INTRODUCTION
Owing to the development of early detection tools and various cancer treatment modalities, the detection rate of thyroid cancer has tremendous increased [1,2]. In Korea, thyroid cancer is commonly diagnosed at young age (age 40s), with a favorable 5-year relative survival rate of approximately 90% to 100% [3]. Generally, total thyroidectomy or lobectomy is the first-line treatment for patients with thyroid cancer. After surgical removal, all patients who undergo total thyroidectomy and some who undergo lobectomy require levothyroxine (T4) replacement. Patients with a high risk of recurrence receive a high dose of T4 to suppress and maintain thyrotropin (TSH) levels below the lower normal range, that is, exogenous subclinical hyperthyroidism, to prevent recurrence for a certain period [4]. Patients with a low-to-intermediate risk receive a relatively lower T4 dose to maintain TSH levels in lower normal range. Nearly 50% patients undergo total thyroidectomy, followed by radioactive iodine (RAI) ablation according to the risk of recurrence [1], and a subsequent iodine scan would be required during cancer surveillance. RAI therapy and iodine scans require a low-iodine diet, which is uncommon in Korea. In patients undergoing thyroid hormone withdrawal for TSH stimulation before RAI, iatrogenic overt hypothyroidism can cause generalized edema and general weakness. In addition to surgery and surgery-related complications, patients with thyroid cancer can experience inconvenience (taking medicine) and an iatrogenic pathologic state (subclinical hyperthyroidism or overt hypothyroidism) due to the effects of thyroid cancer treatment and follow-up strategies. Thus, unfavorable health outcomes due to primary cancer treatment have received much attention [5]. In addition, outcomes such as cardiovascular disease [6]; osteoporosis [7]; and metabolic syndrome, including insulin resistance, obesity, and hypertension [8] are commonly observed after surgical therapy, followed by RAI or TSH suppression. Given the relatively young age at diagnosis and high survival rate among patients with thyroid cancer, quality of life (QoL) and long-term adverse health outcomes are of interest to many physicians. Generally, cancer survivors experience physical and mental difficulties after cancer treatment. Therefore, it is reasonable to suspect that QoL in cancer survivors may be poorer than that in subjects without cancer.
Previous studies have reported that thyroid cancer survivors experience physical or psychological distress [9-14]. These studies had small sample sizes and short-term follow-up period [10,11,13] or lacked a healthy control group [9-12,15]. These studies also lacked valid tools to evaluate health-related (HR) QoL [9-11]. Moreover, limited studies have evaluated long-term HRQoL in a large sample of thyroid cancer survivors. Thus, we aimed to investigate HRQoL using validated questionnaires in thyroid cancer survivors and compared it to that in healthy subjects to better tailor support for thyroid cancer survivors with high survival rates.
METHODS
Study population and data collection
This population-based cohort study used data of the Korea National Health and Nutrition Examination Survey (KNHANES) from KNHANES IV to VII (2007 to 2017). The KNHANES conducted by Korean Centers for Disease Control and Prevention for Health Statistics is an independent dataset obtained from the general South Korean population. The selection of participants for the survey is based on a stratified multistage probability sampling design for the South Korean population and a two-stage stratified systematic sampling method. The KNHANES monitors the health and nutritional status of the Korean population, and assessments are conducted by trained interviewers. The details of the database are described elsewhere (https://knhanes.cdc.go.kr). This study complied with the ethical standards of the Declaration of Helsinki and was approved by the Catholic University of Korea, Catholic Medical Center, Eunpyeong St. Mary’s Hospital Institutional Review Board (IRB approval No. PC20ZASI0091). The requirement for written informed consent was waived due to the use of previously collected and anonymized data. Among the 89,630 subjects who participated from KNHANES IV to KNHANES VII, 38,285 subjects were excluded based on the following criteria: age <19 years (n=20,691), self-reported history of chronic renal disease (n=261), cerebrovascular disease (n=1,369), rheumatic or osteoarthritis (n=8,846), moderate depression (n=678), or other cancers except thyroid cancer (n=15), and missing data on EuroQoL (EQ)-5 dimension (5D) utility scores (n=27,389). In this study, data from 50,278 participants were analyzed. To increase sensitivity, reduce sampling bias, and overcome confounding and selection bias when utilizing large and representative datasets, sampling weights were adjusted according to demographic factors. Propensity score matching (PSM) was performed between 312 thyroid cancer survivors and 49,966 healthy subjects from 50,278 participants from 2007 to 2017. We used the following algorithms for forming matched pairs of thyroid cancer survivors and healthy subjects: optimal matching and greedy nearest neighbor matching within specified caliper widths (0.25). Covariates for PSM matching were sex, age, smoking status, drinking behavior, marital status, education level, income level, physical activity, and self-perceived stress. After PSM matching, 183 subjects with thyroid cancer and 915 subjects without any cancer were extracted. Finally, a total of 1,098 subjects were analyzed in this study. The propensity score for thyroid cancer probability was estimated using a multivariate logistic regression model. The greedy nearest neighbor algorithm was used to match thyroid cancer survivors to five healthy subjects (1:5 matching). After matching, the standardized mean differences (SD) were calculated to verify the quality of matching; if the SD was <10%, matching was confirmed as the best balance between the two groups.
Demographic variables and lifestyle factors
Demographic and lifestyle variables of the subjects were collected using a questionnaire. The following variables were assessed: sex, age, smoking status (stratified into current smokers and non-smokers, including ex-smokers), drinking behavior (classified as non-to-moderate drinking and risky drinking; >30 g/day of alcohol), marital status (married or separated; categorical), educational status (divided into two groups—less than middle school and greater than middle school), income level (categorized into two groups—low/mid-low and mid-high/high), physical activity (defined as activity consisting of at least 30 minutes of walking for at least 5 days a week), and self-perceived stress according to the response (yes or no; categorical).
Sub-classification by duration of thyroid cancer
Follow-up and management strategies for patients with thyroid cancer, including T4 dosing and modality or frequency of laboratory/imaging studies, are subject to change over time (dynamic or ongoing risk stratification) [16]. Thus, the QoL of thyroid cancer survivors can differ according to the time interval after disease diagnosis. According to American and Korean Thyroid Association guidelines, additional treatment and surveillance tools beyond thyroid ultrasonography (US) are rarely indicated if there is no suspected remnant disease or recurrent disease in over 10 years [4,17]. Therefore, we categorized thyroid cancer survivors into groups based on time of diagnosis—<5, 5-9, and ≥10 years after diagnosis.
Assessment of QoL
HRQoL was assessed using the Korean version of the EQ-5D, which has been included to evaluate QoL in the questionnaire since 2005 (KNHANES III) [18]. The EQ-visual analog scale (VAS) score was also measured in the KNHANES from 2007 to 2012. The EQ-5D comprises five targeted questions—mobility, self-care, usual activities, pain/discomfort, and anxiety/depression—to represent the current health status, EQ-5D score, and EQ-VAS score [19]. Each dimension is scored using three severity levels (no problem, some problems, and severe problems). Population-specific value sets are used for comparisons across different populations. The EQ-5D index is a continuous value that ranges from −0.171 (indicating poorer health status) to 1 (indicating the best health status) [20]. The EQ-VAS derives information from the perception of subjective health status, scored on a VAS with endpoints labeled “the best health you can imagine” and “the worst health you can imagine.” The EQ-VAS comprises a scale ranging from 0 (worst health status) to 100 (best health status). The detail of questionnaire of HRQoL were shown in section of Supplemental Methods with Supplemental Table S1.
Statistical analysis
Statistical analysis was performed to reflect the complex sampling design and sampling weights of KNHANES to provide nationally representative prevalence estimates. The SSAS® PROC SURVEY module was used to consider strata, clusters, and weights. Demographic variables including sex, smoking status, drinking behavior, marital status, educational status, income levels, physical activity, and self-perceived stress were analyzed by a complex sample analysis using Pearson’s chi-square test. The characteristics of each group were compared using independent t tests for continuous variables. Based on the data characteristics, the results are expressed as mean±standard error (SE), geometric means (95% confidence interval), or percentages, as appropriate. Adjusted linear regression analysis was performed to test for significant differences in the EQ-5Q score. To determine the association between thyroid cancer and EQ-5D scores, complex sample logistic regression analysis was performed. All statistical analyses were performed using SAS software version 9.4. (SAS Institute Inc., Cary, NC, USA). A P value of <0.005 was considered statistically significant.
RESULTS
Baseline characteristics between patients with thyroid cancer and healthy subjects
Table 1 shows baseline characteristics between thyroid cancer survivors and healthy subjects before and after 1:5 PSM matching. Before matching, there were differences in sex, age, behavior, and education level between thyroid cancer survivors and healthy subjects. Data on 183 thyroid cancer survivors and 915 healthy subjects were analyzed. The mean age of thyroid cancer survivors was 51.2±0.9 years, and 76.4% patients were women. Smoking status, drinking habit, marital status, education status, income levels, and physical activity were similar between the groups after PSM matching. Absence of self-perceived stress did not differ between the groups. The unadjusted mean EQ-5D index score was 1.0±0.0 in both groups (P=0.057). No difference in EQ-VAS scores was observed between thyroid cancer survivors and healthy subjects (74.9±2.5 vs. 76.5±1.0, P=0.503). Thus, the weighted and matched variables were not significantly different between the groups.
Frequencies of reporting problems according to EQ-5D categories in patients with thyroid cancer and healthy subjects using the weighted and matched model
The proportion of subjects who reported problems in the five categories of the EQ was analyzed (Table 2). The percentage of participants who answered “mild-to-severe problem” for the mobility, self-care, and usual activity; “none” for pain/discomfort; and “none” for anxiety/depression dimensions of the EQ-5D was similar between the groups. None of the thyroid cancer survivors responded to the “mild-to-severe problem” in terms of self-care.
Multivariate linear regression model for EQ-5D score and EQ-5D index in patients with thyroid cancer versus healthy subjects
In linear regression, the presence of thyroid cancer was positively correlated with a better EQ-5D index score (β coefficient=0.010, P=0.0460). Though EQ-VAS showed negative correlation, no statistical significance was observed, (β coefficient=-1.670, P=0.4636) (Table 3). After adjustment for age; sex; income level; education level; marital status; and lifestyle habits including smoking, risky drinking, and physical activity, an increase in the EQ-5D index score in patients with thyroid cancer was observed compared to that in healthy subjects (β coefficient=0.010, P=0.046). Among the factors comprising the EQ-5D index, thyroid cancer survivors showed a lower risk of having problems with self-care.
Correlation between EQ-5D index score and thyroid cancer duration
The EQ-5D index score was negatively correlated with thyroid cancer duration. However, this difference was not significant (β coefficient=-0.001, P=0.2887). Considering the different treatment and follow-up strategies based on the thyroid cancer duration, we performed subgroup analysis according to the thyroid cancer duration in thyroid cancer survivors. The thyroid cancer duration in 40,292, 38,251, and 16,981 patients was <5 years (mean, 2.8±1.0 years), 5-9 years (mean, 6.4±1.3 years), and >10 years (mean, 13.9±4.1 years), respectively. The EQ-5D index score showed a positive correlation for the duration of 5-9 years (β coefficient=0.003, P=0.008). It decreased in thyroid cancer survivors with a disease duration of >10 years (β coefficient=-0.026, P=0.168), although the association was not significant (Table 4).
DISCUSSION
In the analysis of a nationwide population-based database, the QoL of thyroid cancer survivors, estimated by the EQ-5D index score, was non-inferior to that of healthy subjects without any cancer. The QoL of thyroid cancer survivors showed even better scores due to the increased self-care index compared to non-cancer subjects. Additionally, no difference was observed in the QoL of thyroid cancer survivors according to thyroid cancer duration.
Differentiated thyroid cancer has an excellent prognosis with a high long-term survival rate [21,22]. However, despite the favorable long-term prognosis, there have been concerns regarding HRQoL in thyroid cancer survivors arising from their unique treatment modalities. HRQoL in thyroid cancer survivors in relation with treatment modalities such as surgery including total thyroidectomy, and lobectomy, RAI therapy, and TSH suppressive therapy depending on disease stages has been established [4,17,23-25]. With respect to surgery, one study showed lower HRQoL in thyroid cancer survivors due to tightening and pain around surgical scars [26] while another study in the United States showed no impact of surgical wounds on HRQoL 2 years after surgery [27]. The differences in HRQoL among total thyroidectomy, lobectomy, and robot-associated thyroidectomy remain inconclusive [5,28-30]. The risk of sialadenitis and dysphagia after RAI, anxiety before RAI, iatrogenic hypothyroidism, and hyponatremia during a low-iodine diet with thyroid hormone withdrawal were associated with low HRQoL [11,31-33]. In contrast to thyroid hormone withdrawal, preparation for RAI therapy with recombinant TSH was positively correlated with HRQoL [5,34,35]. HRQoL negatively correlated with TSH suppression [10,36-39]. In addition to the total HRQoL, chronic subclinical hyperthyroidism or hyperthyroidism can cause medical problems such as cardiovascular complications [6], bone disease [7], muscle dysfunction [40], and metabolic disease [8].
Most studies previously evaluated HRQoL according to treatment modality in only thyroid cancer survivors without comparative data and focused on depression, anxiety, fatigue, and concern about recurrence of thyroid cancer. Some articles evaluated HRQoL without valid tools or normative data [9-11,15]. After developing, validating, and translating questionnaires such as the EQ-5D [41], Thyroid Cancer Survivors’ Association (ThyCA)QoL [42], Thyroid-Specific Patient Reported Outcome (ThyPRO) [43], and European Organization for Research and Treatment of Cancer (EORTC)-Quality of Life Questionnaire-30 (EORTC-QLQ-C30) [44], several studies have adapted these tools. A cross-sectional study (n=316) by Lee et al. [45] and a German cohort study (n=121) [46] showed that HRQoL in survivors of thyroid cancer was lower than that in the general population. Additionally, anxiety, depression, and levels of fatigue significantly determined the lower HRQoL based on the EORTC-QLQ-C30. A recent study using EQ-5D or EQ-VAS concluded that HRQoL in thyroid cancer survivors was lower than that in the reference or general population [47]. A Chinese population-based survey using the EORTC-QLQ-C30 combined with SF-36 showed that thyroid cancer survivors had impaired HRQoL [48]. However, these studies included only subjects with relatively short duration after surgery (mean 37.3±28.8 months [45]) or a short time since diagnosis (mean 5 years with range from 1 to 27 years [47], 2.6±3.6 years [48]) Also, subjects could not represent the population.
In terms of the duration of thyroid cancer, there are reports that impaired HRQoL is correlated with short-term follow-up (<5 years) [10,14]. HRQoL was restored to normal after 12 to 20 years of follow-up [14]. However, no significant correlation was observed between the duration of thyroid cancer and HRQoL by EQ-5D in this study. Additionally, compared to subjects with a thyroid cancer duration of less than 5 years, HRQoL was not inferior in thyroid cancer survivors with a longer duration. The different results of our study with those of previous studies might be due to the population distribution. In our study, more than half of the population (59.0%) had a disease duration >5 years.
Our study has several advantages over previous studies. First, it included a relatively large number of subjects who were representative of the nationwide population. Second, a validated tool for QoL assessment was used. Finally, there were a substantial number of long-term survivors more than 5 years after the diagnosis of the disease. Surprisingly, the EQ-5D and EQ-VAS scores were not significantly different between the thyroid cancer survivor and non-cancer control groups. One possible explanation for the similar HRQoL is the high percentage of education level (77.8%) in the study population. A lower educational level is associated with a poor HRQoL index [9,13,42]. Highly educated subjects have a better understanding of the disease and are more satisfied [49]. Another possible explanation is the marital status. Among thyroid cancer survivors, 83.4% were married and received family support [50].
To our knowledge, this is the first study to utilize data from the third to seventh editions of the KNHANES to investigate the association between thyroid cancer and health-related QoL. This study was based on a large sample size from a nationally representative database. The use of PSM to match thyroid cancer survivors to healthy individuals enhanced the reliability of our results and minimized selection bias. Thyroid cancer survivors were matched by sex, age, education, and income level to five healthy subjects (non-cancer group) to intensify comparability. However, this study had several limitations.
First, there have been the specific instruments with cross-cultural validity for evaluating HRQoL in thyroid cancer such as THYCA-QoL [42] and ThyPRO [43]. The KNHANES has used EQ-5D to assess HRQoL in general population and EQ-5D and EQ- VAS might not be the best available HRQoL for thyroid cancer patients. Thus, thyroid cancer-specific tool is necessary to evaluate the health outcome in general population. Second, the causal relationship between thyroid cancer and related QoL could not be assessed because of the cross-sectional study design. Third, underreporting could not be excluded from the self-reporting system. Fourth, we could not exclude the comorbidities such as chronic pulmonary disease and diabetes mellitus which are meaningfully associated with HRQoL due to sample size. Finally, we could not obtain information on thyroid cancer stages, histology of thyroid cancer, and types of thyroid cancer treatment (T4 dose, extent of surgery, additional post-operative medications, and RAI therapy) because the KNHANES data lack specific clinical features. Therefore, the factors affecting HRQoL could not be identified. The combination of representative and clinical data could provide an excellent result for QoL in thyroid cancer survivors according to multiple clinical features.
HRQoL in long-term thyroid cancer survivors of was not inferior to that in healthy subjects. Preserved HRQoL, after long-term survivors of thyroid cancer presumably underwent standard treatment, and follow-up modalities should be considered in decision-making for cancer treatment, including both surgery and active surveillance.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
CrossRef TDM
Supplement1
Print
