Hashimoto's Thyroiditis, Risk of Coronary Heart Disease, and l-Thyroxine Treatment: A Nationwide Cohort Study.

Hashimoto’s Thyroiditis, Risk of Coronary Heart

Disease, and

-Thyroxine Treatment: A Nationwide

Cohort Study

Wei-Hung Chen, Yen-Kung Chen, Cheng-Li Lin, Jiann-Horng Yeh, and Chia-Hung Kao

H

cause of noniatrogenic hypothyroidism (1). Hypothyroidism has been observed to increase the risk of cardiovascular disease and atherosclerosis (2– 4). A recent metaanalysis reported an increased risk of approximately 20%

for coronary heart disease (CHD) in patients with subclinical hypothyroidism (5).

Most studies have focused on hypothyroidism, which has been presumed to causeCHDthrough hyperlipidemia,

hypertension, diabetes, and obesity. Few studies have specifically

examined autoimmune thyroid disease, which can precipitate cardiovascular disease through chronic inflammation

in addition to hypothyroidism (6, 7).

The aim of this study was to investigate the risk ofCHD among HT patients in a large cohort and determine the interaction between other cardiovascular risk factors and the effect of T4 treatment.

Subjects and Methods

Data source

The National Health Insurance (NHI) program in Taiwan was established on March 1, 1995. Under the NHI program, 99% of the island’s population receives all forms of healthcare service, including ambulatory care, outpatient and inpatient treatment, dental services, and physician services (http://www.

nhi.gov.tw/english/index.aspx). We conducted a retrospective

cohort study by using the Longitudinal Health Insurance Database 2000 (LHID 2000) released by the Taiwan National Health

Research Institutes. The LHID 2000 is composed of historical claims data for 1 million people selected randomly from the

23.74 million people in the NHI registry of beneficiaries. In the NHI program, diseases are diagnosed according to the International Classification of Diseases, Ninth Revision, Clinical

Modification (ICD-9-CM). The NHI program has a rigorous

monitoring system to ensure that the claims for healthcare reimbursement are based on valid diagnoses. Personal information

is scrambled to protect privacy to ensure that all data are deidentified and analyzed anonymously. This study was approved

by the China Medical University Institute Review Board (CMU-REC-101-012).

Sampled participants

From the LHID 2000, the HT cohort consisted of patients aged 20 years and older newly diagnosed with HT (ICD-9-CM code 245.2) from 2000 to 2010. The index date for anHTpatient was the date of the first clinic visit after theHTdiagnosis. Patients with a history of CHD (ICD-9-CM codes 410–414) before the index date were excluded. The non-HT cohort was formed by randomly selecting patients without a history of HT and CHD from the claims data, frequency matched with theHTpatients by

sex; age (in 5-year bands); baseline comorbidities including hypertension, diabetes, hyperlipidemia, stroke, rheumatic heart

disease, chronic kidney disease, and heart failure; and index year, at a ratio of 1:4.

Outcome and relevant variables

Patients in both the HT and non-HT cohorts were followed up until they were diagnosed with CHD or censored because of loss of follow-up, withdrawal from insurance, death, or the end of 2011. The baseline comorbidities considered were hypertension (ICD-9-CM codes 401–405), diabetes (ICD-9-CM code

250), hyperlipidemia (ICD-9-CM code 272), stroke (ICD-9-CM codes 430–438), rheumatic heart disease (ICD-9-CM codes 393–398), chronic kidney disease (ICD-9-CM code 585), and heart failure (ICD-9-CM code 428).

Statistical analysis

The distributions of demographic characteristics were compared between the HT cohort and the non-HT cohort, and the

differences were examined. The _2 test was used for categorical

follow-up person-years were used to estimate the incidence density for both cohorts. Univariable and multivariable Cox’s proportional hazards regressions were used to assess the risk of

developingCHDthat was associated withHT.The multivariable

models were simultaneously adjusted for demographic characteristics and the comorbidities of hypertension, diabetes hyperlipidemia, stroke, chronic kidney disease, and heart failure. The

hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using the Cox model. Further analysis was performed to assess whether L_-T4 treatment played a role in the CHD

outcomes.

We assessed the cumulative incidence of CHD by using the Kaplan-Meier method between the HT cohort and the non-HT cohort and estimated their differences using the log-rank test. All analyses were performed using the SAS statistical package (version 9.2 for Windows; SAS Institute). A 2-tailed P value _.05

was considered statistically significant.

Results

This cohort study consisted of 1165 newly diagnosed HT patients and 4660 non-HT controls from 2000 to 2010, with a median follow-up time of 5.46 (range 0.07–12.0) years. Table 1 shows the baseline demographic data and comorbidities. Most patients were women (90.8%) and

_49 years of age (77.0%). The mean ages of the non-HT cohort and the HT cohort were 40.9 _ 12.9 and 41.1 _

12.6 years, respectively. The HT cohort and non-HT cohort were well-matched for comorbidities.

Figure 1 shows that the cumulative incidence of CHD was higher in the HT cohort than that in the non-HT

cohort by 2.30% at the end of the follow-up period (logrank test P _ .02).

Table 2 shows the incidence rates of CHD for both cohorts. Compared with the non-HT cohort, the crude HRofCHDin theHTcohort was 1.44 times higher (a rate of 7.56 vs 5.22 per 1000 person-years), with an adjusted HRof 1.44(95%CI_1.05–1.99) (Table 2). Although the CHD incidence was higher in men than in women in both cohorts, the sex-specific HT to non-HT cohort adjusted

HR of CHD was significant for women (adjusted HR _ 1.45, 95% CI _ 1.03–2.04). The age-specific incidence rate ofCHDshowed that the incidence rate increased with age in both cohorts; however, theHRwas significant only for those_49 years of age (adjustedHR_1.74,95%CI_ 1.10–2.75).

The comorbidity-specific analysis showed that although the incidence rate of CHD was consistently higher in patients with comorbidities, the risk for developing

CHD was significant only in HT patients without comorbidities (adjusted HR _ 2.42, 95% CI _ 1.46–4.01).

Table 3 shows the results of univariate and multivariate

Cox proportional hazards analyses for the association between CHD and HT. After multivariable analysis for comorbidities, HT (HR _ 1.44, 95% CI _ 1.09–1.99), hypertension

(HR _ 2.06, 95% CI _ 1.46–2.92),

hyperlipidemia (HR _ 1.83, 95% CI _ 1.31–2.55), and heart failure(HR_2.21,95%CI_1.09–4.49) remained independent risk factors for CHD.

Table 4 shows that the joint effects of hypertension or hyperlipidemia and HT increased the overall risk for

CHD. Compared with the non-HT cohort without hypertension, the HT cohort with hypertension had a significantly

increased risk of CHD (adjusted HR _ 2.78, 95%

CI _ 1.58–4.90). Similarly, the HT cohort with hyper- lipidemia had a higher risk of CHD than did the non-HT

cohort without hyperlipidemia (adjusted HR _ 2.85, 95% CI _ 1.69–4.80).

The results of T4 treatment related to CHD risk are

shown in Table 5. Compared with the non-HT cohort,HT without treatment and HT with treatment for less than 1 year were associated with higher risk of CHD (adjusted HR _ 1.55, 95% CI _ 0.98–2.46, and adjusted HR _ 2.42, 95% CI_1.43–3.97). By contrast, the risk of CHD decreased after treatment withT4 for more than 1 year and

did not differ from that of the non-HT cohort (adjusted HR _ 0.84, 95% CI _ 0.47–1.52).

Our study consisted of a cohort of 1165 HT patients and a control cohort of 4660 non-HT patients, matched by age, sex, comorbidities, and index year at a ratio of 1:4. This study provides further evidence thatHTis associated with an increased risk of CHD irrespective of baseline comorbidities, including diabetes mellitus, hypertension, hyperlipidemia, stroke, chronic kidney disease, and heart disease. The adjusted HR for developing CHD in the current HT cohort was 1.44, a figure higher than that reported in previous studies. Compared with previous studies on subclinical hypothyroidism and CHD, our study

was focused specifically on HT patients. The study group is unique because of its female predominance (90.8%) and younger age at onset (mean age of 41.1 year), which is consistent with demographic data from the HT population (1, 8).

When examining the sex-specific HR, only female HT patients showed a significantly higher risk for CHD, despite the CHD incidence being higher in men in both cohorts. The age-specific HR was significant only for those

_49 years of age with an adjustedHRof 1.74, higher than those of 50 to 64 and _65 years of age (adjusted HR _ 1.53 and 0.87, respectively). Few studies have examined the sex- and age-related difference; however, data from a meta-analysis revealed that women have a higher risk for CHD than men do (adjusted HR _ 1.23 vs 1.06), and

those of younger age (18–49 years) have a higher risk than those of older age do (50–64 years) (adjusted HR _ 1.55 vs 1.11) (9). Both age and sex are possibly related toCHD in patients with subclinical hypothyroidism, and the effect is more significant in HT patients.

HT patients are known to have a higher incidence of diabetes and hyperlipidemia, which reflects the metabolic effects of hypothyroidism (10–12). Nevertheless, HT remains a significant risk factor for CHD after adjusting for

comorbidities, as well as in patients without comorbidities, suggesting a mechanism independent of hypothyroidism. Chronic inflammation in autoimmune diseases has

been recognized as a contributing factor to atherosclerosis and cardiovascular disease (13, 14). A nationwide study from Sweden investigated 31 immune-mediated diseases and the risk of subsequent CHD, among which HT was listed as having the fourth highest risk, below chorea minor, systemic lupus erythematosus, and rheumatic fever,

with a standardized incidence ratio of 4.30 (95% CI _

3.87–4.75) (15). The results suggested that HT, an autoimmune disease, is a crucial risk factor for CHD. Several

studies have shown endothelial dysfunction in HT patients, which might further contribute to atherosclerosis

(16, 17). Although HT is a risk factor for future CHD, comorbidities with conventional cardiovascular risk factors,

such as hypertension or hyperlipidemia, further increase the risk (HT with hypertension HR _ 2.06 and HT with hyperlipidemia HR _ 1.83, respectively). Recognizing

and treating the concomitant hypertension and hyperlipidemia in HT patients is therefore critical.

The effect of T4 treatment in preventing CHD and reducing

mortality is controversial. T4 therapy has been

shown to improve coronary flow (18), reverse cardiac dysfunction (19), and decrease hyperlipidemia inHTpatients

(20); all are beneficial for HT patients at risk for CHD. However, Flynn et al (21) reported that nonfatal ischemic

heart disease and dysrhythmias increased in treated hypothyroidism when adjusted for age, sex, diabetic status,

and previous vascular disease. A recent study that examined the risk for stroke in autoimmune thyroiditis disclosed

that the effect of autoimmune thyroiditis on stroke risk was highest in the first year after diagnosis, and the risk reduced after thyroid hormone replacement (22). This is because a prediagnosis of hypothyroidism can take time to reduce the risk after treatment. Alternatively, T4 treatment

might increase coagulation factor levels and inhibit fibrinolysis, which in turn leads to an increased ischemic stroke risk (23). Our results are consistent with their findings thatHTpatients withT4 treatment for less than 1 year

more than 1 year significantly reduces the risk. Atherogenesis in HT patients might be reversible after treatment,

although it takes time. However, an insufficient period of treatment adversely affects the outcome.

This study had a few limitations. First, the National

Health Research Institutes does not provide detailed patient information, such as smoking habits, alcohol consumption, body mass index, physical activity, socioeconomic

status, and family history ofCHD.These covariates might have affected the study results because of their strong relationship with risk of CHD. Second, the results

derived from a cohort study are generally of lower methodological quality than those derived from randomized

trials because a cohort study design is subject to bias related to adjustments for confounders. Despite our meticulous study design, including adequate control of confounding factors, bias might have remained because of

possible unmeasured or unknown confounders. Third, the diagnoses recorded in NHI claims are used primarily for administrative billing; therefore, they are not subject to verification for scientific purposes. We were unable to contact the patients directly to obtain additional information

because of the anonymity ensured by the identification numbers. However, the NHI program does have a rigorous monitoring system to ensure that the claims for healthcare

reimbursement are based on valid diagnoses. Medical reimbursement specialists and peer review should scrutinize all

insurance claims. The diagnoses ofHTandCHDwere based on the ICD-9 codes, which were judged and determined by related specialists and physicians according to the standard criteria. Therefore, the diagnosesandcodes forHTandCHD used in this cohort study should be correct and reliable. Of course, misclassification of cases basedonICD-9codes alone without other more definitive criteria such as individual laboratory and imaging data is still one of the study limitations.

Finally, we obtained no information regarding the serum thyroid hormone level or the compliance of the patients to document the adequacy of T4 treatment. However, a therapeutic

supplement of T4 is well-tolerated with few adverse

effects, and patients takingT4 for more than 1 year suggested

positive compliance.

In conclusion, this paper provides valuable information regarding the association betweenHTandCHD.This was

a large-scale nationwide cohort study of an Asian population. Patients with HT, particularly women and younger

patients, are at higher risk of developing CHD compared

with the general population. For patients with HT, the comorbidities of hypertension

or hyperlipidemia further increase the risk; treatments for the concomitant hypertension and hyperlipidemia are crucial in preventing future CHD. Furthermore,

our results confirm that treatment with T4 reduces the risk

of CHD, and most importantly, the treatment should be implemented early and continue for more than 1 year.