Page 1 of 26
Accepted Preprint first posted on 12 September 2014 as Manuscript ERC-14-0258
1 1
Title: Association of Hashimoto’s Thyroiditis with Thyroid Cancer
2
By: G Azizi, MD1; JM Keller, MD2; M Lewis, PA-C1; K Piper, MS3; D Puett, MD4; KM
3
Rivenbark, BA1; CD Malchoff, MD5
4
1
Wilmington Endocrinology
1717 Shipyard Boulevard, Wilmington, NC 28403
5
2
6
Wilmington Pathology Associates
1915 South 17th Street, Suite 100 Wilmington, NC 28401
7
3
8
Children’s Hospital Colorado
13123 East 16th Avenue, Aurora, CO 80045
9
4
10
Carolina Arthritis
1710 South 17th Street, Wilmington, NC 28401
11
5
University of Connecticut Health Center
12
263 Farmington Avenue, Farmington, CT 06030
13
14
Corresponding Author:
15
Ghobad Azizi
16
Wilmington Endocrinology
17
1717 Shipyard Boulevard, Suite 220, Wilmington, NC 20403
18
Email: [email protected]
19
20
Running Title: Hashimoto’s Thyroiditis and Thyroid Cancer
21
Keywords: Hashimoto’s thyroiditis, Thyroid Cancer, Thyroid Nodule, FNAB, TgAb
22
Word Count: 3396
23
Copyright © 2014 by the Society for Endocrinology.
Page 2 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 2 24
Abstract
25
Objective: This prospective study investigates the relationship between
26
Hashimoto’s Thyroiditis (HT) and thyroid cancer (TC) in patients with thyroid nodules
27
(TNs). Methods: We prospectively examined 2100 patients with 2753 TNs between
28
January 5, 2010 and August 15, 2013. A total of 2023 patients with 2669 TNs met the
29
inclusion criteria of TN >5 mm and age >18 years (y). Each patient had blood drawn
30
prior to fine-needle aspiration biopsy (FNAB) for the following measurements: TSH, free
31
T4, free T3, thyroid peroxidase antibody (TPO Ab), and antithyroglobulin antibody
32
(TgAb). Diagnosis of TC was based on pathology analysis of thyroidectomy tissue.
33
Associations of TC with the independent variables were determined by univariate and
34
multivariate logistic regression analysis and reported as adjusted odds ratio (OR) with
35
95% confidence interval (CI). Results: 248 malignant nodules were found in 233
36
patients. There was an association of TC with both increased serum TgAb
37
concentration and age 1 µIU/ml are independent predictors for TC. The
44
association between HT and TC is antibody specific.
Page 3 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 3 45
46
Introduction
Hashimoto’s Thyroiditis (HT) is the most common autoimmune thyroid disease
47
48
and the most common cause of hypothyroidism (Jankovic et al. 2013). Epidemiological
49
and histological data indicate that thyroid cancer (TC) frequently occurs in the context of
50
one of the most common autoimmune thyroid diseases, HT, and that TC is frequently
51
infiltrated by inflammatory-immune cells (Guarino et al. 2010). HT is characterized by infiltration of the thyroid gland by inflammatory cells. This
52 53
often leads to hypothyroidism due to destruction and eventual fibrous replacement of
54
the parenchymal tissue. The relationship between HT and papillary carcinoma (PC)
55
was first proposed by Dailey et al. in 1955 (Dailey et al. 1955). Since this initial
56
description, the association between the diseases has been repeatedly reported and
57
highly debated in the literature and remains controversial (Cunha et al. 2011).
A relationship between chronic inflammation and cancer was first proposed by
58
59
Virchow in 1863 and has been sustained by clinical and epidemiological evidence. The
60
most compelling evidence is the association between: (a) intestinal chronic
61
inflammatory diseases (Crohn’s disease and ulcerative rectocolitis) and
62
adenocarcinoma of the colon; (b) chronic HBV or HCV hepatitis and liver carcinoma; (c)
63
Heliobacter pylori-induced chronic gastritis and gastric carcinoma; (d) asbestosis and
64
mesothelioma; (e) chronic obstructive pulmonary disease (COPD) and lung cancer; (f)
65
scleroderma and carcinoma of the breast and lung (Guarino et al. 2010; Joseph et al.
66
2014).
Page 4 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 4 67
Two recent retrospective studies reported a positive association between HT and
68
TC in patients presenting with thyroid nodules (TNs). In these two studies TgAb was an
69
independent risk factor for TC, but TPO Ab was not (Azizi & Malchoff 2011; Kim et al.
70
2010).
71
Based on these previous retrospective studies, we hypothesized that the
72
association of HT with TC is antibody specific. Since a retrospective study’s design
73
may be subject to ascertainment bias and assay heterogeneity, we chose to further
74
investigate the relationship between HT and TC by prospectively examining this
75
relationship in TN patients using a multivariate approach to account for the possible
76
influence of other variables including levels of thyroid stimulating hormone (TSH).
77
Materials and Methods
78
Patients
79
80
81
The inclusion criteria were the presence of a single or multiple TNs >5mm and
age >18 years (y).
We prospectively evaluated 2100 patients with 2753 TNs from January 5, 2010
82
to August 15, 2013. All TNs were evaluated with a high-resolution ultrasound (US),
83
fine-needle aspiration biopsy (FNAB), and, when indicated, by tissue pathology. All
84
patients were evaluated by a single endocrinologist (G.A.) with more than 15 years of
85
experience in thyroid US and ultrasound-guided FNAB.
86
The study protocol was approved by the Institutional Review Board of the New
87
Hanover Regional Medical Center, Wilmington, NC. All patients gave written informed
88
consent.
89
Biochemical Assays
Page 5 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 5 90
At the time of initial US exam, the following factors were ascertained in all study
91
subjects: gender, age, and number of US-determined TN. Treatment with any TSH-
92
altering medication was documented. All patients had their blood drawn prior to FNAB
93
for the following measurements: TSH, free T4, free T3, thyroid peroxidase antibody
94
(TPO Ab) and antithyroglobulin antibody (TgAb). All measurements were performed by
95
LabCorp (LabCorp Research Triangle Park-Center for Molecular Biology & Pathology,
96
Esoteric Division, Research Triangle Park, North Carolina). Both TPO Ab and TgAb
97
were measured using Immulite 2000 technology by Siemens Medical solutions
98
diagnostics (Brewster, NY) utilizing purified antigen (TPO) immobilized to insoluble
99
beads in a solid phase, enzyme-labeled chemiluminescent sequential immunometric
100
assay. The reference range for TPO Ab and TgAb were 0 to 34 IU/ml and 0 to 40 IU/ml,
101
respectively. The analytical sensitivity for TPO Ab and TgAb was 5 IU/ml and 2.2 IU/ml,
102
respectively. Serum TSH was measured at LabCorp by an electrochemiluminescence
103
immunoassay developed by Roche Diagnostic. The reference range for TSH was 0.45-
104
4.5 µIU/ml. Hypothyroidism is defined as TSH >4.5 µIU/ml without thyroid medication
105
and in patients with known history of hypothyroidism on thyroid hormone replacement.
106
Hyperthyroidism is defined as TSH 1 µIU/mL was a predictor for higher risk for cancer (OR (95% CI))
202
OR: 1.49 (1.09, 2.03). In an age and gender adjusted multivariate logistic regression
Page 10 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 10 203
analysis of subjects who were not taking any TSH-altering medication, TgAb remained a
204
significant predictor of cancer (OR (95% CI)): OR: 2.28 (1.42, 3.66).
205
Table 4 shows a detailed breakdown of all patients included in the study. We
206
divided patients with normal TgAb by those with undetectable TgAb (Ref: 0) and those
207
with detectable TgAb, but still within normal range (>0 to 40 to 200.
212
In the first group we had a total of 122 patients with 24 cancer and 98 without TC. In the
213
final group with TgAb >200, we had 106 patients. Among those patients, 19 had TC,
214
and 87 had no cancer. Table 5 contains a detailed breakdown of the number of patients
215
by TSH concentration.
216
Discussion
217
Our prospective study of patients with TN demonstrates that an elevated serum
218
TgAb is an independent predictor for TC. Serum TPO Ab was not a predictor of TC in
219
either univariate or multivariate analyses.
220
This prospective study was undertaken to further investigate the risk of thyroid
221
inflammation and TC. It confirms two retrospective studies that suggested this
222
association (2, 22). The normal reference range for TgAb for this study’s assay was 0-
223
40 IU/mL. In multivariate analysis, we found that patients with detectable but normal
224
TgAb concentrations (>0 to 1
445
Page 26 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 26 446 447
Table 4: Breakdown of patients by TgAb level in Age and Gender Adjusted Multivariate
448
Logistic Regression
449
A detailed breakdown of all patients included in the study. All patients were divided into those
450
with undetectable TgAb (Ref: 0), and those with detectable but normal range TgAb (>0-40-200.
453
Table of cancer by TgAb Level
Frequency
Ref: 0
>0 to 40 to 200
Total
No Cancer
1531
74
98
87
1790
Cancer
175
15
24
19
233
Total
1706
89
122
106
2023
454
455
Table 5: Breakdown of patients by TSH level in Age and Gender Adjusted Multivariate
456
Logistic Regression
457
A detailed breakdown of all patients included in the study. All patients were divided based on
458
TSH level.
459
Table of cancer by TSH Level
Frequency
No Cancer
Cancer
Total
460
Ref:
0 to 0.45 >0.45 to 2 to 4.5
Total
232
995
438
125
1790
26
128
61
18
233
258
1123
499
143
2023
Accepted Preprint first posted on 12 September 2014 as Manuscript ERC-14-0258
1 1
Title: Association of Hashimoto’s Thyroiditis with Thyroid Cancer
2
By: G Azizi, MD1; JM Keller, MD2; M Lewis, PA-C1; K Piper, MS3; D Puett, MD4; KM
3
Rivenbark, BA1; CD Malchoff, MD5
4
1
Wilmington Endocrinology
1717 Shipyard Boulevard, Wilmington, NC 28403
5
2
6
Wilmington Pathology Associates
1915 South 17th Street, Suite 100 Wilmington, NC 28401
7
3
8
Children’s Hospital Colorado
13123 East 16th Avenue, Aurora, CO 80045
9
4
10
Carolina Arthritis
1710 South 17th Street, Wilmington, NC 28401
11
5
University of Connecticut Health Center
12
263 Farmington Avenue, Farmington, CT 06030
13
14
Corresponding Author:
15
Ghobad Azizi
16
Wilmington Endocrinology
17
1717 Shipyard Boulevard, Suite 220, Wilmington, NC 20403
18
Email: [email protected]
19
20
Running Title: Hashimoto’s Thyroiditis and Thyroid Cancer
21
Keywords: Hashimoto’s thyroiditis, Thyroid Cancer, Thyroid Nodule, FNAB, TgAb
22
Word Count: 3396
23
Copyright © 2014 by the Society for Endocrinology.
Page 2 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 2 24
Abstract
25
Objective: This prospective study investigates the relationship between
26
Hashimoto’s Thyroiditis (HT) and thyroid cancer (TC) in patients with thyroid nodules
27
(TNs). Methods: We prospectively examined 2100 patients with 2753 TNs between
28
January 5, 2010 and August 15, 2013. A total of 2023 patients with 2669 TNs met the
29
inclusion criteria of TN >5 mm and age >18 years (y). Each patient had blood drawn
30
prior to fine-needle aspiration biopsy (FNAB) for the following measurements: TSH, free
31
T4, free T3, thyroid peroxidase antibody (TPO Ab), and antithyroglobulin antibody
32
(TgAb). Diagnosis of TC was based on pathology analysis of thyroidectomy tissue.
33
Associations of TC with the independent variables were determined by univariate and
34
multivariate logistic regression analysis and reported as adjusted odds ratio (OR) with
35
95% confidence interval (CI). Results: 248 malignant nodules were found in 233
36
patients. There was an association of TC with both increased serum TgAb
37
concentration and age 1 µIU/ml are independent predictors for TC. The
44
association between HT and TC is antibody specific.
Page 3 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 3 45
46
Introduction
Hashimoto’s Thyroiditis (HT) is the most common autoimmune thyroid disease
47
48
and the most common cause of hypothyroidism (Jankovic et al. 2013). Epidemiological
49
and histological data indicate that thyroid cancer (TC) frequently occurs in the context of
50
one of the most common autoimmune thyroid diseases, HT, and that TC is frequently
51
infiltrated by inflammatory-immune cells (Guarino et al. 2010). HT is characterized by infiltration of the thyroid gland by inflammatory cells. This
52 53
often leads to hypothyroidism due to destruction and eventual fibrous replacement of
54
the parenchymal tissue. The relationship between HT and papillary carcinoma (PC)
55
was first proposed by Dailey et al. in 1955 (Dailey et al. 1955). Since this initial
56
description, the association between the diseases has been repeatedly reported and
57
highly debated in the literature and remains controversial (Cunha et al. 2011).
A relationship between chronic inflammation and cancer was first proposed by
58
59
Virchow in 1863 and has been sustained by clinical and epidemiological evidence. The
60
most compelling evidence is the association between: (a) intestinal chronic
61
inflammatory diseases (Crohn’s disease and ulcerative rectocolitis) and
62
adenocarcinoma of the colon; (b) chronic HBV or HCV hepatitis and liver carcinoma; (c)
63
Heliobacter pylori-induced chronic gastritis and gastric carcinoma; (d) asbestosis and
64
mesothelioma; (e) chronic obstructive pulmonary disease (COPD) and lung cancer; (f)
65
scleroderma and carcinoma of the breast and lung (Guarino et al. 2010; Joseph et al.
66
2014).
Page 4 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 4 67
Two recent retrospective studies reported a positive association between HT and
68
TC in patients presenting with thyroid nodules (TNs). In these two studies TgAb was an
69
independent risk factor for TC, but TPO Ab was not (Azizi & Malchoff 2011; Kim et al.
70
2010).
71
Based on these previous retrospective studies, we hypothesized that the
72
association of HT with TC is antibody specific. Since a retrospective study’s design
73
may be subject to ascertainment bias and assay heterogeneity, we chose to further
74
investigate the relationship between HT and TC by prospectively examining this
75
relationship in TN patients using a multivariate approach to account for the possible
76
influence of other variables including levels of thyroid stimulating hormone (TSH).
77
Materials and Methods
78
Patients
79
80
81
The inclusion criteria were the presence of a single or multiple TNs >5mm and
age >18 years (y).
We prospectively evaluated 2100 patients with 2753 TNs from January 5, 2010
82
to August 15, 2013. All TNs were evaluated with a high-resolution ultrasound (US),
83
fine-needle aspiration biopsy (FNAB), and, when indicated, by tissue pathology. All
84
patients were evaluated by a single endocrinologist (G.A.) with more than 15 years of
85
experience in thyroid US and ultrasound-guided FNAB.
86
The study protocol was approved by the Institutional Review Board of the New
87
Hanover Regional Medical Center, Wilmington, NC. All patients gave written informed
88
consent.
89
Biochemical Assays
Page 5 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 5 90
At the time of initial US exam, the following factors were ascertained in all study
91
subjects: gender, age, and number of US-determined TN. Treatment with any TSH-
92
altering medication was documented. All patients had their blood drawn prior to FNAB
93
for the following measurements: TSH, free T4, free T3, thyroid peroxidase antibody
94
(TPO Ab) and antithyroglobulin antibody (TgAb). All measurements were performed by
95
LabCorp (LabCorp Research Triangle Park-Center for Molecular Biology & Pathology,
96
Esoteric Division, Research Triangle Park, North Carolina). Both TPO Ab and TgAb
97
were measured using Immulite 2000 technology by Siemens Medical solutions
98
diagnostics (Brewster, NY) utilizing purified antigen (TPO) immobilized to insoluble
99
beads in a solid phase, enzyme-labeled chemiluminescent sequential immunometric
100
assay. The reference range for TPO Ab and TgAb were 0 to 34 IU/ml and 0 to 40 IU/ml,
101
respectively. The analytical sensitivity for TPO Ab and TgAb was 5 IU/ml and 2.2 IU/ml,
102
respectively. Serum TSH was measured at LabCorp by an electrochemiluminescence
103
immunoassay developed by Roche Diagnostic. The reference range for TSH was 0.45-
104
4.5 µIU/ml. Hypothyroidism is defined as TSH >4.5 µIU/ml without thyroid medication
105
and in patients with known history of hypothyroidism on thyroid hormone replacement.
106
Hyperthyroidism is defined as TSH 1 µIU/mL was a predictor for higher risk for cancer (OR (95% CI))
202
OR: 1.49 (1.09, 2.03). In an age and gender adjusted multivariate logistic regression
Page 10 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 10 203
analysis of subjects who were not taking any TSH-altering medication, TgAb remained a
204
significant predictor of cancer (OR (95% CI)): OR: 2.28 (1.42, 3.66).
205
Table 4 shows a detailed breakdown of all patients included in the study. We
206
divided patients with normal TgAb by those with undetectable TgAb (Ref: 0) and those
207
with detectable TgAb, but still within normal range (>0 to 40 to 200.
212
In the first group we had a total of 122 patients with 24 cancer and 98 without TC. In the
213
final group with TgAb >200, we had 106 patients. Among those patients, 19 had TC,
214
and 87 had no cancer. Table 5 contains a detailed breakdown of the number of patients
215
by TSH concentration.
216
Discussion
217
Our prospective study of patients with TN demonstrates that an elevated serum
218
TgAb is an independent predictor for TC. Serum TPO Ab was not a predictor of TC in
219
either univariate or multivariate analyses.
220
This prospective study was undertaken to further investigate the risk of thyroid
221
inflammation and TC. It confirms two retrospective studies that suggested this
222
association (2, 22). The normal reference range for TgAb for this study’s assay was 0-
223
40 IU/mL. In multivariate analysis, we found that patients with detectable but normal
224
TgAb concentrations (>0 to 1
445
Page 26 of 26
Hashimoto’s Thyroiditis and Thyroid Cancer 26 446 447
Table 4: Breakdown of patients by TgAb level in Age and Gender Adjusted Multivariate
448
Logistic Regression
449
A detailed breakdown of all patients included in the study. All patients were divided into those
450
with undetectable TgAb (Ref: 0), and those with detectable but normal range TgAb (>0-40-200.
453
Table of cancer by TgAb Level
Frequency
Ref: 0
>0 to 40 to 200
Total
No Cancer
1531
74
98
87
1790
Cancer
175
15
24
19
233
Total
1706
89
122
106
2023
454
455
Table 5: Breakdown of patients by TSH level in Age and Gender Adjusted Multivariate
456
Logistic Regression
457
A detailed breakdown of all patients included in the study. All patients were divided based on
458
TSH level.
459
Table of cancer by TSH Level
Frequency
No Cancer
Cancer
Total
460
Ref:
0 to 0.45 >0.45 to 2 to 4.5
Total
232
995
438
125
1790
26
128
61
18
233
258
1123
499
143
2023
