Thyroid function and life expectancy with and without noncommunicable diseases: A population-based study

Abstract

Background

Variations in thyroid function within reference ranges are associated with increased risk of diseases and death. However, the impact of thyroid function on life expectancy (LE) with and without noncommunicable diseases (NCDs) remains unknown. We therefore aimed to investigate the association of thyroid function with total LE and LE with and without NCD among euthyroid individuals.

Methods and findings

The study was embedded in the Rotterdam Study, a prospective population-based study carried out in the Netherlands. In total, 7,644 participants without known thyroid disease and with thyroid-stimulating hormone (TSH) and free thyroxine (FT4) levels within reference ranges were eligible. NCDs were defined as presence of cardiovascular disease, diabetes mellitus type 2, or cancer. We used the demographic tool of multistate life tables to calculate LE estimates at the age of 50 years, using prevalence, incidence rates, and hazard ratios for three transitions (healthy to NCD, healthy to death, and NCD to death). The total LE and LE with and without NCD among TSH and FT4 tertiles were calculated separately in men and women. Analyses were adjusted for sociodemographic and cardiovascular risk factors. The mean (standard deviation) age of the participants was 64.5 (9.7) years, and 52.3% were women. Over a median follow-up of 8 years (interquartile range 2.7–9.9 years), 1,396 incident NCD events and 1,422 deaths occurred. Compared with those in the lowest TSH tertile, men and women in the highest TSH tertile were expected to live 1.5 years (95% confidence interval [CI] 0.8–2.3, p < 0.001) and 1.5 years (CI 0.8–2.2, p < 0.001) longer, respectively, of which 1.4 years (CI 0.5–2.3, p = 0.002) and 1.3 years (CI 0.3–2.1, p = 0.004) with NCD. Compared with those in the lowest FT4 tertile, the difference in LE for men and women in the highest FT4 tertile was −3.7 years (CI −5.1 to −2.2, p < 0.001) and −3.3 years (CI −4.7 to −1.9, p < 0.001), respectively, of which −1.8 years (CI −3.1 to −0.7, p = 0.003) and −2.0 years (CI −3.4 to −0.7, p = 0.003) without NCD. A limitation of the study is the observational design. Thus, the possibility of residual confounding cannot be entirely ruled out.

Conclusions

In this study, we found that people with low–normal thyroid function (i.e., highest tertile of TSH and lowest tertile of FT4 reference ranges) are expected to live more years with and without NCD than those with high–normal thyroid function (i.e., lowest tertile of TSH and highest tertile of FT4 reference ranges). These findings provide support for a re-evaluation of the current reference ranges of thyroid function.

About the Authors:

Arjola Bano

Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

Affiliations Department of Internal Medicine and Academic Center for Thyroid Diseases, Erasmus Medical Center, Rotterdam, the Netherlands, Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands, Institute of Social and Preventive Medicine (ISPM), University of Bern, Bern, Switzerland, Department of Cardiology, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland

ORCID logo http://orcid.org/0000-0003-0956-7145

Layal Chaker

Roles Conceptualization, Data curation, Investigation, Methodology, Project administration, Supervision, Visualization, Writing – review & editing

Affiliations Department of Internal Medicine and Academic Center for Thyroid Diseases, Erasmus Medical Center, Rotterdam, the Netherlands, Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands

Francesco U. S. Mattace-Raso

Roles Conceptualization, Data curation, Funding acquisition, Investigation, Project administration, Resources, Supervision, Visualization, Writing – review & editing

Affiliation: Section of Geriatric Medicine, Erasmus Medical Center, Rotterdam, the Netherlands

Natalie Terzikhan

Roles Data curation, Investigation, Methodology, Resources, Writing – review & editing

Affiliation: Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands

Maryam Kavousi

Roles Data curation, Investigation, Methodology, Resources, Validation, Writing – review & editing

Affiliation: Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands

M. Arfan Ikram

Roles Data curation, Funding acquisition, Investigation, Methodology, Resources, Software, Supervision, Writing – review & editing

Affiliation: Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands

ORCID logo http://orcid.org/0000-0003-0372-8585

Robin P. Peeters

Roles Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – review & editing

* E-mail: [email protected]

Affiliations Department of Internal Medicine and Academic Center for Thyroid Diseases, Erasmus Medical Center, Rotterdam, the Netherlands, Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands

Oscar H. Franco

Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – review & editing

Affiliations Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands, Institute of Social and Preventive Medicine (ISPM), University of Bern, Bern, Switzerland

Introduction

Noncommunicable diseases (NCDs) pose a global health threat, inflicting high disability rates and a huge economic burden [1,2]. According to the World Health Organization, NCDs account for approximately 41 million deaths worldwide each year [2,3]. Global preventive strategies can be beneficial for cardiovascular disease (CVD), diabetes, cancer, and chronic kidney disease (CKD), which account for large proportions of deaths and a large number of years lived with disability [4]. Among other factors, clinical and subclinical thyroid dysfunction have been associated with CVD, diabetes, cancer, and NCD mortality [5–9].

Thyroid dysfunction is defined based on the circulating levels of thyroid-stimulating hormone (TSH) and free thyroxine (FT4) [10]. The secretion of these hormones is regulated by the hypothalamus–pituitary–thyroid axis via a negative-feedback mechanism [10]. The reference ranges of TSH and FT4 levels are crucial for the diagnosis and treatment of thyroid disease. However, over the past years, there has been an ongoing debate on whether the reference ranges of TSH and FT4 levels should be re-evaluated [11,12]. The debate has arisen from studies indicating that not only thyroid disease but also variations in thyroid function within the reference ranges can contribute to the occurrence of chronic conditions and deaths [11,13–15]. Prospective investigations in euthyroid individuals have suggested that high–normal thyroid function increases the risk of CVD, cancer, or CKD, whereas low–normal thyroid function has been associated with an increased risk of chronic metabolic diseases, such as diabetes [11,14–19]. Based on this evidence, it is challenging to determine the balance of overall benefits and risks for specific cutoffs of TSH and FT4 levels within the reference ranges [20,21]. In order to capture the pleiotropic effects of thyroid hormones, there is a need for utilizing composite measurements that integrate multiple aspects of health and disease rather than focusing on one adverse outcome alone. A recent study utilizing life expectancy (LE) measurements showed that middle-aged and older adults with low–normal thyroid function have a longer total LE and a longer LE without CVD than those with high–normal thyroid function [22]. Still, it remains unclear whether the differences in LE without CVD reflect healthy years or years lived with other chronic diseases. To address this, multidimensional measures such as LE with and without NCD can provide useful information on the qualitative and quantitative impact of thyroid function on general health.

In a large prospective population-based cohort study, we investigated (1) the association of thyroid function within the reference range with the risk of incident NCD and (2) whether there are differences in the number of years lived with and without NCD, within the reference range of thyroid function. NCDs were defined as the presence of CVD, diabetes, or cancer, which have been highlighted as a global threat by the United Nations and have been prospectively associated with thyroid function [13,14,22,23]. CKD was added to the definition of NCD in a secondary analysis.

Methods

Study population

This study was embedded within the Rotterdam Study, a prospective population-based cohort study. The Rotterdam Study investigates the determinants, occurrence, and progression of chronic diseases in middle-aged and older adults from the Netherlands (overall response rate of the study, 72%). The objectives and design have been described in detail previously [24]. The study was initiated in 1989, including 7,983 participants aged 55 years or older. In 2000, the study was extended with a second cohort of 3,011 individuals. In 2006, a third cohort of 3,932 individuals aged 45 years or older was added. Study participants underwent extensive follow-up medical examinations every 3–5 years. Baseline measurements for our study were performed during the third visit of the first cohort (RS-I.3, 1997–1999, n = 4,797) and the first visits of the second (RS-II.1, 2000–2001, n = 3,011) and third (RS-III.1, 2006–2008, n = 3,932) cohorts of the Rotterdam Study. The original cohort during these three visits included a total of 11,740 participants, of which 10,050 had available blood measurements. Thyroid function measurements were performed in a random sample of 9,702 participants. Of these, we excluded 836 participants with past thyroid disease or who were taking thyroid medications, 182 participants without complete information on prevalent or incident NCD, and 1,040 participants with TSH or FT4 outside the normal reference ranges. The remaining 7,644 participants with data available on NCD, without known thyroid disease, and with TSH and FT4 levels within the reference ranges were eligible. The Rotterdam Study was approved by the Medical Ethics Committee of Erasmus University and by the Ministry of Health, Welfare and Sport of the Netherlands, implementing the Population Study Act Rotterdam Study. In accordance with the Declaration of Helsinki, all participants provided written informed consent. This study is reported per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 STROBE Checklist).

Assessment of thyroid parameters

Thyroid function was assessed at baseline (RS-I.3 1997–1999; RS-II.1 2000–2001, and RS-III.1 2006–2008) using the same method, assay, and reference ranges. Measurements of TSH and FT4 were performed in baseline serum samples stored at −80°C using the Roche electrochemiluminescence immunoassay (ECLIA). The reference ranges of TSH (0.4–4 mIU/L) and FT4 (0.86–1.94 ng/dL, alternatively 11–25 pmol/L) were determined based on national guidelines and our previous studies [18,19]. Participants and family physicians were not informed about the results of thyroid function measurements. The study investigators assessing the outcomes of interest were blinded to the thyroid status of participants. Thyroid peroxidase antibodies (TPOAbs) were assessed with the ECLIA for TPOAbs (Roche Diagnostics International, Rotkreuz, Switzerland). TPOAbs > 35 kU/mL were considered positive, as recommended by the assay manufacturer.

NCD, mortality, and additional measurements

Outcome measures were incident nonfatal NCD, mortality among those with NCD, and overall mortality. NCDs were defined as presence of at least one disease out of CVD, diabetes mellitus type 2, or cancer. All events were independently assessed by two research physicians and further validated by a medical specialist. Information on mortality was obtained from municipality records, general practitioners, and reports of medical specialists [24,25]. CVD was defined as presence of coronary heart disease (CHD), stroke, or heart failure. CHD was defined as coronary revascularization, myocardial infarction, or fatal CHD [25]. Based on the World Health Organization criteria, stroke was defined as a syndrome of rapidly developing symptoms with an apparent vascular cause of focal or global cerebral dysfunction lasting 24 hours or longer or leading to death [26]. Based on the European Society of Cardiology criteria, heart failure was defined as the presence of typical symptoms and signs—breathlessness at rest or during exertion, ankle edema, and pulmonary crepitations—confirmed by the objective evidence of cardiac dysfunction (i.e., chest X-ray, echocardiography) or a positive response to the initiated treatment [27]. Prevalent CVD was assessed at baseline through interview and medical records. After enrollment, participants were monitored for incident CVD through linkage of the study database with files from general practitioners and hospital records. Diabetes mellitus was defined as a fasting serum glucose level ≥ 7 mmol/L, a nonfasting plasma glucose level ≥ 11.1 mmol/L (when fasting samples were absent), or the use of blood glucose–lowering medications [28]. Cases of type 2 diabetes were ascertained at baseline and during follow-up through general practitioners, hospital discharge letters, and serum glucose measurements from the Rotterdam Study visits. Information regarding the use of blood glucose–lowering medications was derived from both structured home interviews and linkage to pharmacy records [28]. Cancer events were classified according to the International Classification of Diseases (ICD) 10th Edition [29]. Cases of cancer were determined through general practitioners and hospital discharge letters and by linkage with a nationwide registry of histopathology and cytopathology in the Netherlands, Pathologisch-Anatomisch Landelijk Geautomatiseerd Archief (PALGA) [29]. All types of cancer events were confirmed by pathology records. All cancer events were combined. The most common cancers in the Rotterdam Study are breast cancer, prostate cancer, pancreatic cancer, lung cancer, and colorectal cancer.

Additional measurements

The baseline home interview provided information on medical history, medication use, tobacco smoking, alcohol consumption, education level, and marital status [24]. Smoking habits were categorized as current, former, and never smoking. Education level was categorized as elementary, lower secondary, higher secondary, and tertiary. Marital status was categorized as single, married, widowed, and divorced/separated. Lipid levels were measured by an automated enzymatic procedure (Mannheim system). Anthropometric measurements were performed in the research center by trained medical staff. Body mass index was calculated as weight in kilograms divided by height in meters squared. Blood pressure was measured in the sitting position on the right arm and calculated as the mean of two measurements using a random-zero sphygmomanometer. In order to perform sensitivity analyses, we also evaluated CKD and chronic obstructive pulmonary disease (COPD). CKD was defined as estimated glomerular filtration rate (eGFR) < 60 ml/min per 1.73 m2. eGFR was calculated at baseline and follow-up visits according to the CDK epidemiology collaboration equation [30]. Patients with incident CKD were defined as individuals free of CKD at baseline (eGFR > 60 ml/min per 1.73 m2) who had a decline in eGFR to <60 ml/min per 1.73 m2 between the two periodic examinations [31]. To estimate the censoring date of the cases, we assumed a linear decrease in eGFR. The date that each case had passed the eGFR threshold of 60 ml/min per 1.73 m2 was taken as the censoring date, and it was used to calculate the follow-up time for incident cases. For participants who remained free from CKD, the time spent between the two examinations was used as the follow-up time. COPD was diagnosed based on an obstructive prebronchodilator spirometry (FEV1/FVC < 0.70) according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines [32]. The incident date of COPD was defined as the date of the first obstructive lung function examination, the date of COPD diagnosis in the medical records, or the date of the first prescription of COPD medications, whichever came first.

Statistical analyses

An outline of the analysis plan of this study is provided in S1 Text. We did not publish or preregister the analysis plan. The multistate life table is a demographic tool that can estimate the total LE and disease-specific life expectancies using information from people at different ages moving between different health states (e.g., healthy, diseased, dead) [33]. We used multistate lifetables to calculate estimates of total LE and LE with and without NCD among tertiles of TSH and FT4 [34]. Multistate life tables combined information from participants in three possible health states—namely, “free of NCD,” “NCD,” and “death.” Possible transitions of participants were (1) from free of NCD to NCD (incident NCD), (2) from free of NCD to death (mortality among those without NCD), and (3) from NCD to death (mortality among those with NCD). Backflows were not allowed, and only the first event into a state was considered [33]. The analyses were performed separately for TSH and FT4 tertiles. Because of the known gender differences in LE, analyses were performed separately among men and women. To calculate LE, we followed a similar approach to previous studies [34,35]. We first calculated the prevalence of TSH tertiles among participants with and without NCD, categorized in 10-year age groups. In each transition, we calculated age-specific incidence rates. Next, we applied Poisson regression with Gompertz distribution to compute hazard ratios (HRs) of the association between TSH tertiles and incident NCD or mortality, adjusting for potential confounders. Differences in LE were evaluated using the lowest TSH tertile as a reference. The 95% confidence intervals (CIs) of LE estimates were calculated using the Monte Carlo method with 10,000 bootstrap simulations [36]. We repeated the analyses for the FT4 tertiles. Differences in LE were evaluated using the lowest FT4 tertile as a reference. Confounders were selected based on biological plausibility and previous literature. Model 1 was adjusted for age, sex, and cohort. Model 2 was adjusted for age, sex, cohort, smoking, alcohol intake, education level, marital status, body mass index, systolic blood pressure, total cholesterol, triglycerides, use of antihypertensive medications, and use of lipid-lowering medications. Multiple imputations were performed in case of missing covariates (<5% for all covariates). Statistical analyses were conducted using IBM SPSS version 21 (IBM), STATA version 13 for Windows (StataCorp), and @RISK software (Palisade).

Sensitivity analyses

(1) To account for potential reverse causation, we excluded NCD events or deaths that occurred during the first 2 years of follow-up. (2) We excluded participants using thyroid function–altering medications, such as amiodarone or corticosteroids. (3) To detect a potential influence of follow-up duration on our results, we performed the analyses restricting the length of follow-up to 8 years (median follow-up time). (4) CKD was defined by eGFR, measured up to two times within our cohort, and was added to the definition of NCD in a sensitivity analysis. (5) COPD was not included in the primary outcome of NCD, because of the lack of evidence of a pathophysiological association between thyroid function and incident COPD. However, we performed a sensitivity analysis adding COPD to the definition of NCD. (6) We assessed the differences in LE with and without CVD, diabetes, and cancer, separately, among the TSH and FT4 tertiles. (7) We calculated unadjusted HRs for incident NCD and death among TSH and FT4 tertiles. (8) To account for the potential role of thyroid autoimmunity, we additionally adjusted our analyses for TPOAb positivity. Furthermore, we investigated the association of TPOAb positivity with incident NCD and NCD mortality among those with and without disease.

Results

A total of 7,644 participants with complete data available on NCD, without known thyroid disease, and with TSH and FT4 levels within the reference ranges were considered eligible. Baseline characteristics of eligible participants are presented in Table 1. The mean (standard deviation) age of participants was 64.5 (9.7) years, and 52.3% were women. Over a median follow-up time of 8 years (interquartile range 2.7–9.9 years), 1,396 incident NCD events and 1,422 deaths occurred. Both models yielded similar estimates; therefore, we further report the results of the most adjusted model (model 2), which accounted for age, sex, cohort, smoking, alcohol intake, education level, marital status, body mass index, systolic blood pressure, total cholesterol, triglycerides, use of antihypertensive medications, and use of lipid-lowering medications.

[Figure omitted. See PDF.]

Table 1. Baseline characteristics of 7,644 participants.

https://doi.org/10.1371/journal.pmed.1002957.t001

Association of thyroid function within the reference range with the risk of NCD and death

The association of TSH tertiles with the risk of incident NCD was not statistically significant (Table 2). Among participants without NCD, the highest TSH tertile was associated with a lower risk of mortality compared with the lowest TSH tertile (HR, 0.67; 95% CI 0.54–0.83; p-value < 0.001) (Table 2). Among participants with NCD, the highest TSH tertile had a borderline statistically significant association with a lower risk of mortality than the lowest TSH tertile (HR, 0.88; 95% CI 0.75–1.03; p-value = 0.1) (Table 2).

[Figure omitted. See PDF.]

Table 2. HRs for incident NCD and death among TSH and FT4 tertiles.

https://doi.org/10.1371/journal.pmed.1002957.t002

The highest FT4 tertile was associated with a 1.17-times-higher risk of incident NCD than the lowest FT4 tertile (HR, 1.17; 95% CI 1.02–1.34; p-value = 0.02) (Table 2). The highest FT4 tertile was also associated with a 1.56-times-higher risk of mortality among participants with NCD (HR, 1.56; 95% CI 1.32–1.85; p-value < 0.001) and a 1.44-times-higher risk of mortality among participants without NCD (HR, 1.44; 95% CI 1.15–1.79; p-value = 0.001) (Table 2), compared with the lowest FT4 tertile.

Results for TSH and FT4 analyses remained similar in the raw models, after excluding the events that occurred during the first 2 years of follow-up, after excluding users of thyroid function–altering medications, and after additionally adjusting for TPOAb positivity (S1 Table, S2 Table, S3 Table). TPOAb positivity was not associated with the risk of incident NCD, mortality among those without NCD, and mortality among those with NCD (HR, 1.20; 95% CI 0.99–1.44; p-value 0.06; HR, 1.24; 95% CI 0.92–1.68, p-value = 0.2; HR, 0.97; 95% CI 0.76–1.25; p-value = 0.8; respectively).

Association of thyroid function within the reference range with total LE and LE with and without NCD

Total LE was higher in the middle than in the lowest TSH tertile and did not change substantially from the middle to the highest TSH tertile (Fig 1). Compared with those in the lowest TSH tertile, men in the highest TSH tertile were expected to live 1.5 years (95% CI 0.8–2.3; p-value < 0.001) longer overall; of which 0.1 years (95% CI −0.8 to 1.2; p-value = 0.8) longer without NCD and 1.4 years (95% CI 0.5–2.3; p-value = 0.002) longer with NCD (Table 3). Compared with those in the lowest TSH tertile, women in the highest TSH tertile were expected to live 1.5 years (95% CI 0.8–2.2; p-value < 0.001) longer overall; of which 0.2 years (95% CI −0.8 to 1.3; p-value 0.7) longer without NCD, and 1.3 years (95% CI 0.3–2.1; p-value 0.004) longer with NCD (Table 3).

[Figure omitted. See PDF.]

Fig 1. LE with and without NCD at age 50 years among TSH and FT4 tertiles in men and women.

(A) Estimates of LE with and without NCD (in years) are plotted against TSH tertiles in men. (B) Estimates of LE with and without NCD (in years) are plotted against TSH tertiles in women. (C) Estimates of LE with and without NCD (in years) are plotted against FT4 tertiles in men. (D) Estimates of LE with and without NCD (in years) are plotted against FT4 tertiles in women. NCDs include cardiovascular disease, diabetes mellitus, and cancer. FT4, free thyroxine; LE, life expectancy; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.g001

[Figure omitted. See PDF.]

Table 3. LE with and without NCD at age 50 among TSH and FT4 tertiles in men and women.

https://doi.org/10.1371/journal.pmed.1002957.t003

Total LE decreased progressively with increasing FT4 tertiles (Fig 1). Compared with those in the lowest FT4 tertile, the difference in LE for men in the highest FT4 tertile was −3.7 years (95% CI −5.1 to −2.2; p-value < 0.001) overall; of which −1.8 years (95% CI −3.1 to −0.7; p-value 0.003) without NCD, and −1.9 years (95% CI −3.4 to −0.3; p-value 0.01) with NCD (Table 3). Compared with those in the lowest FT4 tertile, the difference in LE for women in the highest FT4 tertile was −3.3 years (95% CI −4.7 to −1.9; p-value < 0.001) overall; of which −2.0 years (95% CI −3.4 to −0.7; p-value 0.003) without NCD and −1.3 years (95% CI −2.7 to 0.2; p-value 0.07) with NCD (Table 3). Results remained similar over the length of follow-up of 8 years (S4 Table) or after including CKD in the definition of NCD (S5 Table) or after including COPD in the definition of NCD (S6 Table). Compared with those in the lowest FT4 tertile, men and women in the highest FT4 tertile were expected to live fewer years free of CVD, diabetes, and cancer, respectively (S7 Table, S8 Table, S9 Table).

Discussion

In this large prospective population-based cohort study among middle-aged and older adults, we found differences in LE with and without NCD within the reference ranges of thyroid function. Participants with low–normal FT4 levels were expected to live up to 3.7 years longer overall, of which up to 1.9 years longer with NCD than those with high–normal FT4 levels. Participants with high–normal TSH levels were expected to live up to 1.5 years longer overall, of which up to 1.4 years longer with NCD than those with low–normal TSH levels. No meaningful sex differences throughout the TSH and FT4 tertiles were observed.

Previous studies have suggested that the beneficial effects of high–normal thyroid function on metabolism can be counterbalanced by detrimental effects on other systems, such as the cardiovascular system [13,18,19,22]. In this context, our study sheds light on the resultant system-specific effects of thyroid function, suggesting that the overall risk of NCD increases in the high–normal range of FT4 levels. However, the LE estimates not only are attributable to the risk of developing the diseases but also depend on mortality risk. Similar to previous studies conducted in middle-aged and older adults, we showed that high–normal thyroid function is associated with an increased risk of mortality [22,37,38]. This was further translated into a decreased LE with and without NCD for participants with high–normal thyroid function.

More specifically, LE without NCD is the result of two components: risk of incident NCD (transition 1) and risk of mortality among participants without NCD (transition 2). In our study, high–normal FT4 levels were associated with a higher risk of incident NCD compared with low–normal FT4 levels, leading to earlier clinical manifestation of NCD and fewer years lived without NCD. High–normal FT4 levels were also associated with an increased mortality risk among participants without NCD, resulting in a further decrease in total LE and LE without NCD. LE with NCD reflects the combined risk of incident NCD (transition 1) and risk of mortality among participants with NCD (transition 3). We showed that high–normal FT4 levels were associated with a 1.17-times-higher risk of incident NCD compared with low–normal FT4 levels, meaning an earlier occurrence of NCD across the life span and more years lived with NCD. However, participants with NCD and high–normal FT4 levels had an even higher risk of mortality (i.e., 1.56 times higher) than those with low–normal FT4 levels, which explains the decrease in the number of years lived with NCD.

Another study from our group previously reported that individuals with low–normal thyroid function were expected to live more years free of CVD than those with high–normal thyroid function [22]. Given the pleiotropic effects of thyroid hormones, the present study provides a broader perspective by using the multidimensional measure of LE with and without NCD, taking into account multiple diseases involving a wide array of mechanisms. Our results indicate that the estimates of LE with and without NCD reflect the combination of all NCDs and are not driven by one chronic disease alone. In a consistent manner, participants with low–normal thyroid function were expected to live more years with CVD, diabetes, and cancer, respectively, than those with high–normal thyroid function. Also, participants with low–normal thyroid function were expected to live more years free of CVD, diabetes, and cancer, respectively, than those with high–normal thyroid function. Together, these differences in the number of years lived with and without diseases contributed to the total differences in LE within the reference ranges of thyroid function.

The added values of this population-based cohort study include the investigation of the association between thyroid function and incident NCD, as well as the evaluation of differences in LE with and without NCD within the reference ranges of thyroid function. Other strengths of the study are the large sample size, the prospective design with a long follow-up period, and the utilization of a multidimensional measure as LE with and without NCD. Participants had extensive and detailed information on covariates including exposures, outcomes, and potential confounders. Events were adjudicated using standardized criteria. Thyroid function measurements were performed before the occurrence of incident NCD events, and event assessors were blinded to the thyroid status of participants. The possibility of reverse causation was taken into account by excluding the events that occurred during the first 2 years of follow-up. Multiple sensitivity analyses provided consistent findings.

Several limitations should also be considered. We did not have data available on triiodothyronine, the active form of thyroid hormone. Nevertheless, TSH and FT4 represent the most relevant measurements of thyroid function in clinical practice. We also lacked repeated measurements of thyroid function, even though the normal reference ranges of TSH and FT4 levels are shown to be stable over time [39]. In certain conditions (e.g., pregnancy or critical illnesses), the affinity of thyroid hormones to plasma proteins can be affected by substances interfering with the FT4 immunoassay. We lacked information on thyroid hormone binding proteins. However, our population consists of community-dwelling middle-aged and elderly people, and therefore, the concentrations of thyroid hormone binding proteins are not expected to have been altered. One can also assume that NCD events may have occurred before the date of diagnosis. To address this, we excluded the events that occurred within the first two years of follow-up; and results remained consistent. Furthermore, the exact date of incident CKD in our study was uncertain because eGFR measurements were performed only twice. We could therefore not include CKD in the primary outcome of NCD. However, we used the repeated measurements of eGFR to determine the slope of eGFR changes over time. Results remained similar after adding CKD to the definition of NCD. Finally, the possibility of residual confounding cannot be entirely ruled out because of the observational character of our study. The Rotterdam Study includes predominantly participants of European descent older than 45 years. Therefore, our findings need to be confirmed in other populations with similar characteristics to our population, as well as in other ethnicities and other age categories.

Our study provides novel insights about the qualitative and quantitative impact of thyroid function on LE, revealing meaningful differences in LE with and without NCD, within the reference range of thyroid function. These results add valuable information to the ongoing discussion on the reference range of thyroid function. Furthermore, these results support a re-evaluation of the reference ranges of thyroid function in middle-aged and older adults, implying the possibility of a downward shift of the FT4 current limits. This can, in turn, have further implications on the diagnosis and treatment of thyroid disease.

Conclusions

In a population of middle-aged and older euthyroid individuals, we found that high–normal FT4 levels were associated with an increased risk of incident NCD. Furthermore, we found meaningful differences in LE with and without NCD within the reference ranges of TSH and FT4 levels. People with low–normal thyroid function were expected to live more years with and without NCD than those with high–normal thyroid function. These results provide support for a re-evaluation of the current reference ranges of thyroid function. Future studies in other populations are warranted to support the generalization of our findings. Additional research is needed to establish causality and elucidate the underlying mechanisms.

Supporting information

S1 STROBE checklist. STROBE checklist.

STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.

https://doi.org/10.1371/journal.pmed.1002957.s001

(DOCX)

S1 Table. Unadjusted HRs for incident NCD and death among TSH and FT4 tertiles.

FT4, free thyroxine; HR, hazard ratio; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s002

(DOCX)

S2 Table. HRs for incident NCD and death among TSH and FT4 tertiles, additionally adjusted for thyroid peroxidase antibody positivity.

FT4, free thyroxine; HR, hazard ratio; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s003

(DOCX)

S3 Table. HRs for incident NCD and death among TSH and FT4 tertiles after excluding the first 2 years of follow-up for NCD and death or after excluding the users of thyroid function–altering medications.

FT4, free thyroxine; HR, hazard ratio; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s004

(DOCX)

S4 Table. LE with and without NCD at age 50 years among TSH and FT4 tertiles in men and women over 8 years of follow-up.

FT4, free thyroxine; LE, life expectancy; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s005

(DOCX)

S5 Table. LE with and without NCD at age 50 years among TSH and FT4 tertiles in men and women.

FT4, free thyroxine; LE, life expectancy; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s006

(DOCX)

S6 Table. LE with and without NCD at age 50 years among TSH and FT4 tertiles in men and women.

FT4, free thyroxine; LE, life expectancy; NCD, noncommunicable disease; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s007

(DOCX)

S7 Table. LE with and without CVD at age 50 years among TSH and FT4 tertiles in men and women.

CVD, cardiovascular disease; FT4, free thyroxine; LE, life expectancy; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s008

(DOCX)

S8 Table. LE with and without diabetes at age 50 years among TSH and FT4 tertiles in men and women.

FT4, free thyroxine; LE, life expectancy; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s009

(DOCX)

S9 Table. LE with and without cancer at age 50 years among TSH and FT4 tertiles in men and women.

FT4, free thyroxine; LE, life expectancy; TSH, thyroid-stimulating hormone.

https://doi.org/10.1371/journal.pmed.1002957.s010

(DOCX)

S1 Text. Outline of changes made to the analysis plan.

https://doi.org/10.1371/journal.pmed.1002957.s011

(DOCX)

Acknowledgments

We thank the study participants, the staff from the Rotterdam Study, and the participating general practitioners and pharmacists.

Citation: Bano A, Chaker L, Mattace-Raso FUS, Terzikhan N, Kavousi M, Ikram MA, et al. (2019) Thyroid function and life expectancy with and without noncommunicable diseases: A population-based study. PLoS Med 16(10): e1002957. https://doi.org/10.1371/journal.pmed.1002957

References

1. Suhrcke M, Nugent RA, Stuckler D, Rocco L. Chronic Disease: An Economic Perspective. London: Oxford Health Alliance; 2006.

2. WHO. Global status report on noncommunicable diseases 2014. Geneva, Switzerland: WHO; 2014.

3. GBD 2015 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1659–724. pmid:27733284

4. Murray CJ, Vos T, Lozano R, Naghavi M, Flaxman AD, Michaud C, et al. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2197–223. pmid:23245608

5. Rodondi N, den Elzen WP, Bauer DC, Cappola AR, Razvi S, Walsh JP, et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12):1365–74. pmid:20858880

6. Collet TH, Gussekloo J, Bauer DC, den Elzen WP, Cappola AR, Balmer P, et al. Subclinical hyperthyroidism and the risk of coronary heart disease and mortality. Archives of internal medicine. 2012;172(10):799–809. pmid:22529182

7. Gencer B, Collet TH, Virgini V, Bauer DC, Gussekloo J, Cappola AR, et al. Subclinical thyroid dysfunction and the risk of heart failure events: an individual participant data analysis from 6 prospective cohorts. Circulation. 2012;126(9):1040–9. pmid:22821943

8. Hellevik AI, Asvold BO, Bjoro T, Romundstad PR, Nilsen TI, Vatten LJ. Thyroid function and cancer risk: a prospective population study. Cancer epidemiology, biomarkers & prevention: a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2009;18(2):570–4.

9. Gronich N, Deftereos SN, Lavi I, Persidis AS, Abernethy DR, Rennert G. Hypothyroidism is a Risk Factor for New-Onset Diabetes: A Cohort Study. Diabetes care. 2015;38(9):1657–64. pmid:26070591

10. Chiamolera MI, Wondisford FE. Minireview: Thyrotropin-releasing hormone and the thyroid hormone feedback mechanism. Endocrinology. 2009;150(3):1091–6. pmid:19179434

11. Chaker L, Korevaar TIM, Rizopoulos D, Collet TH, Volzke H, Hofman A, et al. Defining Optimal Health Range for Thyroid Function Based on the Risk of Cardiovascular Disease. The Journal of clinical endocrinology and metabolism. 2017;102(8):2853–61. pmid:28520952

12. Wartofsky L, Dickey RA. The evidence for a narrower thyrotropin reference range is compelling. The Journal of clinical endocrinology and metabolism. 2005;90(9):5483–8. pmid:16148345

13. Chaker L, Ligthart S, Korevaar TI, Hofman A, Franco OH, Peeters RP, et al. Thyroid function and risk of type 2 diabetes: a population-based prospective cohort study. BMC Med. 2016;14(1):150. pmid:27686165

14. Khan SR, Chaker L, Ruiter R, Aerts JG, Hofman A, Dehghan A, et al. Thyroid Function and Cancer Risk: The Rotterdam Study. The Journal of clinical endocrinology and metabolism. 2016;101(12):5030–6. pmid:27648963

15. Brabant G, Beck-Peccoz P, Jarzab B, Laurberg P, Orgiazzi J, Szabolcs I, et al. Is there a need to redefine the upper normal limit of TSH? Eur J Endocrinol. 2006;154(5):633–7. pmid:16645008

16. Chaker L, Sedaghat S, Hoorn EJ, Elzen WP, Gussekloo J, Hofman A, et al. The association of thyroid function and the risk of kidney function decline: a population-based cohort study. Eur J Endocrinol. 2016;175(6):653–60. pmid:27926474

17. Cappola AR, Arnold AM, Wulczyn K, Carlson M, Robbins J, Psaty BM. Thyroid function in the euthyroid range and adverse outcomes in older adults. The Journal of clinical endocrinology and metabolism. 2015;100(3):1088–96. pmid:25514105

18. Bano A, Chaker L, Mattace-Raso FUS, van der Lugt A, Ikram MA, Franco OH, et al. Thyroid Function and the Risk of Atherosclerotic Cardiovascular Morbidity and Mortality: The Rotterdam Study. Circ Res. 2017;121(12):1392–400. pmid:29089349

19. Bano A, Chaker L, Plompen EP, Hofman A, Dehghan A, Franco OH, et al. Thyroid Function and the Risk of Nonalcoholic Fatty Liver Disease: The Rotterdam Study. The Journal of clinical endocrinology and metabolism. 2016;101(8):3204–11. pmid:27270473

20. Chaker L, Cappola AR, Mooijaart SP, Peeters RP. Clinical aspects of thyroid function during ageing. The lancet Diabetes & endocrinology. 2018;6(9):733–42.

21. Taylor PN, Razvi S, Pearce SH, Dayan CM. Clinical review: A review of the clinical consequences of variation in thyroid function within the reference range. The Journal of clinical endocrinology and metabolism. 2013;98(9):3562–71. pmid:23824418

22. Bano A, Dhana K, Chaker L, Kavousi M, Ikram MA, Mattace-Raso FUS, et al. Association of Thyroid Function With Life Expectancy With and Without Cardiovascular Disease: The Rotterdam Study. JAMA Intern Med. 2017;177(11):1650–7. pmid:28975207

23. Beaglehole R, Bonita R, Alleyne G, Horton R, Li L, Lincoln P, et al. UN High-Level Meeting on Non-Communicable Diseases: addressing four questions. Lancet. 2011;378(9789):449–55. pmid:21665266

24. Ikram MA, Brusselle GGO, Murad SD, van Duijn CM, Franco OH, Goedegebure A, et al. The Rotterdam Study: 2018 update on objectives, design and main results. European journal of epidemiology. 2017;32(9):807–50. pmid:29064009

25. Leening MJ, Kavousi M, Heeringa J, van Rooij FJ, Verkroost-van Heemst J, Deckers JW, et al. Methods of data collection and definitions of cardiac outcomes in the Rotterdam Study. European journal of epidemiology. 2012;27(3):173–85. pmid:22388767

26. Bos MJ, Koudstaal PJ, Hofman A, Ikram MA. Modifiable etiological factors and the burden of stroke from the Rotterdam study: a population-based cohort study. PLoS Med. 2014;11(4):e1001634. pmid:24781247

27. Swedberg K, Cleland J, Dargie H, Drexler H, Follath F, Komajda M, et al. Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005): The Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology. Eur Heart J. 2005;26(11):1115–40. pmid:15901669

28. Ligthart S, van Herpt TT, Leening MJ, Kavousi M, Hofman A, Stricker BH, et al. Lifetime risk of developing impaired glucose metabolism and eventual progression from prediabetes to type 2 diabetes: a prospective cohort study. The lancet Diabetes & endocrinology. 2016;4(1):44–51.

29. De Bruijn KM, Ruiter R, de Keyser CE, Hofman A, Stricker BH, van Eijck CH. Detection bias may be the main cause of increased cancer incidence among diabetics: results from the Rotterdam Study. Eur J Cancer. 2014;50(14):2449–55. pmid:25047425

30. Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Greene T, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. The New England journal of medicine. 2012;367(1):20–9. pmid:22762315

31. Perrone RD, Madias NE, Levey AS. Serum creatinine as an index of renal function: new insights into old concepts. Clinical chemistry. 1992;38(10):1933–53. pmid:1394976

32. Vestbo J, Hurd SS, Agusti AG, Jones PW, Vogelmeier C, Anzueto A, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med. 2013;187(4):347–65. pmid:22878278

33. Peeters A, Mamun AA, Willekens F, Bonneux L. A cardiovascular life history. A life course analysis of the original Framingham Heart Study cohort. Eur Heart J. 2002;23(6):458–66. pmid:11863348

34. Franco OH, de Laet C, Peeters A, Jonker J, Mackenbach J, Nusselder W. Effects of physical activity on life expectancy with cardiovascular disease. Archives of internal medicine. 2005;165(20):2355–60. pmid:16287764

35. Franco OH, Steyerberg EW, Hu FB, Mackenbach J, Nusselder W. Associations of diabetes mellitus with total life expectancy and life expectancy with and without cardiovascular disease. Arch Intern Med. 2007;167(11):1145–51. pmid:17563022

36. Efron B, Tibshirani RJ. An introduction to the bootstrap. New York: Chapman and Hall; 1993.

37. Yeap BB, Alfonso H, Hankey GJ, Flicker L, Golledge J, Norman PE, et al. Higher free thyroxine levels are associated with all-cause mortality in euthyroid older men: the Health In Men Study. Eur J Endocrinol. 2013;169(4):401–8. pmid:23853210

38. Gussekloo J, van Exel E, de Craen AJ, Meinders AE, Frolich M, Westendorp RG. Thyroid status, disability and cognitive function, and survival in old age. JAMA. 2004;292(21):2591–9. pmid:15572717

39. Andersen S, Pedersen KM, Bruun NH, Laurberg P. Narrow individual variations in serum T(4) and T(3) in normal patients: a clue to the understanding of subclinical thyroid disease. The Journal of clinical endocrinology and metabolism. 2002;87(3):1068–72. pmid:11889165