Transient hyperthyroidism and hyperemesis gravidarum: Clinical aspects

T. Murphy Goodwin, MD, Martin Montoro, MD, and Jorge H. Mestman, MD

Los Angeles, California

OBJECTIVES: Our objectives were to describe the presentation and course of hyperemesis gravidarum

with respect to thyroid function and to test the hypothesis that patients with biochemical hyperthyroidism differ in clinical presentation from euthyroid hyperemesis patients.

STUDY DESIGN: Sixty-seven patients seen at Los Angeles County Women's Hospital over a 10-month period with hyperemesis gravidarum were studied prospectively with respect to thyroid function.

RESULTS: Forty-four patients (66%) had biochemical hyperthyroidism (increased free thyroxine index

[n = 39) or suppressed thyroid-stimulating hormone [n = 40)) that was self-limited, resolving by 18 weeks' gestation. Hyperthyroid patients were more likely than euthyroid patients to have abnormal electrolyte

levels (23/39 [59%) vs 6/28 [21%) and increased liver enzyme levels (23/59 [59%) vs 5/28 [18%),

P < 0.01). The severity of hyperemesis was found to vary directly with the degree of hyperthyroidism. CONCLUSIONS: Hyperthyroidism is a common, self-limited finding in hyperemesis. The cause of the hyperthyroidism may be linked to the cause of hyperemesis itself. (AM J OBSTET GVNECOL 1992;167: 648-52.)

Key words: Hyperthyroidism, hyperemesis gravidarum, pregnancy, thyroid, vomiting

Over the past 10 years no more than 70 patients with transient hyperthyroidism and hyperemesis gravida­rum have been described in several case reports l

-3 and

four series,,-7 (Two other groups, Lao et al. 8 and Chin and Lao,9 apparently described different aspects of the same group of patients reported on by Swaminathan et al. 7) In spite of the recognition of this association important questions regarding the presentation and management of hyperthyroidism and hyperemesis are unresolved. Incidence data differ widely (21 % to 70%). Overlap with subclinical intrinsic thyroid disease has not been systematically ruled out. Antithyroid medi­cation has been advocated for a subset of hyperthy­roxinemic hyperemesis patients, although with scant justification .. ' 8, 9 Little is known about the status of ul­trasensitive thyroid-stimulating hormone (TSH) in hy­peremesis patients, although this test is now used rou­tinely in the study of patients with thyroid disorders. Finally, although Mori et al. 10 have shown a relationship between the severity of vomiting and the degree of thyroid stimulation in simple "morning sickness," such a relationship has not been found among hyperemesis patients .. ' 5, 8, 9 This may be due to the difficulty in ob-

From the Division of Maternal-Fetal Medicine, University of South­ern California. Received for publication October 18, 1991; revised March 2, 1992; accepted March 13, 1992. Reprint requests: T. Murphy Goodwin, MD, Division of Maternal­Fetal Medicine, Women's Hospital, Room 5K-40, 1240 N. Mission Road, Los Angeles, CA 90033. 6/1/37861

648

jectively distinguishing grades of severity of vomiting. This association may be important in understanding the cause of hyperemesis.

A prospective study was undertaken to describe the clinical and biochemical presentation and natural his­tory of hyperemesis gravidarum with respect to thyroid function and to test the hypothesis that patients with evidence of biochemical hyperthyroidism differ in clin­ical presentation from those without this finding_

Material and methods

Hyperemesis gravidarum was defined as persistent vomiting with large ketonuria on dipstick examination and >5% weight loss in women presenting at < 16 weeks' gestation. All patients had onset of vomiting in the first trimester and no other cause for the vomiting. All patients were well before the onset of vomiting. Patients with a history of thyroid disease, abnormal results of thyroid examination or presence of thyroid antibodies, fever, evidence of other condition respon­sible for vomiting, or multiple gestation were excluded from this study.

Patients with a preliminary diagnosis of hyperemesis gravidarum admitted to the emergency room holding area for intravenous hydration were interviewed at ran­dom and examined by a single physician (T.M.G.) to determine eligibility for the study. No patient was tak­ing proprietary antiemetic formulations at the time of initial evaluation. Pelvic ultrasonography was per­formed in each case to confirm gestational age and a

Volume 167 Number 3

normal-appearing singleton gestation. The fOllowing tests were performed at the time of admission: serum sodium, potassium, bicarbonate, blood urea nitrogen, aspartate aminotransferase, alanine aminotransferase, total bilirubin, and serum amylase. Total thyroxine, to­tal triiodothyronine, resin triiodothyronine uptake, and sensitive TSH (Tandem High Sensitivity, Hybritech, San Diego) were drawn after rehydration. Patients with abnormal electrolyte levels remained in hospital until those levels normalized. Patients with elevated free thy­roxine index or suppressed TSH were tested for an­timicrosomal antibodies, antithyroglobulin antibodies (Direct addition radioimmunoassay kit, Kronus, Dana Point, Calif.), and TSH receptor antibodies (competi­tive binding kit, Kronus). These patients were evalu­ated serially with thyroid function studies. Patients with elevated aminotransferase levels were studied for hep­atitis B surface antigen, anti-hepatitis A immunoglob­ulin M, and hepatitis B core antibody immunoglobulin M. All patients with abnormal results of thyroid or liver studies were given follow-up appointments in the high­risk obstetrics clinic within 2 weeks.

Comparisons between groups were analyzed with the Student t test or Mann-Whitney U test for nonnormal distributions. For analysis of the relation among groups according to the severity of vomiting and the degree of thyroid stimulation, one-way analysis of variance was used. The study was approved by the Institutional re­view board of the University of Southern California. Unless otherwise stated, all laboratory data refer to the date of initial presentation.

Results

In the lO-month period during 1989 and 1990,92 patients were examined. Seventy-four were found to satisfy the study criteria, and 67 had data collection suitable for analysis. Of the original 92 patients, two were excluded for evidence of thyroid illness (Grave's disease, primary hypothyroidism), two had multiple gestation, and 15 had evidence of other acute illness. One patient was excluded because of the presence of anti microsomal antibodies. Three patients refused to have the required blood drawn, and specimens were misplaced in two cases.

Table I shows the characteristics of the population. Hyperthyroid hyperemesis patients did not differ from others with respect to gestational age at presentation, parity, number of days of vomiting, or amount of weight loss. Table II shows the spectrum of biochemical abnormalities in patients with hyperemesis at initial presentation. Of the 39 patients with elevated free thy­roxine index at initial presentation, two had elective abortions in the first trimester and five were lost to follow-up. Thirty-two patients underwent follow-up to

Hyperthyroidism and hyperemesis gravidarum 649

Table I. Population characteristics (n = 67)

Age (yr) Mean ± SO Range

Estimated gestational age (wk) Mean ± SO Range

Height (in) Mean ± SO Range

Weight (lb) Mean ± SD Range

Prepregnancy weight by patient history (lb) Mean ± SO Range

Race (No. and %) Hispanic Asian Black

Parity (No. and %) o 1 2 >2

24.0 ± 5.0 16-37

9.7 ± 2.2 6-15.9

61.6 ± 3.0 55-70

116.9 ± 22.0 82-192

127.7 ± 22.2 96-202

63 (94%) 1 (2%) 3 (4%)

39 (58%) 9 (13%) 8 (12%)

11(17%)

normalization of the free thyroxine index over a period of 1 to 10 weeks. No patient had physical findings spe­cific for thyrotoxicosis (goiter, tremor, lid lag, proximal muscle weakness). None received antithyroid medica­tion. All patients for whom follow-up was available had documentation of normal free thyroxine index by 18 weeks' gestation. In no case did evidence of hyperthy­roidism persist after resolution of vomiting, although seven patients had continued vomiting for 2 to 14 weeks after normalization of thyroid studies.

Of 28 patients with elevated aminotransferase levels, two had elective abortions and seven were lost to follow­up. Nineteen underwent follow-up to resolution over 2 to 9 weeks. None had serologic evidence of active viral hepatits. Six patients had a mildly elevated amylase level but no other signs or symptoms of pancreatitis. The amylase level returned to normal within one week in five of six cases, and the clinical course of these patients did not differ from other patients with hyper­emesis gravidarum.

Hyperthyroid patients with hyperemesis had a higher mean arterial pressure (89 ± 11 vs 81 ± 10 mm) and pulse (97 ± 15 vs 82 ± 9 beats/min, p < 0.01) but the difference resolved promptly with hydration in all but two patients. Those with biochem­ical evidence of hyperthyroidism were significantly more likely to exhibit biochemical abnormalities on ad­mission, as shown in Table Ill. Conversely, patients with more severe hyperemesis exhibited a greater de­gree of biochemical hyperthyroidism. This is shown in Fig. 1, where patients are grouped according to the

650 Goodwin, Montoro, and Mestman September 1992 Am J Obstet Gynecol

Table II. Biochemical abnormalities in hyperemesis gravida rum (n = 67)

Thyroid Free thyroxine index

>13.2 Free triiodothyronine in­

dex >225 . TSH (fLU / ml)

Undetectable Suppressed

Free thyroxine index> 13.2 or TSH <0.4 fLU/ml

Electrol ytes Sodiu'm <135 mEq/L Potassium <3.2 mEq/L Chloride <99 mEq/L Bicarbonate >26 mEg/L

Liver / gastrointestinal Aspartate aminotransferase

or alanine aminotransferase >40 U/L

Total bilirubin> 1.0 mg/dl Amylase> 150 U / dl

39

6

20 40 44

19 \0 16 \0

28

14 6

'l<

59

9

30 60 66

28 15 24 15

42

21 9

1:\.2-37.6

225-350

<0.04 <0.4

125-134 2.3-3.1 8(J-98 27-34

41-324

1.1-5.:1 151-391

Table III. Electrolyte and liver function abnormalities in patients with hyperemesis with and without hyperthyroidism

Free th~ruxine Fret' thyroxine index >13.2 index <13.2

(u = 39) (/I = 28)

1\'0. I o/c No. I % P Value

Sodium <135 mEq/L 13 33 4 14 0.16 Potassium <3.2 mEg/L 9 23 I 4 <(J.OI Chloride <99 mEq/L 14 36 I 4 <(J.OI Bicarbonate >26 mEq/L 8 21 2 7 0.13

TOTAL* 23 59 6 21 <0.01

Alanine aminotransferase or aspartate aminotransferase >40 U / L 23 59 5 18 <0.01 Total bilirubin> 1.0 mg/dl Amylase >150 U/dl

*Total number of patients with any electrolyte abnormality.

severity of hyperemesis. Patients with hyperemesis who had normal electrolyte levels and liver function test results (n = 23) were classified as having mild hyper­emesis. Those with either abnormal electrolyte levels or abnormal liver function test results (n = 32) were classified as having moderate hyperemesis and those with both liver function abnormalities and electrolyte abnormalities that persisted after initial hydration, ne­cessitating hospitalization, were classified as severe. These groups did not differ significantly with respect to gestational age.

Comment

In this, the largest series of hyperemesis subjects studied prospectively with respect to thyroid function, we have attempted to clarify certain aspects of the clin­ical presentation and management in the syndrome of

14 36 0 0 <0.01 5 13 I 4 0.19

transient hyperthyroidism of hyperemesis gravidarum. The 70% incidence of biochemical hyperthyroidism in our study is in agreement with the findings of Bouillon et al! In other studies, where weight loss was not a criterion for the diagnosis of hyperemesis;- 7 the inci­dence was considerably lower. Once the abnormal thy­roid studies have been obtained, the differential di­

agnosis can be difficult. Patients with transient hyper­thyroidism of hyperemesis gravidarum will have no history consistent with thyroid illness before pregnancy. They lack findings suggestive of tissue thyrotoxicosis and thyroid antibodies.

The sensitive TSH level was abnormal in 60% of our patients. In the only other report to consider ultrasen­sitive TSH in hyperemesis, Bober et aI.' found widely discrepant results, depending on the type of kit used (80% concordance with free thyroxine vs 20%). We

Volume 167 Number 3

found 85% concordance between an elevated free thy­roxine and suppressed TSH in subjects with hyper­emesis.

Consistent with prior reports, patients with hyper­thyroid hyperemesis were found to lack specific signs of thyrotoxicosis. This has been attributed by some to the relatively low triiodothyronine reported in this con­dition. We have confirmed that the free triiodothyro­nine index is less commonly elevated in hyperemesis than is the free thyroxine index. Nevertheless, the mean free triiodothyronine index was higher than that seen in the sick euthyroid syndrome and highest among those patients most severely affected (Fig. 1). II This would suggest either that peripheral conversion to tri­iodothyronine is spared or that the thyroid is being stimulated directly.

The time course to resolution of hyperthyroidism among our patients (l to 10 weeks) agrees with what has been reported for the majority of those described in the literature. In contrast, however, where 15% to

25% of patients in other series" ,. 8 had persistence of hyperthyroidism past midgestation or were otherwise thought to require specific antithyroid therapy, none of 32 patients with follow-up until resolution of hy­perthyroidism in the current study required such treat­ment. It is possible that the subset of patients with per­sistent biochemical hyperthyroidism in other reports had other underlying thyroid disease that would have been excluded by the methods of the current study.

That thyroid function may vary with the severity of vomiting has been debated in the literature. Evans et al. I~ observed no relation between thyroid function and the severity of vomiting in 342 pregnant women. In contrast, Mori et al. lo found greater evidence of thyroid stimulation (lower TSH, increased free thyroxine) among those patients with more severe "morning sick­ness ." Neither author studied patients with hyper­emesis. Among patients with hyperemesis, no relation­ship between the degree of hyperthyroidism and the severity of vomiting has been previously described.··g

We have identified a significant association between the degree of thyroid stimulation and the severity of vom­iting, as indicated principally by disturbances in serum electrolyte levels and liver function test results. Elec­trolyte alterations may acutely reflect the volume of emesis whereas alterations in liver status are thought to reflect the more chronic process of starvation-in­duced hepatic injury. The latter changes have been associated with more severe hyperemesis. I' Hyperthy­roid hyperemesis patients were more likely to have ab­normalities in electrolyte levels and liver status test re­sults, and conversely, those patients with most severe aberrations in these parameters demonstrated the greatest degree of biochemical hyperthyroidism.

In view of the latter association, it may well be asked whether hyperthyroidism is a cause of the vomiting and

25

20

15

10

250

200

150

100

1.5

1.0

"U/ml

0.5

Hyperthyroidism and hyperemesis gravidarum 651

*

TSH

MILD (N=23) MODERATE (N=32) SEVER!: (N=12)

0.0+--- • 1-Fig. 1. Hormone levels are stratified by severity of vomiting (mean ± SEM). Mild, Normal electrolyte, alanine aminotrans­ferase, and aspartate aminotransferase levels; moderate. ab­normal electrolyte, alanine aminotransferase, or aspartate aminotransferase levels; severe. abnormal electrolyte levels and abnormal aspartate aminotransferase or alanine aminotrans­ferase levels plus admission > 24 hours fo r persistent electro­lyte abnormalities. }T41, Free thyroxine index; FT31, free tri­iodothyronine index; asterisk. each of three groups differed from others with respect to free thyroxine index, free tri­iodothyronine index, and TSH (p < 0.05).

whether the course of hyperthyroid heperemesis might be ameliorated by antithyroid therapy. There are sev­eral problems with this proposition. Although patients with overt hyperthyroidism may be first seen for vom­iting,14 it is distinctly uncommon. Most patients with vomiting in pregnancy have normal thyroid function , 10

and even in those with evidence of elevated thyroid function tests vomiting can persist after resolution of the biochemical hyperthyroidism. In addition, overtly hypothyroid patients who become pregnant may, in our experience, present with severe vomiting. Therefore

652 Goodwin, Montoro, and Mestman

treatment with antithyroid medication would not seem justified.

The converse, that vomiting itself might induce the thyroid stimulation, would serve to explain the asso­ciation between severity of vomiting and degree of bio­chemical hyperthyroidism, but evidence for such an effect of vomiting is lacking. In fact, we have found no evidence of biochemical hyperthyroidism among 10 first-trimester pregnant women with severe vomiting not caused by hyperemesis. Thus it seems likely that that the source of the hyperthyroidism is closely linked to the cause of the vomiting but is not causative in and of itself. It is possible that human chorionic go­nadotropin, acting through distinct mechanisms, plays a role in the causation of both hyperthyroidism and hyperemesis.

REFERENCES

I. Valentine BH, jones C, Tyack A]. Hyperemesis gravi­darum due to thyrotoxicosis. Postgrad Med 1980;56: 746-7.

2. Dozeman R, Kaiser FE, Cass 0, Pries]. Hyperthyroidism appearing as hyperemesis gravidarum. Arch Intern Med 1983; 143:2202-3.

3. jeffcoate WJ, Bain C. Recurrent pregnancy-induced thy­rotoxicosis presenting as hyperemesis gravidarum. Br j Obstet Gynaecol 1985;92:413-5.

September 1992 Am J Obstet Gyneco1

4. Bouillon R, Naesens M, van Assche FA, et al. Thyroid function in patients with hyperemesi, gravidarum. AM j OBSTET GYNECOL 1982; 143:922-96.

5. Bober SA, McGill AC, Turnbridge WMG. Thyroid func­tion in hyperemesis gravidarum. Acta Endocrinol 1986; III :404-10.

6. Montoro M, Spencer C, Jacobson S, Alexander D, Mest­man .IH, Nicoloff ]. Evidence for a physiologic role of hCG as a thyroid stimulator in hyperemesis gravidarum. Clin Res 1984;32:20A.

7. Swaminathan R. Chin RK, Lao TTH, Mak YT, Panesar NS, Cockram CS. Thyroid function in hyperemesis grav­idarum. Acta Endocrinol 1989;120:155-60.

8. Lao TT, Chin RKH, Chang AMZ. The outcome of hy­peremetic pregnancies complicated by transient hyper­thyroidism. Aust N ZJ Obstet GynaecoI1987;27:99-101.

9. Chin RKH, Lao TTH. Thyroxine concentration and out­come of hyperemetic pregnancies. Br .I Obstet Gynaecol 1988;95:507 -9.

10. Mori M, Amino M, Tamaki O. Morning sickness and thy­roid function in normal pregnancy. Obstet Gynecol 1988;72:355-9.

II. Chopra Ij, Hershman .1M, Pard ridge WM, et al. UCLA conference: thyroid function in non thyroidal illnesses. Ann Intern Med 1983;98:946-50.

12. Evans Aj, Li TC, Selby C,Jeffcoate W]. Morning sickness and thyroid function. Br j Obstet Gynaecol 1986;93: 520-2.

13. Adams RH, Gordon j, Combes B. Hyperemesis gravi­darum. I. Evidence of hepatic dysfunction. Obstet Gy­necol 1968;31 :659-64.

14. Rosenthal FD, Jones C, Lewis SI. Thyrotoxic vomiting. BMj 1976;2:209-11.