OBJECTIVE: To evaluate the efficacy of a targeted thyroid testing approach during pregnancy in clinical practice.
METHODS: This is a retrospective cohort study performed within Uppsala County, Sweden. Data were derived from the population-based Uppsala Biobank of Pregnant Women, in which blood samples are collected in conjunction with the routine ultrasound screening in gestational week 17–19. For this study, 5,254 pregnant women with an estimated date of delivery between January 1, 2009, and December 31, 2011, were included. On review of their medical records, women who were tested for thyroid dysfunction during pregnancy in clinical practice were identified (n=891). From the remaining untested women, 1,006 women were randomly selected for analyses of thyrotropin (TSH), free thyroxine levels, and thyroid peroxidase antibodies. Thyroid-stimulating hormone levels in both groups were analyzed with regard to trimester-specific upper reference levels as recommended by the International Endocrine Society Guidelines.
RESULTS: The proportion of trimester-specific TSH elevation was 12.6% in the targeted thyroid testing group and 12.1% in the untested group (P=.8; odds ratio [OR] 1.04, 95% confidence interval [CI] 0.79–1.37). The proportion of overt hypothyroidism was 1.1% and 0.7% in the groups, respectively (P=.4; OR 1.57, 95% CI 0.55–4.45).
CONCLUSIONS: The prevalence of trimester-specific elevated TSH and overt hypothyroidism was equal in targeted thyroid tested and untested women. When implemented in clinical practice, targeted thyroid testing is unsatisfactory. If ongoing studies provide support for treatment of pregnant women with elevated TSH, universal thyroid testing appears the most reasonable approach.
LEVEL OF EVIDENCE: II
Targeted thyroid testing during pregnancy is unsatisfactory in clinical practice because the prevalence of hypothyroidism is equal in targeted tested and untested women.
Departments of Women's and Children's Health and Medical Sciences, Uppsala University, Uppsala, Sweden.
Corresponding author: Michaela Granfors, MD, Department of Women's and Children's Health, University Hospital, SE 751 85 Uppsala, Sweden; e-mail: firstname.lastname@example.org.
Supported by grants from Gillbergska stiftelsen, Uppsala, Sweden, and grants from the Department of Women's and Children's Health, University Hospital, Uppsala, Sweden.
The authors thank Lena Moby for skillful technical assistance.
Financial Disclosure The authors did not report any potential conflicts of interest.
Whether to apply targeted or universal testing for thyroid dysfunction during pregnancy is a matter of controversy. The American Thyroid Association Guidelines state that there is insufficient evidence for or against universal thyrotropin (TSH) screening at the first-trimester visit.1 Similarly, the most recent International Guidelines of the Endocrine Society has a split recommendation, advocating either universal TSH screening or aggressive case finding.2 Thus far, there is limited evidence to suggest any of these approaches.3
In the research setting, when detailed study protocols are used to identify pregnant women with risk factors for thyroid disease, targeted thyroid function testing has been shown to miss between 33% and 67% of pregnant women with overt or subclinical hypothyroidism.3,4 However, the efficacy of targeted thyroid testing when implemented in clinical practice is currently unknown.
We and others have previously shown that the implementation of targeted thyroid testing in clinical practice is not optimal. In a survey from Wisconsin, only 11.5% of care providers had read the Endocrine Society's Guidelines.5 In another district in the United States, less than 20% of the high-risk pregnant women were tested for thyroid disease despite the development and circulation of local guidelines.6 The poor implementation of targeted thyroid testing was recently confirmed in a nationwide Swedish study.7
Hence, the aim of this retrospective cohort study was to examine the real-life efficacy of a targeted thyroid testing approach in identifying women with elevated TSH and overt hypothyroidism during pregnancy.
MATERIALS AND METHODS
Antenatal care in Sweden is standardized, free of charge, and decentralized to maternity care districts. Within each maternity care district, a consultant in obstetrics is responsible for development and implementation of guidelines. In the guidelines of Uppsala County, a targeted thyroid testing approach was implemented in 2004.
Data for this study were derived from the population-based Uppsala Biobank of Pregnant Women, in which blood samples are collected in conjunction with the routine ultrasound screening. Eligible women are 18 years or older, Swedish-speaking, and without bloodborne disease. In Sweden, all pregnant women are invited to an ultrasound examination at 17–19 weeks of gestation, and approximately 97% of the Swedish pregnant population participates.8 In Uppsala County, all routine ultrasound examinations are performed at Uppsala University Hospital, which is also the only available delivery ward within the county. Hence, the Biobank participants represent a population-based sample. The Biobank is a convenience sample based on availability of a research nurse. Approximately 30% of the respondents decline participation, and it is estimated that the Biobank covers approximately half of the pregnant population of Uppsala County.7
For this study, 5,254 pregnant women from the Biobank with an estimated date of delivery between January 1, 2009, and December 31, 2011, were included. Brief demographic data were collected, including ongoing chronic disorders, medication, smoking, and height and weight. A blood sample was collected before the routine ultrasound examination of participants and the sample was stored at −70°C. The participating women gave written informed consent and the ethical review board at Uppsala University, Uppsala, Sweden, approved the study, Dnr 2012/377.
The medical records of the 5,254 women were reviewed between February and May 2012. Any TSH testing performed during pregnancy was defined as targeted thyroid testing (from gestational week 4+0 until delivery based on the ultrasound-estimated date of delivery). In case of repeated TSH testing, only the first TSH test sample was considered. We also recorded the gestational age at targeted thyroid testing, the results of all analyzed thyroid function tests, and the prepregnancy use of levothyroxine.
Targeted thyroid testing was performed on 1,054 women (Fig. 1). All women who were on levothyroxine treatment at conception (before gestational week 2 according to ultrasound-estimated date of delivery; n=163) were excluded from the study. Thus, the targeted thyroid testing group consisted of 891 women. On April 23, 2009, there was a change in laboratory methods for free thyroxine (T4), affecting the levels and reference ranges of free T4. The 235 women tested before this date have been kept in the material but were not used for analysis and comparison of free T4 levels.
In the guidelines from the district of Uppsala County, valid during the entire study period, five reasons for targeted thyroid testing were given: personal history of thyroid dysfunction with or without ongoing levothyroxine treatment; family history of thyroid dysfunction; goiter; insulin-dependent diabetes mellitus; and other autoimmune diseases (Addison's disease, celiac disease, and atrophic gastritis with vitamin B12 deficiency were specified). The guidelines recommended targeted thyroid testing by measurement of TSH, free T4, and thyroid peroxidase antibodies, but even so, free T4 and thyroid peroxidase antibodies were lacking in a substantial proportion of women (n=75 and n=282, respectively).
Four thousand two hundred of the 5,254 included pregnant women were not tested for thyroid dysfunction in clinical practice. From these 4,200 untested women, 1,006 were randomly selected in three blocks (representing equal samples of women from the years 2009, 2010, and 2011, respectively). The random selection was made by the Biobank manager, who was unaware of the purpose and intention of the study. A certain degree of oversampling was used to ensure sufficient numbers following the laboratory analyses. The women's stored Biobank samples were assayed for levels of TSH, free T4, and thyroid peroxidase antibodies in December 2012. The prevalence of primary criteria for targeted thyroid testing such as family history of thyroid dysfunction could not be established with certainty in the untested population. Because the primary outcome in this study was to evaluate the real-life efficacy of targeted thyroid testing, it was considered acceptable that an unknown fraction of untested women, in theory, should have been tested.
In line with International Guidelines from the Endocrine Society,2 trimester-specific TSH upper reference limits for intervention of 2.5 milliunits/L in the first and 3.0 milliunits/L in the second and third trimester were used. Overt hypothyroidism was defined as trimester-specific elevated TSH in conjunction with free T4 less than 9.7 pmol/L (according to local reference range; see subsequently) or TSH greater than 10.0 milliunits/L irrespective of free T4. The remaining women with trimester-specific elevated TSH levels were considered to have subclinical hypothyroidism. According to local reference ranges, women with thyroid peroxidase antibody levels 34 kIE/L or greater were considered to be thyroid peroxidase antibody-positive. Pregnancy trimesters were defined as follows: first trimester until 13 completed weeks of gestation; second trimester gestational week 14–27; third trimester 28 completed weeks of gestation or later.
The second trimester-specific free T4 reference range was calculated in the untested group among those who had no detectable thyroid peroxidase antibodies. Hence, samples from 949 untested women were available for determination of the free T4 reference range. The second trimester 2.5th–97.5th percentile for free T4 was 9.7–15.7 pmol/L.
Thyrotropin, free T4, and thyroid peroxidase antibody analyses were performed by use of the same laboratory methods in the targeted thyroid testing group and the untested group. Nonfasting blood samples were used. Free T4, TSH, and thyroid peroxidase antibody analyses were run on an automatic immune analyzer. The total assay variation in the individual assays was less than 5%. All analyses were performed at the routine laboratory of the Department of Clinical Chemistry at the University Hospital in Uppsala. The laboratory is certified by a Swedish government authority (Swedac).
The power calculation was based on Negro et al.3 Because we knew that the proportion of women with subclinical hypothyroidism in the targeted thyroid testing group was 12.6%, we assumed, based on Negro et al,3 that the prevalence of subclinical hypothyroidism in the untested group would be approximately half of that in the targeted thyroid testing group. Thus, a sample size of 1,000 in the untested group would give a power of 99.6% at a significance level of .05 to reject the null hypothesis of zero correlation.
Comparisons between the targeted thyroid testing group and the untested group were performed by χ2 tests or Fisher’s exact test. Because all continuous variables were normally distributed, independent t tests were used for groupwise comparisons. The statistical software package SPSS was used for all analyses.
Women in the untested group were younger, had lower body mass indexes (BMIs, calculated as weight (kg)/[height (m)]2), and were less often smokers than women in the targeted thyroid testing group (Table 1). Mean gestational age at targeted thyroid testing was 15.6±8.1 weeks (median 11.9 weeks, range 4.4–40.3 weeks), whereas untested women donated their blood sample at a mean gestational age of 17.9±1.2 weeks (median 18 weeks, range 10.3–24.6 weeks).
The proportion of women with thyroid disturbances is shown in Table 2. The proportion of trimester-specific TSH elevation was 12.6% in the targeted thyroid testing group and 12.1% in the untested group (P=.8; odds ratio [OR] 1.04, 95% confidence interval [CI] 0.79–1.37). Two women in each group had TSH levels greater than 10 milliunits/L. Rates of overt and subclinical hypothyroidism were similar in the targeted thyroid testing and untested groups. Free T4 and thyroid peroxidase antibodies had not been analyzed in all women in the targeted thyroid testing group (75 and 282 women, respectively), making further comparisons impossible.
International Guidelines suggest that pregnant women age 30 years or older and with BMIs 40 or greater should undergo targeted thyroid testing.1 Because these risk factors were not included in the Uppsala County guidelines during the study period, further analyses in women with these additional risk factors were conducted. In Table 3, rates of thyroid dysfunction in untested women with or without additional risk factors for thyroid disease are shown. The majority of women with additional risk factors were age 30 years or older (n=548), whereas only a few women younger than 30 years had exclusively BMIs of 40 or greater (n=3) or autoimmune disorders (n=7). The addition of these risk factors did not result in improved efficacy because no difference in rates of thyroid dysfunction between untested women with or without additional risk factors for thyroid disease were found (Table 3).
Pregnancy outcomes in untested women who had elevated or normal TSH levels are shown in Table 4. Bearing in mind that the study was unpowered for this analysis, no increased prevalence of adverse pregnancy outcomes was noted in untested women who had elevated TSH levels.
In this population-based sample of pregnant women, the prevalence of trimester-specific elevated TSH and overt hypothyroidism was equal in targeted thyroid tested and untested women. Extrapolating to all 5,254 women included in our study, approximately 82% of women with elevated TSH were missed by targeted thyroid testing. Thus, real-life efficacy of the targeted thyroid testing approach is not as good as when performed in the research setting.3,4,9
When adding extra indications for targeted case finding, as exemplified in Table 3, targeted thyroid testing should have had to be performed in 63% of women in the study population. In line with earlier studies, addition of these criteria to the targeted thyroid testing strategy would improve its efficiency substantially, but only as a result of the increased number of women being tested.10 It is questionable whether it is worthwhile to apply targeted thyroid testing when the goal is to test almost two thirds of the pregnant population while still missing a significant number of women with hypothyroidism.
Although conducted in Sweden, data suggest that the results from our study might be generalized also to other settings.5,6 In a survey from Wisconsin, International Guidelines reached only a minority (11.5%) of obstetric providers.5 Similar to our results, in a large national sample of 500,000 pregnant women in the United States, 23% underwent targeted thyroid testing during pregnancy.11 Of those, 15.5% tested positive for gestational hypothyroidism.11 Furthermore, in a clinical audit, less than 20% of high-risk women were tested for thyroid dysfunction during pregnancy despite available local guidelines.6 Thus, the implementation of targeted thyroid testing into everyday clinical practice seems to be equally difficult independent of the setting.
Importantly, the intended target for detection and treatment of hypothyroidism during pregnancy is to improve offspring and pregnancy outcomes. Assuming the commonly reported incidence of hypothyroidism during pregnancy of 0.3–0.5% and 3–5% for subclinical hypothyroidism,12 a recent systematic review reported insufficient evidence for treatment of subclinical hypothyroidism during pregnancy.13 In a recent randomized trial, antenatal thyroid testing and maternal treatment for hypothyroidism did not result in improved cognitive function in children at 3 years of age or, as a secondary outcome, in improved pregnancy outcomes.14 If ongoing randomized controlled trials provide support for treatment of pregnant women with elevated TSH, universal thyroid testing thus appears the most reasonable approach.
The prevalence of women with trimester-specific elevated TSH and especially of those with subclinical hypothyroidism was much higher in our study than commonly reported.15 However, similar and even higher prevalence rates of subclinical hypothyroidism have been described in a few studies that also applied trimester-specific upper reference levels for TSH as recommended by the International Guidelines.11,16 Whether the high prevalence of elevated trimester-specific TSH in our study reflects a more affected population (despite the fact that Sweden has been considered to be an iodine-sufficient country17), or unrealistically low trimester-specific reference levels for TSH in our setting, remains unclear. Although this study did not have enough power to evaluate pregnancy outcomes, it is interesting to note that there were no differences in pregnancy outcomes between women with elevated and normal TSH in the untested group. Establishment of trimester-specific TSH reference ranges for each laboratory would probably help to improve detection of those women who really benefit from treatment with levothyroxine.
One major strength of our study is the population-based design and that data were derived from clinical practice. Implementation of guidelines into clinical practice is known to be associated with difficulties,5–7,18 and earlier studies on targeted thyroid testing might have overlooked this point by using “expert panels” for classification of pregnant women into low-risk or high-risk groups.3,4,9 A limitation of our study is that blood samples in the untested group were not taken in the first trimester of pregnancy, but in the early second trimester. Moreover, despite clear recommendations in the local guidelines, free T4 and thyroid peroxidase antibodies were not analyzed in all women who underwent targeted thyroid testing, making comparisons other than presence of subclinical and overt hypothyroidism impossible.
In conclusion, a targeted thyroid testing approach in clinical practice was unsatisfactory, missing more than 80% of pregnant women with elevated TSH. Once, and if, evidence to support treatment of pregnant women with elevated TSH becomes available, universal thyroid testing appears the most reasonable approach.
1. Stagnaro-Green A, Abalovich M, Alexander E, Azizi F, Mestman J, Negro R, et al.. Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and postpartum. Thyroid 2011;21:1081–125.
2. De Groot L, Abalovich M, Alexander EK, Amino N, Barbour L, Cobin RH, et al.. Management of thyroid dysfunction during pregnancy and postpartum: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2012;97:2543–65.
3. Negro R, Schwartz A, Gismondi R, Tinelli A, Mangieri T, Stagnaro-Green A. Universal screening versus case finding for detection and treatment of thyroid hormonal dysfunction during pregnancy. J Clin Endocrinol Metab 2010;95:1699–707.
4. Vaidya B, Anthony S, Bilous M, Shields B, Drury J, Hutchison S, et al.. Detection of thyroid dysfunction in early pregnancy: universal screening or targeted high-risk case finding? J Clin Endocrinol Metab 2007;92:203–7.
5. Haymart MR. The role of clinical guidelines in patient care: thyroid hormone replacement in women of reproductive age. Thyroid 2010;20:301–7.
6. Vaidya B, Bilous M, Hutchinson RS, Connolly V, Jones S, Kelly WF, et al.. Screening for thyroid disease in pregnancy: an audit. Clin Med 2002;2:599–600.
7. Granfors M, Åkerud H, Berglund A, Skogö J, Sundström-Poromaa I, Wikström AK. Thyroid testing and management of hypothyroidism during pregnancy: a population based study. J Clin Endocrinol Metab 2013;98:2687–92.
8. Reports from the Swedish Council on Technology Assessment in Health Care (SBU). Int J Technol Assess Health Care 1999;15:424–36.
9. Chang DL, Leung AM, Braverman LE, Pearce EN. Thyroid testing during pregnancy at an academic Boston Area Medical Center. J Clin Endocrinol Metab 2011;96:E1452–6.
10. Potlukova E, Potluka O, Jiskra J, Limanova Z, Telicka Z, Bartakova J, et al.. Is age a risk factor for hypothyroidism in pregnancy? An analysis of 5223 pregnant women. J Clin Endocrinol Metab 2012;97:1945–52.
11. Blatt AJ, Nakamoto JM, Kaufman HW. National status of testing for hypothyroidism during pregnancy and postpartum. J Clin Endocrinol Metab 2012;97:777–84.
12. Casey BM, Dashe JS, Spong CY, McIntire DD, Leveno KJ, Cunningham GF. Perinatal significance of isolated maternal hypothyroxinemia identified in the first half of pregnancy. Obstet Gynecol 2007;109:1129–35.
13. Vissenberg R, van den Boogaard E, van Wely M, van der Post JA, Fliers E, Bisschop PH, et al.. Treatment of thyroid disorders before conception and in early pregnancy: a systematic review. Hum Reprod Update 2012;18:360–73.
14. Lazarus JH, Bestwick JP, Channon S, Paradice R, Maina A, Rees R, et al.. Antenatal thyroid screening and childhood cognitive function. N Engl J Med 2012;366:493–501.
15. Casey BM, Leveno KJ. Thyroid disease in pregnancy. Obstet Gynecol 2006;108:1283–92.
16. Altomare M, La Vignera S, Asero P, Recupero D, Condorelli RA, Scollo P, et al.. High prevalence of thyroid dysfunction in pregnant women. J Endocrinol Invest 2013;36:407–11.
17. Andersson M, Berg G, Eggertsen R, Filipsson H, Gramatkovski E, Hansson M, et al.. Adequate iodine nutrition in Sweden: a cross-sectional national study of urinary iodine concentration in school-age children. Eur J Clin Nutr 2009;63:828–34.
18. Vaidya B, Hubalewska-Dydejczyk A, Laurberg P, Negro R, Vermiglio F, Poppe K. Treatment and screening of hypothyroidism in pregnancy: results of a European survey. Eur J Endocrinol 2012;166:49–54.