THYROID

Volume 13, Number 5, 2003

© Mary Ann Liebert, Inc.

Impact of Previous Thyroid Autoimmune Diseases

on Prognosis of Patients with Well-Differentiated
Thyroid Cancer

Suzikelli Lisboa Souza, Lígia Vera Montalli da Assumpção, and Laura Sterian Ward

Autoimmune phenomena are frequently associated with differentiated thyroid carcinomas. However, the sig-
nificance of thyroid gland autoimmune aggression on the outcome of these patients is still controversial. To
address this issue, we studied 173 patients (123 with papillary and 50 with follicular carcinomas) who under-
went surgery complemented by radioiodine ablation and followed up for 0.5–29 (6 6 5.76) years. Analysis of
the prognostic factors revealed that higher age, male gender, larger nodule size, follicular tumors, presence of
metastases at diagnosis, grade of differentiation, and stage correlated positively with the occurrence of death,
metastasis and/or recurrence, while the presence of antibodies and the previous history of autoimmune dis-
ease correlated negatively with these events. Long distant metastases increased the odds for a lower disease-
free rate for patients with papillary (8.366 times) and follicular (7.373 times) carcinoma. However, univariate
and multivariate analysis failed to demonstrate that neck node involvement could influence the outcome for
patients with well-differentiated thyroid carcinoma. The odds for patients with previous history of thyroid au-
toimmune disease (p , 0.02) or with thyroid autoantibodies (p , 0.001) to have a worse outcome were lower
than for patients with no evidence of autoimmune activity, suggesting that autoimmune activity against the
gland may exert a protective effect on the outcome of differentiated thyroid carcinoma patients.

Introduction finding in patients with papillary carcinomas suggesting that

autoimmune thyroiditis may be a predisposing factor for

F ACTORS INFLUENCING PROGNO SIS and long-term outcome of

thyroid cancer have been described by several groups;
however, the influence of previous autoimmune diseases
and antithyroid antibodies on outcome for patients with dif-
ferentiated thyroid cancer has not been sufficiently investi-
gated. Most retrospective studies on antithyroid antibodies
focus on thyroglobulin antibody (TgAb), which has been
viewed as acting as a “tumor marker,” because its presence
may indicate tumor persistence or recurrence (1). On the
other hand, it has been suggested that thyroid nodules found
in patients with Graves’ disease have a higher likelihood of
being malignant, and that thyroid cancer behaves more ag-
gressively when associated with Graves’ disease although
both of these assertions remain controversial (2–6). A recent
publication was unable to find a correlation between thy-
roid-stimulating antibodies and the aggressiveness of well-
differentiated thyroid carcinomas (7). Also, there is an
association between chronic lymphocytic thyroiditis and
papillary carcinomas (8–10). In fact, lymphocytic infiltrate,
which is indicative of autoimmune thyroiditis, is a frequent
papillary thyroid carcinoma (10).
There are very few reports on the influence of thyroid per-
oxidase antibody (TPOAb) on the outcome for patients with
differentiated thyroid carcinoma. An early report by Pacini
et al. (11) studying TgAb and TPOAb measured by hemag-
glutination technique was not able to detect any difference
in the clinical course or in the mortality rate of antibody-pos-
itive compared to antibody-negative patients.
The goal of this study was to investigate the role of pre-
vious autoimmune injuries on the thyroid gland and of the
presence of autoantibodies as detected by sensitive methods
in the outcome for patients with differentiated thyroid car-
cinoma.
Subjects and Methods
The charts of 173 consecutive cases of well-differentiated
thyroid carcinoma treated between 1977 and 2002 were care-
fully reviewed. All cases were managed according to a stan-
dard protocol. The diagnosis of thyroid carcinoma was es-

Laboratory of Cancer Molecular Genetics, Department of Medicine, State University of Campinas, São Paulo, Brazil.

491
492 SOUZA ET AL.

tablished or suspected by fine-needle aspiration cytologic
study and/or by the histologic analysis of thyroid tissues
from patients who underwent surgery because of thyroid
nodules that presented clinical or epidemiologic suspicions
of cancer. All patients underwent total or near-total thy-
roidectomy. Patients with neck node metastases palpable
preoperatively or intraoperatively underwent regional neck
dissection. Four to 6 weeks after the operations, whole-body
scans (WBS) were performed. All patients received 30–100
mCi 131I. Long-term levothyroxine suppressive doses were
given after the WBS in order to keep serum thyrotropin
(TSH) at low/normal levels.
Clinical and pathologic features of the patients with thy-
roid carcinomas are detailed in Table 1. Hospital records
were reviewed and data were collected on each patient’s age,
gender, family and personal history of thyroid disease, and
radiation exposure. The primary tumor size, fine-needle as-
piration cytologic results, thyroid function tests, and thyroid
autoantibodies measurements were compiled, as well as the
surgical findings. Tumor stage and degree of differentiation
were obtained from surgical and pathologic records. Expe-
rienced pathologists of the University Hospital confirmed all
diagnoses. The histologic classification was made according
to the World Health Organization (WHO) criteria (12). One
hundred twenty-three patients (71%) were classified as pre-
senting with papillary thyroid carcinomas and 50 patients
(29%) with follicular thyroid cancer; the latter group in-
cluded 13 patients with the Hürthle cell variant of follicular
cancer. The degree of differentiation was stated as either well
differentiated or poorly differentiated. Tumor staging was
based on clinical staging of DeGroot (1995). Stage I is a tu-
mor with single or multiple intrathyroidal foci. Stage II is a
tumor with limited cervical metastasis only. Histopatholog-
ically proven cervical lymph node metastases were identi-
fied in all the patients. Stage III is a thyroid tumor with lo-
cal cervical metastasis or fixed cervical metastasis. Stage IV
is a lesion metastasis outside the neck.
Serum thyrotropin (TSH) was measured with a sensitive
chemiluminescent assay using a commercial kit (Immulite,
Diagnostic Products Company, Gwynedd, UK) with a func-
tional sensitivity of 0.05 mU/L, and intra-assay and interas-
say coefficients of variation of 5.7% and 7.9%, respectively.
The normal range in our laboratory was 0.38 to 4.0 mU/L.
Thyroglobulin (Tg) levels were determined using a Tg kit
(Immulite) with a functional sensitivity of 0.9 ng/mL, in-
terassay and intra-assay coefficients of variation of 8.9% and
4.2%, respectively. TgAb and TPOAb were measured with
sensitive immunoradiometric quantitative assays (Biodata-
Serono Diagnostics, Rome, Italy); coefficients of intra-assay
and interassay variation of 6%, sensitivity of 20 U/mL, pos-
itive values 100 U/mL or more for TgAb; coefficients of in-
tra-assay and interassay variation 5% and 6%, respectively,
sensitivity of 1.0 U/mL, positive values 100 U/mL or more
for TPOAb.
All patients were followed with WBS, serum TSH and Tg
measurements under suppression according to a routine fol-
low-up protocol that includes x-ray, ultrasonography, com-
puted tomography (CT) scan, magnetic resonance imaging
(MRI), and other procedures to detect recurrence or distant
metastasis. Patients with high serum Tg levels (.2 mg/dL)
and/or suspicious WBSs underwent a thorough image
search. Prognostic outcomes were obtained from follow-up
examinations. Follow-up duration was calculated from the
time of last evaluation or the time of death. We defined tu-
mors as recurrent when there was serum Tg increase and/or
new evidence of loco-regional disease or distant metastases
occurring more than 6 months after successful 131I ablative
therapy.
Data analysis
Data were analyzed using SAS 8.01 statistical software
(SAS Institute, Cary, NC). Time-dependent variables were
analyzed by the Cox proportional hazard models and the
Kaplan-Meier product limit estimate of survival curves. Mul-
tivariate analysis and a log-rank test (LR) for univariate anal-
ysis were performed to assess the independent effect of the
collected variables using the Cox model. A log-rank test was
used for comparison of survival curves. The differences be-
tween groups with and without autoantibodies and with and
without previous autoimmune disease were evaluated using
the x2 test for independence or Fisher’s exact test (F). The oc-
currence of metastasis, recurrence or death relative to
whether a risk factor was present or not is described by the
odds ratio (OR). A 95% confidence interval (CI) for the OR
was used to determine whether estimations of the odds were
statistically significant. Multivariate analysis was carried out
using multiple logistic regression. Comparison between the
disease-free and overall survival rates in patients with stages
I–IV were calculated using the Kaplan-Meier method and

TABLE 1. DISTRIBUTION OF PATIENTS W ITH THYROID CARCINOMA a

Diagnosis
Clinical characteristics (presence of Follow-up
metastasis)
Gender History Antibodies Recurrence
Lymph and/or distant
Histology Age M F N G H TPOAb TgAb node Distant metastasis

Papillary carcinomas 38 6 15 26 97 88 9 10 14 12 49 9 32
Follicular carcinomas 52 6 17 9 41 23 0 5 1 0 4 16 30
a
Patients are grouped according to their histology, clinical features including age (X 6 standard deviation [SD] in years), gender (M, male;
F, female), the history of previous thyroid benign diseases (N 5 nodules, G 5 Graves’ disease, H 5 Hashimoto’s disease), the presence of
thyroid autoantibodies at diagnosis (TPOAb, TgAb), the presence of lymph node involvement and distant metastasis by the time of the
diagnosis and the diagnosis of recurrence and/or distant metastasis during the follow-up.
AUTOIMMUNITY: THYROID CARCINOMA RISK FACTOR
log-rank test for univariate analysis and the Cox proportional
hazards model for multivariate analysis. The Kaplan-Meier
method and the log-rank test were used to carry out a sep-
arate analysis comparing disease-free survival in patients
with previous history of autoimmune disease or not. The ob-
served differences are assumed to be statistically significant
if the probability of chance occurrence is p , 0.05.
Results
Median follow-up of the group was 79 months with a
range of 6 months to 29 years. Table 1 summarizes clinical,
pathologic, and follow-up data of the differentiated thyroid
carcinoma patients. A total of 69 patients with stage I dis-
ease, 40 with stage II, 31 with stage III, and 31 with stage IV
disease were identified. Twelve patients died during our fol-
low-up, 6 of these as a result of well-differentiated thyroid
carcinomas, 4 because of concurrent diseases, including heart
disease (1 patient) and other types of cancer (3 patients); 3
causes of death were unknown.
Previous history of autoimmune thyroid disease was con-
firmed in 25 (14.45%) patients including 9 with Graves’ dis-
ease and 16 with Hashimoto’s thyroiditis. The odds for a pa-
tient with no history of autoimmune thyroid disease to
present recurrence, metastasis, or death (F; p 5 0.0185) were
4363 times higher (95% CI: 1.238–15.373) than a patient with
this antecedent. A previous diagnosis of Hashimoto’s dis-
ease was associated with a better outcome (LR 5 5.81, p 5
A B
493
0.0160). Thyroid autoantibodies were detected in 19 of 56 pa-
tients at diagnosis including 15 patients with positive TPOAb
and 12 with positive TgAb. The presence of positive TPOAb
was associated with disease-free survival (LR 5 7.18; p 5
0.0074). The odds for a patient with negative TPOAB or neg-
ative TPOAb and TgAb were 6.75 (95% CI: 0.9854–46.2376)
and 17,053 (95% CI: 2.057–141.34) times higher, respectively,
than for a patient with positive autoantibodies to present re-
currence, metastasis or death (F; p 5 0.0010). Univariate lo-
gistic regression analysis of the 173 patients with differenti-
ated thyroid carcinoma revealed that the variables histology
(p 5 0.0001, odds 4.266, 95% CI: 2.130–8.542), presence of
metastases at diagnosis (p 5 0.0001, odds 37.235, 95% CI:
10.543–131.505), grade of differentiation (p 5 0.0001, odds
5.786, 95% CI: 2.452–13.650), and stage (p 5 0.0001, odds
32.625, 95% CI: 9.325–114.149) correlated positively with the
occurrence of death, metastasis and/or recurrence during
follow-up, while the presence of antibodies (p 5 0.0086, odds
0.059, 95% CI: 0.007–0.486) and the previous history of auto-
immune disease (p 5 0.0219, odds 0.229, 95% CI: 0.065–
0.808) correlated negatively with these events. The chance
for patients with papillary carcinomas and those with follic-
ular carcinoma with positive antibodies to have a disease-
free evolution were 12.82 (p 5 0.0199; CI 0.012; 0.646) and
16,949 (p 5 0.0086; CI: 0.007; 0.486) times higher than the
chances of patients with papillary carcinoma and follicular
carcinoma with negative antibodies. Figure 1A represents the
Kaplan-Meier curve for the free-of-disease time of differen-
FIG. 1. Disease-free survival Kaplan-Meier curves in pa-
tients. A: Stage I (n 5 69), stage II (n 5 40), stage III (n 5 31),
and stage IV (n 5 32). Log rank 5 87.14; p 5 0.0001. B: With
previous autoimmune (n 5 25) and nodular (n 5 126) thy-
roid disease. Log rank 5 6.12; p 5 0.0133. C: With positive
(n 5 19) and negative (n 5 37) thyroid autoantibodies. Log
rank 5 10.57; p 5 0.0012. TPOAb, thyroid peroxidase anti-
body, TgAb, thyroglobulin antibody.
494
tiated carcinoma patients according to the stage of the tu-
mor; Figure 1B shows the Kaplan-Meier curve for the free-
of-disease time of differentiated carcinoma patients with
positive and negative antecedents of thyroid autoimmune
diseases; in Figure 1C we represent the influence of the pres-
ence of antibodies on the Kaplan-Meier curve for the free-
of-disease time of patients with differentiated carcinoma.
Cox univariate regression analysis also identified age (p 5
0.0022, odds 2.224, 95% CI: 1.335–3.705), gender (p 5 0.0270,
odds 1.866, 95% CI: 1.073–3.244), nodule size (p 5 0.0009,
odds 2.434, 95% CI: 1.439–4.115), the tumor grade of differ-
entiation (p 5 0.0001, odds 6.351, 95% CI: 2.881–13.999 for
poorly differentiated carcinomas), and the presence of me-
tastasis at diagnosis (p 5 0.0001, odds 8.743, 95% CI:
4.996–15.392) as important factors of risk.
The presence of long distant metastases by the time of the
diagnosis increased the odds for a lower disease-free rate for
both patients with papillary carcinoma (8.366 times) and for
patients with follicular carcinoma (7.373 times). Lymph node
metastases were detected in 62 patients: 49% of the patients
with papillary carcinoma and 10% of the patients with fol-
licular carcinoma. Disease-free rates were similar in uni-
variate analysis for patients with papillary and with follicu-
lar carcinoma, with and without neck node metastases (p 5
0.2 and p 5 0.09 for papillary and follicular carcinoma, re-
spectively). Multivariate analysis for disease-free survival
and overall survival rates also failed to demonstrate that the
presence of initial neck node involvement could influence
outcome for patients with well-differentiated thyroid carci-
noma.
Discussion
Patients with thyroid cancer present a spectrum of diverse
behavior. Although the overall survival of patients with
well-managed well-differentiated thyroid carcinomas is con-
sidered excellent, 7%–20% of these patients die at 20 years
(13,14). Various series have identified prognostic factors,
which essentially have divided the thyroid cancer into low-
risk and high-risk groups, in an attempt to optimize the sur-
gical approach and the follow-up strategies (15,16). Signifi-
cant prognostic indicators of poor survival include age
greater than 45 years, male gender, tumor size greater than
4 cm, follicular disease, extrathyroidal extension, and distant
metastasis, with the presence of lymph node metastasis also
significant in some series (17,18). The importance of previ-
ous benign autoimmune thyroid disorders in thyroid cancer
incidence has been investigated by register linkage studies.
Although the general cancer risk was not significantly ele-
vated, an increased risk of thyroid cancer was found among
women with myxedema, thyrotoxicosis, and nontoxic goiter
of a large cohort of individuals who were discharged from
a Danish hospital with a diagnosis of a benign thyroid dis-
ease (19). A similar increase was found in men, but the num-
bers were small. Further analysis of these data, however, re-
vealed that only 5 patients had a diffuse goiter and a
diagnosis of Hashimoto’s thyroiditis was established in only
1 case. Moreover, no data were available on thyroid au-
toantibodies (20). A pooled analysis of 12 case control stud-
ies including 2519 patients and 4176 controls found an in-
creased risk of thyroid cancer among patients with a history
of goiter (women’s OR 5 5.9, 95% CI: 4.2–8.1 and men’s
SOUZA ET AL.
OR 5 38.3, 95% CI: 5.0–291.2). The risk for thyrotoxicosis was
only marginally increased (OR 5 1.4, 95% CI: 1.0–2.1) among
women and among men (OR 5 3.1, 95% CI: 1.0–9.8). How-
ever, in the absence of or after adjustment for a history of
goiter, thyrotoxicosis was not associated with an increased
risk of thyroid cancer among women (21). Myxedema was
also not associated with an increased risk (21). Again, data
on thyroid autoantibodies were missing making it difficult
to interpret these results because autoimmune diseases were
not defined. Some studies have suggested a higher incidence
of malignancy in thyroid nodules found in patients with
Graves’ disease and hyperthyroidism, and that thyroid can-
cer behaves more aggressively when associated with Graves’
disease but this is still controversial (22,23). On the other
hand, although data on thyroid autoantibodies are scarce,
there are compelling evidences that chronic thyroiditis is a
favorable prognostic factor in papillary thyroid carcinoma
(24–27).
Our data indicate that previous autoimmune diseases and
the presence of thyroid autoantibodies exert a favorable in-
fluence on the outcome of differentiated carcinoma patients.
Patients with a previous history of autoimmune thyroid dis-
ease have a 3.75 times higher chance, and patients with pos-
itive antibodies a 12.19 times higher chance to become dis-
ease-free than patients with no evidence of autoimmune
injury of the gland.
We suggest that previous or concurrent autoimmune in-
jury exerts a protective effect on differentiated thyroid car-
cinoma patient outcome.
Acknowledgments
We are indebted to the statistics team of the School of Med-
ical Sciences, FCM/UNICAMP for useful discussion and im-
portant suggestions. This work was partially supported by
a grant from FAPESP #99/03379-1.
References
1. Hjiyiannakis P, Mundy J, Harmer C 1999 Thyroglobulin an-
tibodies in differentiated thyroid cancer. Clin Oncol (R Coll
Radiol) 11:240–244.
2. Stocker DJ, Foster SS, Solomon BL, Shriver CD, Burch HB
2002 Thyroid cancer yield in patients with Graves’ disease
selected for surgery on the basis of cold scintiscan defects.
Thyroid 12:305–311.
3. Pinchera A, Bartalena L, Marcocci C 1995 Therapeutic con-
troversies. Radioiodine may be bad for Graves’ ophthal-
mopathy, but . . . J Clin Endocrinol Metab 80:342–345.
4. Pellegriti G, Belfiore A, Giuffrida D, Lupo L, Vigneri R 1998
Outcome of differentiated thyroid cancer in Graves’ pa-
tients. J Clin Endocrinol Metab 83:2805–2809.
5. Mazzaferri EL 1990 Thyroid cancer and Graves’ disease. J
Clin Endocrinol Metab 70:826–829.
6. Belfiore A, Russo D, Vigneri R, Filetti S 2001 Graves’ dis-
ease, thyroid nodules and thyroid cancer. Clin Endocrinol
(Oxf) 55:711–718.
7. Sachmechi I, Bitton R 2000 Role of thyroid-stimulating im-
munoglobulin in aggressiveness of well-differentiated thy-
roid cancer. Endocr Pract 6:139–142.
8. Tamimi D 2002 The association between chronic lympho-
cytic thyroiditis and thyroid tumors. Int J Surg Pathol
10:141–146.
9. Kebebew E, Treseler PA, Ituarte PH, Clark OH 2001 Coex-
AUTOIMMUNITY: THYROID CARCINOMA RISK FACTOR
isting chronic lymphocytic thyroiditis and papillary thyroid
cancer revisited. World J Surg 25:632–637.
10. Okayasu I 1997 The relationship of lymphocytic thyroiditis
to the development of thyroid carcinoma. Endocr Pathol Au-
tunn 8:225–230.
11. Pacini F, Mariotti S, Formica N, Elisei R, Anelli S, Capotorti
E, Pinchera A 1988 Thyroid autoantibodies in thyroid can-
cer: incidence and relationship with tumour outcome. Acta
Endocrinol (Copenh) 119:373–380.
12. Neracher H, Hedinger C 1975 Classification of thyroid gland
tumors in accordance with 1974 WHO nomenclature. His-
tological follow-up examination of 327 malignant thyroid
gland tumors. Schweiz Med Wochenschr 105:1000–1006.
13. Mazzaferri EL, Jhiang SM 1994 Long-term impact of initial
surgical and medical therapy on papillary thyroid cancer.
Am J Med 97:418–428.
14. Schlumberger MJ, Torlantano M 2000 Papillary and follicu-
lar thyroid carcinoma. Baillieres Best Pract Res Clin En-
docrinol Metab 14:601–613.
15. Shaha AR 1998 Thyroid carcinoma: implications of prog-
nostic factors. Cancer 83:401–402.
16. Hadjieva T 2001 Scoring patients’ risk in differentiated thy-
roid cancer. Onkologie 24:561–568.
17. Beasley NJ, Lee J, Eski S, Walfish P, Witterick I, Freeman JL
2002 Impact of nodal metastases on prognosis in patients
with well-differentiated thyroid cancer. Arch Otolaryngol
Head Neck Surg 128:825–828.
18. Witte J, Goretzki PE, Dieken J, Simon D, Roher HD 2002 Im-
portance of lymph node metastases in follicular thyroid can-
cer. World J Surg 26:8:1017–1022.
19. Mellemgaard A, From G, Jorgensen T, Johansen C, Olsen JH,
Perrild H 1998 Cancer risk in individuals with benign thy-
roid disorders. Thyroid 8:751–754.
20. From G, Mellemgaard A, Knudsen N, Jorgensen T, Perrild
H 2000 Review of thyroid cancer cases among patients with
previous benign thyroid disorders. Thyroid 10:697–700.
495
21. Franceschi S, Preston-Martin S, Dal Maso L, La Vecchia C,
Mack WJ, Mc Tiernan A, Kolonel L, Mark SD, Mabuchi K,
Jin F, Wingren G, Galanti R, Hallquist A, Glattre E, Lund E,
Levi F, Linos D, Ron E 1999 A pooled analysis of case-con-
trol studies of thyroid cancer. IV. Benign thyroid diseases.
Cancer Causes Control 10:583–595.
22. Mazzaferri EL 1990 Thyroid cancer and Graves’ disease.
J Clin Endocrinol Metab 70:826–829.
23. Mazzaferri EL 2000 Thyroid cancer and Graves’ disease: The
controversy ten years later. Endocr Pract 6:221–225.
24. Kashima K, Yokoyama S, Noguchi S, Murakami N, Yamashita
H, Watanabe S, Uchino S, Toda M, Sasaki A, Daa T, Nakayama
I 1998 Chronic thyroiditis as a favorable prognostic factor in
papillary thyroid carcinoma. Thyroid 8:197–202.
25. Loh KC, Greenspan FS, Dong F, Miller TR, Yeo PP 1999 In-
fluence of lymphocytic thyroiditis on the prognostic out-
come of patients with papillary thyroid carcinoma. J Clin
Endocrinol Metab 84:458–463.
26. Singh B, Shaha AR, Trivedi H, Carew JF, Poluri A, Shah JP
1999 Coexistent Hashimoto’s thyroiditis with papillary thy-
roid carcinoma: Impact on presentation, management, and
outcome. Surgery 26:1070–1076.
27. Gupta S, Patel A, Folstad A, Fenton C, Dinauer CA, Tuttle
RM, Conran R, Francis GL 2001 Infiltration of differentiated
thyroid carcinoma by proliferating lymphocytes is associ-
ated with improved disease-free survival for children and
young adults. J Clin Endocrinol Metab 86:1346–1354.
Address reprint requests to:
Suzikelli Lisboa Souza
Laboratory of Cancer Molecular Genetics
Department of Medicine
State University of Campinas
São Paulo
Brazil