Family history of cancer and risk of sporadic differentiated thyroid carcinoma†
We thank Margaret Lung, Kathyrn Patterson, Liliana Mugartegui, and Jenny Vo for their help with subject recruitment and Stephanie Deming for manuscript editing.
Thyroid cancer incidence in the United States, particularly in women, has increased dramatically since the 1980s. Although the causes of thyroid cancer in most patients remain largely unknown, evidence suggests the existence of an inherited predisposition to development of differentiated thyroid carcinoma (DTC). Therefore, the authors explored the association between sporadic DTC and family history of cancer.
In a retrospective hospital-based case-control study of prospectively recruited subjects who completed the study questionnaire upon enrollment, unconditional logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) as estimates of the DTC risk associated with first-degree family history of cancer.
The study included 288 patients with sporadic DTC and 591 cancer-free controls. Family history of thyroid cancer in first-degree relatives was associated with increased DTC risk (adjusted OR, 4.1; 95% CI, 1.7-9.9). All DTC cases in patients with a first-degree family history of thyroid cancer were cases of papillary thyroid carcinoma (PTC) (adjusted OR, 4.6; 95% CI, 1.9-11.1). Notably, the risk of PTC was highest in subjects with a family history of thyroid cancer in siblings (OR, 7.4; 95% CI, 1.8-30.4). In addition, multifocal primary tumor was more common among PTC patients with first-degree family history of thyroid cancer than among PTC patients with no first-degree family history of thyroid cancer (68.8% vs 35.5%, P = .01).
The study suggests that family history of thyroid cancer in first-degree relatives, particularly in siblings, is associated with an increased risk of sporadic PTC. Cancer 2012;. © 2011 American Cancer Society.
Thyroid cancer is the most prevalent endocrine malignancy.1 In the United States, it was predicted to account for nearly 3% of cancer diagnoses in 2010 (45,000 new cases), and it is the fifth most common cancer in women.1 The incidence of thyroid cancer is increasing more rapidly than that of any other cancer in the United States; the incidence of thyroid cancer today is 2.4× what it was 3 decades ago.2, 3 It is argued that the rising incidence of thyroid cancer may largely be attributed to improved diagnostic procedures and advanced screening.3-5 However, because there has been an increase in the incidence of not only small, clinically insignificant tumors but also larger, clinically significant tumors, improved screening and medical practices alone do not fully explain the observed increase.4, 5
Differentiated thyroid carcinoma (DTC) comprises approximately 90% of all thyroid cancers and consists of 3 distinct histological types: papillary (80%-90% of cases), follicular (10%), and Hurthle cell.6, 7 The etiology of DTC is largely unknown and may vary according to histological type.8 DTC risk has been examined in relation to family history of cancer in several epidemiologic studies, many of which reported a family cluster of thyroid cancer,9-19 suggesting a potential interplay of genetic and environmental factors in thyroid carcinogenesis. However, most of these studies did not provide histology-specific risk. Although several families with a cluster of thyroid cancer reported a more aggressive clinical course,7, 20 epidemiological studies of the association between family history of thyroid cancer and pathoclinical feature of DTC are limited.
Thus, to address these issues, we carried out a hospital-based case-control study. To exclusively investigate the association with sporadic DTC, we excluded familial DTC cases according to the standard definition of ≥3 first-degree family members with thyroid cancer in a kindred.7, 20 In the analysis reported here, to assess whether first-degree family history of cancer was associated with risk of sporadic DTC, particularly sporadic papillary thyroid carcinoma (PTC), we retrospectively compared the proportion of first-degree family history of cancer among 3 groups including: patients with incident DTC, patients with incident benign thyroid disease, and cancer-free controls. Pathologic characteristics of DTC were also examined to explore the clinical relevance of family history of cancer.
MATERIALS AND METHODS
In this retrospective hospital-based case-control study, cases and controls were prospectively recruited to and enrolled in institutional review board-approved studies of cancer susceptibility at The University of Texas MD Anderson Cancer Center, and each participant prospectively provided written informed consent and completed the questionnaire described below. The study included 2 case groups and 1 cancer-free control group. Incident cases with papillary, follicular, or Hurthle cell carcinoma of the thyroid composed the DTC case group, and incident cases with thyroid benign mass composed the benign thyroid disease case group (an intermediate-risk comparison group).
Between November 1999 and November 2010, we prospectively recruited patients who presented to our institution for evaluation of a thyroid gland mass. The final diagnoses of DTC and benign thyroid disease were assigned by histologic review of the surgical specimen. Patients who did not fit the recruitment criteria (>17 years of age, no prior cancer history except non-melanoma skin cancer, no current use of steroids or immunosuppressive medication, no blood transfusion in the previous 6 months) were excluded.
Controls were visitors to our institution using the same exclusion criteria who were recruited in a molecular epidemiologic study of head and neck squamous cell carcinoma between October 2001 and November 2009. For the current study, controls were retrospectively frequency-matched with patients with DTC by sex.
DTC cases, benign thyroid disease patients, and controls completed the same self-administered questionnaires. Data were obtained on demographic information, environmental exposure, and personal and family history. The family history of cancer included questions about subjects' adoption status; number of first-degree relatives (parents, siblings, and children); current age or age at death of those first-degree relatives; and age at diagnosis of cancer and site of cancer in first-degree relatives with cancer. Subjects who had cumulatively smoked >100 cigarettes in their lifetimes were defined as smokers (current or former). Subjects who had drunk alcoholic beverages at least once a week for >1 year were defined as drinkers (current or former). Former smokers and former drinkers were those subjects who had quit smoking or drinking at least 1 year before study enrollment. Radiation exposure was defined as previous whole-body or head-and-neck–specific radiotherapy.
We used chi-square test, Fisher exact test, or t test to compare distributions of characteristics between case and control groups as appropriate. Unconditional logistic regression models were used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) associated with first-degree family history of cancer, using subjects with no family history of cancer as the reference group. To adjust for potential confounders, the variables were included in the final models if they changed the target OR by 10% or more when added to the unadjusted model,21 including education, family income level, smoking status, and alcohol drinking status. In addition, age at reference date (age at diagnosis for cases and age at recruitment for controls), sex, race, and number of family members (or siblings or children, as appropriate) were fitted into the final regression model. Furthermore, we estimated the adjusted OR and 95% CI for PTC risk in association with a family history of thyroid cancer in different types of relatives and stratified by sex. A P value of <.05 (2-tailed) was considered statistically significant. Statistical analysis was performed using SAS version 9.2 software (SAS Institute Inc., Cary, NC).
We successfully enrolled 290 patients with DTC and 197 patients with benign thyroid disease. Of these, 1 patient with DTC and 9 patients with benign thyroid disease had not completed questionnaires. After review of the questionnaire data, 1 patient with PTC was excluded because 2 first-degree relatives had been diagnosed with thyroid cancer. The final study included 288 patients with DTC, 188 patients with benign thyroid disease, and 591 cancer-free controls.
Compared with cancer-free controls, DTC patients had a higher family income and were less likely to smoke (Table 1). Also, the proportion of non-Hispanic whites was lower in the DTC group than in the control group. DTC and benign thyroid disease patients were similar in distribution of sex, ethnicity, education, family income, smoking, and alcohol drinking status. The age at recruitment was significantly greater for controls and benign thyroid disease patients than the age at diagnosis for DTC patients (mean ± standard deviation: 53.6 ± 13.0 years, 50.3 ± 13.7 years, and 44.8 ± 14.1 years, respectively, P < .01). Only 4 DTC patients, 4 benign thyroid disease patients, and 8 controls reported a history of radiotherapy, and these differences were not statistically different. The great majority (91.3%) of patients with DTC had PTC, and classic PTC was the most common subtype (71.5%).
Table 1. Demographic and Clinical Characteristics of Cases and Controls
|Sex|| || || |
| Male||95 (33.0)||83 (31.6)a||228 (38.6)|
| Female||193 (67.0)||180 (68.4)||363 (61.4)|
|Age, years|| || || |
| <45||155 (53.8)a||147 (55.9)a||148 (25.0)|
| ≥45||133 (46.2)||116 (44.1)||443 (75.0)|
|Ethnicity|| || || |
| Non-Hispanic whites||198 (68.7)a||181 (68.8)a||449 (76.0)|
| Other||90 (31.3)||82 (31.2)||142 (24.0)|
|Residence|| || || |
| Texas||229 (79.5)||207 (78.7)||488 (82.6)|
| Others||59 (20.5)||56 (21.3)||103 (17.4)|
|Educational level|| || || |
| High school graduate or less||68 (23.7)||61 (23.3)||178 (30.1)|
| Some college||89 (31.0)||83 (31.7)||178 (30.1)|
| College graduate or advanced||130 (45.3)||118 (45.0)||235 (39.8)|
|Family income, $/y|| || || |
| <35,000||64 (23.3)a||54 (21.3)a||122 (21.7)|
| 35,000-75,000||74 (26.9)||73 (28.9)||214 (38.0)|
| >75,000||137 (49.8)||126 (49.8)||227 (40.3)|
| Never||193 (67.0)a||176 (66.9)a||334 (56.5)|
| Former||58 (20.1)||51 (19.4)||163 (27.6)|
| Current||37 (12.9)||36 (13.7)||94 (15.9)|
|Alcohol drinking status|
| Never||165 (57.3)||149 (56.6)||330 (55.9)|
| Former||25 (8.7)||21 (8.0)||80 (13.5)|
| Current||98 (34.0)||93 (35.4)a||181 (30.6)|
|Radiotherapy history|| || || |
| Yes||4 (1.4)||4 (1.5)||8 (1.4)|
| No||284 (98.6)||259 (98.5)||583 (98.6)|
|Histological type|| || || |
| Papillary||263 (91.0)|| || |
| Classic PTC|| ||188 (72.9)|| |
| Follicular variant PTC|| ||60 (23.2)|| |
| Others|| ||10 (3.9)|| |
| Follicular/Hurthle cell||26 (9.0)|| || |
|Multifocal primary|| || || |
| Yes||100 (35.3)||97 (37.6)|| |
| No||183 (64.7)||161 (62.4)|| |
|Thyroiditis|| || || |
| Yes||63 (22.3)||57 (22.1)|| |
| No||220 (77.7)||201 (77.9)|| |
Family history of cancer in first-degree relatives was reported by 141 (49.0%) of the patients with DTC, 103 (54.8%) of the patients with benign thyroid disease, and 343 (58.0%) of the cancer-free controls. No significant association with family history of cancer in first-degree relatives was found for DTC (adjusted OR, 1.0; 95% CI, 0.7-1.4) or benign thyroid disease (adjusted OR, 0.9; 95% CI, 0.7-1.3).
Table 2 shows the association between self-reported family history of cancer at various sites in first-degree relatives and risk of DTC. The only cancers for which significant associations were found were thyroid cancer and liver cancer. Patients with DTC were significantly more likely than controls to report a family history of thyroid cancer in first-degree relatives (6.3% vs 1.4%; adjusted OR, 4.1; 95% CI, 1.7-9.9). A subgroup analysis of the subjects with Texas residence revealed a similar association between a first-degree family history of thyroid cancer and DTC risk (adjusted OR, 4.8; 95% CI, 1.6-14.7). All patients with DTC who had a first-degree family history of thyroid cancer (18 patients) had PTC. The association with first-degree family history of thyroid cancer was slightly stronger for PTC (adjusted OR, 4.6; 95% CI, 1.9-11.1) (Table 3) than for DTC overall.
Table 2. Family History of Cancer in First-Degree Relatives in DTC Cases and Controls, by Cancer Type
|Salivary gland||1 (0.4)||2 (0.3)||1.3 [0.1-14.5]|
|Mouth||2 (0.7)||5 (0.9)||0.7 [0.1-3.6]|
|Pharynx and throat||2 (0.7)||8 (1.4)||0.6 [0.1-2.8]|
|Esophagus||2 (0.7)||8 (1.4)||0.7 [0.1-3.5]|
|Stomach||8 (2.8)||9 (1.5)||1.9 [0.7-5.4]|
|Small intestine||0||2 (0.3)||—|
|Colon and rectum||11 (3.8)||41 (6.9)||0.8 [0.4-1.5]|
|Liver||11 (3.8)||12 (2.0)||2.6 [1.1-6.3]|
|Gallbladder||1 (0.4)||1 (0.2)||8.8 [0.5-148.7]|
|Pancreas||5 (1.7)||13 (2.2)||1.2 [0.4-3.4]|
|Lung||27 (9.4)||67 (11.3)||1.1 [0.7- 1.9]|
|Bone and joints||2 (0.7)||4 (0.7)||1.9 [0.3-11.0]|
|Soft tissue||0||5 (0.9)||—|
|Melanoma||8 (2.8)||30 (5.1)||0.6 [0.3-1.4]|
|Nonmelanoma skin||9 (3.1)||33 (5.6)||0.7 [0.3-1.6]|
|Breast||25 (8.7)||76 (12.9)||0.8 [0.5-1.4]|
|Uterus||6 (2.1)||8 (1.4)||1.6 [0.5-4.9]|
|Cervix||4 (1.4)||10 (1.7)||0.8 [0.2-2.7]|
|Ovary||5 (1.7)||10 (1.7)||1.0 [0.3-3.1]|
|Prostate||24 (8.3)||53 (9.0)||1.2 [0.7-2.0]|
|Testis||2 (0.7)||4 (0.7)||1.1 [0.2-6.1]|
|Urinary bladder||4 (1.4)||3 (0.5)||2.5 [0.5-11.6]|
|Kidney||6 (2.1)||12 (2.0)||1.4 [0.5-4.0]|
|Ureter and other urinary organs||0||1 (0.2)||—|
|Eye and orbit||0||1 (0.2)||—|
|Brain and other nervous system||5 (1.7)||14 (2.4)||0.9 [0.3-2.7]|
|Thyroid||18 (6.3)||8 (1.4)||4.1 [1.7-9.9]|
|Lymphoma||3 (1.0)||20 (3.4)||0.4 [0.1-1.4]|
|Leukemia||6 (2.1)||16 (2.7)||1.1 [0.4-3.1]|
|Other and unspecified sites||6 (2.1)||10 (1.7)||2.2 [0.7-6.4]|
|All||141 (49.0)||343 (58.0)||1.0 [0.7-1.4]|
Table 3. Family History of Thyroid Cancer in First-Degree Relatives in PTC Cases and Controls, by Type of Relative
|First-degree relative||18/8||4.6 (1.9-11.1)||14/6||4.4 (1.6-12.2)||4/2||5.3 (0.9-30.8)|
| Parent||10/4||3.3 (1.0-11.0)||9/3||3.4 (0.9-13.3)||1/1||1.9 (0.1-32.6)|
| Sibling||7/3||7.4 (1.8-30.4)||5/3||5.9 (1.3-26.2)||2/0||—|
| Child||1/1||2.4 (0.1-57.7)||0/0||—||1/1||2.8 (0.2-49.8)|
|First-degree female relative||14/7||4.5 (1.7-11.8)||11/6||3.6 (1.2-10.4)||3/1||9.9 (1.0-102.6)|
| Mother||8/4||2.8 (0.8-9.7)||7/3||2.9 (0.7-11.6)||1/1||1.9 (0.1-32.6)|
| Sister||5/3||5.9 (1.3-26.2)||4/3||4.8 (1.0-23.1)||1/0||—|
|First-degree male relative||4/1||4.6 (0.5-42.9)||3/0||—||1/1||1.7 (0.1-29.0)|
Family history of thyroid cancer in first-degree relatives was associated with an increased risk of benign thyroid disease (adjusted OR, 3.2; 95% CI, 1.2-8.7). Of the 188 patients with benign thyroid disease, 9 (4.8%) reported a first-degree family history of thyroid cancer; this proportion was nonsignificantly lower than the proportion of patients with DTC who reported a first-degree family history of thyroid cancer (6.3%, P = .53).
An analysis of the association between first-degree family history of thyroid cancer and PTC risk according to the type of relative was carried out in all subjects and subsequently in women and men separately (Table 3). The PTC risk was more evident in subjects with a family history of thyroid cancer in siblings (OR, 7.4; 95% CI, 1.8-30.4), and the same pattern was found when the analysis was limited to women (OR, 5.9; 95% CI, 1.3-26.2), although the sample sizes of these individual subgroups were limited.
Results of a descriptive analysis of the 18 DTC patients with a first-degree relative diagnosed with thyroid cancer are presented in Table 4. As previously mentioned, all of the patients in this group had histologically confirmed PTC, although the proportion of patients with PTC was not significantly different between DTC patients with and without family history of thyroid cancer (P = .38). More interestingly, 68.8% (11 of 16) of PTC patients with a first-degree relative with thyroid cancer had multifocal primary tumors, compared with only 35.2% (86 of 244) of the PTC patients with no family history of thyroid cancer (P = .01). There were no significant differences between patients with PTC with and without a first-degree family history of thyroid cancer in age at diagnosis, prevalence of thyroiditis, or distribution of PTC histological subtype. In addition, none of the PTC patients with family history of thyroid cancer in first-degree relatives reported prior history of radiation exposure.
Table 4. Description of Patients With DTC and First-Degree Family History of Thyroid Cancer
|Mother||34||Female||Follicular variant PTC||No||No|
|Mother||37||Female||Follicular variant PTC||No||No|
|Mother||44||Female||Follicular variant PTC||Yes||No|
|Sister||42||Female||Follicular variant PTC||Yes||No|
|Brother||31||Male||Follicular variant PTC||Yes||No|
This study confirms an association between family history of thyroid cancer in first-degree relatives and risk of sporadic PTC. The excess risk of PTC was greater in subjects who reported a family history of thyroid cancer in siblings. In addition, among patients with sporadic PTC, those with a family history of thyroid cancer developed multifocal primary tumor more frequently than those without a family history of thyroid cancer.
Increased risk of DTC associated with a family history of thyroid cancer has been observed in most previous case-control9-16 and cancer family registry17-19 studies, but some results, particularly those from case-control studies, were not statistically significant.10, 13, 16 The reported excess risk usually ranges from 2- to 10-fold. This variation in cancer risk estimates is in part because of variations in the types of relatives assessed: parents only,11, 19 parents and children,12 first-degree relatives plus grandparents,10 and all relatives without regard to degree of relationship.9, 12, 16 Also, these previous studies often did not distinguish between sporadic thyroid cancer and familial thyroid cancer. It has been reported that familial nonmedullary thyroid cancer accounts for 5% of thyroid cancers and might be more aggressive in clinical behavior than sporadic cases.7, 20 We limited our study to patients with sporadic DTC (we excluded 1 patient with PTC with 2 first-degree family members with thyroid cancer from analysis) and reported that the risk of sporadic PTC was 4.6× as high in subjects with a first-degree family history of thyroid cancer as in those without. This excess risk is in agreement with a recent case-control study in a radiation-exposed population that reported an overall 4.5-fold increase in DTC risk associated with family history of thyroid cancer in first-degree relatives.15 Moreover, we observed that the risk of PTC associated with a family history of thyroid cancer in siblings was more than double the risk of PTC associated with a family history of thyroid cancer in parents. Although this relatives-type effect was not reported in previous case-control studies,9-16 in the nationwide Swedish Family-Cancer Database, which recorded nonmedullary thyroid cancers between 1986 and 2002, the standardized incidence ratio of PTC was more than twice as high in individuals with siblings diagnosed with thyroid cancer as in those with parents diagnosed with thyroid cancer.22 In our study, further stratified analysis showed that this relatives-type effect was confined to women. Because of the relatively small number of participants in our study who reported a family history of thyroid cancer in first-degree relatives, we cannot conclude whether there is a sex difference in PTC risk associated with a family history of thyroid cancer by type of relative. We also observed an association between family history of liver cancer and DTC risk, whereas this finding may be subject to chance /misclassification and has not been observed in previous studies.
Besides family history, in this study, patients with DTC had higher family income and were less like to smoke than controls. Although the association between DTC and socioeconomic factors is plausible, because it is suggested that the rising incidence of thyroid cancer, particularly PTC, is largely attributable to improved diagnostic practices, only weak associations have been found.3-5, 23 Consistently, in this study, we did not find strong association with this factor, and the significance disappeared when it was entered into a multivariate analysis. Similarly, the association between smoking status and DTC risk was not significant in the multivariate risk model. The protective effect of cigarette smoking for thyroid cancer has been consistently observed. Given that hormonal and reproductive factors may be involved in thyroid carcinogenesis,24 cigarette smoking is suspected to exhibit its protective effect by lowering the endogenous thyroid-stimulating hormone level,25, 26 although the exact mechanism remains unclear.
Our finding of an elevated risk of sporadic DTC associated with a family history of thyroid cancer in first-degree relatives may indicate a genetic component in the etiology of thyroid cancer. Previous studies in a subset of the same case-control population revealed that common variants in the RET proto-oncogene,27 DNA repair genes XRCC128 and XRCC3,29 and xenobiotic metabolizing genes GSTT1 and GSTM130 were significantly associated with DTC risk. The significance of a genetic component in DTC development is also supported by a large cancer registry-based study,18 which found a higher familial relative risk of thyroid cancer than other major cancer types, including environmental exposure-related cancers such as lung cancer. In this study, we did not observe the expected difference in age at diagnosis between PTC patients with and without a first-degree family history of thyroid cancer, possibly because the genetic predisposition in these cases may not be strong enough to trigger an early onset of disease as in familial DTC cases. Alternatively, the shared environment of families may play a role in these observed excess risks. It is well established that ionizing radiation exposure is an environmental risk factor for PTC.31 However, only a few participants in our study reported a radiotherapy history, and the prevalence in patients with DTC was not different from that in controls. Also, although we were unable to obtain exposure information for the relatives of study subjects, adjustment of possible environmental factors, including socioeconomic status, smoking, alcohol drinking, and radiotherapy history, did not modify the association between first-degree family history of thyroid cancer and DTC (or PTC) risk. Therefore, we believe that environmental components are less likely to contribute to the excess DTC (and PTC) risk associated with family history of thyroid cancer in this study. However, the possibility should be recognized that the familial association of thyroid cancer may be caused by unsuspected or unidentified local environmental factors.
Multifocal primary tumors are common among patients with PTC; previously reported proportions range from 18% to 87%.32, 33 In this study, 36.9% of newly diagnosed patients had multifocal PTCs. The multiple cancer loci often appear to be a consequence of independent clonal events, implicating a predisposing influence in development of multifocal PTC, including both genetic susceptibility and environmental insult.34 As expected, multifocal PTC was more common among PTC patients with a first-degree family history of thyroid cancer than among those with no family history of thyroid cancer. Moreover, because multifocal PTC is more likely to have lymph node and pulmonary metastases,32, 35 a prognostic difference might exist between patients with and without family history of cancer, although we have not explored this possibility here.
Our study has several limitations. First, the study sample size was moderate. As we included only sporadic DTC cases from a mixed background, the small number of subjects with a first-degree family history of thyroid cancer further limited statistical power, especially for stratification analysis. Second, the family history of cancer was self-reported by participants, which might result in recall bias. However, although no attempt was made to verify the diagnoses of cancer in first-degree relatives, those diagnoses are likely to be accurate, because several studies have shown a high accuracy of diagnosis in first-degree relatives, and the accuracy proportion was comparable between cases and controls.36-38 Third, because we used a hospital-based case-control design, the possibility of selection bias, especially among controls (visitors to our institution), which could result in the lack of association between family history of cancer and DTC risk should be considered. Forth, there was a mismatch in age between our patients and controls. However, because our control population was older with a resulting greater possibility for having a positive family history, we may have underestimated the risk of PTC associated with family history of thyroid cancer. Finally, the possibility of screening bias is certainly possible, whereby our cases (both patients with DTC and those with benign thyroid disease) would have been more likely than controls to have sought thyroid examination and subsequently to have been diagnosed with PTC or benign thyroid disease because of a relative (particularly a sibling) with a prevalent thyroid cancer than would a control population with no such family history.
In summary, here we provide evidence suggesting that family history of thyroid cancer in first-degree relatives is associated with a significant increase in sporadic PTC risk and that risk is greater for people whose siblings were diagnosed with thyroid cancer. Such results should be interpreted with caution, and confirmation by larger prospective studies is warranted.
This work was supported by an American Thyroid Association Thyroid Cancer grant (principal investigator, E.M.S.), a National Institutes of Health U01 grant (DE019765-01; principal investigator, A.K.E.-N.), and Cancer Center support grant CA-16672 to The University of Texas MD Anderson Cancer Center (principal investigator, Dr. John Mendelsohn). L.X. is currently a Cancer Prevention Postdoctoral Fellow at The University of Texas MD Anderson Cancer Center supported by Halliburton Employees Fellow in Cancer Prevention funds.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.