that the mean TSH level
measured before or at 5 weeks was 2.66 mIU/l with a decline to a
nadir of 1.59 mIU/l at 8–9 weeks and a subsequent rise
until 20 weeks of gestation increasing to a maximum of
2.39 mIU/l. After this stage, TSH remained steady with minor
fluctuations until term. FT4 values reached a maximum mean value of
14.95 pmol/l at 8–9 weeks and then decreased gradually
to 12.66 pmol/l at term. FT3 values reached a maximum value of
5.10 pmol/l at weeks 8–9 and then remained overall
steady with only a minor decrease until term.
Fig. 2 Thyroid stimulating hormone (TSH), free thyroxine (FT4),
and triiodothyronine (FT3) levels throughout gestation.
Table 2 TSH and thyroid hormones levels throughout
gestation.
n
TSH (mIU/l)
n
FT4 (pmol/l)
n
FT3 (pmol/l)
Gestational week
1st Trimester
19374
1.93 ±1.72 (1.61)
3076
14.53 ±2.39 (14.40)
2865
4.92 ±0.78 (4.80)
≤5 weeks
3856
2.66±2.09 (2.20)
1188
14.34±2.11 (14.30)
803
4.83±0.69 (4.80)
6–7 weeks
5310
1.97±1.70 (1.66)
699
14.56±2.40 (14.31)
571
4.94±0.81 (4.90)
8–9 weeks
4901
1.59±1.67 (1.28)
616
14.95±2.80 (14.81)
772
5.10±0.83 (5.00)
10–11 weeks
3313
1.62±1.27 (1.38)
362
14.79±2.44 (14.50)
487
4.86±0.76 (4.77)
12–13 weeks
1994
1.78±1.24 (1.55)
211
13.81±2.18 (13.61)
232
4.76±0.81 (4.60)
2nd trimester
7281
2.21 ±1.77 (1.86)
1716
12.99 ±1.90 (12.89)
1257
4.21 ±0.61 (4.17)
14–15 weeks
1797
2.13±1.77 (1.76)
457
13.65±2.03 (13.58)
373
4.42±0.67 (4.35)
16–17 weeks
1547
2.26±1.80 (1.90)
418
13.07±1.79 (12.95)
303
4.25±0.65 (4.19)
18–19 weeks
995
2.39±1.68 (2.02)
283
12.92±1.73 (12.89)
197
4.15±0.53 (4.10)
20–21 weeks
658
2.31±2.05 (1.91)
158
12.47±1.84 (12.27)
107
4.09±0.59 (4.08)
22–23 weeks
537
2.39±2.05 (1.99)
125
12.76±1.75 (12.60)
85
4.11±0.44 (4.20)
24–25 weeks
1106
2.06±1.68 (1.76)
166
12.27±1.64 (12.11)
125
3.94±0.44 (3.91)
26–27 weeks
641
2.08±1.27 (1.77)
109
12.25±1.78 (12.10)
67
3.93±0.40 (4.00)
3rd trimester
2567
2.38 ±1.69 (2.09)
849
12.49 ±2.05 (12.30)
620
3.98 ±0.64 (3.90)
28–29 weeks
676
2.31±1.94 (1.90)
243
12.48±2.06 (12.30)
183
4.01±0.69 (3.90)
30–31 weeks
607
2.34±1.72 (2.07)
218
12.59±2.24 (12.40)
163
3.99±0.65 (3.90)
32–33 weeks
486
2.39±1.44 (2.18)
139
12.50±1.86 (12.30)
101
4.02±0.71 (3.90)
34–35 weeks
435
2.48±1.70 (2.15)
141
12.38±1.80 (12.26)
95
3.94±0.47 (3.85)
36–37 weeks
262
2.41±1.26 (2.25)
81
12.49±2.13 (12.36)
61
3.92±0.57 (3.90)
38–39 weeks
91
2.42±1.16 (2.18)
23
12.66±2.22 (12.01)
16
3.78±0.55 (3.67)
≥40 weeks
10
3.85±4.11 (2.15)
4
10.62±2.87 (10.81)
1
3.27±0 (3.27)
Data presented as mean ± SD (median).
Thyroid hormones levels according to TSH cut-off levels
In [Table 3 ], presenting data of patients
with a TSH measurement plus a FT4 or FT3 test, the percentage of patients
with FT4 and FT3 below the lower normal limits (–2 standard deviations
for the study population, taken from the study data) are presented for different
TSH recommended cut-off levels in the three gestational trimesters. As the TSH
levels rise, the percentage of patients with FT4 below the lower normal limit
increases. However, even for TSH levels above 10 mIU/l, FT4
levels are below the normal limit in only 12.8% in the first trimester.
This trend was also shown for FT3, in which the percentage of measurements below
the normal limit was higher for TSH levels above 10 mIU/l in
both the first and second trimester (3.8% and 6.6%,
respectively).
Table 3 Free thyroxine (FT4) and free triiodothyronine
(FT3) levels across the different thyrotropin (TSH) cutoffs
according to trimesters of pregnancy.
1st trimester
FT4
TSH<=0.1
0.1<TSH<=2.5
2.5<TSH<=4
4<TSH<=10
10<TSH
n
121
1371
426
992
101
Mean1 ±SD
17.38±3.06
14.74±2.01
14.25±1.9
13.99±2.06
12.37±2.5
Median
17.21
14.50
14.18
13.99
12.30
2.5–97.5%
12.44–25.44
11.3–19.52
10.44–18.05
9.99–18.00
7.62–18.17
% below LNL*
0
0.07
1.43
2.0
12.8
FT3
TSH<=0.1
0.1<TSH<=2.5
2.5<TSH<=4
4<TSH<=10
10<TSH
n
210
1768
322
315
47
Mean2 ±SD
5.46±0.79
4.86±0.58
4.67±0.57
4.51±0.65
4.23±0.94
Median
5.40
4.80
4.70
4.50
4.10
2.5–97.5%
4.06–7.00
3.9–6.12
3.61–5.89
3.39–5.94
2.83–7.48
% below LNL*
0
0.16
0.6
1.6
15.3
2nd trimester
FT4
TSH<=0.2
0.1<TSH<=3.0
3.0<TSH<=4
4<TSH<=10
TSH>10
n
60
852
188
511
46
Mean3 ±SD
13.25±2.07
12.83±1.68
13.15±1.87
12.99±1.86
12.91±2.30
Median
12.95
12.72
12.99
12.90
13.21
2.5–97.5%
9.46–18.23
9.69–16.38
9.51–17.53
9.51–16.81
7.18–16.97
% below LNL*
1.5
0.7
1.4
0.8
3.8
FT3
TSH<=0.2
0.1<TSH<=3.0
3.0<TSH<=4
4<TSH<=10
TSH>10
n
78
718
157
198
30
Mean4 ±SD
4.44±0.58
4.17±0.47
4.11±0.54
4.04±0.54
3.75±0.91
Median
4.40
4.12
4.05
4.00
3.58
2.5–97.5%
3.19–5.95
3.3–5.2
3.23–5.81
3.1–5.29
2.31–6.80+
% below LNL*
1.2
0.7
1.6
0.4
14.7
3rd trimester
FT4
TSH<=0.3
0.3<TSH<=3.0
3.0<TSH<=4
4<TSH<=10
TSH>10
n
31
462
108
212
15
Mean5 ±SD
13.28±2.13
12.41±1.74
12.72±2.07
12.13±1.87
11.79±2.65
Median
12.79
12.29
12.49
11.92
11.34
2.5–97.5%
10.18–17.40+
9.24–16.02
9.4–17.95
8.74–15.89
6.95–18.05+
% below LNL*
0
0
1.6
1.2
6.6
FT3
TSH<=0.3
0.3<TSH<=3.0
3.0<TSH<=4
4<TSH<=10
TSH>10
n
38
379
86
83
10
Mean⁶±SD
4.33±0.39
3.92±0.47
3.83±0.38
3.86±0.45
3.51±0.66
Median
4.28
3.90
3.74
3.80
3.40
2.5–97.5%
3.58–5.20+
3.01–4.99
3.21–4.79
3.12–4.97
2.8–4.8+
% below LNL*
0
0.21
0
0
0
1 p<0.05 in comparison between each consecutive TSH
group except between 2.5<TSH<=4 and
4<TSH<=10, which was non-significant.
2 p<0.05 in comparison between each consecutive
TSH group all the TSH groups except between the fourth and fifth
group.3 T here is no difference between the
groups.4 p<0.05 in comparison between each
consecutive TSH group except the comparison between the third and the
adjacent groups.5 p<0.05 in comparison only between
the first and fourth group. ⁶ p<0.05 in comparison to the first
group (the last group is too small for comparsion)
+ Highest number (due to low sample number).
* LNL: Lower normal limit, this figure was
derived after including hypothyroid women (before treatment) who were
previously excluded from the cohort in order to derive normal
values.
As expected, a negative correlation (R=Pearson's correlation
coefficient) between TSH and both FT4 and FT3 was shown in the first trimester
(R=–0.31 and –0.25, p<0.001 for both). However,
this was true only for FT3 in both the second (R=–0.02, NS and
–0.18, p<0.001), and third trimester (R=–0.05,
NS and 0.20, p<0.001).
Discussion
The major finding in this study was that TSH does not appear to be a good index of
thyroid function during pregnancy. In fact, for the majority of women who would have
been considered hypothyroid by TSH values, the diagnosis would have been subclinical
hypothyroidism (because they had normal levels of free thyroid hormone levels, from
same-population based data), which is a debatable issue with differing views
regarding its significance [14 ]
[15 ] and this study can determine the final
verdict. The correlations between TSH and FT4/FT3 are rather low at first
trimester: R=–0.25 to –0.31, which means a R2
between 6 and 9%. During the second/third trimester there are no
correlations at all. These R’s show a very low effect size, meaning they are
clinically not relevant. It is unlikely that this was due solely to an effect of HCG
[5 ] because the correlations between TSH
and FT4/FT3 seem to disappear with increasing term when the hCG
concentrations largely decrease.
In this study, our large cohort enabled us to establish population specific reference
ranges for TSH, FT4 and FT3 and study the relationship between TSH and the thyroid
hormones FT3 and FT4. We included a population of several thousand healthy women
whose pregnancy ended in delivery of a live child, who did not have preexisting
treated thyroid disease and who were not taking chronic medication. Our healthy
population-based data are in agreement with previous studies [4 ]
[16 ]
[17 ]
[18 ] showing similar patterns of changes in TSH,
FT4 and FT3 levels through gestation. TSH decreased dramatically until 8–9
weeks and later (12 weeks) began increasing gradually and stabilized, the FT4
pattern reflected a quite similar though less drastic "mirror image"
pattern. FT3 levels remained rather steady with a gradual minor decrease throughout
gestation.
Thyroid physiology changes significantly during gestation and particularly in the
first trimester, so the interpretation of thyroid function tests are challenging
[8 ]
[13 ]
[19 ]. Since the first trimester
is a critical period for fetal neurological development [3 ]
[7 ],
maternal thyroid function and its pathologies need to be evaluated and addressed
early and accurately. For this purpose, understanding the relationship of TSH to the
FT4 and FT3 levels is important. As expected, our findings indicate lower mean
levels of FT4 and FT3 with higher levels of TSH.
When we applied recommended TSH cut-off levels – i. e.,
2.5 mIU/l or 4.0 mIU/l [8 ], we surprisingly found that with TSH above
each of these cut-off levels, less than 10% of samples were associated with
low FT3 or low FT4. In fact, even at TSH above 10 mIU/l only
15.3% of samples were associated with low FT3 and 12.8% were
associated with low FT4 in the first trimester. The strengths of this study are the
source, i. e., the fact that women are sampled for TSH, FT4 and FT3 without
prior selection, thus allowing for a large number of healthy women for whom all
three parameters were available. Secondly, by producing our own population-based
normal thyroid hormone data we were able to calculate the relationship between TSH
and normal thyroid hormone levels in the same population.
Apart from its retrospective nature, this study has several limitations: presence of
FT4 and FT3 levels in some but not all of the women, lack of concomitant hCG levels
and exclusion of women whose pregnancy ended in spontaneous miscarriage.
Conclusion
Though our results do not eliminate TSH from acting as part of the thyroid function
tests during pregnancy, based on our findings, it is imperative to test FT4 and FT3
levels and not to make decisions based on TSH levels alone when diagnosing
hypothyroidism during pregnancy.
Data Availability
Data may be provided upon request to the corresponding author.
Author Contributions
GK, LDS and DS designed the study, collected the data, analyzed the data, and drafted
the manuscript. CM analyzed the data, drafted the manuscript and reviewed it. DG
contributed to the study design, data analysis and drafting the manuscript. All
authors approved the final version of the manuscript.
