1814 THERAPEUTIC CONTROVERSY JCE & M • 1999
Vol 84 • No 6
FIG. 1. The prevalence of antithyroid
MCAb and the occurrence of postpar-
tum thyroid dysfunction among preg-
nant women in the general population.
FIG. 2. A protocol to screen and follow
MCAb-positive mothers. MCAb, anti-
thyroid microsomal antibody; FT4, free
thyroxine; FT3, free triiodothyronine;
TSH, thyrotropin; TRAb, anti-TSH re-
4 – 8 weeks. We believe that this close observation protocol ance system is well-established. It should be modified
for MCAb-positive subjects is practical in only a few according to the situation and health system of each
countries, such as Japan, where the national health insur- country.
THERAPEUTIC CONTROVERSY 1815
The benefits of screening postpartum thyroid dysfunction concluded that MCAb measurement is the most cost-effec-
Although postpartum thyroid dysfunction may be tran- tive, but their report compares only the screening methods
sient in many cases (20), we should pay close attention to the for postpartum thyroid dysfunction; they did not analyze the
possibility that mothers with thyroid dysfunction already benefit of mothers’ quality of life, nor did they discuss cost-
tired from the birth and from taking care of a newborn, may effectiveness compared with the screening for other diseases
feel terribly worn out and helpless or may become ill in bed. or other health services. Indeed, there seems no analysis of
Postpartum depression, which in severe cases may result in the cost-effectiveness of screening for postpartum thyroid
a forced double suicide of mother and baby (called “Shinjuh” dysfunction (23, 30). However, the observation protocol we
in Japanese), can be screened by MCAb measurement, al- describe can be modified to be less expensive, and screening
though postpartum depression may occur not in association subjects may be restricted to certain high-risk groups, such
with thyroid dysfunction (17, 21–23) but with antithyroid as the patients with IDDM. We believe an optimized system
MCAb (24, 25). The main benefit of screening for postpartum of screening will be found whose costs are acceptable for each
thyroid dysfunction is the opportunity to improve the qual- society.
ity of life of mothers who may suffer from the above symp- We experimentally applied TSAb to the second-line
toms. Even when the patient does not want drug therapy, she screening of MCAb-positive mothers and found that TSAb
benefits by being informed about what is happening to her. gives predictive information on potential development of
From our experience, patients with mild to moderate thyroid postpartum Graves’ disease. TSAb measurement is obvi-
dysfunction can live well with the support of family without ously too expensive, and the procedure is too complicated to
medication. Further, prevention of future episodes of post- apply screening to the general population. However, we
partum thyroid dysfunction may be possible. We have ex- expect the development of a more sensitive radioreceptor
perienced one case of successful prevention. The patient had assay may change the situation (31).
developed severe postpartum thyroid dysfunction in each of
her previous parturitions, and she did not want to repeat the Conclusion
condition after the next delivery. Moderate doses of glu- The benefit of screening for postpartum thyroid dysfunc-
cocorticoid, gradually decreased and stopped in one month, tion mainly concerns elusive, nonquantitative parameters—
suppressed her thyroid dysfunction to only a small fluctu- the quality of life of mothers—that can be difficult to assess.
ation within the normal range. However, we would hope to find an acceptable, cost-effec-
Postpartum Graves’ disease accounts for 11.4% of post- tive system of screening for postpartum thyroid dysfunction.
partum thyroid dysfunctions (occurring in 0.54% of all moth-
ers in the general population) (19). Conversely, 40% of
Graves’ patients 20 –39 yr old, who have had one or more References
deliveries, developed their disease during the postpartum 1. Amino N, Tada H, Hidaka Y. 1996 Autoimmune thyroid disease and preg-
period (26). Diagnosis of postpartum Graves’ disease early, nancy. J Endocrinol Invest. 19:59 –70
2. Amino N, Mori H, Iwatani Y, et al. 1982 High prevalence of transient post-
while it is mild, may easily lead to remission and may reduce partum thyrotoxicosis and hypothyroidism. N Engl J Med. 306:849 – 852.
Graves’ disease in older age. In our experience, the early start 3. Jansson R, Bernander S, Karlsson A, Levin K, Nilsson G. 1984 Autoimmune
thyroid dysfunction in the postpartum period. J Clin Endocrinol Metab.
of antithyroid therapy reduces the period of therapy by half 58:681– 687.
(19). Antithyroid therapy may be a good choice for first-line 4. Walfish P, Chan J. 1985 Postpartum hyperthyroidism. Clin Endocrinol Metab.
therapy (27) because postpartum Graves’ hyperthyroidism is 14:417– 447.
5. Freeman R, Rosen H, Thysen B. 1986 Incidence of thyroid dysfunction in an
often transient and because mothers may not want to inter- unselected postpartum population. Arch Intern Med. 146:1361–1364.
rupt breast-feeding to undergo radioiodine therapy. 6. Nikolai TF, Turney SL, Roberts RC. 1987 Postpartum lymphocytic thyroiditis.
An additional benefit to screening with anti-MCAb anti- Prevalence, clinical course, and long-term follow-up. Arch Intern Med.
bodies in early pregnancy may be not only to find postpar- 7. Lervang HH, Pryds O, Kristensen HP. 1987 Thyroid dysfunction after deliv-
tum thyroid dysfunction, but also to find mothers at high risk ery: incidence and clinical course. Acta Med Scand. 222:369 –374.
for spontaneous abortion. In our prospective study, the spon- 8. Fung HYM, Kologlu M, Collison K, et al. 1988 Postpartum thyroid dysfunc-
tion in Mid Glamorgan. Br Med J. 296:241–244.
taneous abortion rate in MCAb-positive mothers was twice 9. Rasmussen NG, Hornnes PJ, Hoiter-Madsen M, Feldt-Rasmussen U, Hege-
as high as that in MCAb-negative mothers (28). dus L. 1990 Thyroid size and function in healthy pregnant women with thyroid
autoantibodies. Acta Endocrinol (Copenh). 123:395– 401.
10. Rajatanavin R, Chailurkit LO, Tirarungsikul K, Chalayondeja W, Jittivanich
There has been no analysis of the cost-effectiveness of U, Puapradit W. 1990 Postpartum thyroid dysfunction in Bangkok: a geo-
graphical variation in the prevalence. Acta Endocrinol (Copenh). 122:283–287.
screening for postpartum thyroid dysfunction 11. Roti E, Bianconi L, Gardini E, et al. 1991 Postpartum thyroid dysfunction in
an Italian population residing in an area of mild iodine deficiency. J Endocrinol
The cost of MCAb measurement to screen for the occur- Invest. 14:669 – 674.
rence of postpartum thyroid dysfunction is low when we use 12. Lobig H, Bohn W, Mau J, Schatz H. 1991 Prevalence of postpartum thyroiditis
semiquantitative particle agglutination tests. However, thy- in two iodine-deficient regions of Germany. In: Scherbaum W, Bogner U,
Weinheimer B, and Bottazzo G, eds. Autoimmune thyroiditis. Berlin: Springer-
roid function tests and anti-TSHR antibody measurement for Verlag; 185–193.
MCAb-positive subjects are not cheap, especially when pa- 13. Walfish PG, Meyerson J, Provias JP, Vargas MT, Papsis FR. 1992 Prevalence
and characteristics of post-partum thyroid dysfunction: Results of a survey
tients are observed closely and are tested repeatedly. On the from Toronto, Canada. J Endocrinol Invest. 15:265–272.
other hand, the main benefit of screening for postpartum 14. Stagnaro-Green A, Roman SH, Cobin RH, el-Harazy E, Wallenstein S, Da-
thyroid dysfunction is improving the mother’s quality of life, vies TF. 1992 A prospective study of lymphocyte-initiated immunosuppres-
sion in normal pregnancy: evidence of a T-cell etiology for postpartum thyroid
which is difficult to numerically assess. Heyslip et al. (29) dysfunction. J Clin Endocrinol Metab. 74:645– 653.
tried to estimate the costs of several screening methods and 15. Yoshida H, Amino N, Yagawa K, et al. 1978 Association of serum antithyroid
1816 THERAPEUTIC CONTROVERSY JCE & M • 1999
Vol 84 • No 6
antibodies with lymphocytic infiltration of the thyroid gland: studies of sev- illness are often mild in degree and brief in duration and,
enty autopsied cases. J Clin Endocrinol Metab. 46:859 – 862.
16. Solomon BL, Fein HG, Smallridge RC. 1993 Usefulness of antimicrosomal consequently, may not require treatment. When symptoms
antibody titers in the diagnosis and treatment of postpartum thyroiditis. J Fam are moderate or severe in degree they should be recognized
Pract. 36:177–182. clinically, although this will require more education of pri-
17. Pop VJM, de Rooy HA, Vader HL, van der Heide D, van Son MM, Komproe
IH. 1993 Microsomal antibodies during gestation in relation to postpartum mary care physicians and postpartum patients about PPT.
thyroid dysfunction and depression. Acta Endocrinol (Copenh). 129:26 –30. Which thyroid autoantibody assay should be used for screen-
18. Tamaki H, Amino N, Aozasa M, Mori M, Tanazawa O, Miyai K. 1987 Serial
changes in thyroid-stimulating antibody and thyrotropin binding inhibitor
ing and when the assay should be performed have not been
immunoglobulin at the time of postpartum occurrence of thyrotoxicosis in completely resolved. Furthermore, there is a lack of knowl-
Graves’ disease. J Clin Endocrinol Metab. 65:324 –330. edge about how to follow patients with a positive antibody
19. Hidaka Y, Tamaki H, Iwatani Y, Tada H, Mitsuda N, Amino N. 1994 Pre-
diction of postpartum onset of Graves’ thyrotoxicosis by measurement of assay. Should a serum TSH level be done only when they are
thyroid stimulating antibody in early pregnancy. Clin Endocrinol (Oxf). symptomatic, or should serial TSH levels be done and, if so,
41:15–20. at what intervals? Finally and most importantly, there has
20. Amino N, Tada H, Hidaka Y. 1996 The spectrum of postpartum thyroid
dysfunction: diagnosis, management, and long-term prognosis. Endocr Pract. been no prospective diagnostic and therapeutic trial to date
2:406 – 410. that tells us whether or not a screening program would be
21. Harris B, Fung H, Johns S, et al. 1989 Transient postpartum thyroid dysfunc-
tion and postnatal depression. J Affect Disord. 17:243–249.
22. Pop VJM, Rooy HAM, Vader HL, et al. 1991 Postpartum thyroid dysfunction We will initially discuss screening for thyroid disease in
and depression in an unselected population. New Engl J Med. 324:1815–1816. general to set the stage for a detailed discussion of screening
23. Lazarus JH, Harris B, Parkes AB. 1996 Antenatal screening of thyroid anti-
bodies. Lancet. 348:1516 –1517. for PPT. Available data and studies on these important issues
24. Harris B, Othman S, Davies JA, et al. 1992 Association between postpartum will lead to the conclusion that the time is not yet ripe to
thyroid dysfunction and thyroid antibodies and depression. BMJ. 305:152–156. recommend a screening program for all pregnant or post-
25. Pop VJ, Maartens LH, Leusink G, et al. 1998 Are autoimmune thyroid dys-
function and depression related? J Clin Endocrinol Metab. 83:3194 –3197. partum women to identify those who are at risk for or who
26. Tada H, Hidaka Y, Itoh E, et al. 1994 Prevalence of postpartum onset of disease have PPT.
within patients with Graves’ disease of child-bearing age. Endocr J. 41:325–327.
27. Amino N, Tada H. 1997 Postpartum thyroid disease. In: Bardin CW, ed.
Current therapy in endocrinology and metabolism, 6th ed. Philadelphia:
Mosby-Year Book, Inc.; 327–330. Screening for thyroid disease in general
28. Iijama T, Tada H, Hidaka Y, Mitsuda N, Murata Y, Amino N. 1997 Effects of
autoantibodies on the course of pregnancy and fatal growth. Obstet Gynecol. In general, screening may be defined as testing for the
90:364 –369. presence of a disease or the risk of a disease when no known
29. Haylip CC, Fein HG, O’Donnel VM, et al. 1988 The value of serum antimi- signs or symptoms of the disease are present, with the pur-
crosomal antibody testing in screening for symptomatic postpartum thyroid
dysfunction. Am J Obstet Gynecol. 159:203–209. pose of improving health outcomes in the target population
30. Ball S. 1996 Antenatal screening of thyroid antibodies. Lancet. 348:906 –907. (6). Screening recommendations for thyroid disease involve
31. Watabnabe Y, Tada H, Hidaka Y, et al. 1999 Polyethylene glycol increases the
detection of anti-thyrotropin (TSH) receptor antibodies by a radioreceptor
an assessment of age, gender, family history, and associated
assay. Clin Chem. 45:407– 409. physiologic or pathophysiologic conditions. In addition to
specifying who should be screened, such recommendations
The Time Is Not Ripe To Screen also need to specify which tests should be employed, how
Lawrence M. Crapo frequently they should be employed, and where they should
Division of Endocrinology be employed.
Stanford University School of Medicine and Presently, there is only one universally agreed-upon strat-
Santa Clara Valley Medical Center egy in screening for thyroid disease, which is that neonates
San Jose, California 95128 should be screened for congenital hypothyroidism with a
A T FIRST GLANCE, screening for postpartum thyroid-
itis (PPT) appears to be an attractive and possibly even
a cost-effective strategy. The disorder is quite common, af-
blood TSH or total T4 assay shortly after birth in the hospital.
In 1990, the American Thyroid Association (ATA) recom-
mended against screening for thyroid disease in asymptom-
fecting in the range of 4 – 6% of all postpartum women (1). In atic subjects in the general population who were not at high
the United States, where there are approximately 4 million risk for thyroid disease and its consequences (7). The Cana-
live births annually (2), this means that PPT afflicts nearly dian Task Force on the Periodic Health Examination in 1994
200,000 women each year. The symptoms associated with (8), and the United States Preventative Services Task Force
PPT are often subtle and difficult to distinguish from symp- (USPSTF) in 1996 (9), concurred in this recommendation
toms frequently present during the postpartum period. against general screening. All agencies recommend screen-
Those women at high risk for PPT are easy to identify be- ing for hypothyroidism in neonates even though it is infre-
cause PPT is an autoimmune disease in which most patients quent (1 case per 4,000 births). The ATA recommends screen-
have elevated serum levels of thyroid autoantibodies. It is ing with tests of thyroid function in certain higher risk
generally assumed that over 20% of women with PPT will situations, such as in elderly patients, those with a family
eventually develop permanent hypothyroidism within a 5-yr history of thyroid disease, individuals with autoimmune
period (3, 4). Finally, PPT is easy to treat when symptoms are diseases such as type I diabetes mellitus, and in postpartum
present by employing beta-blockers for the hyperthyroid women at 4 – 8 weeks (7). The USPSTF adds the high risk
phase and levothyroxine for the hypothyroid phase of the group of persons with Down’s syndrome, and notes that at
disorder (5). the present time there is insufficient evidence to recommend
On closer scrutiny, however, the screening strategy loses for or against screening for thyroid disease in any of the
some of its initial appeal. Although PPT is common, symp- childhood or adult high risk groups, including postpartum
toms during the hyperthyroid and hypothyroid phases of the women (9). They do recommend that clinicians should re-
THERAPEUTIC CONTROVERSY 1817
main alert for subtle signs and symptoms of thyroid dys- The epidemiology of PPT is presented in Table 1, where the
function and keep a low threshold for evaluating thyroid outcome of 1,000 hypothetical unselected postpartum
status in such high risk subjects. women is traced, assuming that 5% of the women will de-
Recently, one study has demonstrated by cost-utility de- velop PPT, 10% of the women will be positive for serum
cision analysis that it is cost-effective to screen for mild thy- thyroid peroxidase autoantibodies (TPO-Ab), nearly 50% of
roid failure with a serum TSH assay at 5-yr intervals starting the women with a positive TPO-Ab titer will develop PPT,
at age 35 yr for both men and women, and that such screening and that 90% of women with PPT will have a positive
is especially recommended for elderly women (10). More TPO-Ab titer. The data in this table have been derived from
recently, the American College of Physicians has recom- several reviews of numerous studies on the incidence of PPT
mended screening for unsuspected but symptomatic hypo- and the prevalence of positive TPO-Ab titers in postpartum
thyroidism and hyperthyroidism in women over 50 yr of age, women (1, 14). The incidence of PPT in these studies varies
employing a sensitive serum TSH assay (11, 12). They further by over an order of magnitude from 1.1% to 16.7%, which
note that there is insufficient evidence at present to recom- may be explained in part by variations in geographical lo-
mend for or against screening for subclinical hypothyroidism cation, definition of PPT, duration of studies, and frequency
or subclinical hyperthyroidism, although this latter recom- of testing. A critical evaluation of many of the studies sug-
mendation has been criticized (13). gests that the incidence of symptomatic PPT is about 5% (1).
Thus, from the general literature on screening for thyroid The incidence data on permanent overt hypothyroidism in
disease, there is little to help guide us in a decision about Table 1 and Table 2 has been derived from 5 separate studies
screening for PPT, and what there is remains controversial. in which patients with PPT have been followed for 3–5 yr (3,
In the presence of conflicting recommendations from pres- 4, 15–17).
tigious national agencies, and in the absence of published What is the purpose of screening for PPT? The simple
cost-benefit analyses or prospective controlled clinical trials, answer to this question is that the purpose is to alleviate
we are compelled to form a conclusion about the merits of symptoms from hypothyroidism and transient hyperthy-
screening for PPT on other grounds, employing available roidism and to identify women who are at risk for subse-
studies on the epidemiologic, laboratory, and clinical char- quent permanent hypothyroidism, as well as PPT in future
acteristics of PPT. pregnancies. But is screening really necessary to carry out
this purpose, or could the purpose be accomplished without
Screening for postpartum thyroiditis screening through careful attention to clinical detail by phy-
In order to develop a strategy of screening for PPT it is sicians who care for postpartum women? Unfortunately, no
necessary to review the epidemiology of the disorder and prospective studies have yet been conducted that answer this
assess its varied clinical manifestations so we can answer the question. The symptoms of hyperthyroidism and hypothy-
questions of what are we screening for, what tests should we roidism blend with and are notoriously difficult to separate
use for screening, when we should screen, and whether or from other symptoms associated with the postpartum state.
not screening is superior to clinical ascertainment of PPT. Some of the symptoms of hyperthyroidism and hypothy-
PPT is an autoimmune disorder characterized by a de- roidism may be found more frequently in postpartum
structive lymphocytic infiltration of the thyroid gland, very women with PPT and/or positive TPO-Ab titers, compared
often with accompanying circulating thyroid autoantibodies, to those without PPT (18 –20). However, the clinical distinc-
which can manifest itself as transient hyperthyroidism, tran- tion between these two groups of postpartum women can be
sient hypothyroidism, or permanent hypothyroidism. These quite difficult. Depression may be more frequent and severe
manifestations result from activation of immunological in women with PPT (21–23). Nevertheless, when symptoms
changes that occur in the first postpartum year. The symp- are mild in degree and transient in duration, it is unlikely that
toms of PPT are those of hyperthyroidism or hypothyroidism treatment is necessary, and equally unlikely that patients will
added to a background of symptoms commonly found in the seek medical attention. If symptoms are severe enough in
postpartum period, such as fatigue, tiredness, depression, degree and duration to bring patients to medical attention,
emotional lability, and anxiety. These symptoms can be mild then clinicians should test for thyroid dysfunction anyway,
or severe and transient or prolonged. Thus, PPT is a protean and screening would not be necessary. All postpartum
disorder with varied presentations at varied times hiding in women, whether or not screening is done, should be en-
the shadow of common postpartum symptoms, which ren- couraged by their physicians to seek medical attention if they
ders the development of a screening strategy very difficult. have troublesome symptoms and not automatically attribute
TABLE 1. Epidemiology of postpartum thyroiditis
Total number Women with positive Women with negative
of women TPO-Ab titers TPO-Ab titers
Postpartum women 1000 100 900
Women with PPT 50 45 5
Women with permanent hypothyroidism by 5 yr 9 9 0
The data in this table assumes that, for 1,000 unselected postpartum women, 5% will develop postpartum thyroiditis (PPT), and 90% of these
women will be positive for serum thyroid peroxidase autoantibodies (TPO-Ab). It is further assumed that nearly 50% of women with positive
serum TPO-Ab titers will develop PPT, and that, in these women, permanent hypothyroidism develops at an average rate of 3.6% per yr, with
an overall incidence rate of 1.8 cases/1000 women/yr. See the text and Table 2 for references and calculations to support these assumptions.
1818 THERAPEUTIC CONTROVERSY JCE & M • 1999
Vol 84 • No 6
TABLE 2. Progression of postpartum thyroiditis to permanent overt hypothyroidism
Patients Patients with Years of Progression to permanent HYPO Progression to permanent HYPO Study
with PPT permanent HYPO follow-up (per 100 women with PPT/yr) (per 1000 postpartum women/yr) (ref)
25 3 3 4.0 2.0 Nikolai (16)
44 9 5 4.1 2.1 Taichi (3)
31 5 5 3.2 1.6 Jansson (15)
43 4 3.5 2.7 1.4 Othman (4)
40 6 4 3.7 1.9 Soloman (17)
This table presents the rate of progression from postpartum thyroiditis (PPT) to permanent overt hypothyroidism (HYPO) in percent per
year based on long-term follow-up data from five separate studies. It also presents the rate of progression of 1000 normal postpartum women
to permanent hypothyroidism in cases per year by assuming that patients with PPT represent about 5% of their original cohort of all postpartum
a lack of well-being to the postpartum state. In particular, all assessment in the diagnosis and treatment of symptomatic
postpartum women with depression should be tested for transient hyperthyroidism or hypothyroidism, should
thyroid dysfunction as depression can be a life-threatening screening be employed to identify postpartum women who
disorder and needs to be treated promptly. are at risk for permanent overt hypothyroidism? The long-
If a screening program for PPT is employed, what tests term sequelae of PPT have been recently reviewed (29), and
should be used and when should they be done? If all post- the results of those studies that have followed patients with
partum women were screened with a sensitive TSH assay at PPT for 3 yr or longer are summarized in Table 2. Several
2- to 3-month intervals, then nearly all cases of PPT would important observations can be extracted from this table. First,
be detected. However, this strategy is clearly impractical and the mean rate of progression from PPT to permanent overt
too costly. In the United States, where there are 4 million live hypothyroidism over a 5-yr period in the 5 studies is 3.6% per
births per year, this would entail approximately 12 million year. This is very close to the rate at which all adult women
TSH assays at a cost of about 250 million dollars per year. A in the community with a positive TPO-Ab titer progress to
more reasonable strategy would be to screen all postpartum spontaneous overt hypothyroidism as determined in the re-
women with a sensitive thyroid autoantibody assay, such as nowned Wickham survey, which observed rates of 4.3% per
TPO-Ab, and then follow the women testing positive with year when the serum TSH level was elevated, and 2.1% per
several TSH assays during the 1-yr postpartum period. From year when the TSH level was normal (30). Second, the mean
the data in Table 1, this would entail 4 million TPO-Ab assays rate of progression of all postpartum women to permanent
and about 1.2 million TSH assays per year at a cost of about overt hypothyroidism over a 5-yr period in the 5 studies is
100 million dollars per year to detect 200,000 cases of PPT at 1.8 cases/1000 women/yr, very close to the rate at which all
a cost of $500 per case detected, many of which would be women develop spontaneous hypothyroidism (3.5 cases/
mild and would not need treatment. A clinical approach to 1000 women/yr for all women, and 1.4 cases/1000 wom-
detection of PPT would cost 30 –50 million dollars per year, en/yr for young women in their child-bearing years) seen in
assuming that 30 –50% of 4 million postpartum women were the Wickham survey (30). Thus, the rate of progression to
symptomatic and tested with a single TSH assay, and that all spontaneous permanent hypothyroidism in postpartum
postpartum women received a $2 pamphlet alerting them to women with and without PPT is virtually the same as the rate
the symptoms of PPT. It is possible that benefits from a seen in the Wickham survey in all young women with and
program employing clinical astuteness would be comparable without positive TPO-Ab titers.
to a screening program. However, an accurate cost-benefit is
not possible until good studies are available detailing the
clinical presentation of PPT and how they respond to ther-
apy. Unfortunately, no study has yet been published to dem- The time is not yet ripe to recommend a screening program
onstrate that such a screening strategy would be superior in for all postpartum women to detect PPT, even though the
cost or in benefit to thorough clinical assessment of post- disorder is common. The basic problem is that there is a lack
partum women. Although the sensitivity of the TPO-Ab as- of knowledge at present about how to screen, when to screen,
say in identifying women who will develop PPT is generally and whether or not treatment makes a difference for those
about 90%, the range in a number of studies is from 50 –100% women with unrecognized mild or moderate thyroid dys-
(14). In a recent, thorough study from The Netherlands the function. Permanent hypothyroidism occurs at about the
sensitivity was 67%. This means that screening with a same rate of incidence in postpartum women as in all young
TPO-Ab assay would miss 33% of the women who develop women, and in both groups is confined almost exclusively to
PPT and, thus, casts doubt on the recommendation of this those with preexisting AITD. Whether or not young women,
strategy for all countries (24, 25). Certain women with a including postpartum women, should be screened to detect
known high risk for PPT, such as those with type I diabetes those who have or are at risk for permanent hypothyroidism
mellitus, previous autoimmune thyroid disease (AITD), or a remains an open question until further studies clarify the
strong family history of AITD should be followed closely issue.
under any circumstances by thorough clinical and, if neces- Finally, a recommendation against screening postpartum
sary, biochemical scrutiny (26 –28). women for PPT at the present time should not be construed
Even if screening for PPT is not superior to careful clinical to mean that clinical vigilance is not warranted. In general,
THERAPEUTIC CONTROVERSY 1819
the thorough care that women receive during their preg- women with insulin dependent diabetes mellitus. J Clin Endocrinol Metab.
nancy has no counterpart during their first postpartum year, 28. Weetman AP. 1994 Editorial: Insulin-dependent diabetes mellitus and post-
as attention in the family and in the medical system shifts partum thyroiditis: an important association. J Clin Endocrinol Metab. 79:7–9.
from the mother to the infant. Thus, there is a compelling 29. Smallridge RC. 1996 Postpartum thyroid dysfunction: a frequently undiag-
nosed endocrine disorder. The Endocrinologist. 6:44 –50.
need to educate postpartum women and the physicians who 30. Vanderpump MPJ, Tunbridge WMG, French JM, et al. 1995 The incidence of
care for them about PPT and other important postpartum thyroid disorders in the community: a twenty-year follow-up of the Wickham
afflictions. survey. Clin Endocrinol. 43:55– 68.
31. Lazarus JH. 1998 Prediction of postpartum thyroiditis. Eur J Endocrinol.
32. Ball S. 1996 Antenatal screening of thyroid antibodies. Lancet. 349:906 –907.
1. Gerstein HC. 1990 How common is postpartum thyroiditis. A methodologic
overview of the literature. Arch Intern Med. 150:1397–1400. Postpartum Thyroiditis:
2. Guyer B, Martin JA, MacDorman MF, Anderson RN, Strobino DM. 1997 The Case For Selective Screening
Annual summary of vital statistics—1996. Pediatrics. 100:905–918.
3. Tachi J, Amino N, Tamaki H, Aozasa M, Iwatani Y, Miyai K. 1988 Long-term Alex Stagnaro-Green
follow-up and HLA association in patients with postpartum hypothyroidism. Division of Endocrinology and Metabolism
J Clin Endocrinol Metab. 66:480 – 484. Department of Medicine
4. Othman S, Phillips DIW, Parkes JH, et al. 1990 A long-term follow-up of
postpartum thyroiditis. Clin Endocrinol. 32:559 –564. Mount Sinai School of Medicine
5. Roti E, Emerson CH. 1992 Clinical review 29: postpartum thyroiditis. J Clin New York, New York 10029
Endocrinol Metab. 74:3–5.
6. Eddy DM. 1991 How to think about screening. In: Eddy DM, ed. Common OSTPARTUM thyroiditis is one of the most common
screening tests. Philadelphia: American College of Physicians; 1–21. diseases of the thyroid and has important clinical se-
7. Surks MI, Chopra IJ, Mariash CN, Nicoloff JT, Solomon DH. 1990 American
Thyroid Association guidelines for use of laboratory tests in thyroid disorders.
quelae. In studies in North America postpartum thyroiditis
JAMA. 263:1529 –1532. affects between 6.0 – 8.8% of all pregnant women (1–3). Post-
8. Canadian Task Force on the Periodic Health Examination. 1994 The Canadian partum thyroiditis also has clearly defined morbidities.
guide to clinical preventive health care. Ottawa: Canada Communication
Group; 1994:611– 618.
Women with postpartum thyroiditis may experience symp-
9. Report of the U.S. Preventive Services Task Force. 2nd ed. Screening for toms in both the hyperthyroid and hypothyroid phase, and
thyroid disease. 1996 In: Guide to clinical preventive services: an assessment may have an increased incidence of postpartum depression
of the effectiveness of 169 interventions. Washington, DC: US Gov Pr Office;
209 –218. (4, 5). Long-term follow-up reveals that approximately 25%
10. Danese MD, Powe NR, Sawin CT, Ladenson PW. 1996 Screening for mild of women with postpartum thyroiditis develop permanent
thyroid failure at the periodic health examination—a decision and cost-effec- primary hypothyroidism within 5 yr of delivery (6, 7). Fur-
tiveness analysis. JAMA. 276:285–292.
11. American College of Physicians. 1998 Clinical guideline, part 1. Screening for thermore, the presence of thyroid autoantibodies in the first
thyroid disease. Ann Intern Med. 129:141–143. trimester of pregnancy is associated with an increased rate of
12. Helfand M, Redfern CC. 1998 Clinical guideline, part 2. Screening for thyroid
disease: an update. Ann Intern Med. 129:144 –158.
spontaneous abortion (8 –11). Given both the prevalence of
13. Cooper DS. 1998 Subclinical thyroid disease: a clinician’s perspective. Ann the disorder and the associated morbidities, the argument for
Intern Med. 129:135–138. screening for postpartum thyroiditis seems iron clad, and
14. Stagnaro-Green A. 1993 Postpartum thyroiditis—prevalence, etiology, and
clinical implications. Thyroid Today. 16:1–11. one can only wonder why a screening program for postpar-
15. Jansson R, Dahlberg PA, Karlson FA. 1988 Postpartum thyroiditis. Bailliere’s tum thyroiditis has not already been instituted.
Clin Endocrinol Metab. 2:619 – 635. On the other hand, screening for postpartum thyroiditis is
16. Nikolai TF, Turney SL, Roberts RC. 1987 Postpartum lymphocytic thyroiditis-
prevalence, clinical course, and long-term follow-up. Arch Intern Med. an enormous undertaking with tremendous fiscal implica-
147:221–224. tions. Every pregnant women would need to be screened for
17. Solomon BL, Fein HG, Smallridge RC. 1993 Usefulness of antimicrosomal
antibody titers in the diagnosis and treatment of postpartum thyroiditis. J Fam
the presence of thyroid peroxidase autoantibodies. Women
Pract. 36:177–182. who test positive would require a minimum of two mea-
18. Amino N, Mori H, Iwatani Y, et al. 1982 High prevalence of transient post- surements of TSH in the postpartum (3 and 6 months). Fur-
partum thyrotoxicosis and hypothyroidism. N Engl J Med. 306:849 – 852.
19. Hayslip CC, Fein HG, O’Donnell VM, et al. 1988 The value of serum anti- thermore, the following three unproven assumptions under-
microsomal antibody testing in screening for symptomatic postpartum thyroid lie the argument for screening; a) that the optimal screening
dysfunction. Amer J Obstet Gynecol. 159:203–209. strategy is known, b) that treating women with postpartum
20. Lazarus JH, Hall R, Othman S, et al. 1996 The clinical spectrum of postpartum
thyroid disease. QJ Med. 89:429 – 435. thyroiditis would decrease the incidence or severity of post-
21. Pop VJM, de Rooy HAM, Vader HL, et al. 1991 Postpartum thyroid dys- partum depression, and c) that treating postpartum thyroid-
function and depression in an unselected population. N Engl J Med. 324:1815–
itis would decrease the incidence of, or the symptoms asso-
22. Harris B, Othman S, Davies JA, et al. 1992 Association between postpartum ciated with, long-term primary hypothyroidism. Given these
thyroid dysfunction and thyroid antibodies and depression. BMJ. 305:152–156. arguments, it now appears that screening for postpartum
23. Pop VJM, de Rooy HAM, Vader HL, van der Heide D, van Son MM, Kom-
proe IH. 1993 Microsomal antibodies during gestation in relation to postpar- thyroiditis is premature and a questionable use of limited
tum thyroid dysfunction and depression. Acta Endocrinol. 129:26 –30. fiscal resources.
24. Kuijpens JL, Pop VJ, Vader HL, Drexhage HA, Wiersinga WM. 1998 Pre- Given the powerful but conflicting, scientific, fiscal, and
diction of postpartum thyroid dysfunction: can it be improved? Eur J Endo-
crinol. 139:36 – 43. emotional arguments for screening for postpartum thyroid-
25. Kuijpens JL, de Haan-Meulman M, Vader HL, Pop VJ, Wiersinga WM, itis the clinician is left in a quandary. A rational approach to
Dreshage HA. 1998 Cell-mediated immunity and postpartum thyroid dys-
function: a possibility for the prediction of disease? J Clin Endocrinol Metab.
resolve the screening debate requires the following four
83:1959 –1966. questions be addressed:
26. Gerstein HC. 1993 Incidence of postpartum thyroid dysfunction in patients
with type I diabetes mellitus. Ann Intern Med. 118:419 – 423.
27. Alvarez-Marfany M, Roman SH, Drexler AJ, Robertson C, Stagnaro-Green 1. Is postpartum thyroiditis a common enough clinical
A. 1994 Long-term prospective study of postpartum thyroid dysfunction in entity to warrant screening?
1820 THERAPEUTIC CONTROVERSY JCE & M • 1999
Vol 84 • No 6
2. Are there important morbidities associated with post- creased incidence of thyroid autoantibodies in women with
partum thyroiditis? recurrent abortion (18 –22).
3. Can the morbidities of postpartum thyroiditis be pre-
vented with levothyroxine therapy?
4. Is there an inexpensive, easily available, and accurate Intervention/prevention
screening test? The impact of levothyroxine therapy on the short- and
long-term complications of postpartum thyroiditis is largely
Prevalence unknown. Beta blockers during the hyperthyroid phase and
levothyroxine therapy in women who are hypothyroid are
Worldwide, the prevalence of postpartum thyroiditis var-
effective interventions. However, whether or not levothy-
ies widely, from 1.1% in Thailand to 16.7% in the United
roxine therapy alters the incidence or severity of the depres-
Kingdom (12, 13). The disparity is explained both by true
sion associated with postpartum thyroiditis has not been
geographic differences in prevalence rates as well as by crit-
ical study design issues such as the length of follow-up in the
Studies attempting to prevent the occurrence of postpar-
postpartum period and the frequency of screening. Studies
tum thyroiditis through the administration of iodide or levo-
performed in the United States and Canada that extend be-
thyroxine in antibody positive women have been unsuccess-
yond three months postpartum have revealed similar prev-
ful (23). Furthermore, at present there are no known
alence rates (6% in Canada, 6.7% in Wisconsin, and 8.8% in
interventions which, administered during the hyperthyroid
the New York Metropolitan area) (1–3). In women with Type
or hypothyroid phase, would result in a decrease in the high
I diabetes mellitus, another autoimmune disorder, the prev-
rate of permanent hypothyroidism.
alence in North America is substantially higher at 25%
A recent study has demonstrated a decrease in the rate of
recurrent abortion in women who were thyroid antibody
The highest prevalence rates of postpartum thyroiditis are
positive through the administration of intravenous immu-
found in women with a prior history of postpartum thy-
noglobulin before conception and throughout the first 8
roiditis. Lazarus et al. (16) found that 69% of women who had
months of pregnancy (24). Replication of these results are
a history of postpartum thyroiditis developed a recurrence
required before their implementation outside of a research
with a subsequent pregnancy. Twenty-five percent of
women who were antibody positive in the initial pregnancy,
but euthyroid in the postpartum period, developed postpar-
tum thyroiditis after the next delivery. Screening
A successful screening strategy for postpartum thyroiditis
Associated morbidities depends on the availability of an accurate test. As defined in
Women with postpartum thyroiditis develop both short- the Guide to Clinical Preventive Services, accuracy relates to the
term and long-term complications. Although few women sensitivity, specificity, and reliability of a screening test. In
require treatment in the hyperthyroid phase, and in fact short, “The test must be able to detect the target condition
many women are not diagnosed until they develop hypo- earlier than without screening and with sufficient accuracy
thyroidism, studies have revealed an increase in palpitations, to avoid producing large numbers of false-positive and false-
heat intolerance, and tremulousness during the hyperthy- negative results. . .” (25).
roid period (3). Classical symptoms of hypothyroidism, as Thyroid peroxidase antibody is the potential screening test
well as concentration and memory deficits, are present in the of choice for postpartum thyroiditis. It is widely available,
hypothyroid phase of postpartum thyroiditis (5). The final easily reproducible, and relatively cheap. Screening for thy-
short-term complication is an increased incidence of post- roid peroxidase antibodies would need to occur early in
partum depression. Studies have indicated a 38 –53% inci- pregnancy in order to identify women who were going to
dence of nonpsychotic depression in women affected with develop postpartum thyroiditis before its manifestation.
postpartum thyroiditis (4, 5). One study has even found an Given the dramatic and well documented decrease (often to
increased incidence of postpartum depression in women undetectable levels) in thyroid antibodies during pregnancy,
who were thyroid autoantibody positive but who did not screening at delivery would miss a large percentage of the
develop postpartum thyroiditis (17). Finally, long-term fol- cases and consequently would have an unacceptably high
low of women with postpartum thyroiditis has revealed an false negative rate.
incidence of long term primary hypothyroidism of approx- Five studies have prospectively followed a cohort of an-
imately 25% (6, 7). tibody positive and antibody negative women during preg-
The correlation between thyroid autoantibodies and an nancy and into the postpartum (1, 13, 26 –28). The positive
increase in spontaneous abortions is also pertinent to the predictive value of thyroid peroxidase (defined as the per-
screening debate. Four studies have revealed a 2- to 3-fold centage of women who have thyroid antibodies during
doubling in the rate of spontaneous miscarriage in un- pregnancy and who subsequently develop postpartum thy-
selected women who screened positive for thyroid autoan- roiditis) was relatively low and ranged from 30 –52%. Fur-
tibodies in the first trimester of pregnancy (8 –11). In women thermore, 9 –39% of the women who developed postpartum
with recurrent abortion, defined as three or more spontane- thyroiditis in these studies were antibody negative during
ous miscarriages without an intervening live birth, the data pregnancy. The positive predictive value of thyroid perox-
has been mixed. Three of five studies have shown an in- idase could be increased by focusing on women with high
THERAPEUTIC CONTROVERSY 1821
titers of thyroid peroxidase, but this would result in an in- function and depression in an unselected population. N Engl Med.
creased false negative rate. 5. Hayslip CC, Fein HG, O’Donnell VM, Friedman DS, Klein TA, Smallridge
The positive predictive value of thyroid peroxidase in RC. 1988 The value of serum antimicrosomal antibody testing in screening for
women with Type I diabetes mellitus was also limited, rang- symptomatic postpartum thyroid dysfunction. Am J Obstet Gynecol.
ing from 33– 67% (14, 15, 29). Similarly, 17–57% of women 6. Tachi, J, Amino, N, Tamaki, H, Aozasa M, Iwatani Y, Miyai K. Long term
with Type I diabetes who developed postpartum thyroiditis, Follow-Up and HLA Association in Patients With Postpartum Hypothyroid-
tested negative for thyroid peroxidase antibodies during ism. Journal of Clinical Endocrinology and Metabolism. 1988;66:480 – 484.
7. Othman S, Phillips DIW, Parkes AG, et al. 1990 A long-term follow-up of
pregnancy. postpartum thyroiditis. Clin Endocrinol. 32:559 –564.
8. Stagnaro-Green A, Roman SH, Cobin, El-Harazy E, Alvarez-Marfany M,
Davies, TF. 1990 Detection of at-risk pregnancy by means of highly sensitive
Conclusion assays for thyroid autoantibodies. JAMA. 264:1422–1425.
Despite the high prevalence of postpartum thyroiditis in 9. Glinoer D, Soto MF, Bourdoux P, et al. 1991 Pregnancy in patients with mild
thyroid abnormalities: maternal and neonatal repercussions. J Clin Endocrinol
the general population, screening of all women cannot be Metab. 73:421– 427.
justified at present. The limited positive predictive value of 10. Singh A, Dantas NZ, Stone SC, Asch RH. 1995 Presence of thyroid antibodies
in early reproductive failure: biochemical versus clinical pregnancies. Fertil
thyroid peroxidase antibodies, the presence of a substantial Steril. 63:277–281.
number of antibody negative women who develop postpar- 11. Iijima T, Tada H, Hidaka Y, Mitsuda N, Murata Y, Amino N. 1997 Effects of
tum thyroiditis, and the unproven efficacy of levothyroxine autoantibodies on the course of pregnancy and fetal growth. Obstet Gynecol.
in preventing postpartum depression or long-term primary 12. Rajatanavin R, Chailurkit L, Tirarungsikul K, Chalayondeja W, Jittivanich
hypothyroidism, makes screening all pregnant women un- U, Puapradit W. 1990 Postpartum thyroid dysfunction in Bangkok: A geo-
tenable. Furthermore, the conflicting data on the association graphical variation in the prevalence. Acta Endocrinol. 122:283–287.
13. Fung HYM, Kologlu M, Collison K, et al. 1988 Postpartum thyroid dysfunc-
of thyroid antibodies and recurrent abortion (despite the tion in Mid Glamorgan. BMJ. 296:241–244.
clear association between thyroid antibodies and spontane- 14. Alverez-Marfany M, Roman SH, Drexler AJ, Robertson C, Stagnaro-Green
A. 1994 Long-term prospective study of postpartum thyroid dysfunction in
ous miscarriage in unselected women) requires further elu- women with insulin dependent diabetes mellitus. J Clin Endocrinol Metab.
cidation before it could be used as a rationale for screening. 79:10 –16.
Two specific populations however, would clearly benefit 15. Gerstein HC. 1993 Incidence of postpartum thyroid dysfunction in physicians.
Ann Intern Med. 118:419 – 423.
from testing for postpartum thyroiditis. Specifically, women 16. Lazarus JH, Ammari F, Oretti R, Parkes AB, Richards CJ, Harris B. 1997
with a prior history of postpartum thyroiditis (prevalence of Clinical aspects of recurrent postpartum thyroiditis. British J Gen Pract.
recurrent thyroiditis— 69%) and individuals with Type I di- 47:305–308.
17. Harris B, Othman S, Davies JA. 1992 Association between postpartum thyroid
abetes mellitus (prevalence of postpartum thyroiditis—25%) dysfunction and thyroid antibodies and depression. BMJ. 305:152–156.
should have TSH determinations at 3 and 6 months post- 18. Bussen S, Steck T. 1995 Thyroid autoantibodies in euthyroid non-pregnant
women with recurrent spontaneous abortions. Hum Reprod. 10:2983–2940.
partum. In these select groups screening with thyroid per- 19. Pratt D, Novotny M, Kaberlein G, Dudkiewicz A, Gleicher N. 1993 Anti-
oxidase antibodies offers no advantages. Women who ex- thyroid antibodies and the association with non-organ specific antibodies in
hibit either suppressed or elevated TSH levels would require recurrent pregnancy loss. Am J Obstet Gynecol. 168:837– 841.
20. Pratt DE, Kaberlein G, Dudkiewicz A, Karande V, Gleicher N. 1993 The
further evaluation. Whether or not women with a history of association of antithyroid antibodies in euthyroid nonpregnant women with
another autoimmune disorder (such as systemic lupus ery- recurrent first trimester abortions in the next pregnancy. Fertil Steril.
thematosus or Sjogren’s disease), or who have a strong fam- 60:1001–1005.
21. Roberts J, Jenkins C, Wilson R. 1996 Recurrent miscarriage is associated with
ily history of autoimmune disease, should be screened is at increased numbers of CD5/20 positive lymphocytes and an increased inci-
present an unresolved question. dence of thyroid antibodies. Eur Endocrinol. 134:84 – 86.
22. Esplin MS, Branch DW, Silver R, Stagnaro-Green A. 1998 Thyroid autoan-
Research in the future should focus on refining screening tibodies are not associated with recurrent pregnancy loss. Am J Obstet Gy-
strategies for postpartum thyroiditis and determining the necol. 179:1583–1586.
efficacy of levothyroxine in preventing or ameliorating post- 23. Kampe O, Jansson R, Karlsson FA. 1990 Effects of l-thyroxine and iodide on
the development of autoimmune postpartum thyroiditis. J Clin Endocrinol
partum depression and primary hypothyroidism. In the in- Metab. 70:1014 –1018.
terim, a dedicated effort to educate obstetricians, internists, 24. Kiprov DD, Nachtigall RD, Weaver RC, Jacobson A, Main EK, Garovoy MR.
pediatricians, family practitioners, and psychiatrists is 1996 The use of intravenous immunoglobulin in recurrent pregnancy loss
associated with combined alloimmune and autoimmune abnormalities. Am J
needed so that the acute symptoms associated with the hy- Reprod Immunol. 36:228 –234.
perthyroid and hypothyroid phase can be recognized, diag- 25. Report of the U.S. Preventive Services Task Force. Guide to Clinical Pre-
ventive Services, 2nd ed. Baltimore: Williams and Wilkins; 1996.
nosed, and appropriately treated. 26. Kuijpens JL, Hann-Meulman MDH, Vader HL, Pop VJ, Wiersinga WM,
Drexhage HA. 1998 Cell-mediated immunity and postpartum thyroid dys-
References function: a possibility for the prediction of disease? J Clin Endocrinol Metab.
1. Stagnaro-Green A, Roman, SH, Cobin, RH, El-Harazy, E, Wallenstein S, 27. Pop VJM, de rooy HAM, Vader HL, van der Heide D, van Son MM, Komproe
Davies, TF. 1992 A Prospective study of lymphocyte-initiated immunosup- IH. 1993 Microsomal antibodies during gestation in relation to postpartum
pression in normal pregnancy: evidence of a T-cell etiology for postpartum thyroid dysfunction and depression. Acta Endocrinol. 129:26 –30.
thyroid dysfunction. J Clin Endocrinol Metab. 74:645– 653. 28. Rasmussen NG, Hornnes PJ, Hoier-Madsen M, Feldt-Ramussen U, Hegedus
2. Nikolai TF, Turney SL, Roberts RC. 1987 Postpartum lymphocytic thyroiditis: Laszlo. 1990 Thyroid size and function in healthy pregnant women with
prevalence, clinical course, and long-term follow-up. Arch Intern Med. thyroid autoantibodies. Relation to development of postpartum thyroiditis.
147:221–224. Acta Endocrinol. 123:395– 401.
3. Walfish PG, Meyerson J, Provias JP, Vargas MT, Papsin FR. 1992 Prevalence 29. Bech K, Høier-Madsen M, Feldt-Rasmussen U, Jensen BM, Mølsted-
and characteristics of post-partum thyroid dysfunction: results of a survey Pedersen L, Kuhl C. 1991 Thyroid function and autoimmune manifestations
from Toronto, Canada. J Endocrinol Invest. 15:265–272. in insulin-dependent diabetes mellitus during and after pregnancy. Acta En-
4. Pop VJM, De Rooy HAM, Vader HL, et al. 1991 Postpartum thyroid dys- docrinologica. 124:534 –539.