Introduction
Female infertility accounts for 37% of infertile couples, with 25% of these women
having ovulatory disorders.[1] Women with thyroid dysfunction may experience ovulatory disorders, menstrual irregularities,
infertility and even increased pregnancy morbidity.[2] Elevated concentrations of thyroid stimulating hormone (TSH) in asymptomatic patients
may also increase the likelihood of pregnancy complications,[3]
[4]
[5] neonatal mortality[4] and miscarriage.[6]
[7]
Until recently, human reproduction societies did not recommend the measurement of
TSH concentration in asymptomatic ovulatory women,[8]
[9] but the new guideline published by the American Society for Reproductive Medicine
(ASRM) endorses dosing TSH in infertile women attempting pregnancy.[7] The American Thyroid Association (ATA) recognizes that the prevalence of thyroid
dysfunction is higher in subfertile women,[5] while the Endocrine Society recommends measuring TSH in patients at “high risk”
of thyroid illness, including asymptomatic infertile ones.[3] Furthermore, there is no consensus among endocrinologists and gynecologists regarding
the cut-off TSH concentrations for patients pursuing pregnancy, whether < 2.5 mIU/L3 or < 4.0/4.5 mIU/L.[7]
[10]
[11]
[12]
[13]
[14]
For patients who will undergo assisted reproduction techniques (ARTs), the high estradiol
levels induced by controlled ovarian stimulation (COS) may alter thyroid function
by increasing the concentrations of TSH, with hypothyroid women being most affected.[15]
[16]
[17] Nevertheless, it is unclear whether the preconceptional concentration of TSH or
the elevations in TSH concentration during COS are relevant in infertile women undergoing
ARTs.[18] This study, therefore, has compared reproductive outcomes in patients with serum
TSH concentrations of < 2.5 mIU/L, 2.5–4.0 mIU/L, and 4.0–10.0 mIU/L undergoing COS
for in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI).
Methods
Study Design
This retrospective cohort study evaluated all women who underwent IVF/ICSI at the
fertility clinic of the university hospital of the Faculdade de Medicina de Ribeirão
Preto, Universidade de São Paulo, Brazil, between January 2011 and December 2012.
Data were obtained from medical records. The study protocol was approved by the Institutional
Review Board.
Participants
Women who had undergone IVF/ICSI between January 2011 and December 2012 were eligible
if they had serum TSH concentrations evaluated by a third-generation assay reported
in their medical records. Patients in treatment with levothyroxine for hypothyroidism
were excluded from the analysis.
In spite of the fact that the concentrations of TSH ≥ 4.0 mIU/L and < 10 mIU/L in
asymptomatic patients are classically used to define subclinical hypothyroidism,[6]
[10]
[11]
[12]
[13]
[14]
[19] patients with TSH < 4.0 mIU/L were divided into two subgroups (TSH < 2.5 mIU/L and
TSH 2.5–4.0 mIU/L). For comparison, three groups were divided: those with TSH < 2.5
(group 1); TSH > 2.5 mIU/L and > 4.0 mIU/L (group 2); and TSH > 4.0 mIU/L and < 10.0
mIU/L (group 3).
COS, Oocyte Retrieval, Fertilization and Embryo Transfer
Menstruation was programmed with combined oral contraceptives administered in the
previous menstrual cycle. All women were subjected to COS monitored by transvaginal
ultrasound (TVUS)[20] according to one of the following protocols:
In the standard long protocol, gonadotropin-releasing hormone (GnRH) agonists (leuprolide
acetate 0.5 mg/day, Lupron, Famar L'Aigle, Saint-Remy-Sur-Avre, France) were introduced
during the mid-luteal phase of the previous cycle, followed by 150–300 IU/day of gonadotropins
(Gonal-F, Merck Serono, Geneva, Switzerland, or Puregon, Organon, Oss, Holland) for
the first six days. Subsequently, the daily dose of gonadotropins was adjusted according
to follicular growth.
In the flexible antagonist protocol, gonadotropins (150–300 IU/day) were administered
for the first six days, with a subsequent daily dose adjusted according to follicular
growth. Gonadotropin-releasing hormone antagonists 0.25mg/day (Ganirelix, Orgalutran,
Organon, Dublin, Ireland, or Cetrorelix, Cetrotide, Merck Serono, Geneva, Switzerland)
were introduced on the day the average diameter of the largest follicle was ≥ 14 mm.
In the clomiphene citrate (CC) (Clomid, Medley, Campinas, Brazil, or Indux, EMS, Campinas,
Brazil) plus gonadotropins (Menopur, Ferring GmbH, Kiel, Germany) plus antagonist
protocol, offered to women with a low antral follicle count (AFC ≤ 6),[21] CC (100 mg/day) was administered for the first five days. Gonadotropins (150 IU/day)
were administered on days 2 and 4, and daily after day 6. Gonadotropin-releasing hormone
antagonists were introduced when the average diameter of the largest follicle was ≥ 14
mm.
Recombinant (Ovidrel, Merck Serono, Modugno, Italy) or urinary (Choriomon, IBSA Institut
Biochimique, S.A., Lamone, Switzerland) human chorionic gonadotropin (hCG) was administered
when at least two follicles measuring 18 mm in diameter were present. Oocytes were
retrieved 34 to 36 hours after hCG administration, and the luteal phase was supported
by micronized progesterone, 600 mg/day (Utrogestan, Besins Manufacturing Belgium,
Drogenbos, Belgium). Cycles were suspended due to poor response when no follicle reached
a diameter of 10 mm after ten days of COS or after 4 to 5 days of additional treatment
without any follicle measuring at least 18 mm in diameter.[22]
The obtained oocytes were incubated at 37°C in 5% CO2 and 95% humidity for 2–3 hours, and later on denuded by hyaluronidase. Mature oocytes
were subjected to ICSI 3 to 4 hours after oocyte retrieval. Fertilization was evaluated
16 to 18 hours after the ICSI and defined as the presence of two pronuclei and two
polar bodies.
Embryo quality was assessed 43 to 45 hours after the ICSI (day 2 embryos with 4 symmetrical
blastomeres of normal size, < 10% of fragmentation, without multinucleation, were
considered of top quality), or 67 to 69 hours after the ICSI (day 3 embryos with 8
symmetrical blastomeres of normal size, < 10% of fragmentation, without multinucleation,
were considered of top quality).[23]
The rate of clinical pregnancy was defined as the number of patients with embryos
exhibiting a heartbeat on TVUS 4 to 5 weeks after embryo transfer divided by the number
of cycles × 100. The miscarriage rate was defined as the percentage of patients with
clinical pregnancy who could not continue a pregnancy at 20 weeks of gestation. The
live birth rate was defined as the percentage of patients with clinical pregnancy
who delivered a live birth. Multiple pregnancy rates were defined as the percentage
of patients with clinical pregnancy who delivered more than one live birth.
Variables and Data Sources
For most of the assessed parameters, the unit of analysis was a woman who started
COS. However, the unit of analysis was pregnant women for subjects who presented miscarriage,
live birth, and multiple pregnancies. The primary endpoints of this study were clinical
pregnancy, miscarriage, live birth and multiple pregnancy rates. The following parameters
were assessed: age, body mass index (BMI), duration of infertility, indication for
ICSI, baseline AFC, number of retrieved oocytes, and number of mature oocytes (MII).
The characteristics of COS included duration, total dose of recombinant follicle stimulating
hormone (r-FSH), GnRH agonist versus antagonist protocols, and cycles with CC. Other
parameters included number of suspended cycles due to poor response, numbers of cycles
with embryo transfer and top quality embryo transfer, as well as numbers of cycles
with oocyte and embryo cryopreservation, and number of positive pregnancy tests.
Potential Sources of Bias
In order to avoid selection bias, we considered eligible all women starting COS for
IVF/ICSI during the study period who had TSH serum concentrations in their medical
records, and that were not using levothyroxine. For patients undergoing more than
one cycle during the study period, only data from the first cycle was included. All
included women were analyzed.
Statistical Analyses
The number of women undergoing COS for IVF/ICSI during the study period who fulfilled
the eligibility criteria determined the sample size. The normality of distribution
of continuous variables was analyzed by the Kolmogorov-Smirnov test. Continuous variables
with normal distribution were summarized as mean ± standard deviation (SD) and compared
among groups by ANOVA. Continuous variables without normal distribution were summarized
as median (interquartile range) and compared by Kruskal-Wallis tests. Binary data
were presented as ratio and proportion, and compared by Chi-squared (χ2) tests. The level of significance was defined as p < 0.05. Statistical analyzes were performed using the Statistical Package for Social
Sciences software (SPSS, version 18.0, SPSS Inc., Chicago, IL, US).
As an additional analysis, the power of this study to detect differences of 10% in
clinical pregnancy and live birth rates among women with TSH < 2.5 mIU/L and TSH ≥ 2.5
mIU/L was determined.
Results
Participants
During the study period, 787 women underwent IVF/ICSI in our clinic. Sixty patients
were excluded due to the absence of TSH measurements, and 77 were excluded because
they were using levothyroxine for the treatment of hypothyroidism. All 650 women were
followed until a negative pregnancy test or the end of pregnancy.
Descriptive Data
The prevalence of patients presenting elevated concentrations of TSH was of 5.07%
(using a TSH threshold of 4.0 mIU/L) and 29.99% (using a threshold of 2.5 mIU/L).
Of the 650 included patients, 455 (70.0%) had TSH < 2.5 mIU/L, 162 (24.92%) had TSH
concentrations of 2.5–4.0 mIU/L, and 33 (5.07%) had TSH > 4.0 mIU/L and < 10.0 mIU/L
([Table 1]). None of the patients had TSH concentrations > 10.0 mIU/L. The time between TSH
measurement and the cycle was reported in [Table 1]. Indications for ICSI included ovulatory disorders (2.92%); endometriosis (15.69%);
male infertility (31.12%); tubal factor (10.15%); combined factors of subfertility
(29.07%); and unexplained infertility (12.46%) ([Table 1]).
Table 1
Characteristics of the patients included in the study
|
Group 1
TSH < 2.5 mIU/L
|
Group 2
TSH 2.5–4.0 mIU/L
|
Group 3
TSH 4.0–10 mIU/L
|
|
|
(n = 455)
|
(n = 162)
|
(n = 33)
|
p
|
|
TSH (mIU/L)
|
1.5 (1.1–1.9)
|
2.9 (2.7–3.4)
|
4.4 (4.1–4.6)
|
< 0.001
|
|
Interval between TSH dosage and cycle (months)
|
8 (4–18)
|
8 (3–13)
|
6 (1–12)
|
0.23
|
|
Age (years)
|
35 (31–38)
|
35 (32–38)
|
35 (31–39)
|
0.36
|
|
Duration of subfertility (months)
|
60 (38–92)
|
54 (30.3–86)
|
60.5 (36.3–91.8)
|
0.26
|
|
Weight (kg)
|
64 (57–73)
|
64 (57.9–72.3)
|
64.7 (53.3–74.1)
|
0.86
|
|
Height (m)
|
1.63 (1.58–1.67)
|
1.64 (1.59–1.68)
|
1.62 (1.6–1.65)
|
0.46
|
|
BMI (kg/m2)
|
24.2 (21.6–27.4)
|
24 (21.9–27.3)
|
24.2 (21.2–27.3)
|
0.94
|
|
AFC
|
10 (6–18)
|
10 (6 -17)
|
12 (6–18)
|
0.62
|
|
Causes of subfertility
|
n (%)
|
n (%)
|
n (%)
|
0.90
|
|
Ovulatory factor
|
15 (3.3)
|
3 (1.9)
|
1 (3)
|
|
|
Endometriosis
|
71 (15.6)
|
28 (17.3)
|
3 (9.1)
|
|
|
Male Factor
|
136 (29.9)
|
46 (28.4)
|
11 (33.3)
|
|
|
Tubal factor
|
42 (9.2)
|
20 (12.3)
|
4 (12.1)
|
|
|
Combined factors
|
136 (29.9)
|
43 (26.5)
|
10 (30.3)
|
|
|
Unexplained
|
55 (12.1)
|
22 (13.6)
|
4 (12.1)
|
|
Abbreviations: AFC, antral follicle count; BMI, body mass index; TSH, thyroid-stimulating
hormone.
Notes: TSH, age, duration of subfertility, weight, height, BMI and AFC are reported
as median (interquartile range).
p determined by either Kruskal-Wallis or Chi-squared test.
The three subgroups were similar in age, weight, height, BMI, duration of infertility,
indications for ICSI and AFC ([Table 1]). The distribution of COS protocols did not differ significantly among the three
groups ([Table 2]).
Table 2
Characteristics of controlled ovarian stimulation (COS) and parameters related to
ovarian response
|
Group 1
TSH < 2.5 mIU/L
|
Group 2
TSH 2.5–4.0 mIU/L
|
Group 3
TSH 4.0–10 mIU/L
|
|
|
(n = 455)
|
(n = 162)
|
(n = 33)
|
p
|
|
COS protocol
|
|
|
|
0.77
|
|
Antagonist, n (%)
|
292 (64.2)
|
108 (66.7)
|
21 (63.6)
|
|
|
Agonist, n (%)
|
100 (22)
|
32 (19.8)
|
9 (27.3)
|
|
|
CC + Antagonist, n (%)
|
63 (13.8)
|
22 (13.6)
|
3 (9.1)
|
|
|
Total dose of FSH (IU)
|
1650 (1,200–2,325)
|
1,600 (1,213–2,225)
|
1,427 (1,206–1,856)
|
0.08
|
|
Length of COS (days)
|
9 (8–11)
|
9 (8–11)
|
9 (8–10.8)
|
0.83
|
|
Number of retrieved oocytes
|
5 (3–10)
|
6 (2–10)
|
5 (3–9)
|
0.88
|
|
Number of mature oocytes (MII)
|
4 (2–7)
|
4 (2–7.5)
|
3.5 (2–6)
|
0.67
|
|
Suspension due to poor response, n (%)
|
19 (4.2)
|
5 (3.1)
|
2 (6.1)
|
0.69
|
Abbreviations: CC, clomiphene citrate; COS, controlled ovarian stimulation; FSH, follicle-stimulating
hormone; IU, international unities; MII, mature oocytes in metaphase II; TSH, thyroid-stimulating
hormone.
Notes: Total dose of FSH, length of COS, suspension due to poor response, oocytes
retrieved, and mature oocytes reported as median (interquartile range).
p determined either by Kruskal-Wallis or Chi-squared test.
Main Results
Parameters of ovarian response were similar in the three subgroups, including total
dose of r-FSH used during the COS, length of COS, number of retrieved oocytes, and
number of mature oocytes (MII) ([Table 2]). Moreover, there was no significant difference in cycles suspended due to poor
response, cycles with embryo transfer, cycles with top quality embryo transfer, cycles
with oocyte cryopreservation and cycles with embryo cryopreservation ([Table 3]). Similar reproductive outcomes were observed among these three groups, including
clinical pregnancy, live birth, miscarriage, and multiple pregnancy rates ([Table 3]).
Table 3
Reproductive outcomes of assisted reproduction cycles
|
Group 1
TSH < 2.5 mIU/L
|
Group 2
TSH 2.5–4.0 mIU/L
|
Group 3
TSH 4.0–10 mIU/L
|
|
|
(n = 455)
|
(n = 162)
|
(n = 33)
|
p
|
|
Cycles with fresh embryo transfer, n (%)
|
369 (81.1)
|
136 (84)
|
24 (72.7)
|
0.31
|
|
Cycles with fresh top quality embryo transfer, n (%)
|
214 (47)
|
75 (46.3)
|
14 (42.4)
|
0.87
|
|
Cycles with oocyte cryopreservation, n (%)
|
20 (4.4)
|
4 (2.5)
|
2 (6.1)
|
0.46
|
|
Cycles with embryo cryopreservation, n (%)
|
183 (40.2)
|
63 (38.9)
|
12 (36.4)
|
0.88
|
|
Positive pregnancy test, n (%)
|
133 (29.2)
|
52 (32.1)
|
9 (27.3)
|
0.75
|
|
Clinical pregnancy rate, n (%)
|
111 (24.4)
|
42 (25.9)
|
8 (24.2)
|
0.93
|
|
Live birth rate, n (%)
|
92 (20.2)
|
36 (22.2)
|
7 (21.2)
|
0.86
|
|
Miscarriage rate, n (%)
|
19 (17.1)
|
6 (14.3)
|
1 (12.5)
|
0.93
|
|
Multiple pregnancy rate, n (%)
|
30 (27)
|
9 (21.4)
|
2 (25)
|
0.90
|
Abbreviation: TSH: thyroid-stimulating hormone.
Notes: Clinical pregnancy rate: number of patients with embryos exhibiting a heartbeat
on transvaginal ultrasound (TVUS) 4 to 5 weeks after embryo transfer divided by the
number of cycles × 100.
Miscarriage rate: number of patients who miscarried divided by the number of patients
who had clinical pregnancy × 100.
Live birth rate: number of patients who delivered a live birth divided by the number
of patients who had clinical pregnancy × 100.
Multiple pregnancy rates: number of patients who delivered more than one live birth
divided by the number of patients who had clinical pregnancy × 100.
p determined by Chi-squared test.
Additional Analysis
The present study had a power of 80% to detect a difference of 10% in clinical pregnancy
rates and live birth rates among women with TSH < 2.5 mIU/L and ≥ 2.5 mIU/L.
Comment
This study showed that women undergoing COS for ARTs had similar reproductive outcomes,
including live birth, clinical pregnancy, miscarriage, and multiple pregnancy rates,
regardless of serum TSH concentrations (< 2.5 mIU/L; 2.5–4.0 mIU/L; and 4.0–10.0 mIU/L).
The response to COS, as evaluated by the total dosage of FSH, the duration of COS
and the number of retrieved and mature oocytes, was also similar among these groups.
Concentrations of TSH > 4.0 mIU/L were considered altered in the present study, reflecting
concerns regarding overdiagnosis and overtreatment when the threshold of 2.5 mIU/L
is chosen. Moreover, there is insufficient evidence that TSH concentrations between
2.5 and 4.0 mIU/L affect pregnancy and miscarriage rates.[7]
[24] Data to support treating patients with TSH concentrations between 2.5 mIU/L and
4.0–4.5 mIU/L are scarce,[7] and only a few studies have evaluated the association between periconceptional TSH
concentration and pregnancy outcomes among euthyroid patients.[25] In the present study, clinical pregnancy, miscarriage and live birth rates were
similar when 2.5 mIU/L and 4.0 mIU/L were used as TSH cut-off values, which was consistent
with previous findings.[14]
[26]
[27]
[28]
[29]
[30] Regarding miscarriage, the ASRM's new guideline points that there is evidence that
TSH > 4.0 mIU/L is associated with an increase in miscarriage rates.[7] As seen in the present study, Karmon et al[25] found similar clinical pregnancy and delivery rates among women with TSH concentrations
of 0.4–2.49 mIU/L and 2.5–4.99 mIU/L undergoing intrauterine insemination.
Guidelines of the American Association of Clinical Endocrinologists (AACE) and the
ATA,[31] and those of the Endocrine Society[3] recommend that infertile patients undergoing ARTs must be treated with levothyroxine
to maintain TSH concentrations < 2.5 mIU/L. In this study, 24.92% of women undergoing
IVF/ICSI had TSH concentrations between 2.5 and 4.0 mIU/L. Although retrospective
studies have shown that untreated women were at an increased risk of adverse pregnancy
outcomes, few well-designed studies have evaluated whether these patients would have
benefited from levothyroxine treatment.[4]
[5] Moreover, until very recently there was no consensus about measuring TSH in asymptomatic
infertile women.[8]
[9] In the present study, reproductive outcomes were not compromised in women with TSH
≥2.5 and < 10 mIU/L undergoing ARTs. Nothing can be said about reproductive outcomes
in patients presenting TSH concentrations above 10.0 mIU/L, once none of the included
patients in this study presented them.
Although human reproduction societies have reported that subfertile women are no more
likely to have thyroid disease,[7]
[9] the Endocrine Society has stated that the prevalence of thyroid disease is 1 to
43% higher in subfertile patients.[3] In our study, the prevalence of patients presenting concentrations of TSH between
4.0 and 10.0 mIU/L was of 5.07% (33/650). If the threshold of TSH is lowered to 2.5
mIU/L, it will result in a prevalence of 30%, more than a 5-fold increase.[14]
Our study is limited by its retrospective design. Serum TSH concentrations were not
available in ∼ 8% of the women who underwent ARTs in the study period. However, until
very recently, there was no consensus about the routine assessment of TSH concentrations
for asymptomatic infertile patients according to the treatment guidelines for ARTs.[8]
[9] In the present study, assay bias should be taken into account, once measurements
of TSH were not standardized.[32] TSH concentrations were measured by ultra-sensitive methodology, but once data were
obtained by reviewing medical records, it was not possible to mention information
regarding the coefficient of variation of the assays. Ideally, for each patient, TSH
concentration should have been assessed more than once to exclude laboratory errors
or transient elevations of TSH to make an adequate diagnosis. Free thyroxine concentrations
were not available in the medical records of the majority of the patients in this
cohort, but in asymptomatic patients with mildly elevated TSH concentrations the yield
of this test is low.[27] Our findings cannot be generalized, but they may encourage large prospective studies
evaluating the potential impact of TSH concentrations on reproductive outcomes after
ARTs.
In conclusion, response to COS, live birth, and miscarriage rates were not worse in
women with elevated TSH concentrations (2.5–10.0 mIU/L) undergoing IVF/ICSI. These
findings reinforce the uncertainties related to the impact of elevated concentrations
of TSH on reproductive outcomes in women undergoing COS for ARTs, and the need for
well-designed prospective studies evaluating this important issue.
