- Email: [email protected]
Relationship between obesity, diabetes and the risk of thyroid cancer
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
Available online at www.sciencedirect.com
ScienceDirect www.elsevier.com/locate/amjoto
Original contribution
Relationship between obesity, diabetes and the risk of thyroid cancer☆,☆☆,★ Benjamin Oberman, MD a , Aliasgher Khaku, BS a , Fabian Camacho, MS, MA b , David Goldenberg, MD a,⁎ a
Department of Surgery, Division of Otolaryngology–Head and Neck Surgery, The Pennsylvania State University, College of Medicine, Hershey, PA, USA Department of Public Health Sciences, Division of Health Services and Behavioral Research, The Pennsylvania State University, College of Medicine, Hershey, PA, USA b
ARTI CLE I NFO
A BS TRACT
Article history:
Purpose: Analyze the relationship between obesity and type-2 diabetes mellitus (DM) and
Received 5 January 2015
the development of differentiated thyroid cancer (DTC). Materials and methods: A randomized case-controlled retrospective chart review of outpatient clinic patients at an academic medical center between January 2005 and December 2012. DTC patients were compared to two control groups: primary hyperparathyroidism (PHPTH) patients with euthyroid state and Internal Medicine (IM) patients. Exposure variables included historical body-mass-index (BMI), most recent BMI within 6 months and DM. Multivariate logistic regressions adjusting for gender, age, and year of BMI assessed the adjusted Odds Ratio (OR) of DTC with both BMI and DM. Results: Comparison of means showed a statistically significant higher BMI in DTC (BMI = 37.83) than PHPTH, IM, and pooled controls, BMI = 30.36 p = <0.0001, BMI = 28.96 p = <0.0001, BMI = 29.53 p = <0.0001, respectively. When compared to PHPTH, DM was more frequent in DTC (29% vs. 16%) and prevalence trended towards significance (p = 0.0829, 95% CI = 0.902–5.407). BMI adjusted OR was significant when compared to PHPTH, IM and pooled controls: 1.125 (p = 0.0001), 1.154 (p = <0.0001), and 1.113 (p = <0.0001), respectively. DM adjusted OR was significant when compared to PHPTH and pooled controls at 3.178 (95% 1.202,8.404, p = 0.0198) and 2.237 (95% 1.033,4.844, p = 0.0410), respectively. Conclusion: Our results show that obesity and, to a lesser degree, DM are significantly associated with DTC. BMI in particular was a strong predictive variable for DTC (C = 0.82 bivariate, C = 0.84 multivariate). © 2015 Elsevier Inc. All rights reserved.
☆
This paper was presented at the AAO-HNSF 2014 Annual Meeting and OTO Expo, Orlando, FL, September 23, 2014. Funding: Wells Fargo Jane Barsumian and Mary Lyons Trust Grant. ★ Conflicts of Interest: None. ⁎ Corresponding author at: The Pennsylvania State University, College of Medicine, Department of Surgery, Division of Otolaryngology– Head and Neck Surgery, 500 University Drive, MC H091, Hershey, PA 17033-0850. Tel.: + 1 717 531 8946; fax: +1 717 531 6160. E-mail address: [email protected] (D. Goldenberg). ☆☆
http://dx.doi.org/10.1016/j.amjoto.2015.02.015 0196-0709/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
2
AMER ICA N JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D N E CK M EDI CI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
1.
Introduction
In recent decades there has been a notable trend in the increase of thyroid cancer, although the specific reasons for this increase have not yet been well defined. Recent reports show that the incidence of thyroid cancer in 2009 increased to 14.3 cases per 100,000 persons in the United States, nearly triple the 1975 rate of 4.9 cases per 100,000 individuals. More than fifty percent of that increased rate has occurred since 2007. The increase has primarily been secondary to a rise in papillary thyroid cancer, which has increased by 9.1 per 100,000 persons in the past 30 years. When considering gender as a factor, the increase has also been more prevalent in women than men [1]. National data compiled in 2010 shows that Pennsylvania has the highest incidence of thyroid cancer amongst all states [2]. Ongoing research suggests that the rising incidence of thyroid cancer in Pennsylvania is statistically more significant than the national correlate [3]. When considering the overall increases in thyroid cancer, much of the rise in incidence may be due to improved detection methods, however this does not sufficiently explain the increased incidence in larger tumors, in all size categories [4]. It has been hypothesized that factors such as diabetes mellitus (DM) and obesity may be factors playing a role in the increase of thyroid cancer. The prevalence of obesity has increased worldwide in recent decades and studies have shown that an increased body-massindex (BMI) is a risk for other cancers as well: esophageal, colon, and renal cancers in men as well as endometrial, gallbladder, esophageal and renal cancers in women [5]. In the past 20 years the prevalence of obesity in U.S. adults has doubled, and overweight children and adolescents have tripled [6]. A recent study by Xu et al. showed that the prevalence of obesity increased from 13.4% to 35.7% from 1960–1962 to 2009–2010. This increase was not subject to influence by age, sex, ethnicity, or socioeconomic status [7]. The prevalence of obesity in Pennsylvania in 2010 is reported as 29.2% of the adult population versus 27.6% nationally as recorded by the Behavioral Risk Factor Surveillance System [8]. Various case–control and prospective studies have shown obesity to be an independent risk factor for thyroid cancer in men and women [6,9,10]. Additionally, a positive association between BMI and thyroid cancer risk has been shown in adults of all ages, including particularly young adults aged 18–20 [6]. In concert with the rise in thyroid cancer and obesity, DM has markedly increased in recent decades. Obese people have a 10-fold increased risk of diabetes and over 80% of type 2 diabetics are obese. Moreover, diabetics have a higher prevalence of thyroid disorders than the general population [11–13]. A recent study reported that there are over 250 million people globally with DM which is expected to reach 380 million in the next twenty years [13]. The factors driving the increase in thyroid cancer are not fully understood. In the present study, we sought to distinguish the interplay between obesity, DM, and well-differentiated thyroid cancer. This study took a novel approach by investigating the relationship between obesity, DM, and thyroid cancer by using data from three case-controlled populations, one with evidence of a euthyroid state.
2.
Materials and methods
2.1.
Study population
A randomized, case–control retrospective chart review was conducted for patients presenting with differentiated thyroid carcinoma (DTC) to the Penn State Hershey Medical Center Otolaryngology outpatient clinic between January 1, 2005 and December 31, 2012. DTC patients were selected by using ICD-9 code 193 “malignant neoplasm of the thyroid gland.” Patients were excluded from the study for the following reasons: DTC could not be confirmed pathologically or cytologically, any evidence of undifferentiated thyroid cancer, or if height or weight was not documented. The study included 107 patients diagnosed with DTC from which 48 (9 males, 39 females) cases fit the study criteria. Characteristics of each patient’s tumor, the date of diagnosis, age at diagnosis, sex, ethnicity, height, weight, history of diabetes, and other co-morbidities were recorded. The case group was compared to two different control groups for a total of 172 control patients. The control groups consisted of two independent populations within the study timeline. Group 1 consisted of 70 patients diagnosed primary hyperparathyroidism (PHPTH) who presented to the Otolaryngology outpatient clinic with documented euthyroid state. Group 2 consisted of 102 systematically sampled patients selected from the Internal Medicine (IM) clinic. PHPTH patients were selected using ICD9 codes 227 and 252 (benign neoplasm of the parathyroid gland and primary hyperparathyroidism, respectively). Parathyroid disease type, patient age and date of PHPTH diagnosis, serum TSH levels, sex, race, height, weight, history of diabetes, and other comorbidities were documented. Patients with history of thyroid dysfunction, thyroid cancer, unconfirmed diagnosis of primary hyperparathyroidism, a history of gastrointestinal bypass or gastrointestinal cancer, or a history of multiple endocrine neoplasia were excluded. The Internal Medicine control patients were systemically sampled at a rate of 2 controls to 1 DTC case. A randomly selected calendar date was chosen, then every 5th patient presenting to the IM outpatient clinic from that date in each year (2005 through 2012) was selected until the matching quota was reached. A history of diabetes, age, sex, race, height, weight, and other comorbidities were recorded. The only exclusion criterion for this group was any history of thyroid cancer. This study was approved by the organization's institutional review board.
2.2.
Anthropometric analysis
Body Mass Index (BMI) was determined by dividing weight in kilograms by height in meters squared (kg/m2). BMI data were collected to as far back as 2 years prior to the diagnosis of DTC and as recently as the most recent available BMI data after diagnosis of DTC. The weight data for both control groups were collected to as many as 3 years preceding diagnosis of hyperparathyroidism (PHPTH group) or 3 years preceding year of presentation (IM group). The most recently available weight was recorded as well. BMI was categorized as underweight
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
(< 18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25.0–29.9 kg/m2), or obese (≥30 kg/m2) according to the World Health Organization definition. The body fat percentage was estimated from BMI by using the formula of Deurenberg et al. [14]: body fat percentage = (1.20 × BMI) + (0.23 × Age) − (10.8 × Sex) − 5.4, where age is in years and sex is set to 0 for women and 1 for men. Body surface area (BSA) was calculated from height and weight by using the formula of Du Bois and Du Bois [15]: BSA = (0.007184) × (Weight0.425) × Height0.725).
2.3.
Statistical analysis
Exposure variables consisted of the most recent BMI within 6 months prior to diagnosis of DTC and any diagnosis of DM. Also evaluated was a “historical BMI” variable assessing whether a patient had met clinical criteria for obesity 1 to 2 years prior to diagnosis of PHPTH or 1 to 2 years prior to the control reference date. The time frame for this measure was determined in order to preserve sufficient sample size in the case group. Multivariate logistic regressions adjusting for gender, age, race, and year of BMI were used to assess the odds ratio (OR) of DTC with both BMI and DM. Furthermore, multivariate logistic regression was used to analyze the historical relationship of BMI with DTC. Anthropometric measurements were examined as continuous variables and treated as categorical variables either as quartiles or tertiles following prior usage in literature. BMI histograms of DTC case patients were used as comparisons against both control groups. The Mann–Whitney U test was used to compare continuous variables between cases and controls as appropriate. Percentages may not total 100% as a result of rounding.
3.
3
Table 2 shows the odds ratios for the DTC cases compared to each control group as well as the pooled analysis for the anthropometric measurements and the association with diabetes. When considering BMI, BSA, and body fat, the risk of DTC was increased for obese patients when comparing each control group and the pooled data. The percentage of patients with diabetes was increased in the DTC group compared to each individual control group as well as pooled data. Although diabetes was not significantly more prevalent in the case group versus the PHPTH control group, it did trend toward significance (p = 0.0829, 95% CI = 0.902–5.407). Using multivariate analysis, the odds ratio when comparing cases versus PHPTH controls was significant (OR = 3.178, p = 0.0198, 95% CI = 1.202, 8.404). Comparing DTC cases to the IM control group, diabetes was not significantly associated with the DTC group in either bivariate or multivariate analysis. However, when comparing cases versus the pooled control data, the odds ratio showed diabetes was significantly associated with the risk of having DTC (OR = 2.237, p = 0.0410, 95% CI = 1.033, 4.844). Table 3 shows the odds ratios broken into quartiles for anthropometric data. As one might predict from our aforementioned results, patients with a greater BMI had statistically significantly increased risk for DTC. Interquartile comparisons of BMI, BSA, and body fat percentage showed, respectively, that lower BMI, BSA, and body fat percentage were associated with a lower risk of DTC. The number of patients with historical BMI was 26. The risk of DTC when including the historical data for BMI continued remained significantly increased when compared to the PHPTH group, with an OR of 1.112 (95% CI 1.039, 1.189). Additionally, there was a persistently increased risk when comparing DTC to the IM group, with an OR of 1.122 (95% CI 1.058, 1.190). Pooled data showed a significant, but slightly less strong risk, of DTC with history of obesity, with an OR of 1.087 (95% CI 1.036, 1.142). The odds ratios, using both BMI and historical BMI values, were significantly less than 1.0 for patients with BMI less than 30 when DTC patients were compared to each control group.
Results
A total of 48 DTC cases, 70 PHPTH controls and 102 IM controls were used in the analyses. Gender distribution among the DTC cases showed 39 (81.3%) patients were female and 9 (18.8%) were male. Table 1 shows the characteristics of each subject group as well as their anthropometric measurements. Fig. 1 shows an overall view of the distribution of each subject group when comparing BMI and DM. Additionally, 42 of the case group patients (87.5%) were found to have a BMI greater than 30. The IM control group had 36 patients (35.3%) with BMI greater than 30, whereas the PHPTH control group had 25 patients (35.7%) with BMI greater than 30.
4.
Discussion
There has been a great deal of focus on the epidemics of obesity and diabetes in the United States in recent years, and the number of affected people continues to rise each year. Thyroid cancer is on the rise; the Commonwealth of Pennsylvania has the highest incidence of thyroid cancer in the country as well as a high rate of obesity [3]. Therefore a large academic institution in Central PA would be an appropriate setting for a study of this kind. Determining any associations between thyroid cancer
Table 1 – Characteristics of differentiated thyroid cancer (DTC) cases and control study. Characteristics
DTC Case (n = 48)
PHPTH Control Group (n = 70)
p value
IM Control Group (n = 102)
p value
All Controls (n = 172)
p value
Age (years) Height (cm) Weight (kg) Body surface area (m2) Body fat (%) BMI (kg/m2) Type-II diabetes Race (% of Caucasians) Female gender
49.13 (14.73) 164.81 (9.25) 103.22 (22.61) 7.38 (0.90) 49.27 (9.92) 37.83 (6.95) 29.17% 89.58% 81.25%
56.63 (12.82) 165.82 (9.25) 84.06 (26.68) 6.76 (1.02) 41.28 (11.13) 30.36 (8.50) 15.71% 95.71% 74.29%
0.0082 0.5196 <0.0001 0.0002 0.0001 <0.0001 0.0809 0.1972 0.3804
51.88 (16.44) 165.90 (10.64) 80.28 (24.88) 6.63 (1.04) 37.69 (11.55) 28.96 (8.08) 21.57% 83.33% 66.67%
0.4110 0.4379 <0.0001 <0.0001 <0.0001 <0.0001 0.3124 0.3160 0.0667
53.81 (15.22) 165.87 (10.07) 81.82 (25.62) 6.68 (1.03) 39.15 (11.48) 29.53 (8.26) 19.19% 88.37% 69.77%
0.0794 0.4258 <0.0001 <0.0001 <0.0001 <0.0001 0.1371 0.8174 0.1173
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
4
AMER ICA N JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D N E CK M EDI CI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
Fig. 1 – This graph shows the prevalence rates of normal weight, overweight, obese, and type-II diabetic patients amongst the study groups, highlighting the significance of obesity in the case group.
and obesity may assist with future prevention or reduction in thyroid cancer as well as enhanced screening in an at risk population for early detection. The unique features of this study are seen with the use of two control groups, objective values for weight and height instead of surveyed estimates, and evaluation of patients over the course of years instead of a single time-point. Notably, one of the control groups had a documented euthyroid state which, to our knowledge, has not been included in prior evaluations. In regard to the evaluation of obesity and its relationship to DTC, our results are consistent with prior reports. Specifically, the risk of DTC is significantly increased in patients who are obese compared to overweight or normal weight patients. This was true even amongst all three types of mathematical comparisons: BMI, body surface area, or body fat percentage. Evaluating patients' body surface area, which reportedly is less affected by abnormal adipose mass than the BMI calculation, led to an even greater association of DTC with increases in mass than DTC's association to obesity. Further support for this association has been shown when accounting for the way men and women distribute fat, thus eliminating the limitations of using BMI [9]. The accumulation of visceral adipose tissue is related to dyslipidemia and insulin resistance. Insulin may directly contribute to proliferation of thyroid cancer cells by stimulating insulin-like growth factor. This relation further supports the involvement of insulin resistance biologic mechanisms in thyroid cancer development [5,6,9,10,12]. Estrogen may be another associative factor
between overall body fat and thyroid cancer risk. Estradiol, which is converted from testosterone by fat cells, is an estrogen which has been shown to be a promoter of benign and malignant thyroid tumor cells [9]. Recent evaluation of excess body weight and papillary thyroid cancer has shown a significantly positive relationship with larger tumor size, presence of microscopic extrathyroidal invasion, and higher TNM staging [5]. Hormonal changes associated with obesity implicate leptin and its receptor with tumor aggressiveness as well. Adipokines such as leptin may promote pituitary production of thyroid stimulating hormone (TSH) [9]. Leptin levels were shown to be elevated in patients with papillary thyroid cancer [16]. Additionally, leptin stimulates cell proliferation and inhibits apoptosis via activation of the phosphatidylinositide 3-kinase/protein kinase B (PI3K/ AKT) pathway, which appears to promote angiogenesis and tumor invasion [5]. Our study also evaluates the association of type 2 diabetes with DTC. Comparisons were performed using multivariate analysis, controlling for age and gender, showing significant association of DM with DTC and thereby suggesting that DM does contribute to DTC. Age was controlled to reduce the chances of confounding results due the effect age has on incidence of each disease. Additionally, gender was considered a potential confounder because of the propensity of women to have DTC. Support for DM's association with DTC has been suggested by multiple metabolic and hormonal mechanisms.
Table 2 – Study odds ratio (OR) and 95% confidence intervals for differentiated thyroid cancer (DTC) risk. Variable
2
Body surface area (m ) Body fat (%) BMI (kg/m2) > 30.0 Historical BMI (kg/m2) > 30.0 Type-II diabetes
DTC compared to control group 1 (PHPTH)
DTC compared to control group 2 (IM)
OR (95% CI)
p value
OR (95% CI)
p value
OR (95% CI)
p value
2.299 1.103 1.125 1.112 3.178
0.0011 0.0001 0.0001 0.0021 0.0198
3.363 1.127 1.154 1.122 1.860
<0.0001 <0.0001 <0.0001 0.0001 0.1490
2.538 1.093 1.113 1.087 2.237
<0.0001 <0.0001 <0.0001 0.0007 0.0410
(1.396, (1.049, (1.059, (1.039, (1.202,
3.786) 1.160) 1.195) 1.189) 8.404)
(2.037, (1.074, (1.089, (1.058, (0.801,
5.552) 1.183) 1.223) 1.190) 4.320)
DTC compared t o pooled controls
(1.713, (1.055, (1.066, (1.036, (1.033,
3.760) 1.133) 1.162) 1.142) 4.844)
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
5
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
Table 3 – Odds ratio (OR) and 95% confidence intervals for DTC risk according to quartiles. Variable
Q1 OR (95% CI)
DTC compared to primary hyperparathyroidism control group Body surface area (m2) 4.644–6.279 0.087 (0.020, 0.373) Body fat (%) 19.160–36.540 0.037 (0.007, 0.194) <24.9 (kg/m2) BMI (kg/m2) 0.046 (0.005, 0.403) Historical BMI (kg/m2) <24.9 (kg/m2) 0.073 (0.004, 1.451)
Q2
Q3
Q4
p value OR (95% CI)
p value OR (95% CI)
p value OR (95% CI)
0.0861
6.291–6.896 0.205 (0.056, 0.749) 36.960–43.280 0.151 (0.043, 0.522) 24.9–29.9 (kg/m2) 0.069 (0.020, 0.236) 24.9–29.9 (kg/m2) 0.082 (0.017, 0.394)
DTC compared to internal medicine (IM) control group Body surface area (m2) 4.200–5.998 0.025 (0.004, 0.148) <0.0001 Body fat (%) 10.560–32.770 0.004 (0.001, 0.043) <0.0001 <24.9 (kg/m2) BMI (kg/m2) 0.015 (0.002, 0.131) 0.0001 Historical BMI (kg/m2) <24.9 (kg/m2) 0.019 (0.001, .300) 0.0048
6.003–6.929 0.267 (0.086, 0.829) 32.950–40.010 0.074 (0.019, 0.280) 24.9–29.9 (kg/m2) 0.134 (0.041, 0.439) 24.9–29.9 (kg/m2) 0.186 (0.043, 0.814)
DTC compared to pooled controls (control groups 1 & 2) Body surface area (m2) 4.200–6.003 0.037 (0.008, 0.186) <0.0001 Body fat (%) 10.560–33.140 0.010 (0.002, 0.062) <0.0001 <24.9 (kg/m2) BMI (kg/m2) 0.025 (0.003, 0.191) 0.0004 Historical BMI (kg/m2) <24.9 (kg/m2) 0.029 (0.002, 0.454) 0.0115
0.0001 0.0001 0.0055
0.0197 0.0029 <.0001 0.0018
6.903–7.541 0.708 (0.210, 2.390) 43.500–51.310 0.792 (0.252, 2.489) >30.0 (kg/m2) 1.125 (1.059, 1.195) >30.0 (kg/m2) 1.112 (1.039, 1.189)
0.6183 0.6900
7.574–9.797 2.299 (1.396, 3.786) 51.580–76.620 1.103 (1.049, 1.160)
p value
0.0011 0.0001
0.0001 0.0021
7.645–9.312 3.363 (2.037, 5.552) <0.0001 50.170–78.720 1.127 (1.074, 1.183) <0.0001
0.0256
6.932–7.596 0.749 (0.266, 2.105) 0.5832 40.040–49.990 0.301 (0.105, 0.866) 0.0260 >30.0 (kg/m2) 1.154 (1.089, 1.223) <0.0001 >30.0 (kg/m2) 1.122 (1.058, 1.190) 0.0001
6.018–6.812 0.210 (0.073, 0.601) 0.0036 33.420–40.600 0.097 (0.032, 0.293) <0.0001 24.9–29.9 (kg/m2) 0.105 (0.036, 0.305) <0.0001 24.9–29.9 (kg/m2) 0.161 (0.038, 0.675) 0.0125
6.825–7.459 0.826 (0.344, 1.984) 0.6697 40.820–48.790 0.234 (0.093, 0.589) 0.002 >30.0 (kg/m2) 1.113 (1.066, 1.162) <0.0001 >30.0 (kg/m2) 1.087 (1.036, 1.142) 0.0007
7.465–9.797 2.538 (1.713, 3.760) <0.0001 48.920–78.720 1.093 (1.055, 1.133) <0.0001
Chronic elevated insulin levels have been observed in DM patients due to either insulin resistance or exogenous sources such as medications. Insulin and insulin growth factor-1 (IGF-1) are important for cell proliferation and apoptosis and have been found to be related to various cancers such as breast and colon cancer [11–13]. Insulin may play a role in thyroid carcinogenesis through its contribution to follicular cell growth. Follicular cells synthesize IGF-1 and have IGF-1 receptors. Insulin may mimic IGF-1, thus contributing to follicular cell growth. Additionally, IGF-1 receptor activates mitogen-activated protein (MAP) kinase and PI3K pathways. MAP kinases, like the PI3K pathway, also regulate gene expression, differentiation, mitosis, and cell apoptosis [12,13,17]. Another risk factor for thyroid cancer amongst diabetics is the chronically mild elevation of TSH [11,12,16,18,19]. A recent study showed that there is a 3.45 times increased risk of primary hypothyroidism in type 2 diabetics compared to a non-diabetic population [18]. Higher TSH levels have been shown to be risk factors for thyroid cancer as well as a more advanced stage of differentiated thyroid cancer [11,19]. One of the major mitogenic pathways in thyroid cancer is mediated by TSH. The hormone receptor adenylate cyclasecAMP protein kinase A system (AC/cAMP/PKA) is primarily
0.0224 0.0001 0.0009
stimulated by TSH. This pathway regulates the function, differentiation, and proliferation of the thyroid gland. Additionally, TSH stimulates the hormone receptor phospholipase C cascade (PLC) pathway which subsequently increased the intracellular calcium and protein kinase C (PKC) activity [12]. Once activated, PKC may lead to an increased expression of oncogenes, thus promoting cancer progression. Patients with DM have metabolic abnormalities that place them in a state of increased oxidative stress and a permanent pro-inflammatory state [13]. Prolonged inflammatory responses reduce the intracellular anti-oxidant capacity which may increase the risk of susceptible cells to malignant transformation. The mechanism is via generation of reactive oxygen species (ROS) produced from free radicals and oxidants which accumulate in a pro-inflammatory state [13]. ROS damage cells by direct oxidation or interfering with DNA repair mechanisms [20]. Another pro-inflammatory cytokine which is correlated with the insulin resistant environment of DM is tumor necrosis factor α (TNFα) [13]. TNFα, a cytokine involved in apoptosis, has been widely shown to be involved with the development and progression of many tumors including: breast, ovarian, colorectal, prostate, bladder, esophageal, renal cell, melanoma, lymphoma, and leukemia
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
6
AMER ICA N JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D N E CK M EDI CI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
[21]. Deregulation of apoptosis, and those cytokines involved, is critical in the tumorigenesis process. In comparison to healthy individuals, elevated serum concentrations and increased expression of TNFα are present in various preneoplastic diseases and malignancies [21]. One last proposed mechanism for an increase in thyroid cancer in diabetic patients is Vitamin D deficiency and deiodinase expression. Vitamin D deficiency has been observed in up to 70% of diabetics [12]. A low Vitamin D level decreases deiodinase II (DIO2) enzyme. Deiodinase enzymes regulate intracellular thyroid hormone levels; specifically, DIO2 regulates intracellular T3 concentration in the pituitary, brain, brown fat, and skeletal muscle. In skeletal muscle it indirectly regulates glucose transporter 4 (GLUT4) expression. Inactivation of DIO2 by a Vitamin D deficient environment in diabetic patients results in decreased GLUT4 transcription by skeletal muscle and adipose tissue, thus contributing to insulin resistance [11,12]. Our study has aimed to advance the understanding of the relationship between obesity, DM, and the risk of differentiated thyroid cancer. The strengths in our approach are seen with the use of two control groups, precise definitions for measurements, and multivariate evaluations of the accrued data. Specifically, we attempted to eliminate the bias that hypothyroidism or hyperthyroidism could play on subsequent thyroid cancer risk or weight change by having a euthyroid control group. The euthyroid state group was best ascertained by having a PHPTH group as they presented to the same clinic and underwent laboratory testing verifying hormone levels at the time of diagnosis. Additionally, we collected historical weight and height data to eliminate misclassification bias. Also, acquiring data over multiple time points shows that obese patients have increased risk of DTC over time, not just at a single time point prior to diagnosis of DTC. Some limitations to this study are partly due to not being able to account for all potential risk factors of DTC. We did not adjust for dietary habits, physical activity level, or genetic predisposition factors to obesity or thyroid cancer in our study. We were unable to obtain height and weight values for most patients beyond five years prior to DTC diagnosis. Additionally, our case group numbers were limited secondary to access to properly recorded height and weight data. Exclusion of many case patients is not likely a source of selection bias in our study, as our case groups show an incidence of obesity greater than that of the state in which the patients reside. It is more likely that our referral pool has a higher rate of obesity than the general population. Importantly, even in the setting of higher rates of obesity, we found significant results. Lastly, our study was performed in an area which is not racially diverse, but we believe our results would still be generally applicable to the general population.
5.
Conclusion
Our results show that obesity, as measured using various anthropometric values, is significantly associated with differentiated thyroid cancer. To a lesser degree, type 2 diabetes shares a close association with differentiated thyroid cancer.
Future studies evaluating obesity and DM in other regions of the U.S. would serve to broaden the application of our results. By identifying the rise in thyroid cancer with the rise in DM and obesity in our nation we hope to raise awareness that by addressing the epidemics of obesity and DM in our nation we may additionally address concerns of a rise in thyroid cancer.
Acknowledgments This research was funded by the Wells Fargo Jane Barsumian and Mary Lyons Trust Grant.
REFERENCES
[1] Davies L, Welch HG. Current thyroid cancer trends in the United States. JAMA Otolaryngol Head Neck Surg 2014;140: 317–22. [2] National Cancer Institute. State cancer profiles: incidence. http://statecancerprofiles.cancer.gov/cgi-bin/quickprofiles/ profile.pl?00&080; 2013. [August 9, 2014]. [3] Bann DV, Goyal N, Camacho F, et al. Increasing incidence of thyroid cancer in the Commonwealth of Pennsylvania. JAMA Otolaryngol Head Neck Surg 2014. http://dx.doi.org/10.1001/ jamaoto.2014.1709 [Epub ahead of print]. [4] Morris LG, Myssiorek D. Improved detection does not fully explain the rising incidence of well-differentiated thyroid cancer: a population-based analysis. Am J Surg 2010;200: 454–61. [5] Kim HJ, Kim NK, Choi JH, et al. Associations between body mass index and clinico-pathological characteristics of papillary thyroid cancer. Clin Endocrinol (Oxf) 2013;78:134–40. [6] Kitahara CM, Platz EA, Freeman LE, et al. Obesity and thyroid cancer risk among U.S. men and women: a pooled analysis of five prospective studies. Cancer Epidemiol Biomarkers Prev 2011;20:464–72. [7] Xu L, Port M, Landi S, et al. Obesity and the risk of papillary thyroid cancer: a pooled analysis of three case–control studies. Thyroid 2014;24:966–74. [8] Pennsylvania Department of Health. Obesity prevention and wellness program. Pennsylvania obesity data and trends. http://www.portal.state.pa.us/portal/server.pt/community/ obesity/14184; 2014. [August 9, 2014]. [9] Kitahara CM, Platz EA, Park Y, et al. Body fat distribution, weight change during adulthood, and thyroid cancer risk in the NIH-AARP Diet and Health Study. Int J Cancer 2012;130: 1411–9. [10] Marcello MA, Sampaio AC, Geloneze B, et al. Obesity and excess protein and carbohydrate consumption are risk factors for thyroid cancer. Nutr Cancer 2012;64:1190–5. [11] Aschebrook-Kilfoy B, Sabra MM, Brenner A, et al. Diabetes and thyroid cancer risk in the National Institutes of Health– AARP Diet and Health Study. Thyroid 2011;21:957–63. [12] Shih SR, Chiu WY, Chang TC, et al. Diabetes and thyroid cancer risk: literature review. Exp Diabetes Res 2012;2012: 578285. [13] Vigneri P, Frasca F, Sciacca L, et al. Diabetes and cancer. Endocr Relat Cancer 2009;16:1103–23. [14] Deurenberg P, Weststrate JA, Seidell JC. Body mass index as a measure of body fatness: age- and sex-specific prediction formulas. Br J Nutr 1991;65:105–14. [15] Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition 1989;5:303–11 [discussion 312–303].
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
[16] Akinci M, Kosova F, Cetin B, et al. Leptin levels in thyroid cancer. Asian J Surg 2009;32:216–23. [17] Frohlich E, Wahl R. Do antidiabetic medications play a specific role in differentiated thyroid cancer compared to other cancer types? Diabetes Obes Metab 2012;14:204–13. [18] Tamez-Perez HE, Martinez E, Quintanilla-Flores DL, et al. The rate of primary hypothyroidism in diabetic patients is greater than in the non-diabetic population. An observational study. Med Clin (Barc) 2012;138:475–7.
7
[19] Haymart MR, Repplinger DJ, Leverson GE, et al. Higher serum thyroid stimulating hormone level in thyroid nodule patients is associated with greater risks of differentiated thyroid cancer and advanced tumor stage. J Clin Endocrinol Metab 2008;93:809–14. [20] Frederico A, Morgillo F, Tuccillo C, et al. Chronic inflammation and oxidative stress in human carcinogenesis. Int J Cancer 2007;121:2381–6. [21] Szlosarek P, Charles KA, Balkwill FR. Tumour necrosis factoralpha as a tumour promoter. Eur J Cancer 2006;42:745–50.
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
Available online at www.sciencedirect.com
ScienceDirect www.elsevier.com/locate/amjoto
Original contribution
Relationship between obesity, diabetes and the risk of thyroid cancer☆,☆☆,★ Benjamin Oberman, MD a , Aliasgher Khaku, BS a , Fabian Camacho, MS, MA b , David Goldenberg, MD a,⁎ a
Department of Surgery, Division of Otolaryngology–Head and Neck Surgery, The Pennsylvania State University, College of Medicine, Hershey, PA, USA Department of Public Health Sciences, Division of Health Services and Behavioral Research, The Pennsylvania State University, College of Medicine, Hershey, PA, USA b
ARTI CLE I NFO
A BS TRACT
Article history:
Purpose: Analyze the relationship between obesity and type-2 diabetes mellitus (DM) and
Received 5 January 2015
the development of differentiated thyroid cancer (DTC). Materials and methods: A randomized case-controlled retrospective chart review of outpatient clinic patients at an academic medical center between January 2005 and December 2012. DTC patients were compared to two control groups: primary hyperparathyroidism (PHPTH) patients with euthyroid state and Internal Medicine (IM) patients. Exposure variables included historical body-mass-index (BMI), most recent BMI within 6 months and DM. Multivariate logistic regressions adjusting for gender, age, and year of BMI assessed the adjusted Odds Ratio (OR) of DTC with both BMI and DM. Results: Comparison of means showed a statistically significant higher BMI in DTC (BMI = 37.83) than PHPTH, IM, and pooled controls, BMI = 30.36 p = <0.0001, BMI = 28.96 p = <0.0001, BMI = 29.53 p = <0.0001, respectively. When compared to PHPTH, DM was more frequent in DTC (29% vs. 16%) and prevalence trended towards significance (p = 0.0829, 95% CI = 0.902–5.407). BMI adjusted OR was significant when compared to PHPTH, IM and pooled controls: 1.125 (p = 0.0001), 1.154 (p = <0.0001), and 1.113 (p = <0.0001), respectively. DM adjusted OR was significant when compared to PHPTH and pooled controls at 3.178 (95% 1.202,8.404, p = 0.0198) and 2.237 (95% 1.033,4.844, p = 0.0410), respectively. Conclusion: Our results show that obesity and, to a lesser degree, DM are significantly associated with DTC. BMI in particular was a strong predictive variable for DTC (C = 0.82 bivariate, C = 0.84 multivariate). © 2015 Elsevier Inc. All rights reserved.
☆
This paper was presented at the AAO-HNSF 2014 Annual Meeting and OTO Expo, Orlando, FL, September 23, 2014. Funding: Wells Fargo Jane Barsumian and Mary Lyons Trust Grant. ★ Conflicts of Interest: None. ⁎ Corresponding author at: The Pennsylvania State University, College of Medicine, Department of Surgery, Division of Otolaryngology– Head and Neck Surgery, 500 University Drive, MC H091, Hershey, PA 17033-0850. Tel.: + 1 717 531 8946; fax: +1 717 531 6160. E-mail address: [email protected] (D. Goldenberg). ☆☆
http://dx.doi.org/10.1016/j.amjoto.2015.02.015 0196-0709/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
2
AMER ICA N JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D N E CK M EDI CI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
1.
Introduction
In recent decades there has been a notable trend in the increase of thyroid cancer, although the specific reasons for this increase have not yet been well defined. Recent reports show that the incidence of thyroid cancer in 2009 increased to 14.3 cases per 100,000 persons in the United States, nearly triple the 1975 rate of 4.9 cases per 100,000 individuals. More than fifty percent of that increased rate has occurred since 2007. The increase has primarily been secondary to a rise in papillary thyroid cancer, which has increased by 9.1 per 100,000 persons in the past 30 years. When considering gender as a factor, the increase has also been more prevalent in women than men [1]. National data compiled in 2010 shows that Pennsylvania has the highest incidence of thyroid cancer amongst all states [2]. Ongoing research suggests that the rising incidence of thyroid cancer in Pennsylvania is statistically more significant than the national correlate [3]. When considering the overall increases in thyroid cancer, much of the rise in incidence may be due to improved detection methods, however this does not sufficiently explain the increased incidence in larger tumors, in all size categories [4]. It has been hypothesized that factors such as diabetes mellitus (DM) and obesity may be factors playing a role in the increase of thyroid cancer. The prevalence of obesity has increased worldwide in recent decades and studies have shown that an increased body-massindex (BMI) is a risk for other cancers as well: esophageal, colon, and renal cancers in men as well as endometrial, gallbladder, esophageal and renal cancers in women [5]. In the past 20 years the prevalence of obesity in U.S. adults has doubled, and overweight children and adolescents have tripled [6]. A recent study by Xu et al. showed that the prevalence of obesity increased from 13.4% to 35.7% from 1960–1962 to 2009–2010. This increase was not subject to influence by age, sex, ethnicity, or socioeconomic status [7]. The prevalence of obesity in Pennsylvania in 2010 is reported as 29.2% of the adult population versus 27.6% nationally as recorded by the Behavioral Risk Factor Surveillance System [8]. Various case–control and prospective studies have shown obesity to be an independent risk factor for thyroid cancer in men and women [6,9,10]. Additionally, a positive association between BMI and thyroid cancer risk has been shown in adults of all ages, including particularly young adults aged 18–20 [6]. In concert with the rise in thyroid cancer and obesity, DM has markedly increased in recent decades. Obese people have a 10-fold increased risk of diabetes and over 80% of type 2 diabetics are obese. Moreover, diabetics have a higher prevalence of thyroid disorders than the general population [11–13]. A recent study reported that there are over 250 million people globally with DM which is expected to reach 380 million in the next twenty years [13]. The factors driving the increase in thyroid cancer are not fully understood. In the present study, we sought to distinguish the interplay between obesity, DM, and well-differentiated thyroid cancer. This study took a novel approach by investigating the relationship between obesity, DM, and thyroid cancer by using data from three case-controlled populations, one with evidence of a euthyroid state.
2.
Materials and methods
2.1.
Study population
A randomized, case–control retrospective chart review was conducted for patients presenting with differentiated thyroid carcinoma (DTC) to the Penn State Hershey Medical Center Otolaryngology outpatient clinic between January 1, 2005 and December 31, 2012. DTC patients were selected by using ICD-9 code 193 “malignant neoplasm of the thyroid gland.” Patients were excluded from the study for the following reasons: DTC could not be confirmed pathologically or cytologically, any evidence of undifferentiated thyroid cancer, or if height or weight was not documented. The study included 107 patients diagnosed with DTC from which 48 (9 males, 39 females) cases fit the study criteria. Characteristics of each patient’s tumor, the date of diagnosis, age at diagnosis, sex, ethnicity, height, weight, history of diabetes, and other co-morbidities were recorded. The case group was compared to two different control groups for a total of 172 control patients. The control groups consisted of two independent populations within the study timeline. Group 1 consisted of 70 patients diagnosed primary hyperparathyroidism (PHPTH) who presented to the Otolaryngology outpatient clinic with documented euthyroid state. Group 2 consisted of 102 systematically sampled patients selected from the Internal Medicine (IM) clinic. PHPTH patients were selected using ICD9 codes 227 and 252 (benign neoplasm of the parathyroid gland and primary hyperparathyroidism, respectively). Parathyroid disease type, patient age and date of PHPTH diagnosis, serum TSH levels, sex, race, height, weight, history of diabetes, and other comorbidities were documented. Patients with history of thyroid dysfunction, thyroid cancer, unconfirmed diagnosis of primary hyperparathyroidism, a history of gastrointestinal bypass or gastrointestinal cancer, or a history of multiple endocrine neoplasia were excluded. The Internal Medicine control patients were systemically sampled at a rate of 2 controls to 1 DTC case. A randomly selected calendar date was chosen, then every 5th patient presenting to the IM outpatient clinic from that date in each year (2005 through 2012) was selected until the matching quota was reached. A history of diabetes, age, sex, race, height, weight, and other comorbidities were recorded. The only exclusion criterion for this group was any history of thyroid cancer. This study was approved by the organization's institutional review board.
2.2.
Anthropometric analysis
Body Mass Index (BMI) was determined by dividing weight in kilograms by height in meters squared (kg/m2). BMI data were collected to as far back as 2 years prior to the diagnosis of DTC and as recently as the most recent available BMI data after diagnosis of DTC. The weight data for both control groups were collected to as many as 3 years preceding diagnosis of hyperparathyroidism (PHPTH group) or 3 years preceding year of presentation (IM group). The most recently available weight was recorded as well. BMI was categorized as underweight
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
(< 18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25.0–29.9 kg/m2), or obese (≥30 kg/m2) according to the World Health Organization definition. The body fat percentage was estimated from BMI by using the formula of Deurenberg et al. [14]: body fat percentage = (1.20 × BMI) + (0.23 × Age) − (10.8 × Sex) − 5.4, where age is in years and sex is set to 0 for women and 1 for men. Body surface area (BSA) was calculated from height and weight by using the formula of Du Bois and Du Bois [15]: BSA = (0.007184) × (Weight0.425) × Height0.725).
2.3.
Statistical analysis
Exposure variables consisted of the most recent BMI within 6 months prior to diagnosis of DTC and any diagnosis of DM. Also evaluated was a “historical BMI” variable assessing whether a patient had met clinical criteria for obesity 1 to 2 years prior to diagnosis of PHPTH or 1 to 2 years prior to the control reference date. The time frame for this measure was determined in order to preserve sufficient sample size in the case group. Multivariate logistic regressions adjusting for gender, age, race, and year of BMI were used to assess the odds ratio (OR) of DTC with both BMI and DM. Furthermore, multivariate logistic regression was used to analyze the historical relationship of BMI with DTC. Anthropometric measurements were examined as continuous variables and treated as categorical variables either as quartiles or tertiles following prior usage in literature. BMI histograms of DTC case patients were used as comparisons against both control groups. The Mann–Whitney U test was used to compare continuous variables between cases and controls as appropriate. Percentages may not total 100% as a result of rounding.
3.
3
Table 2 shows the odds ratios for the DTC cases compared to each control group as well as the pooled analysis for the anthropometric measurements and the association with diabetes. When considering BMI, BSA, and body fat, the risk of DTC was increased for obese patients when comparing each control group and the pooled data. The percentage of patients with diabetes was increased in the DTC group compared to each individual control group as well as pooled data. Although diabetes was not significantly more prevalent in the case group versus the PHPTH control group, it did trend toward significance (p = 0.0829, 95% CI = 0.902–5.407). Using multivariate analysis, the odds ratio when comparing cases versus PHPTH controls was significant (OR = 3.178, p = 0.0198, 95% CI = 1.202, 8.404). Comparing DTC cases to the IM control group, diabetes was not significantly associated with the DTC group in either bivariate or multivariate analysis. However, when comparing cases versus the pooled control data, the odds ratio showed diabetes was significantly associated with the risk of having DTC (OR = 2.237, p = 0.0410, 95% CI = 1.033, 4.844). Table 3 shows the odds ratios broken into quartiles for anthropometric data. As one might predict from our aforementioned results, patients with a greater BMI had statistically significantly increased risk for DTC. Interquartile comparisons of BMI, BSA, and body fat percentage showed, respectively, that lower BMI, BSA, and body fat percentage were associated with a lower risk of DTC. The number of patients with historical BMI was 26. The risk of DTC when including the historical data for BMI continued remained significantly increased when compared to the PHPTH group, with an OR of 1.112 (95% CI 1.039, 1.189). Additionally, there was a persistently increased risk when comparing DTC to the IM group, with an OR of 1.122 (95% CI 1.058, 1.190). Pooled data showed a significant, but slightly less strong risk, of DTC with history of obesity, with an OR of 1.087 (95% CI 1.036, 1.142). The odds ratios, using both BMI and historical BMI values, were significantly less than 1.0 for patients with BMI less than 30 when DTC patients were compared to each control group.
Results
A total of 48 DTC cases, 70 PHPTH controls and 102 IM controls were used in the analyses. Gender distribution among the DTC cases showed 39 (81.3%) patients were female and 9 (18.8%) were male. Table 1 shows the characteristics of each subject group as well as their anthropometric measurements. Fig. 1 shows an overall view of the distribution of each subject group when comparing BMI and DM. Additionally, 42 of the case group patients (87.5%) were found to have a BMI greater than 30. The IM control group had 36 patients (35.3%) with BMI greater than 30, whereas the PHPTH control group had 25 patients (35.7%) with BMI greater than 30.
4.
Discussion
There has been a great deal of focus on the epidemics of obesity and diabetes in the United States in recent years, and the number of affected people continues to rise each year. Thyroid cancer is on the rise; the Commonwealth of Pennsylvania has the highest incidence of thyroid cancer in the country as well as a high rate of obesity [3]. Therefore a large academic institution in Central PA would be an appropriate setting for a study of this kind. Determining any associations between thyroid cancer
Table 1 – Characteristics of differentiated thyroid cancer (DTC) cases and control study. Characteristics
DTC Case (n = 48)
PHPTH Control Group (n = 70)
p value
IM Control Group (n = 102)
p value
All Controls (n = 172)
p value
Age (years) Height (cm) Weight (kg) Body surface area (m2) Body fat (%) BMI (kg/m2) Type-II diabetes Race (% of Caucasians) Female gender
49.13 (14.73) 164.81 (9.25) 103.22 (22.61) 7.38 (0.90) 49.27 (9.92) 37.83 (6.95) 29.17% 89.58% 81.25%
56.63 (12.82) 165.82 (9.25) 84.06 (26.68) 6.76 (1.02) 41.28 (11.13) 30.36 (8.50) 15.71% 95.71% 74.29%
0.0082 0.5196 <0.0001 0.0002 0.0001 <0.0001 0.0809 0.1972 0.3804
51.88 (16.44) 165.90 (10.64) 80.28 (24.88) 6.63 (1.04) 37.69 (11.55) 28.96 (8.08) 21.57% 83.33% 66.67%
0.4110 0.4379 <0.0001 <0.0001 <0.0001 <0.0001 0.3124 0.3160 0.0667
53.81 (15.22) 165.87 (10.07) 81.82 (25.62) 6.68 (1.03) 39.15 (11.48) 29.53 (8.26) 19.19% 88.37% 69.77%
0.0794 0.4258 <0.0001 <0.0001 <0.0001 <0.0001 0.1371 0.8174 0.1173
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
4
AMER ICA N JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D N E CK M EDI CI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
Fig. 1 – This graph shows the prevalence rates of normal weight, overweight, obese, and type-II diabetic patients amongst the study groups, highlighting the significance of obesity in the case group.
and obesity may assist with future prevention or reduction in thyroid cancer as well as enhanced screening in an at risk population for early detection. The unique features of this study are seen with the use of two control groups, objective values for weight and height instead of surveyed estimates, and evaluation of patients over the course of years instead of a single time-point. Notably, one of the control groups had a documented euthyroid state which, to our knowledge, has not been included in prior evaluations. In regard to the evaluation of obesity and its relationship to DTC, our results are consistent with prior reports. Specifically, the risk of DTC is significantly increased in patients who are obese compared to overweight or normal weight patients. This was true even amongst all three types of mathematical comparisons: BMI, body surface area, or body fat percentage. Evaluating patients' body surface area, which reportedly is less affected by abnormal adipose mass than the BMI calculation, led to an even greater association of DTC with increases in mass than DTC's association to obesity. Further support for this association has been shown when accounting for the way men and women distribute fat, thus eliminating the limitations of using BMI [9]. The accumulation of visceral adipose tissue is related to dyslipidemia and insulin resistance. Insulin may directly contribute to proliferation of thyroid cancer cells by stimulating insulin-like growth factor. This relation further supports the involvement of insulin resistance biologic mechanisms in thyroid cancer development [5,6,9,10,12]. Estrogen may be another associative factor
between overall body fat and thyroid cancer risk. Estradiol, which is converted from testosterone by fat cells, is an estrogen which has been shown to be a promoter of benign and malignant thyroid tumor cells [9]. Recent evaluation of excess body weight and papillary thyroid cancer has shown a significantly positive relationship with larger tumor size, presence of microscopic extrathyroidal invasion, and higher TNM staging [5]. Hormonal changes associated with obesity implicate leptin and its receptor with tumor aggressiveness as well. Adipokines such as leptin may promote pituitary production of thyroid stimulating hormone (TSH) [9]. Leptin levels were shown to be elevated in patients with papillary thyroid cancer [16]. Additionally, leptin stimulates cell proliferation and inhibits apoptosis via activation of the phosphatidylinositide 3-kinase/protein kinase B (PI3K/ AKT) pathway, which appears to promote angiogenesis and tumor invasion [5]. Our study also evaluates the association of type 2 diabetes with DTC. Comparisons were performed using multivariate analysis, controlling for age and gender, showing significant association of DM with DTC and thereby suggesting that DM does contribute to DTC. Age was controlled to reduce the chances of confounding results due the effect age has on incidence of each disease. Additionally, gender was considered a potential confounder because of the propensity of women to have DTC. Support for DM's association with DTC has been suggested by multiple metabolic and hormonal mechanisms.
Table 2 – Study odds ratio (OR) and 95% confidence intervals for differentiated thyroid cancer (DTC) risk. Variable
2
Body surface area (m ) Body fat (%) BMI (kg/m2) > 30.0 Historical BMI (kg/m2) > 30.0 Type-II diabetes
DTC compared to control group 1 (PHPTH)
DTC compared to control group 2 (IM)
OR (95% CI)
p value
OR (95% CI)
p value
OR (95% CI)
p value
2.299 1.103 1.125 1.112 3.178
0.0011 0.0001 0.0001 0.0021 0.0198
3.363 1.127 1.154 1.122 1.860
<0.0001 <0.0001 <0.0001 0.0001 0.1490
2.538 1.093 1.113 1.087 2.237
<0.0001 <0.0001 <0.0001 0.0007 0.0410
(1.396, (1.049, (1.059, (1.039, (1.202,
3.786) 1.160) 1.195) 1.189) 8.404)
(2.037, (1.074, (1.089, (1.058, (0.801,
5.552) 1.183) 1.223) 1.190) 4.320)
DTC compared t o pooled controls
(1.713, (1.055, (1.066, (1.036, (1.033,
3.760) 1.133) 1.162) 1.142) 4.844)
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
5
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
Table 3 – Odds ratio (OR) and 95% confidence intervals for DTC risk according to quartiles. Variable
Q1 OR (95% CI)
DTC compared to primary hyperparathyroidism control group Body surface area (m2) 4.644–6.279 0.087 (0.020, 0.373) Body fat (%) 19.160–36.540 0.037 (0.007, 0.194) <24.9 (kg/m2) BMI (kg/m2) 0.046 (0.005, 0.403) Historical BMI (kg/m2) <24.9 (kg/m2) 0.073 (0.004, 1.451)
Q2
Q3
Q4
p value OR (95% CI)
p value OR (95% CI)
p value OR (95% CI)
0.0861
6.291–6.896 0.205 (0.056, 0.749) 36.960–43.280 0.151 (0.043, 0.522) 24.9–29.9 (kg/m2) 0.069 (0.020, 0.236) 24.9–29.9 (kg/m2) 0.082 (0.017, 0.394)
DTC compared to internal medicine (IM) control group Body surface area (m2) 4.200–5.998 0.025 (0.004, 0.148) <0.0001 Body fat (%) 10.560–32.770 0.004 (0.001, 0.043) <0.0001 <24.9 (kg/m2) BMI (kg/m2) 0.015 (0.002, 0.131) 0.0001 Historical BMI (kg/m2) <24.9 (kg/m2) 0.019 (0.001, .300) 0.0048
6.003–6.929 0.267 (0.086, 0.829) 32.950–40.010 0.074 (0.019, 0.280) 24.9–29.9 (kg/m2) 0.134 (0.041, 0.439) 24.9–29.9 (kg/m2) 0.186 (0.043, 0.814)
DTC compared to pooled controls (control groups 1 & 2) Body surface area (m2) 4.200–6.003 0.037 (0.008, 0.186) <0.0001 Body fat (%) 10.560–33.140 0.010 (0.002, 0.062) <0.0001 <24.9 (kg/m2) BMI (kg/m2) 0.025 (0.003, 0.191) 0.0004 Historical BMI (kg/m2) <24.9 (kg/m2) 0.029 (0.002, 0.454) 0.0115
0.0001 0.0001 0.0055
0.0197 0.0029 <.0001 0.0018
6.903–7.541 0.708 (0.210, 2.390) 43.500–51.310 0.792 (0.252, 2.489) >30.0 (kg/m2) 1.125 (1.059, 1.195) >30.0 (kg/m2) 1.112 (1.039, 1.189)
0.6183 0.6900
7.574–9.797 2.299 (1.396, 3.786) 51.580–76.620 1.103 (1.049, 1.160)
p value
0.0011 0.0001
0.0001 0.0021
7.645–9.312 3.363 (2.037, 5.552) <0.0001 50.170–78.720 1.127 (1.074, 1.183) <0.0001
0.0256
6.932–7.596 0.749 (0.266, 2.105) 0.5832 40.040–49.990 0.301 (0.105, 0.866) 0.0260 >30.0 (kg/m2) 1.154 (1.089, 1.223) <0.0001 >30.0 (kg/m2) 1.122 (1.058, 1.190) 0.0001
6.018–6.812 0.210 (0.073, 0.601) 0.0036 33.420–40.600 0.097 (0.032, 0.293) <0.0001 24.9–29.9 (kg/m2) 0.105 (0.036, 0.305) <0.0001 24.9–29.9 (kg/m2) 0.161 (0.038, 0.675) 0.0125
6.825–7.459 0.826 (0.344, 1.984) 0.6697 40.820–48.790 0.234 (0.093, 0.589) 0.002 >30.0 (kg/m2) 1.113 (1.066, 1.162) <0.0001 >30.0 (kg/m2) 1.087 (1.036, 1.142) 0.0007
7.465–9.797 2.538 (1.713, 3.760) <0.0001 48.920–78.720 1.093 (1.055, 1.133) <0.0001
Chronic elevated insulin levels have been observed in DM patients due to either insulin resistance or exogenous sources such as medications. Insulin and insulin growth factor-1 (IGF-1) are important for cell proliferation and apoptosis and have been found to be related to various cancers such as breast and colon cancer [11–13]. Insulin may play a role in thyroid carcinogenesis through its contribution to follicular cell growth. Follicular cells synthesize IGF-1 and have IGF-1 receptors. Insulin may mimic IGF-1, thus contributing to follicular cell growth. Additionally, IGF-1 receptor activates mitogen-activated protein (MAP) kinase and PI3K pathways. MAP kinases, like the PI3K pathway, also regulate gene expression, differentiation, mitosis, and cell apoptosis [12,13,17]. Another risk factor for thyroid cancer amongst diabetics is the chronically mild elevation of TSH [11,12,16,18,19]. A recent study showed that there is a 3.45 times increased risk of primary hypothyroidism in type 2 diabetics compared to a non-diabetic population [18]. Higher TSH levels have been shown to be risk factors for thyroid cancer as well as a more advanced stage of differentiated thyroid cancer [11,19]. One of the major mitogenic pathways in thyroid cancer is mediated by TSH. The hormone receptor adenylate cyclasecAMP protein kinase A system (AC/cAMP/PKA) is primarily
0.0224 0.0001 0.0009
stimulated by TSH. This pathway regulates the function, differentiation, and proliferation of the thyroid gland. Additionally, TSH stimulates the hormone receptor phospholipase C cascade (PLC) pathway which subsequently increased the intracellular calcium and protein kinase C (PKC) activity [12]. Once activated, PKC may lead to an increased expression of oncogenes, thus promoting cancer progression. Patients with DM have metabolic abnormalities that place them in a state of increased oxidative stress and a permanent pro-inflammatory state [13]. Prolonged inflammatory responses reduce the intracellular anti-oxidant capacity which may increase the risk of susceptible cells to malignant transformation. The mechanism is via generation of reactive oxygen species (ROS) produced from free radicals and oxidants which accumulate in a pro-inflammatory state [13]. ROS damage cells by direct oxidation or interfering with DNA repair mechanisms [20]. Another pro-inflammatory cytokine which is correlated with the insulin resistant environment of DM is tumor necrosis factor α (TNFα) [13]. TNFα, a cytokine involved in apoptosis, has been widely shown to be involved with the development and progression of many tumors including: breast, ovarian, colorectal, prostate, bladder, esophageal, renal cell, melanoma, lymphoma, and leukemia
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
6
AMER ICA N JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D N E CK M EDI CI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
[21]. Deregulation of apoptosis, and those cytokines involved, is critical in the tumorigenesis process. In comparison to healthy individuals, elevated serum concentrations and increased expression of TNFα are present in various preneoplastic diseases and malignancies [21]. One last proposed mechanism for an increase in thyroid cancer in diabetic patients is Vitamin D deficiency and deiodinase expression. Vitamin D deficiency has been observed in up to 70% of diabetics [12]. A low Vitamin D level decreases deiodinase II (DIO2) enzyme. Deiodinase enzymes regulate intracellular thyroid hormone levels; specifically, DIO2 regulates intracellular T3 concentration in the pituitary, brain, brown fat, and skeletal muscle. In skeletal muscle it indirectly regulates glucose transporter 4 (GLUT4) expression. Inactivation of DIO2 by a Vitamin D deficient environment in diabetic patients results in decreased GLUT4 transcription by skeletal muscle and adipose tissue, thus contributing to insulin resistance [11,12]. Our study has aimed to advance the understanding of the relationship between obesity, DM, and the risk of differentiated thyroid cancer. The strengths in our approach are seen with the use of two control groups, precise definitions for measurements, and multivariate evaluations of the accrued data. Specifically, we attempted to eliminate the bias that hypothyroidism or hyperthyroidism could play on subsequent thyroid cancer risk or weight change by having a euthyroid control group. The euthyroid state group was best ascertained by having a PHPTH group as they presented to the same clinic and underwent laboratory testing verifying hormone levels at the time of diagnosis. Additionally, we collected historical weight and height data to eliminate misclassification bias. Also, acquiring data over multiple time points shows that obese patients have increased risk of DTC over time, not just at a single time point prior to diagnosis of DTC. Some limitations to this study are partly due to not being able to account for all potential risk factors of DTC. We did not adjust for dietary habits, physical activity level, or genetic predisposition factors to obesity or thyroid cancer in our study. We were unable to obtain height and weight values for most patients beyond five years prior to DTC diagnosis. Additionally, our case group numbers were limited secondary to access to properly recorded height and weight data. Exclusion of many case patients is not likely a source of selection bias in our study, as our case groups show an incidence of obesity greater than that of the state in which the patients reside. It is more likely that our referral pool has a higher rate of obesity than the general population. Importantly, even in the setting of higher rates of obesity, we found significant results. Lastly, our study was performed in an area which is not racially diverse, but we believe our results would still be generally applicable to the general population.
5.
Conclusion
Our results show that obesity, as measured using various anthropometric values, is significantly associated with differentiated thyroid cancer. To a lesser degree, type 2 diabetes shares a close association with differentiated thyroid cancer.
Future studies evaluating obesity and DM in other regions of the U.S. would serve to broaden the application of our results. By identifying the rise in thyroid cancer with the rise in DM and obesity in our nation we hope to raise awareness that by addressing the epidemics of obesity and DM in our nation we may additionally address concerns of a rise in thyroid cancer.
Acknowledgments This research was funded by the Wells Fargo Jane Barsumian and Mary Lyons Trust Grant.
REFERENCES
[1] Davies L, Welch HG. Current thyroid cancer trends in the United States. JAMA Otolaryngol Head Neck Surg 2014;140: 317–22. [2] National Cancer Institute. State cancer profiles: incidence. http://statecancerprofiles.cancer.gov/cgi-bin/quickprofiles/ profile.pl?00&080; 2013. [August 9, 2014]. [3] Bann DV, Goyal N, Camacho F, et al. Increasing incidence of thyroid cancer in the Commonwealth of Pennsylvania. JAMA Otolaryngol Head Neck Surg 2014. http://dx.doi.org/10.1001/ jamaoto.2014.1709 [Epub ahead of print]. [4] Morris LG, Myssiorek D. Improved detection does not fully explain the rising incidence of well-differentiated thyroid cancer: a population-based analysis. Am J Surg 2010;200: 454–61. [5] Kim HJ, Kim NK, Choi JH, et al. Associations between body mass index and clinico-pathological characteristics of papillary thyroid cancer. Clin Endocrinol (Oxf) 2013;78:134–40. [6] Kitahara CM, Platz EA, Freeman LE, et al. Obesity and thyroid cancer risk among U.S. men and women: a pooled analysis of five prospective studies. Cancer Epidemiol Biomarkers Prev 2011;20:464–72. [7] Xu L, Port M, Landi S, et al. Obesity and the risk of papillary thyroid cancer: a pooled analysis of three case–control studies. Thyroid 2014;24:966–74. [8] Pennsylvania Department of Health. Obesity prevention and wellness program. Pennsylvania obesity data and trends. http://www.portal.state.pa.us/portal/server.pt/community/ obesity/14184; 2014. [August 9, 2014]. [9] Kitahara CM, Platz EA, Park Y, et al. Body fat distribution, weight change during adulthood, and thyroid cancer risk in the NIH-AARP Diet and Health Study. Int J Cancer 2012;130: 1411–9. [10] Marcello MA, Sampaio AC, Geloneze B, et al. Obesity and excess protein and carbohydrate consumption are risk factors for thyroid cancer. Nutr Cancer 2012;64:1190–5. [11] Aschebrook-Kilfoy B, Sabra MM, Brenner A, et al. Diabetes and thyroid cancer risk in the National Institutes of Health– AARP Diet and Health Study. Thyroid 2011;21:957–63. [12] Shih SR, Chiu WY, Chang TC, et al. Diabetes and thyroid cancer risk: literature review. Exp Diabetes Res 2012;2012: 578285. [13] Vigneri P, Frasca F, Sciacca L, et al. Diabetes and cancer. Endocr Relat Cancer 2009;16:1103–23. [14] Deurenberg P, Weststrate JA, Seidell JC. Body mass index as a measure of body fatness: age- and sex-specific prediction formulas. Br J Nutr 1991;65:105–14. [15] Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition 1989;5:303–11 [discussion 312–303].
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015
AMER IC AN JOURNAL OF OT OLA RYNGOLOGY– H E A D A N D NE CK M E D ICI N E AN D S U RGE RY X X (2 0 1 5) XXX – XXX
[16] Akinci M, Kosova F, Cetin B, et al. Leptin levels in thyroid cancer. Asian J Surg 2009;32:216–23. [17] Frohlich E, Wahl R. Do antidiabetic medications play a specific role in differentiated thyroid cancer compared to other cancer types? Diabetes Obes Metab 2012;14:204–13. [18] Tamez-Perez HE, Martinez E, Quintanilla-Flores DL, et al. The rate of primary hypothyroidism in diabetic patients is greater than in the non-diabetic population. An observational study. Med Clin (Barc) 2012;138:475–7.
7
[19] Haymart MR, Repplinger DJ, Leverson GE, et al. Higher serum thyroid stimulating hormone level in thyroid nodule patients is associated with greater risks of differentiated thyroid cancer and advanced tumor stage. J Clin Endocrinol Metab 2008;93:809–14. [20] Frederico A, Morgillo F, Tuccillo C, et al. Chronic inflammation and oxidative stress in human carcinogenesis. Int J Cancer 2007;121:2381–6. [21] Szlosarek P, Charles KA, Balkwill FR. Tumour necrosis factoralpha as a tumour promoter. Eur J Cancer 2006;42:745–50.
Please cite this article as: Oberman B, et al, Relationship between obesity, diabetes and the risk of thyroid cancer, Am J Otolaryngol–Head and Neck Med and Surg (2015), http://dx.doi.org/10.1016/j.amjoto.2015.02.015