Duration of Hyperthyroidism and Lack of Sufficient Treatment Are Associated with Increased Cardiovascular Risk

Abstract

Background:

Cardiovascular disease remains the most prevalent cause of death in hyperthyroidism. However, the impact on cardiovascular events of varying thyroid status and of treatment remains unclarified. The aims of this study were to investigate the association between hyperthyroidism and cardiovascular events in treated and untreated hyperthyroid individuals, as well as exploring the impact of cumulative periods of hyperthyroidism as a proxy for undertreatment on cardiovascular events.

Method:

This was a case-control study nested within a population-based cohort of individuals attending health services in Funen County, Denmark, in the period from 1995 to 2011. Data on comorbidities and mortality were collected from The Danish National Patient Register and The Danish Register of Causes of Death. Participants were 275,467 individuals with at least one serum thyrotropin (TSH) measurement in the study period. Hyperthyroidism was defined as at least two measurements of decreased serum TSH within six months, separated by at least 14 days. Incident cases of cardiovascular disease (myocardial infarction, atrial fibrillation, heart failure, stroke, and cardiovascular death) were matched with controls. Conditional logistic regression analyses were performed to calculate odds ratios (OR) for exposure to hyperthyroidism, adjusting for preexisting comorbidities.

Results:

A total of 20,651 individuals experienced a cardiovascular event (9.5% incidence rate 13.2/1000 person-years [confidence interval (CI) 13.0–13.4]) compared to euthyroid individuals, conditional logistic regression showed increased cardiovascular risk in untreated hyperthyroid patients (OR = 1.25 [CI 1.06–1.48], p = 0.007) but not in treated hyperthyroid patients (OR = 1.04 [CI 0.90–1.22], p = 0.57)]. The OR for cardiovascular events per six months of decreased TSH was 1.09 ([CI 1.05–1.14], p < 0.001) in treated hyperthyroid individuals, and 1.10 ([CI 1.05–1.15], p < 0.001) in untreated hyperthyroid individuals.

Conclusions:

The risk of cardiovascular disease was found to be increased in untreated hyperthyroid patients, and the duration of decreased TSH associated with increasing risk of cardiovascular outcomes in both treated and untreated hyperthyroid individuals. This suggests that increased cardiovascular risk is driven not only by lack of treatment but also by insufficient therapy. The results support timely treatment and careful monitoring of hyperthyroid patients in order to reduce cardiovascular risk.

Introduction

Hyperthyroidism, whether due to Graves' disease (GD) (1) or multinodular goiter (MNG) (2), is a common endocrine condition that has been associated with psychiatric disease (3), somatic morbidity (4), as well as an increased risk of disability pension and loss of labor market income (5). Using three different approaches (background population, biochemical data, and disease-discordant twins), it has been shown that all-cause mortality is increased in hyperthyroid patients (6,7), in both GD and MNG (8). Due to the numerous effects of the thyroid hormones thyroxine (T4) and triiodothyronine (T3) on the musculature and conduction system of the heart, as well as vascular musculature and coagulation (9,10), cardiovascular death remains the most prevalent and investigated cause of death in hyperthyroidism. Despite this, uncertainty remains as to whether cardiovascular disease (CVD) is increased in hyperthyroid individuals. While previous studies of hyperthyroid individuals have found an increased risk of CVDs such as atrial fibrillation (AF) (11), myocardial infarction (MI) (12), heart failure (HF) (13) and cardiovascular death (14), others have not (15,16). Interpretation of these studies is hampered by inconsistent definitions of CVD and failure to take into account treatment of hyperthyroidism with regard to development of CVD. Few studies have found improvement of cardiovascular endpoints after treatment (17 –20), but they too are inconsistent with regard to sample size, study design, statistical methods, use of confounder control, and sex and age of the study population. Lastly, the aforementioned studies do not take into account the effect of varying thyroid status. While several studies have shown an association between duration of hyperthyroidism and increased osteoporotic fractures and all-cause mortality (7,21,22), a potential association between duration of hyperthyroidism and development of CVD remains unclear. Using real-world data from a large cohort of individuals with thyrotropin (TSH) measurements, the present study aimed to investigate the association between both treated and untreated biochemically verified hyperthyroidism and cardiovascular risk. Because of the large Danish medical registers, it was possible to provide confounder control in the form of comprehensive adjustment for comorbidities, age, and sex. Additionally, the relationship between duration of hyperthyroidism and cardiovascular risk was investigated.

Methods

Study design

This was a case-control study nested within a population-based registry cohort in order to determine cardiovascular risk as a function of biochemically verified hyperthyroidism.

Data sources

The Odense Patient data Explorative Network (OPEN) contains biochemical data from all blood samples taken on the island of Funen (7,22). The data were linked to the Danish National Patient Registry (DNPR), containing diagnoses based on hospital admissions (according to the International Classification of Diseases [ICD]-8 since 1977 and ICD-10 since 1995), as well as outpatient visits since January 1, 1995 (23 –25). The data were also linked to the Danish National Prescription Registry (DNPrR), which contains data on all prescriptions filled since 1995 (26), and the Danish Register of Causes of Death, which contains data on mortality (27).

Study population

The cohort comprised 275,467 individuals on the island of Funen with at least one measurement of TSH in the period January 1, 1995, to January 1, 2011, as registered in OPEN. After excluding individuals based on a one-year washout period (n = 5644), prior use of antithyroid drugs (ATD; n = 8819), pituitary disease (n = 753), age <18 years (n = 20,260), loss to follow-up or emigration (n = 223), hypothyroidism (n = 2414), and previous cardiovascular disease (n = 18,034), the cohort was narrowed down to 219,320 eligible individuals with at least one TSH measurement.

A case-control study was nested within this cohort in order to examine the risk of development of CVD. All those who developed CVD during the study period were matched to three controls according to sex and age at CVD using incidence density sampling.

Outcomes

Outcomes were incident CVD events registered as diagnoses in the DNPR. CVD was defined as a composite of myocardial infarction (defined by the ICD-8 code 410, and by ICD-10 codes I21–I22), stroke (defined by the ICD-8 codes 430–434, as well as ICD-10 codes I60–I64), atrial fibrillation (defined by the ICD-8 code 427, and the ICD-10 code I48), heart failure (defined by the ICD-8 code 428, and the ICD-10 code I50), and cardiovascular death (defined by the ICD-8 codes 390–458 and ICD-10 codes I00–I99), as registered in the Danish Register of Causes of Death. All outcomes have been used in recent studies of CVD (11,14,28). For each individual, the first date of possible registration was identified. In cases of multiple events, the first registered event was used.

Exposure

Exposure was treated (see operational definition below) or untreated hyperthyroidism, defined as having a minimum of two decreased TSH values within a period of six months from the initial TSH measurement with at least 14 days between measurements, irrespective of subsequent TSH status. As a result, a person with a TSH of <0.3 mIU/L at the first measurement was considered hyperthyroid if a following value within the first half year, at least 14 days from the first TSH, was <0.3 mIU/L. Conversely, an individual with an initial TSH <0.3 mIU/L with either no further measurement of TSH or a TSH >0.3 mIU/L was considered euthyroid.

Treatment was defined as having redeemed at least one prescription of ATD, as registered in the DNPrR, or having had thyroid surgery or radioactive iodine (RAI) treatment, as registered in the DNPR. In the case of a participant having received both ATD and surgery/RAI, treatment date was defined as whichever treatment was initiated first.

Euthyroidism was defined as serum TSH between 0.3 and 4.0 mIU/L with no periods of hyperthyroidism in the period from initial TSH measurement until an outcome, or end of study. All TSH determinations were performed in the same laboratory. Methods of measurement have been described in detail elsewhere (7,22,29).

Due to the minimum of 14 days between TSH measurements, the cohort data were left-truncated by 14 days. Individuals who died, developed CVD, or reached the end of the study (November 30, 2012) prior to this date were excluded from the cohort (n = 2153). This left 217,167 participants. Because being hyperthyroid (exposed) and euthyroid (unexposed) was defined by the first ever measured TSH and the nature of repeat blood measurements, it was possible for euthyroid individuals to become hyper- and hypothyroid during follow-up. The study aimed to compare only persistent hyperthyroid individuals and individuals who were never hyperthyroid. Thus, 3026 unexposed individuals became hyperthyroid during follow-up, while 3492 became hypothyroid. Since these individuals were unexposed at cohort entry, they were censored and exited the study at the time of development of thyroid dysfunction.

Confounders

The Charlson Comorbidity Index (CCI) (30) was calculated at entry into the cohort using the DNPR, excluding the diagnoses of CVD. Hyperlipidemia, hypertension, and diabetes were ascertained using the DNPR and DNPrR (Supplementary Table S1). With regard to the nested case-control study, CCI and comorbidities were ascertained at time of CVD for cases and the corresponding time for matched controls.

Statistical analysis

Baseline characteristics of the cohort were those present at the date of first TSH measurement, while the characteristics of the nested case-control analyses were defined at the time of cardiovascular event in order to adhere to case-control principles. Baseline characteristics of the cohort were analyzed using the chi-square test, Mann–Whitney U-test, and t-test, where applicable. Incidence rates and person-years were calculated from the entire cohort.

For the subsequent case-control study, CVD cases were matched 1:3 by age, sex, and year of birth (±1 year). Matching was done using the sttocc command in Stata. This command calculates age at time of CVD for cases and matches them with controls who at that same time had not experienced a cardiovascular event, thus applying incidence density sampling.

After matching, case-control analyses were performed using conditional logistic regression, with results shown as crude and adjusted odds ratios (ORs) with confidence intervals (CI) and adjusting for comorbidities using the CCI, hyperlipidemia, hypertension, and diabetes. Additional regression analyses were performed stratifying for sex and age <65 years and >65 years at the time of first TSH measurement.

As thyroid status may vary over time, a dynamic model using six-month periods of decreased TSH was applied in the conditional logistic regression in order to analyze thyroid status as a dichotomous time-varying covariate, allowing the association between duration of hyperthyroidism and cardiovascular risk to be investigated. Periods during which no TSH determination had been performed were considered periods of normal TSH.

Sensitivity analyses were performed in order to adjust for the possible influence of postpartum thyroiditis and pregnancy on TSH levels by excluding patients with a hospital contact regarding pregnancy or a diagnosis of postpartum thyroiditis as registered in the DNPR in a nine-month interval leading up to or after the initial measurement of TSH. Additionally, analyses with hyperthyroid individuals defined by having at least 6 or 12 weeks between initial TSH measurements were performed in order to exclude instances of non-thyroidal illness and to evaluate the strength of the case definition. Furthermore, a stratified analysis was performed subdividing hyperthyroid individuals into subclinical (TSH <0.3 mIU/L, T4 < 135.0 nmol/L, and T3 < 2.2 nmol/L) and overt hyperthyroidism (TSH <0.3 mIU/L, and T4 > 135.0 nmol/L, and/or T3 > 2.2 nmol/L). Since the euthyroid group also contained individuals with one initially decreased TSH who did not fulfil the hyperthyroid criteria but could de facto have been hyperthyroid, analyses excluding these individuals from the euthyroid population as well as including them as hyperthyroid individuals were performed. Lastly, regression analysis including the individuals who were previously censored due to development of thyroid dysfunction during the study period, hereby using the date of hyperthyroidism as the new index date, was performed in order to investigate whether the censoring itself led to a selection bias.

Comparisons between study groups were performed using a critical significance level of 5% and two-tailed tests throughout.

Statistical analyses were performed using Stata v14.1 (Stata Corp., Inc., College Station, TX) and SAS v9.4 (SAS Institute, Inc., Cary, NC) through virtual private network access to Statistics Denmark.

Ethical considerations

The data used in the study were anonymized, and the identity of all patients remained unknown to the investigators. The project was approved by the Danish Data Protection Agency. OPEN is an approved research institution permitted to access data hosted by Statistics Denmark (Project 704047).

Results

Cohort description

The 217,167 individuals were followed for the development of CVD for a median of 7.0 years (range 0–16.9 years), constituting a total of 1,561,379 person-years. During this time, 20,651 individuals developed CVD (9.5%; incidence rate 13.2/1000 person-years [CI 13.0–13.4]). Treated hyperthyroid individuals had a significantly higher median number of TSH measurements per year (2.2; interquartile range [IQR] 1.3–3.5) than did the untreated individuals (1.3; IQR 0.7–2.1; p < 0.001). At the first measurement, median TSH was significantly lower and T4 and T3 were significantly higher in the treated individuals than in the untreated individuals (Table 1).

Table 1.

Baseline Characteristics of the Cohort

	Euthyroid (n = 214,672)	Hyperthyroid treated (n = 1432)	Hyperthyroid untreated (n = 1063)
M _age (SD)	49.6 (17.8)	58.1 (17.8)	61.7 (16.8)	ET: p < 0.001 EU: p < 0.001 TU: p < 0.001
Females	124,916 (57.5%)	1179 (82.3%)	800 (75.3%)	ET: p < 0.001 EU: p < 0.001 TU: p < 0.001
Males	92,251 (42.5%)	253 (17.7%)	263 (24.7%)	ET: p < 0.001 EU: p < 0.001 TU: p < 0.001
CCI
0	182,610 (85.1%)	1230 (85.9%)	834 (78.5%)	ET: p = 0.4 EU: p < 0.001 TU: p < 0.001
1	23,760 (11.1%)	149 (10.4%)	168 (15.8%)
2	4728 (2.2%)	36 (2.5%)	41 (3.9%)
≥3	3574 (1.7%)	17 (1.2%)	20 (1.9%)
Hyperlipidemia	8685 (4.1%)	30 (2.1%)	32 (3.0%)	ET: p = 0.4 EU: p = 0.1 TU: p = 0.2
Hypertension	65,265 (30.4%)	612 (42.7%)	496 (46.7%)	ET: p < 0.001 EU: p < 0.001 TU: p = 0.05
Diabetes	8943 (4.2%)	43 (3.0%)	58 (5.5%)	ET: p = 0.03 EU: p < 0.001 TU: p < 0.001
Median follow-up, years (range)	7.0 (0–16.9)	9.0 (0–16.9)	7.2 (0–16.7)	ET: p < 0.001 EU: p = 0.2 TU: p < 0.001
Median TSH measurements per year (IQR)	0.4 (0.2–0.6)	2.2 (1.3–3.5)	1.3 (0.7–2.1)	ET: p < 0.001 EU: p < 0.001 TU: p < 0.001
Median TSH, mIU/L (IQR)	—	0.01 (0–0.03)	0.1 (0.01-0.18)	TU: p < 0.001
Median T4, nM (IQR)	—	n = 1296 174 (71–456)	n = 977 128 (60–286)	TU: p < 0.001
Median T3, nM (IQR)	—	n = 1000 3.03 (1.33–17.4)	n = 852 2.21 (0.82–8.97)	TU: p < 0.001

Apart from follow-up time and TSH measurements per year, measures are reported at time of the initial TSH measurement. Percentage decimals in excess due to rounding. Median TSH, T4, and T3 for the euthyroid group left out.

SD, standard deviation; ET, euthyroid vs. treated; EU, euthyroid vs. untreated; TU, treated vs. untreated; CCI, Charlson Comorbidity Index; TSH, thyrotropin; IQR, interquartile range; T4, thyroxine; T3, triiodothyronine.

Nested case-control study of cardiovascular diseases

Table 2 shows the characteristics of the 20,651 individuals who developed CVD and their 61,145 age, sex, and year-of-birth matched (1:3) controls who did not experience the outcome of interest. In brief, those who developed CVD had a significantly higher CCI and prevalence of hyperlipidemia, hypertension, and diabetes than their matched controls. Of the 20,651 cases, 126 (0.6%) were unable to be matched to any controls and were excluded from the regression analysis.

Table 2.

Characteristics and Comorbidities of Cardiovascular Cases and their Respective Matched Controls at the Time of Event

	Cases (n = 20,651)	Controls (n = 61,145)	p-Values
Mean age (SD)	66.9 (14.7)	66.9 (14.7)	—
Females	10,306 (49.9%)	30,536 (49.9%)	—
CCI
0	12,014 (58.2%)	44,002 (72.0%)	<0.001
1	4992 (24.2%)	11,842 (19.4%)
2	1686 (8.2%)	2841 (4.7%)
≥3	1959 (9.5%)	2460 (4.0%)
Hyperlipidemia	4819 (23.3%)	10,701 (17.5%)	<0.001
Hypertension	16,955 (82.1%)	40,732 (66.6%)	<0.001
Diabetes	2865 (13.9%)	5939 (9.7%)	<0.001

Differences tested by chi-square test and Mann–Whitney U-test. Percentage decimal in excess due to rounding. p-Values for comparison of matched parameters left out.

Conditional logistic regression (Table 3) showed an increased risk of CVD in untreated hyperthyroid patients compared to euthyroid individuals (OR = 1.21 [CI 1.03–1.42], p = 0.019), a result that remained significant when adjusting for CCI at time of CVD (OR = 1.25 [CI 1.06–1.48], p = 0.007). This was not the case in the treated hyperthyroid individuals, whether unadjusted (OR = 1.04 [CI 0.89–1.20], p = 0.64) or adjusted for comorbidities (OR = 1.04 [CI 0.90–1.22], p = 0.57). Results essentially remained unchanged when including hyperlipidemia, hypertension, and diabetes in the analysis (Table 3).

Table 3.

Nested Case-Control Analysis Using Conditional Logistic Regression of Exposure to Treated and Untreated Biochemical Hyperthyroidism as Predictor of Cardiovascular Risk

Statistical model and OR	Treated hyperthyroidism (n = 581)	Untreated hyperthyroidism (n = 470)
OR_CVD [CI]	1.04 [0.89–1.20], p = 0.64	1.21 [1.03–1.42], p = 0.019
OR_CVD [CI], CCI	1.04 [0.90–1.22], p = 0.57	1.25 [1.06–1.48], p = 0.007
OR_CVD [CI], HL + HT + DM	1.00 [0.86–1.17], p = 0.95	1.21 [1.03–1.44], p = 0.02
OR_CVD [CI], both	1.02 [0.87–1.19], p = 0.82	1.25 [1.06–1.48], p = 0.007
OR_CVD [CI], per six months of hyperthyroidism	1.09 [1.05–1.14], p < 0.001	1.10 [1.05–1.15], p < 0.001

Adjusted for CCI, hyperlipidemia, hypertension, and diabetes at point of cardiovascular event in order to adhere to case-control principles.

OR, odds ratio; CI, confidence interval; HL, hyperlipidemia; HT, hypertension; DM, diabetes mellitus.

Taking into account cumulative periods of decreased TSH, the OR for CVD was 1.09 ([CI 1.05–1.14], p < 0.001) per six months of decreased TSH in treated hyperthyroid individuals, while it was 1.10 ([CI 1.05–1.15], p < 0.001) in the untreated individuals. These results correspond to a 137% and 159% increased risk of CVD per five years of hyperthyroidism in the treated and the untreated individuals, respectively. While the treated hyperthyroid individuals had a significantly higher number of TSH measurements per year than the untreated individuals, the median percentage of all TSH measurements that decreased was significantly lower in the treated individuals (47%; IQR 30–73%) than in the untreated individuals (75%; IQR 44–100%; p < 0.001).

Untreated female hyperthyroid patients had an increased risk of developing CVD compared to euthyroid females, while this was not the case for women in the treated group. In the male group, risk estimates were similar to euthyroid individuals for both the treated and the untreated patients, possibly due to the small number of hyperthyroid individuals in this group (Table 4).

Table 4.

Nested Case-Control Analysis of Exposure to Treated and Untreated Biochemical Hyperthyroidism as Predictor of Cardiovascular Risk

Stratification by group	Treated hyperthyroidism	Untreated hyperthyroidism
OR_CVD [CI], males	n = 47, 1.33 [0.96–1.86], p = 0.09	n = 51, 1.00 [0.74–1.35], p > 0.99
OR_CVD [CI], females	n = 141, 0.98 [0.83–1.17], p = 0.84	n = 122, 1.39 [1.15–1.70], p = 0.001
OR_CVD [CI], ≥65 years	n = 138, 1.08 [0.90–1.29], p = 0.40	n = 144, 1.33 [1.10–1.60], p = 0.003
OR_CVD [CI], <65 years	n = 50, 0.95 [0.68–1.33], p = 0.78	n = 36, 0.95 [0.65–1.40], p = 0.81
Sub-outcomes
OR_MI [CI]	n = 29, 0.80 [0.56–1.20], p = 0.28	n = 20, 0.87 [0.53–0.41], p = 0.57
OR_HF [CI]	n = 31, 1.16 [0.77–1.75], p = 0.49	n = 32, 1.26 [0.84–1.88], p = 0.27
OR_AF [CI]	n = 62, 1.17 [0.88–1.56], p = 0.28	n = 67, 1.45 [1.08–1.95], p = 0.02
OR_stroke [CI]	n = 54, 0.91 [0.68–1.21], p = 0.50	n = 40, 0.83 [0.59–1.15], p = 0.25
OR_{CV death} [CI]	n = 35, 0.90 [0.63–1.28], p = 0.55	n = 37, 1.17 [0.82–1.67], p = 0.38

n, number of hyperthyroid individuals in each stratum; MI, myocardial infarction; HF, heart failure; AF, atrial fibrillation; CV death, cardiovascular death.

In the group aged >65 years, the untreated hyperthyroid individuals had an increased risk of developing CVD compared to euthyroid individuals, while the treated patients did not. In the group <65 years, there was no significant difference in risk of CVD between hyperthyroid patients and euthyroid individuals, irrespective of treatment status (Table 4).

When subdividing the composite cardiovascular score into a single outcome, a significantly increased risk of AF was found in untreated hyperthyroidism, while risk of HF and cardiovascular death was increased but not significantly (Table 4).

Sensitivity analyses

When excluding individuals with pregnancy or postpartum thyroiditis, the results remained effectively unchanged (data not shown). In order to evaluate the strength of the definition of hyperthyroidism used in this study and to limit the influence of nonthyroidal illness, regression analyses were performed the with hyperthyroidism defined with the two initial TSH measurements at least 6 weeks apart and also 12 weeks apart. Any study participant who, in the respective period of time, developed CVD was also excluded. This did not alter the results of the original analysis (data not shown). Stratified analysis according to subclinically and overtly hyperthyroid individuals yielded similar results as the original analysis, albeit the increased cardiovascular risk in untreated subclinically hyperthyroid individuals was not statistically significant. This result may be due to limited power, as T4 and T3 measurements were not available for all study participants. The exclusion of individuals whose initial TSH measurement was <0.3 mIU/L but did not fulfil the criteria for hyperthyroidism did not significantly affect the results of the main analysis (treated individuals: OR = 1.05 [CI 0.90–1.23], p = 0.5; untreated individuals: OR = 1.25 [CI 1.06–1.47], p = 0.01), neither did considering them hyperthyroid (treated individuals: OR = 0.83 [CI 0.76–0.92], p < 0.001; untreated individuals: OR = 1.15 [CI 1.07–1.24], p < 0.001). Finally, considering the euthyroid individuals who became hyperthyroid during the study as hyperthyroid from this point onward and including them in the study did not change that untreated hyperthyroid individuals had an increased risk of CVD compared to euthyroid individuals (OR = 1.28 [CI 1.15–1.43], p < 0.001), while this was not the case in the treated hyperthyroid individuals (OR = 1.03 [CI 0.97–1.15], p = 0.6).

Discussion

Principal findings

This large case-control study, based on real-world data and nested within a cohort, explored the association between treated and untreated biochemical hyperthyroidism and cardiovascular outcomes. A 25% increased risk of CVD in untreated hyperthyroid individuals compared to age- and sex-matched unexposed individuals was demonstrated, while the risk in treated hyperthyroid individuals was similar to that of euthyroid individuals. Cardiovascular risk was increased per six months of decreased TSH in both treated and untreated hyperthyroid individuals. Stratifying according to age, cardiovascular risk remained increased in untreated hyperthyroid individuals aged ≥65 years but did not differ from euthyroid individuals aged <65 years Though this is not in itself evidence of a causal link, these results point to a lower risk of CVD events in patients who became euthyroid and the benefit of early and effective treatment. Results were not significantly affected by adjustment for key cardiovascular factors such as diabetes, hypertension, and hyperlipidemia.

Comparison with other studies

Consistent with this study, some studies (4,12,14,31 –34), including meta-analyses (13,35), have found an association between hyperthyroidism and CVD, but not all have done so (15,16,36). Possible reasons for this discrepancy may include considerable heterogeneity between studies as regards the definition of hyperthyroidism, cardiovascular outcomes explored, study design, follow-up, and whether the effect of treatment was taken into account. Notably, some studies lacked sufficient control for comorbidities (15,16,34,35). The current study used the CCI, as well as the presence of hyperlipidemia, hypertension, and diabetes, to control for diseases possibly confounding an association between hyperthyroidism and CVD, allowing for analysis of a more direct association between hyperthyroidism and CVD. Using this validated method and further adjusting for comorbidities at time of development of CVD, an increased risk of development of CVD was found in the untreated hyperthyroid individuals.

A composite score consisting of major adverse cardiovascular events (MI, stroke, cardiovascular death), HF, and AF was defined, since these are the most widely investigated specific outcomes with regard to hyperthyroidism. Due to lack of power, subdividing according to these outcomes added little to the study. However, there was a significantly increased risk of AF, while a trend toward higher risk of HF and CV death was found in accordance with that demonstrated in previous studies (12,14,31 –33).

One limitation when comparing the results of previous studies is the differences in their operational definitions of hyperthyroidism. Ideally, thyroid state should be verified biochemically for the purpose of the operational definition and if possible with repeated measurements. Some studies use diagnosis codes from hospital contacts (4,12,14), which will likely capture moderate and severe cases only, failing to capture cases managed in general practice. Others have used a single measurement of TSH (15,16,31). A single low TSH is not indicative of a longer-lasting hyperthyroid state, as spontaneous recovery is common in thyroiditis and nonthyroidal illness. Accepting that 50% of individuals normalize their TSH on repeat measurements (37), this inflicts potential bias if diagnoses are based on one TSH determination. Access was available in this study to blood samples from both general practitioners and hospitals, thus providing data from all individuals who had their TSH measured. Defining the hyperthyroid cases by at least two measurements of decreased TSH, possible confounding by indication and misclassification of hyperthyroid individuals was minimized. Hereby, the likelihood of nonthyroidal illness and/or transient thyrotoxicosis as the cause of decreased TSH was reduced. On the other hand, the inclusion of individuals with one initially decreased TSH in the euthyroid group is likely to confound the results. However, both inclusion and exclusion of these individuals in the sensitivity analysis shows that the findings are very robust to changes in the definition of euthyroidism.

Although it is known that cardiovascular risk increases with age, not all studies performed age-stratified analyses (4,15,16). In line with other studies demonstrating excess cardiovascular risk only in elderly hyperthyroid patients (31,32), this study demonstrates an increased cardiovascular risk in treated patients aged ≥65 years. Contrary to this, however, several studies have reported increased cardiovascular morbidity, even in younger age groups (12,14). This disparity may possibly be explained by the limitations mentioned earlier. In the present study, only 86 patients with cardiovascular events had hyperthyroidism before the age of 65 years. Thus, failure to recognize increased cardiovascular risk in hyperthyroid patients most likely pertains to lack of statistical power.

While a limited number of studies have found improvement of existing cardiovascular symptoms such as increased heart rate (17), arrhythmias (18), pulmonary hypertension (19,20), and echocardiographic changes (38,39), only few have treated hyperthyroid individuals and evaluated their subsequent risk of experiencing cardiovascular events (14,32,33,36). Studies have demonstrated excess cardiovascular morbidity in MNG patients treated with RAI or thyroid surgery (14,32,33). However, none took treatment with ATD into account. Because of the small number of exposed individuals treated with surgery or RAI as first-line treatment in the current study (n = 30), statistical sub-analyses were without meaning. The results are in line with Flynn et al. (36), reporting that cardiovascular disease, apart from arrhythmias, was not increased in treated hyperthyroid patients. Finally, and most importantly, with the exception of the current study, none of the aforementioned studies took varying thyroid status into consideration.

Strengths and limitations

To the authors' knowledge, this is the first study to evaluate the association between varying thyroid function and cardiovascular risk. The study demonstrates that cardiovascular risk was increased per period of decreased TSH (a proxy for undertreatment) in treated as well as untreated hyperthyroid individuals. Coupled with the fact that treated individuals had significantly fewer measurements of decreased TSH than the untreated individuals, this should serve as a reminder that initiating treatment is not necessarily sufficient unless supported by follow-up to ensure continuous euthyroidism. Any six-month period with no TSH measurement was considered as a euthyroid period. This implies that a hyperthyroid individual may have been considered euthyroid during a period of disease, which would lead to an underestimation rather than inflation of cardiovascular risk. Also, an attempt to alleviate this limitation by carrying the last observed TSH value forward for the remainder of the study period may only serve to dilute any association.

Unfortunately, using the available databases, it was not possible to determine whether hyperthyroidism was due to GD or MNG. Giesecke et al. (14) found increased cardiovascular morbidity in MNG but not in GD, possibly explained by differences in age distribution and burden of comorbidity. In contrast to this, increased cardiovascular mortality was previously found in GD but not in MNG (8).

Adjusting for smoking and alcohol consumption, both of which are associated with thyroid and cardiovascular disease (40,41), was not possible. Instead, these were partly controlled for through use of the CCI, which includes chronic pulmonary disease as a proxy for smoking and liver disease as a proxy for alcohol intake. Lacking information on the reason for initiating or withholding therapy, the data, based on initial TSH, T4, and T3 measurements, suggest that untreated hyperthyroid individuals having milder hyperthyroidism than treated individuals is one possible explanation for the high number of untreated hyperthyroid individuals. Also, hyperthyroid patients have been shown not to adhere to treatment with ATD in 31.3 % of cases (42).

As mortality has been shown to be increased in untreated subclinically hyperthyroid individuals (7), it would be appropriate to investigate whether this is also the case for cardiovascular risk. Although stratified analysis found increased cardiovascular risk in both untreated subclinical and overt hyperthyroidism, the result was statistically insignificant in untreated subclinically hyperthyroid individuals. While this result most likely pertains to lack of power and is a limitation of the present study, it is in line with previous studies (31,35).

Conclusions

The data presented here demonstrate excess cardiovascular risk in untreated but not in treated biochemically verified hyperthyroid patients. Importantly, cumulative periods of decreased TSH increased cardiovascular morbidity in both treated and untreated hyperthyroid individuals. Thus, hyperthyroidism is associated with increased cardiovascular risk, warranting careful monitoring and maintaining euthyroidism. However, although unlikely ever to be carried out, definitive proof of treatment effects awaits a randomized clinical trial investigating the effect of treatment of hyperthyroidism on cardiovascular risk.

Footnotes

Acknowledgments

Dr. Mads Nybo, Department of Clinical Biochemistry, Odense University Hospital, is acknowledged for his role in the initial data collection.

Author Disclosure Statement

B.A. has institutional research contracts with UCB and Novartis outside the current study. M.L.J. is enrolled as a PhD student at the University of Southern Denmark, and is financed by the University of Southern Denmark, and Odense University Hospital. M.L.J. has further received research grants from Inger Goldmanns Fond, The Danish Thyroid Patient Society, Odense University Hospital Research Fund, The Agnes & Knut Mørk Foundation, and Overlægerådets Forskningsfond. The remaining authors declare no conflicts of interest. All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf

Supplementary Material

Supplementary Table S1

References

Smith

, Hegedus

. 2016. Graves' disease. N Engl J Med, 375:1552–1565.

Hegedus

, Bonnema

, Bennedbaek

. 2003. Management of simple nodular goiter: current status and future perspectives. Endocr Rev, 24:102–132.

Brandt

, Thvilum

, Almind

, Christensen

, Green

, Hegedus

, Brix

. 2014. Hyperthyroidism and psychiatric morbidity: evidence from a Danish nationwide register study. Eur J Endocrinol, 170:341–348.

Brandt

, Thvilum

, Almind

, Christensen

, Green

, Hegedus

, Brix

. 2013. Morbidity before and after the diagnosis of hyperthyroidism: a nationwide register-based study. PLoS One, 8:e66711.

Brandt

, Thvilum

, Hegedus

, Brix

. 2015. Hyperthyroidism is associated with work disability and loss of labour market income. A Danish register-based study in singletons and disease-discordant twin pairs. Eur J Endocrinol, 173:595–602.

Brandt

, Almind

, Christensen

, Green

, Brix

, Hegedus

. 2012. Excess mortality in hyperthyroidism: the influence of preexisting comorbidity and genetic confounding: a Danish nationwide register-based cohort study of twins and singletons. J Clin Endocrinol Metab, 97:4123–4129.

Lillevang-Johansen

, Abrahamsen

, Jorgensen

, Brix

, Hegedus

. 2017. Excess mortality in treated and untreated hyperthyroidism is related to cumulative periods of low serum TSH. J Clin Endocrinol Metab, 102:2301–2309.

Brandt

, Thvilum

, Almind

, Christensen

, Green

, Hegedus

, Brix

. 2013. Graves' disease and toxic nodular goiter are both associated with increased mortality but differ with respect to the cause of death: a Danish population-based register study. Thyroid, 23:408–413.

Faber

, Selmer

. 2014. Cardiovascular disease and thyroid function. Front Horm Res, 43:45–56.

10.

Jabbar

, Pingitore

, Pearce

, Zaman

, Iervasi

, Razvi

. 2017. Thyroid hormones and cardiovascular disease. Nat Rev Cardiol, 14:39–55.

11.

Selmer

, Olesen

, Hansen

, Lindhardsen

, Olsen

, Madsen

, Faber

, Hansen

, Pedersen

, Torp-Pedersen

, Gislason

. 2012. The spectrum of thyroid disease and risk of new onset atrial fibrillation: a large population cohort study. BMJ, 345:e7895.

12.

Dekkers

, Horvath-Puho

, Cannegieter

, Vandenbroucke

, Sorensen

, Jorgensen

. 2017. Acute cardiovascular events and all-cause mortality in patients with hyperthyroidism: a population-based cohort study. Eur J Endocrinol, 176:1–9.

13.

Gencer

, Collet

, Virgini

, Bauer

, Gussekloo

, Cappola

, Nanchen

, den Elzen

, Balmer

, Luben

, Iacoviello

, Triggiani

, Cornuz

, Newman

, Khaw

, Jukema

, Westendorp

, Vittinghoff

, Aujesky

, Rodondi N; Thyroid Studies

Collaboration

. 2012. Subclinical thyroid dysfunction and the risk of heart failure events: an individual participant data analysis from 6 prospective cohorts. Circulation, 126:1040–1049.

14.

Giesecke

, Rosenqvist

, Frykman

, Friberg

, Wallin

, Hoijer

, Lonn

, Torring

. 2017. Increased cardiovascular mortality and morbidity in patients treated for toxic nodular goiter compared to Graves' disease and nontoxic goiter. Thyroid, 27:878–885.

15.

Boekholdt

, Titan

, Wiersinga

, Chatterjee

, Basart

, Luben

, Wareham

, Khaw

. 2010. Initial thyroid status and cardiovascular risk factors: the EPIC-Norfolk prospective population study. Clin Endocrinol (Oxf), 72:404–410.

16.

Martin

, Daya

, Lutsey

, Matsushita

, Fretz

, McEvoy

, Blumenthal

, Coresh

, Greenland

, Kottgen

, Selvin

. 2017. Thyroid function, cardiovascular risk factors, and incident atherosclerotic cardiovascular disease: the Atherosclerosis Risk in Communities (ARIC) study. J Clin Endocrinol Metab, 102:3306–3315.

17.

Faber

, Wiinberg

, Schifter

, Mehlsen

. 2001. Haemodynamic changes following treatment of subclinical and overt hyperthyroidism. Eur J Endocrinol, 145:391–396.

18.

Osman

, Franklyn

, Holder

, Sheppard

, Gammage

. 2007. Cardiovascular manifestations of hyperthyroidism before and after antithyroid therapy: a matched case-control study. J Am Coll Cardiol, 49:71–81.

19.

Muthukumar

, Sadacharan

, Ravikumar

, Mohanapriya

, Hussain

, Suresh

. 2016. A prospective study on cardiovascular dysfunction in patients with hyperthyroidism and its reversal after surgical cure. World J Surg, 40:622–628.

20.

Six-Merker

, Meisinger

, Jourdan

, Heier

, Hauner

, Peters

, Linseisen

. 2016. Treatment of thyroid dysfunctions decreases the risk of cerebrovascular events in men but not in women: results of the MONICA/KORA cohort study. PLoS One, 11:e0155499.

21.

Abrahamsen

, Jorgensen

, Laulund

, Nybo

, Brix

, Hegedus

. 2014. Low serum thyrotropin level and duration of suppression as a predictor of major osteoporotic fractures—the OPENTHYRO register cohort. J Bone Miner Res, 29:2040–2050.

22.

Laulund

, Nybo

, Brix

, Abrahamsen

, Jorgensen

, Hegedus

. 2014. Duration of thyroid dysfunction correlates with all-cause mortality. the OPENTHYRO Register Cohort. PLoS One, 9:e110437.

23.

Thygesen

, Daasnes

, Thaulow

, Bronnum-Hansen

. 2011. Introduction to Danish (nationwide) registers on health and social issues: Structure, access, legislation, and archiving. Scand J Public Health, 39:12–16.

24.

Thygesen

, Ersboll

. 2011. Danish population-based registers for public health and health-related welfare research: introduction to the supplement. Scand J Public Health, 39:8–10.

25.

Schmidt

, Schmidt

, Sandegaard

, Ehrenstein

, Pedersen

, Sorensen

. 2015. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol, 7:449–490.

26.

Pottegard

, Schmidt

SAJ

, Wallach-Kildemoes

, Sorensen

, Hallas

, Schmidt

. 2017. Data resource profile: The Danish National Prescription Registry. Int J Epidemiol, 46:798–798f.

27.

Helweg-Larsen

. 2011. The Danish Register of Causes of Death. Scand J Public Health, 39:26–29.

28.

Egeberg

, Gislason

, Hansen

. 2016. Risk of major adverse cardiovascular events and all-cause mortality in patients with hidradenitis suppurativa. JAMA Dermatol, 152:429–434.

29.

Jensen

, Hyltoft Petersen

, Blaabjerg

, Hansen

, Brix

, Kyvik

, Hegedus

. 2004. Establishment of a serum thyroid stimulating hormone (TSH) reference interval in healthy adults. The importance of environmental factors, including thyroid antibodies. Clin Chem Lab Med, 42:824–832.

30.

Charlson ME

, Ales

, MacKenzie

. 1987. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis, 40:373–383.

31.

Selmer

, Olesen

, Hansen

, von Kappelgaard

, Madsen

, Hansen

, Pedersen

, Faber

, Torp-Pedersen

, Gislason

. 2014. Subclinical and overt thyroid dysfunction and risk of all-cause mortality and cardiovascular events: a large population study. J Clin Endocrinol Metab, 99:2372–2382.

32.

Metso

, Auvinen

, Salmi

, Huhtala

, Jaatinen

. 2008. Increased long-term cardiovascular morbidity among patients treated with radioactive iodine for hyperthyroidism. Clin Endocrinol (Oxf), 68:450–457.

33.

Ryodi

, Salmi

, Jaatinen

, Huhtala

, Saaristo

, Valimaki

, Auvinen

, Metso

. 2014. Cardiovascular morbidity and mortality in surgically treated hyperthyroidism—a nation-wide cohort study with a long-term follow-up. Clin Endocrinol (Oxf), 80:743–750.

34.

Nyirenda

, Clark

, Finlayson

, Read

, Elders

, Bain

, Fox

, Toft

. 2005. Thyroid disease and increased cardiovascular risk. Thyroid, 15:718–724.

35.

Collet

, Gussekloo

, Bauer

, den Elzen

, Cappola

, Balmer

, Iervasi

, Asvold

, Sgarbi

, Volzke

, Gencer

, Maciel

, Molinaro

, Bremner

, Luben

, Maisonneuve

, Cornuz

, Newman

, Khaw

, Westendorp

, Franklyn

, Vittinghoff

, Walsh

, Rodondi N; Thyroid Studies

Collaboration

. 2012. Subclinical hyperthyroidism and the risk of coronary heart disease and mortality. Arch Intern Med, 172:799–809.

36.

Flynn

, Macdonald

, Jung

, Morris

, Leese

. 2006. Mortality and vascular outcomes in patients treated for thyroid dysfunction. J Clin Endocrinol Metab, 91:2159–2164.

37.

Meyerovitch

, Rotman-Pikielny

, Sherf

, Battat

, Levy

, Surks

. 2007. Serum thyrotropin measurements in the community: five-year follow-up in a large network of primary care physicians. Arch Intern Med, 167:1533–1538.

38.

Sgarbi

, Villaca

, Garbeline

, Villar

, Romaldini

. 2003. The effects of early antithyroid therapy for endogenous subclinical hyperthyroidism in clinical and heart abnormalities. J Clin Endocrinol Metab, 88:1672–1677.

39.

Kaminski

, Michalkiewicz

, Makowski

, Podgajny

, Szalus

, Ruchala

, Szczepanek

, Gielerak

. 2011. Prospective echocardiographic evaluation of patients with endogenous subclinical hyperthyroidism and after restoring euthyroidism. Clin Endocrinol (Oxf), 74:501–507.

40.

Hegedus

, Rasmussen

, Ravn

, Kastrup

, Krogsgaard

, Aldershvile

. 1988. Independent effects of liver disease and chronic alcoholism on thyroid function and size: the possibility of a toxic effect of alcohol on the thyroid gland. Metabolism, 37:229–233.

41.

Brix

, Hansen

, Kyvik

, Hegedus

. 2000. Cigarette smoking and risk of clinically overt thyroid disease: a population-based twin case-control study. Arch Intern Med, 160:661–666.

42.

Tamblyn

, Eguale

, Huang

, Winslade

, Doran

. 2014. The incidence and determinants of primary nonadherence with prescribed medication in primary care: a cohort study. Ann Intern Med, 160:441–450.