THYROIDVolume 1, Number 4, 1991Mary Ann Liebert, Inc., Publishers

Serum Thyroid-Stimulating Antibody, Thyroglobulin Levels, andThyroid Suppressibility Measurement as Predictors of theOutcome of Combined Methimazole and Triiodothyronine

Therapy in Graves' Disease

RUBENS S. WERNER, JOAO H. ROMALDINI, CHADY S. FARAH, MARISA C. WERNER, andNAZARETH BROMBERG

ABSTRACT

The value of the criteria used to anticipate the outcome of treatment of Graves' hyperthyroid patients withmethimazole (MMI) remains controversial. We have reported that high MMI doses combined with T3administration was correlated with higher remission rates. In this study, we used the lowest MMI dose able tocontrol the hyperthyroidism, keeping the free T4 index (FT4I) values below the normal range throughouttreatment, and compared the results with patients treated with a high MMI regimen. Both groups received T3.We also evaluated the usefulness of goiter size, serum thyroid-stimulating antibody (TSAb: adenylate cyclasestimulation in human thyroid membrane), thyroglobulin (Tg) levels, and the T3 suppressibility of 24 h RAIU as

prognostic markers for the outcome of Graves' disease therapy.Twenty-four Graves' hyperthyroid patients were treated with high MMI dose (mean ± SD 60 ± 19, range

40-120 mg daily), and 25 patients received low MMI dose (17 ± 4.3, 5-2Ö mg daily). T3, 75 µg daily, was givento both groups of patients for 15 ± 4 (13-22) months of treatment. After cessation of drug therapy, 31 patients(63%) remained euthyroid for 18 ± 3 (13-49) months of follow-up, 15 (62.5%) and 16 (64%) patients in the highand low dose groups, respectively. Differences between remission patients and relapse patients at the end oftreatment were in the goiter size (41 ± 11 vs 57 ± 22g, < 0.01), in TSAb activity (83 ± 41 vs 223 ± 150%, < 0.001), in positive TSAb patients (4/23 vs 12/18, < 0.01), in T3 24 h RAIU (10.9 ± 11 vs 21 ± 17%, < 0.05), and in T3 suppressible patients (22/31 vs 6/18, < 0.02). Serum Tg concentration did not differbetween remission (132 ± 242 ng/mL) and relapse (214 ± 235 ng/mL) patients. The combination of serum Tgwith TSAb as well as the combination of serum TSAb with T3 suppressibility improved the predictive value.However, analysis of the specificity and sensitivity showed no significance (<80%) for any predictive markersalone or combined.

T3 suppressed 24 h RAIU correlated with serum TSAb activity (r = 0.515, < 0.001) and FT4I values(r = 0.475, < 0.02). Serum Tg levels correlated with goiter size (r = 0.38, < 0.01) but not with T3 24 hRAIU, TSAb, or FT4I values. In conclusion, the secretion of Tg in Graves' disease was related to goiter size.None of the parameters studied was useful for prediction of the outcome of therapy. No difference between thehigh and the low MMI dose regimens was observed when the MMI was combined with T3 during the entiretreatment period. The remission of Graves' disease might be associated with the maintenance of euthyroidismthroughout treatment.

Department of Endocrinology, Hospital Servidor Publico Estadual-IAMSPE, C.P. 8570, 01000, Sao Paulo, Brazil.

293

294 WERNER ET AL.

INTRODUCTION

The value of the prognostic markers used to predict theoutcome of Graves' hyperthyroid patients treated with

antithyroid drugs (ATD) has been questioned (1,2). Increasedrisk of relapse in patients who are positive for serum thyroid-stimulating antibody (TSAb) at the end of treatment has beenshown (3-5). However, patients without detectable TSAb activ¬ity have also relapsed, decreasing the value of this parameter(1,5,6). The correlation between serum thyroglobulin (Tg)levels and either TSAb or T3 suppressibility in patients whoremitted after ATD therapy suggested that the serum Tg couldsubstitute for TSAb or T3 suppression tests (7). A higher risk ofrelapse was reported in patients with elevated serum Tg concen¬

trations at the end of ATD treatment (8-10), but these data havebeen contradicted (11-14). Van Herle et al. (15) proposed thecombined measurement of serum TSAb and Tg as an idealprognostic index.

We have reported that high ATD doses throughout treatmentwere associated with a high remission rate (16), but this effect ofthe dose of ATD in Graves' disease remains controversial (17).Moreover, using T3 with ATD throughout therapy requires more

evaluation. Therefore, we studied prospectively Graves' hyper¬thyroid patients treated either with high methimazole (MMI)doses or with the lowest MMI able to control the hyperthy¬roidism, keeping the free T4 index (FT4I) below the normalrange. Both groups received additional T3. We compared theoutcome of therapy in these two groups. We assessed the goitersize, serum TSAb, serum Tg value, and T3 suppression of the24 h radioactive iodine uptake (RAIU) before withdrawingdrugs as predictors for the outcome of Graves' disease.

PATIENTS AND METHODS

Patients

Fifty-seven consecutive patients with hyperthyroidism due toGraves' disease were studied; 6 were excluded because ofnoncompliance. In all patients' the diagnosis was based on a

history and signs of hyperthyroidism with diffuse goiter, ele¬vated serum thyroid hormone levels, suppressed serum TSHlevels, and increased RAIU. The remaining 49 patients were

randomly distributed into two groups: 24 patients were treatedwith high (MMI) dose (mean ± SD 60 ± 19,40-120 mg daily),and 25 patients received a low maintenance MMI dose(17 ± 4.3, 5-20 mg/daily). T3, 75 µg daily, was given to bothgroups. The MMI and T3 were discontinued after the patient hadmaintained a FT4I below the normal range (12,16) for at least 6months. The duration of treatment for both groups was 15 ± 4.1(range 13-22) months. The patients were followed up clinicallyand biochemically after cessation of therapy for 18 ± 3 (range13-49) months. Each patient's consent was obtained, and theethical committee of the hospital approved the study. Remissionof Graves' disease was defined as absence of typical symptomsand signs and normal serum thyroid hormone levels during thefollow-up. Relapse was defined as symptoms of hyperthy¬roidism, with high serum concentrations of thyroid hormones.

Methods

The size of the goiter was estimated by palpation. Serum T3and T4 concentrations and T3 uptake values were measured withkits (the T3 and T4 kits were obtained from Clinical Assay(Cambridge, MA) and the T3 uptake kit from DiagnosticProducts Corporation (Los Angeles, CA). Serum FT4I was

calculated from the hormone level and uptake value. Normalranges were 80-210 ng/dL for T3, 4.6-12.5 µg/dL forT4, and4.0-11.5 for FT4I. Serum TSH concentrations were measuredby an IRMA, using a Serono Diagnostic kit (Norwell, MA), andthe normal range was 0.35-5 µ /mL. Interassay and intraassaycoefficient of variation for each assay were as follows: 6.8% and5.3% for T3, 6.1% and 3.0% for T4, 7.1% and 4.5% for T3uptake, 4.6% and 3.2% for TSH, respectively. Serum antithy-roglobulin antibody (TgAb) and antibodies to microsomal frac¬tion (TMAb) were measured by hemagglutination technique(Seratek, Ames Company, Elkart, IN). Serum Tg was measuredin TgAb-negative sera (< 1:100) by an IRMA assay able todetect as little as 1.5 ng/mL of Tg(Sorins.p.a., Saluggia, Italy).The normal range was 3-35 ng/mL, and the interassay andintraassay coefficients of variation for this method were 6.4%and 5.6%, respectively. The 24 h RAIU was done beforestopping the MMI and T3 treatment. In our laboratory, thenormal value of T3 suppressed 24 h RAIU was < 11 % of the '3 ' Idose. ß

TSAb activity was determined by a sensitive method usingadenylate cyclase stimulation in human thyroid plasma mem¬

brane (30,000g fraction) described by Macchia et al. (18).Briefly, the adenylate cyclase activity of human thyroid mem¬

branes was measured in 100 µL of 30 mM Tris-HCl buffer, pH7.8, containing 2.5 mM ATP, 5 mM MgCl2, 10 mM phos-phocreatine, 0.35 mg/mL creatine phosphokinase, 104 Mguanyl-5'-yl imidodiphosphate, 10 mM 3-isobutyl 1-methylx-anthine, and 0.1% bovine serum albumin. The reaction was

started by the addition of about 50 µg of the membranepreparations incubated with 1-3 mg of IgG or bovine TSH as

standard, carried out at 34°C for 1 h, and stopped by addition of2 mL of a cold methanol-ethanol mixture. The assay was

performed in duplicate. Adenylate cyclase activity was ex¬

pressed as the generation of picomoles of cAMP per milligramof membrane protein per min. cAMP was estimated by a

modified competitive binding assay from D.P.C. (Los Angeles,CA). The sensitivity with bovine TSH was 0.1 mU/mL. Anincrease of >2 SD from the mean of cAMP caused by IgG fromnormal subjects (128%) was taken as the demarcation betweenthe presence or absence of TSAb activity. The frequency ofpositive assay was 78.9% in 57 untreated Graves' hyperthyroidpatients. Interassay and intraassay coefficients of variation were

14.5% and 9.2%, respectively. IgG was prepared from 1 mL ofserum with polyethylene glycol at a final concentration of 15%.After centrifugation at 3000g for 20 min at 4°C, the pellet was

dissolved with 1 mL of Tris-HCl, 50 mM NaCl, pH 7.4 buffer.

Statistical analysisData were analyzed using Student's unpaired two-tailed

/-test, linear correlation analysis, or 2 test. Logarithmic trans-

PROGNOSTIC MARKERS IN GRAVES' HYPERTHYROIDISM TREATMENT 295

formations of the serum Tg values were performed. A value of < 0.05 was considered significant.

RESULTS

Remission and relapse rate

After cessation of therapy, 31 (63%) patients continued to bein remission, and no difference was noted between the high(15/24,62.5%) and the low (16/25, 64%) MMI dose groups. Allpatients maintained serum T4 and FT4I concentrations below thenormal range, with normal TSH values for at least 6 months ofMMI therapy. As Table 1 shows, there were no differences inthe parameters studied at the end of MMI therapy except a mean

T3 value that was higher in the group treated with the high MMI

Predictive markersResults of the studied parameters for outcome of Graves'

disease treatment are summarized in Table 2 and Figure 1. Thegoiter size, T3 suppressibility, and serum TSAb values weredifferent between remission and relapse patients. Goiter shrink¬age was observed in 67.7% of patients in the remission groupand in 38% of relapsed patients (p < 0.05). TSAb activity was

negative in 82.6% of remission patients and in 33.3% ofrelapsed patients. On the other hand, TSAb was positive in17.4% of remission patients and in 66.6% of relapsed patients( 2 = 8.35, < 0.001). In addition, 71% of remission patientsshowed 24 h RAIU suppressed values, compared with 33.3% ofrelapsed patients ( 2 = 6.59, < 0.02). Of 16 remission pa¬tients who became T3 suppressible at the end of MMI treatment,14 (87.5%) were TSAb positive, whereas 8 (47%) of 17relapsed patients showed a concordance in TSAb activity and T3

Table 1. Clinical and Laboratory Data in Graves' Disease Patients at theEnd of Combined Methimazole and T3 Therapy

Treatment group

High MMI dose Low MMI dose

PatientsSex (female:male)Age (years)Goiter Size (g)Serum T4 ^g/dL)Serum T3 (ng/dL)Serum FT4ISerum TSH (/iU/mL)T3 24 h RAIU (%)T3 suppressible

(positive/negative)TSAb (% of control)TSAb (positive/negative)Serum Tg (ng/mL)TMAb (reciprocal titer)

2423:1

44 ± 1247.5 ±

2.7 ±228 ±

1.7 ±1.3 ±

10.8 ± 1016/8

123

183.11081.61.1

81(21)a8/13

190 ± 2745180 + 9227(19)

2522:3

42 ± 1447.8 ± 18

2.9 ± 2.7172 ±47*1.9 ± 1.31.3 ±0.9

18.2+ 1713/12

161 ± 164(20)8/20

142 ± 210(23)3845 ± 7142(18)

*p < 0.05."Number of patients in parentheses.

Table 2. Prognostic Markers in Graves' Disease Patients at theEnd of Combined Methimazole and T3 Treatment

Parameters Remission patients Relapse patientsPatientsSex (female:male)Age (years)Goiter size (g)Serum T4 (Mg/dl)Serum FT4IT3 24 h RAIU (%)TSAb (% of control)Serum Tg (ng/mL)Low MMI dose group

(mg/daily)

3129:0

40.6 ± 1041.9 ± 11

2.7 ± 2.81.7 ± 1.2

10.9 ± 1183.3 ± 41(23)a132 ± 242(29)

18 ±2.9

1816:2

41.8 ± 1357.5 ± 22*

2.9 ± 3.01.9 ± 1.8

20.8 ± 17**223 ± 159 (17)***214 ±235(17)

14 ±5.3*

*p < 0.01; **p < 0.05; ***p < 0.001."Number of Patients in Parentheses

296 WERNER ET AL.

suppressibility. The serum Tg concentrations did not differbetween remission and relapsed patients. The use of Tg as a

prognostic marker was improved by the combination of serum

Tg and TSAb ( 2 = 10.8, < 0.001). The combination ofserum TSAb with T3 suppressibility ( 2 = 8.36, < 0.01) alsowas useful.

concentrations was not significant (Fig. 3). Also a relationshipbetween T3 24 h RAIU and serum FT4I was noted (r = 0.512, = 48, < 0.02). The serum Tg levels correlated with goitersize but not with serum TSAb (Fig. 4) or FT4I levels (r = 0.21,„ = 44, NS).

Efficiency of predictive markers

The results shown in Figure 1 suggest that goiter size change,serum TSAb activity, T3 suppressibility, and the combination oftwo parameters are significant prognostic markers of the out¬come of treatment of Graves' disease. To analyze these datafurther, we used the criteria of Galen and Gambino (19) tocalculate specificities and sensitivities. Specificity was definedas true negative parameters divided by true negative plus falsepositive parameters, and sensitivity was defined as true positiveparameters divided by true positive plus false negative parame¬ters. In this study, true negatives were considered patients inremission with no change in goiter size, T3 nonsuppressibility,positive serum TSAb, and elevated Tg concentration. Truepositives were defined as patients in remission with goitershrinkage, T3 suppression, negative serum TSAb, and normalTg concentration. False negatives were found in relapsed pa¬tients with no change in goiter size, T3 nonsuppressibility, andelevated levels of serum TSAb and Tg. False positives were

considered relapsed patients showing decrease in goiter size, T3suppression, negative serum TSAb, and normal Tg concentra¬tion.

The specificity and sensitivity shown in Figure 2 among thestudied parameters revealed no significance of any of thepredicitive markers (all were less than 80%).

Correlations

A correlation between T3 suppressed 24 h RAIU and serum

TSAb values was found, but the correlation with serum Tg

19/23

Relapsed Patients

FIG. 1. Incidence of prognostic markers for outcome ofGraves' disease measured at the end of combined methimazoleand T3 treatment. Filled bars show the results for remissionpatients. Hatched bars show the results for relapsed patients.*/7<0.01; **/><0.02; #goiter size decreased; ##T3S, T3suppressible patients; f"serum Tg concentration in the normalrange (< 35 ng/mL); Serum TSAb, number of patients withnegative TSAb activity.

DISCUSSION

Since small doses of ATD are effective in the control ofhyperthyroidism (17,21) and a higher incidence of adverseeffects was associated with the high ATD dose regimen (20), we

studied a group of patients treated with the lowest MMI dose(10-25 mg daily) able to control the hyperthyroidism andcompared this therapy with patients treated with high MMIdoses. The addition of T3 maintained the euthyroidism, asindicated by the serum TSH, and kept the FT4I values below thenormal range (12,16). Although the only difference between thisstudy and the previous one (16) was the addition of T3 to thetreatment of patients receiving low MMI doses, a similarremission rate was obtained for both groups. It is difficult to findan explanation for the high T3 concentration levels observed inpatients on high MMI doses at the end of therapy.

Volpe et al. (22) have claimed that the immunologie ATDeffect during therapy of Graves' disease is primarily due to therestoration of euthyroidism. We have reported that the decreasein thyroid autoantibody titers was correlated with euthyroidism

Goiter

Serum TSAb

T3S + TSAb

40 60Percent

100

Specificity Sensitivity

FIG. 2. Efficiency of predictive markers for the outcome ofGraves' disease measured before the cessation of methimazoleand T3 treatment. Filled bars show the percentage of specificity,and hatched bars show the percentage of sensitivity.suppressible patients.

T3S, T3

PROGNOSTIC MARKERS IN GRAVES' HYPERTHYROIDISM TREATMENT 297

achieved with the combined regimen (23). An explanation forour present results might be that the maintenance of euthy¬roidism over 6 months of therapy influences the immune systemof Graves' disease patients (5,22,23) and permits a long-termremission.

Murakami et al. (21) reported a low relapse rate in Graves'hyperthyroid patients maintained in euthyroid status with a lowMMI dose ( 10 mg daily ), but the medication was stopped only inpatients who became T3 suppressible, requiring at least 20months of treatment. A high remission rate was obtained alsowhen ATD was withdrawn after the T3 suppression returned to

normal, but it required a longer period (up to 8 years) oftreatment (24). Although no predictive value was noted in thepresent study, most of our remission patients were suppressible,and the T3 suppression test was useful for monitoring combinedtherapy.

Active Graves' disease is closely associated with the presenceof TSAb (3,22), but the clinical significance of TSAb measure¬

ment in patients with Graves' disease during ATD treatmentremains unclear. Our results confirm the data of Macchia et al.(6), who by using a similar method, observed no clear prognos¬tic value for TSAb measurement. The effect of Graves' IgG on

the TSH receptor complexes may be explained by the heteroge¬neity of the autoantibodies present in the sera of patients withGraves' disease, as shown by Zakarija et al. (25). Therefore,TSAb may not be the only thyroid antibody involved in theoutcome of Graves' hyperthyroidism. Blocking antibodies to theTSH receptor with no stimulating activity have been demon¬strated in some patients with Graves' disease (26). Tamai et al.(27) observed that positive TSAb activity became negative andblocking activity became detectable during ATD therapy. Thus,the emergence of blocking-type antibodies might be responsiblefor the remission.

We found that serum Tg concentration is of limited value inpredicting the outcome of therapy. Discordant results in thepredictive value of serum Tg may derive from the differentdietary iodine intake of the patients studied. Interestingly, mostreports of good prognostic value of serum Tg levels were fromcountries with a high iodine intake (8-11). This parameter was

not of predictive value in populations with comparatively lowdietary iodine intake (11-14). The mechanism by which Tg issecreted in Graves' disease patients is not fully understood. VanHerle et al. (28) demonstrated that injection of IgG of Graves'hyperthyroid patients into rats leads to a prompt and prolonged

800

%Serum Tg ( ng/ml )

P«Oj01

eoo

400

200

rai JLs. 9>-m

a a

°_a.40 ß

Goiter Size ( g )80

80

50

40

30

20

10

T3-24-h RAIU (*)

r-0.8«·, n-40 < 0.001

D

««

**!·'XX) 200 300 400

Serum TSAb Activity (* of Normal)600

XX»

800

800-

S«um Tg (ng/ml)

fOJtill·»·

*

,dbHi*

» m-=-1-'—100 200 300 400

Serum TSAb Activity (* of Normal)600

80

so

40

30

20

10

3-24- 1 RAIU ( % )

r-OJtO:HA.

*

-

a

D «o *

»_ _

200 400 800

Serum Tg ( ng/ml

eoo 1000

FIG. 3. Relationship between serum thyroglobulin concentra¬tion and goiter size (top) and serum TSAb activity (bottom) inGraves' hyperthyroid patients at the end of treatment withmethimazole and triiodothyronine. Asterisks indicate relapsedpatients after therapy, and open squares indicate patients inremission after therapy.

FIG. 4. Relationship between T3 24 h RAIU determinationsand serum TSAb activity (top) and serum thyroglobulin concen¬tration (bottom) in Graves' hyperthyroid patients at the end oftreatment with methimazole and triiodothyronine. Asterisksindicate relapsed patients after therapy, and open squares indi¬cate patients in remission after therapy.

298 WERNER ET AL.

release ofTg. Tg was released from human thyroid cells culturedin IgG from Graves' patients (29), suggesting that serum Tglevels reflect TSAb activity. In our study, we were not able toobtain such a relationship. However, the correlation betweenserum Tg concentration and goiter size is in accordance with thedata suggesting that one of the factors that affected the serum Tglevels might be the thyroid mass (13).

In conclusion, we could not predict the outcome of Graves'disease during treatment with MMI plus T3. Even the combina¬tion of two parameters showed no predictive value in terms ofspecificity and sensitivity. These findings are in disagreementwith Kasagi et al. (2), who found that the combination of twoindices improved the prediction of the outcome of therapy ofGraves' disease. However, these authors did not study thespecificity and sensitivity of those parameters. Our results withthe combined regimen showed no difference between the highand the low dose regimen when the MMI was given with T3during the entire treatment period. These data suggest thatremission of Graves' disease might be associated with themaintenance of euthyroidism throughout treatment.

ACKNOWLEDGMENTS

We wish to express our gratitude to Dulce I.A. Figueiredo,Lidia Tanaka-Matsuura, and Humberto Laudari for their techni¬cal support. This work was supported in part by C.N.Pq.

REFERENCES

1. Orgiazzi J, Madec AM, Geneter N, Gueguen M, Allanic H 1987Imunological paramenters in Graves' disease: Are they useful forindication and monitoring of antithyroid drug treatment? Horm Res26:131-136.

2. Kasagi K, Ijda Y, Hatabu H, et al 1988 Evaluation of TSH-receptorantibodies as prognostic markers after cessation of antithyroid drugtreatment in patients with Graves' disease. Acta Endocrinol(Copenh) 117:173-179.

3. Zakarija M, McKenzie JM, Banovac 1980 Clinical significanceof assay of thyroid-stimulating antibody in Graves' disease. AnnIntern Med 93:28-32.

4. Rapoport B, Greenspan FS, Filetti S, Pepitone M 1984 Clinicalexperience with a human thyroid cell bioassay for thyroid-stimulat¬ing immunoglobulins. J Clin Endocrinol Metab 58:332-338.

5. Edan G, Massart C, Hody B, et al 1989 Optimum duration ofantithyroid drug treatment determined by assay of thyroid-stimulat¬ing antibody in patients with Graves' disease. Br Med J298:359-361.

6. Macchia E, Concetti R, Borgoni F, et al 1989 Assays of TSH-receptor antibodies in 576 patients with various thyroid disorders:Their incidence, significance and clinical usefulness. Autoimmu-nity 3:103-112.

7. Isozaki O, Tsushima T, Shizume K, et al 1985 Thyroid stimulatingantibody bioassay using porcine thyroid cells cultured in follicles. JClin Endocrinol Metab 61:1105-1 111.

8. Uller RP, Van Herle AJ 1978 Effect of therapy on serum thyroglo¬bulin levels in patients with Graves' disease. J Clin EndocrinolMetab 46:747-755.

9. Gardner DF, Rothman J, Utiger RD 1979 Serum thyroglobulinlevels in normal subjects and patients with Graves' disease: Effectof T3, iodide, 13lI and antithyroid drugs. Clin Endocrinol (Oxf)11:585-594.

10. Kawamura S, Kishino B, Tajima K, Mashita K, Tarai S 1983Serum thyroglobulin changes in patients with Graves' diseasetreated with long-term antithyroid drag therapy. J Clin EndocrinolMetab 56:507-512.

11. Feldt-Rasmussen U, Bech K, Date J, et al 1982 A prospectivestudy of the differencial changes in serum thyroglobulin and itsautoantibodies during propylthiouracil or radioiodine therapy ofpatients with Graves' disease. Acta Endocrinol (Copenh)99:379-385.

12. Bromberg , Werner RS, Farah CS, Werner MC, Dall'AntoniaRP, Romaldini JH 1986 Graves' disease treated with high dose ofmethimazole: Lack of correlation between serum thyroglobulin andthyroid-stimulating antibody. In: Medeiros-Neto G, Gaitan E (eds)Frontiers in Thyroidology. Plenum, New York, pp 1125-1128.

13. Ericsson UB, Tegler L, Dymling JF, Thorell JI 1987 Effect oftherapy on the serum thyroglobulin concentration in patients withtoxic diffuse goiter, toxic nodule goiter and toxic adenoma. JEndocrinol Invest 10:351-355.

14. Talbot JN, Duron F, Feron R, Aubert P, Milhaud G 1989 Thyro¬globulin, ihyrotropin and thyrotropin binding inhibiting immuno¬globulins assayed at the withdrawal of antithyroid drug therapy as

predictors of relapse of Graves' disease within one year. J Endo¬crinol Invest 12:589-595.

15. VanHerle AJ, VassartG, Dumont JE 1979 Control of thyroglobulinsynthesis and secretion (second of two parts). Engl J Med301:307-314.

16. Romaldini JH. Bromberg , Wemer RS, et al 1983 Comparison ofeffects of high and low dosage regimens of antithyroid drags in themanagement of Graves' hyperthyroidism. J Clin Endocrinol Metab57:563-570.

17. BenkerG, Reinwein D, Creutzig H, et al 1987 Effects of high andlow doses of methimazole in patients with Gravés' thyrotoxicosis.Acta Endocrinol (Copenh) 281 (suppl):312-317.

18. Macchia E, Carayon P, Fenzi GF, Lissitzky S, Pinchera A 1983A sensitive method for assaying thyroid stimulating immunoglob¬ulins of Graves' disease: Use of the guanyl nucleotide-amplifiedthyroid adenylate cyclase assay. Acta Endocrinol (Copenh)103:345-351.

19. Galen RS, Gambino SL 1980 Beyond Normality: The PredictiveValue and Efficiency of Medical Diagnosis. John Wiley and Sons,New York, pp9-14.

20. Werner MC, Romaldini JH, Bromberg , Werner RS, Farah CS1989 Adverse effects related to thionamide drugs and their doseregimen. Am J Med Sci 297:216-220.

21. Murakami M, Koizumi Y, Aizawa T, et al 1989 Studies of thyroidfunction and immune parameters in patients with hyperthyroidGraves' disease in remission. J Clin Endocrinol Metab66:103-108.

22. Volpe R 1988 The immunoregulatory disturbance in autoimmunethyroid disease. Autoimmunity 2:55-72.

23. Romaldini JH, Werner MC Rodrigues HF, et al 1986 Graves'disease and Hashimoto's thyroiditis: Effects of high doses ofantithyroid drugs on thyroid autoantibody levels. J EndocrinolInvest 9:233-237.

24. Yamamoto M, Totsuka Y, Kojima I, et al 1983 Outcome of patientswith Graves' disease after long-term medical treatment guided bytriiodothyronine (T3) suppression test. Clin Endocrinol (Oxf)19:467-476.

25. Zakarija M, Garcia A, McKenzie JM 1985 Studies on multiplethyroid cell membrane-directed antibodies in Graves' disease. JClin Invest 76:1885-1891.

PROGNOSTIC MARKERS IN GRAVES' HYPERTHYROIDISM TREATMENT 299

26. Macchia E, Concetti R, Carone G, Borgoni P, Fenzi GF, PincheraA 1988 Demonstration of blocking immunoglobulins G, having a

heterogeneous behaviour, in sera of patients with Graves' disease:Possible coexistence of different autoantibodies directed to the TSHreceptor. Clin Endocrinol (Oxf) 28:147-156.

27. Tamai H, Hirora Y, Kasagi K, et al 1987 The mechanism ofspontaneous hypothyroidism in patients with Graves' disease afterantithyroid drag treatment. J Clin Endocrinol Metab 64:718-722.

28. Van Herle AJ, Klandorf H, UllerRP 1975 A radioimmunoassay forserum rat thyroglobulin. Physiologic and pharmacological studies.J Clin Invest 56:1073-1081.

29. Fukue Y, Uchimura H, Mitsuhashi Y, et al 1987 Thyroglobulinrelease-stimulating activity in immunoglobulin G from patientswith Graves' disease studied by human thyroid cells in vitro. J ClinEndocrinol Metab 63:261-265.

Address reprint requests to:Joao H. Romaldini, M.D.

Department of EndocrinologyHSPE-IAMSPE, C.Postal 8570

01000, S.Paulo, Brazil