Subclinical Hypothyroidism and the Risk of Cardiovascular Disease and All-Cause Mortality: A Meta-Analysis of Prospective Cohort Studies

Abstract

Objectives:

To determine the impact of subclinical hypothyroidism (SCH) on the risk of cardiovascular disease (CVD) and all-cause mortality, a comprehensive meta-analysis was performed according to the age or coexisting CVD risk status of the participants.

Methods:

Studies regarding the association of SCH with all-cause mortality from PubMed and Embase databases were included. The pooled relative risk (RR) of CVD and all-cause mortality was calculated using the Mantel–Haenszel method. A subgroup analysis of participants with high CVD risk was conducted, including history of coronary, cerebral, or peripheral artery disease; dilated cardiomyopathy; heart failure; atrial fibrillation; venous thromboembolism; diabetes mellitus; or chronic kidney disease.

Results:

In total, 35 eligible articles incorporating 555,530 participants were included. SCH was modestly associated with CVD and all-cause mortality (RR for CVD = 1.33 [confidence interval (CI) 1.14–1.54]; RR for all-cause mortality = 1.20 [CI 1.07–1.34]). However, the association was not observed in participants aged ≥65 years. Subgroup analysis showed that participants with SCH and high CVD risk showed a significantly higher risk of all-cause mortality (RR for CVD = 2.20 [CI 1.28–3.77]; RR for all-cause mortality = 1.66 [CI 1.41–1.94]), whereas those with SCH and low CVD risk did not. Additional subgroup analysis of six studies with a mean participant age of ≥65 years and high CVD risk showed a significant high risk of all-cause mortality in the SCH group (RR = 1.41 [CI 1.08–1.85]; I ² = 0%).

Conclusions:

SCH is associated with an increased CVD risk and all-cause mortality, particularly in participants with high CVD risk.

Introduction

Subclinical hypothyroidism (SCH) is characterized by high thyrotropin (TSH) concentrations and normal free thyroxine (fT4) levels (1). The prevalence of SCH in the general population is between 4% and 20% (2,3). As screening for thyroid disease has become more common, the prevalence of SCH has been increasing in clinical settings (4). Therefore, evidence-based clarification about the influence of SCH on general health, specifically for mortality or morbidities such as cardiovascular disease (CVD), is becoming increasingly important to assist screening or treatment of these conditions (2). Until now, a number of studies have shown an association with various adverse clinical outcomes, including increased blood pressure (5) and lipid levels (6), and risk of coronary artery disease (7 –9). Meta-analyses have revealed an association between SCH and an increased risk of CVD-related mortality (10), congestive heart failure (HF) (11), and fatal stroke (12) in subjects with a high TSH concentration (≥10 mIU/L). Therefore, current clinical guidelines for SCH recommend considering CVD risk when planning commencement of treatment (13). However, there are conflicting reports regarding the impact of SCH according to individual risk factors, such as underlying CVD risks. A meta-analysis with individual participant data from 12 prospective cohorts reported that SCH was not associated with CVD and all-cause mortality in patients with pre-existing CVD (10). However, several recent cohort studies showed that SCH significantly increased risk of CVD and all-cause mortality (14 –18), suggesting that the impact of SCH may vary according to individual CVD risk. Therefore, an updated meta-analysis with recent data is warranted to identify individuals with SCH at high risk for adverse clinical outcomes.

Therefore, an updated comprehensive meta-analysis was performed to clarify the association of SCH with CVD with all-cause mortality according to the age of participants as well as the presence of CVD risk.

Materials and Methods

Search strategy

This meta-analysis follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Checklist (19). A literature search of PubMed and Embase databases was conducted from their inception (January 1, 1990) to November 26, 2017. Two independent investigators (S.M. and M.J.K.) selected articles with a combination of the following relevant terms: “subclinical,” “thyroid,” “hypothyroidism,” “cardiovascular disease,” “heart failure,” “myocardial infarction,” and “mortality.” Only articles published in English were considered.

Study selection

The literature search yielded 2781 potentially relevant articles, of which 238 were screened for further review after excluding duplicate studies (n = 498) or those with a different topic of interest (n = 2045). If studies had multiple reports, the latest or most complete article was used. All articles were electronically downloaded and screened for inclusion using a two-step method. After evaluation of the titles and abstracts according to predefined criteria, 189 articles were further excluded for the following reasons: (i) there was no information on TSH, fT4, or all-cause mortality; (ii) the study was published in the form of an abstract, expert opinion, letter, conference article, or review; or (iii) it was not a cohort study. Subsequently, the full texts of 49 selected articles were reviewed by two independent investigators (S.M. and M.J.K.), and any disagreement was resolved by a third investigator (Y.J.P.). Because the data of 14 of these articles were insufficient for the analysis, 35 articles were finally selected for the meta-analysis (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/thy).

Data extraction and quality assessment

The following variables were independently extracted by the two investigators based on the same rules: first author, publication year, country, number and characteristics of study participants, number of men and women, mean age, coexisting diseases potentially associated with CVD risk, follow-up duration, and mortality. All disagreements were discussed with the third investigator. The Newcastle–Ottawa Scale (NOS) was applied to evaluate the methodological quality of the eligible studies. Accumulated scores ranged from 0 to 9 points; scores from 6 to 9 points were considered to suggest high quality, thus representing a lower risk of bias.

Data analyses and statistical methods

The pooled relative risks (RR) with confidence intervals (CIs) were calculated using the Mantel–Haenszel method. The Higgins' I ² statistic was used to test for heterogeneity. When I ² was ≤50%, the included studies were considered to have little heterogeneity, and a fixed-effects model was used. When I ² was ≥50%, the included studies were considered to have significant heterogeneity, and a random-effects model was used. The potential for publication bias was assessed using a funnel plot analysis. To examine the strength of the outcome, a sensitivity analysis was conducted. Subgroup analyses were used to determine the cause of heterogeneity stratified by age (cutoff 65 years) and CVD risk. High CVD risk was defined as the presence of any disease that could increase CVD risk, including history of coronary, cerebral, or peripheral artery disease; dilated cardiomyopathy; heart failure; atrial fibrillation; venous thromboembolism; diabetes mellitus; or CVD risk factors based on the World Health Organization criteria (20). Participants with chronic kidney disease were also classified as a high CVD risk group based on the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines (21). Population-based studies of the general population were classified as having a subgroup with low CVD risk because they included mainly healthy participants.

All statistical analyses were performed using the statistical program R v3.1.0 (www.r-project.org).

Results

Characteristics of selected studies

In total, 35 articles underwent meta-analysis (14 –18,22 –51), and their main characteristics are summarized in Table 1. In total, 555,530 participants were enrolled, and 21,176 (mean 5.7%; range 1.6–28.8%) had SCH. Sample sizes of these studies ranged from 229 to 228,357 participants. Among them, 14 studies showed a mean age of <65 years, and 18 had a mean age of ≥65 years. Three studies reported age-based subgroup data with a cutoff age of 65 years (29,34,35). A total of 11 studies were defined as studies with participants at high CVD risk based on predefined criteria. Twenty-three studies were population-based studies with participants from a general population. One study reported subgroup data stratified by preexisting CVD (17). All of the included studies were judged as being of relatively high quality according to the NOS assessment tool, with scores ranging from 6 to 9.

Table 1.

Summary of the 35 Cohort Studies Included in the Present Meta-Analysis

Study	Country	Study population	M (SD) or median (range) age	No. of subjects	No of SCH, n	CVD events, n	CVD death, n	All cause-death, n
Tohidi et al. (22)	Iran	Community dwelling	EU: 46.3 ± 12 SCH: 47.6 ± 11.6	3677	189	374
Segna et al. (23)	Switzerland	High CVD risk (venous thromboembolism)	74^a (69–80)^b	535	35			53
Grossman et al. (14)	Israel	Community dwelling	EU:76 (75.9–76.2)SCH: 78 (77.7–78.3)	16,902	1956			5737
Asfar et al. (24)	Turkey	High CVD risk (CKD patients)	EU: 44.8 (12.2)SCH: 49.4 (11.5)	229	66	65
Hayashi et al. (25)	Japan	High CVD risk (heart failure)	EU: 70 (15)SCH: 69 (12)	246	58	84
Martin et al. (26)	United States	Community dwelling	EU: 56.5(5.7)SCH: 58.3 (5.7)	10,520	542	1102
Kovar et al. (15)	Austria	Community dwelling	48^a (34–61)^b	80,490	3934		2577	7322
Wang et al. (16)	China	High CVD risk (dilated cardiomyopathy)	51 (49.6–52.4)	394	41		87	111
Laulund et al. (27)	Denmark	Community dwelling	EU: 50.8 (50.7–50.9)SCH: 55 (54.4–55.4)	228,657	6519			34,494
Li et al. (28)	China	High CVD risk (dilated cardiomyopathy)	52 (51.2–53.0)	895	79			192
Ven et al. (29)	Netherlands	Community dwelling	55.7 (55.2–56.2)	5577	212			710
Drechsler et al. (30)	Germany	High CVD risk (diabetic hemodialysis patients)	EU: 65.3 (64.7–65.9)SCH: 66.4 (62.1–70.7)	797	16		287	372
Yeap et al. (31)	Australia	Community dwelling	Survived: 76.6 (76.5–76.7)Died: 78.5 (78.2–78.7)	3588	416			832
Ceresini et al. (32)	Italy	Community dwelling	Survived: 73.6 (73.1–74.1)Died: 82.2 (81.0–83.4)	848	29			171
Hyland et al. (33)	United States	Community dwelling	EU: 73.4 (5.7)SCH: 74.1 (5.9)	4863	679	925	769
Legrys et al. (34)	United States	Community dwelling	NA	3663	282	738
Rhee et al. (17)	United States	Community dwelling	EU: 45.4 (45.1–45.7)SCH: 59.2 (58.5–59.9)	14,821	691			3337
Tseng et al. (35)	Taiwan	Community dwelling	EU: 42.9 (42.8–43.0)SCH: 47.1 (46.6–47.6)	115,746	1841		680	3669
Molinaro et al. (36)	Italy	High CVD risk (emergence cardiac care)	67.7 (66.9–68.4)	647	68		31	62
Nanchen et al. (37)	NetherlandIrelandScotland	High CVD risk (high CVD risk factors)	EU: 75.3 (75.2–75.4)SCH: 75.5 (75.1–75.9)	5245	199	884	260	540
Waring et al. (38)	United States	Community dwelling	EU: 73.6 (73.3–73.9)SCH: 76.4 (75.3–77.5)	1511	124		150	432
Frey et al. (39)	Germany	High CVD risk (heart failure)	68 (67.0–69.0)	662	34		123	215
Jongh et al. (40)	Netherlands	Community dwelling	EU: 75.5 (75.0–76.0)SCH: 74.9 (74.4–75.4)	1185	64		180	588
Schultz et al. (41)	Denmark	Community dwelling	EU: 67.5 (66.6–68.4)SCH: 69.5 (67.3–71.7)	580	31	74		81
Sgarbi et al. (42)	Brazil	Community dwelling	EU: 56.4 (55.6–57.2)SCH: 58.5 (55.6–61.4)	977	101			68
Razvi et al. (43)	England	Community dwelling	EU: 45.3 (44.5–46.1)SCH: 49.9 (48.6–51.2)	2406	124	465	239	730
Boekholdt et al. (44)	England	Community dwelling	Men: EU 58 (57.7–58.3)SCH 61 (59.9–62.1)Women: EU 58 (57.8–58.2)SCH 59 (58.2–59.7)	12,617	720	1678	453	1824
Asvold et al. (45)	Norway	Community dwelling	55^a (41–98)^c	24,590	814		399	2275
Iervasi et al. (18)	Italy	High CVD risk (cardiac patients)	61.1 (60.7–61.5)	2875	228	168	106	277
Chen et al. (46)	Taiwan	High CVD risk (type 2 diabetic mellitus)	EU: 66.3 (65.4–7.2)SCH: 67.2 (63.4–71.0)	556	41	61	12	31
Cappola et al. (47)	United States	Community dwelling	72.7 (72.5–72.9)	3003	492	1135	440	1824
Rodondi et al. (48)	United States	Community dwelling	EU: 74.7 (74.6–74.8)SHC: 74.8 (74.6–75.0)	2660	335	555	175	791
Walsh et al. (49)	Australia	Community dwelling	EU: 49.2 (48.3–50.1)SCH: 58.1 (56.9–59.3)	1984	89	386	157	515
Gussekloo et al. (50)	Netherland	Community dwelling	85	486	35	83	44	390
Parle et al. (51)	United Kingdom	Community dwelling	Men: 70.1 (69.6–70.6)Women: 70.7 (70.2–71.4)	1098	92		124	467

Median.

Interquartile range.

Range (min–max).

EU, euthyroidism; SCH, subclinical hypothyroidism; CVD, cardiovascular disease.

Effect of SCH on CVD risk and all-cause mortality

In total, 13,415 fatal or non-fatal CVD events occurred in patients among 283,578 participants from 27 studies (15,16,18,22,24 –26,30,32 –41,43 –51). The pooled RR of CVD was 1.33 ([CI 1.14–1.54]; Fig. 1 and Supplementary Fig. S2), with high heterogeneity (I ² = 73%) among studies. Since analysis with the funnel plot was asymmetric, the trim-and-fill method to adjust for publication bias was conducted by adding one estimated missing study (RR = 1.25 [CI 1.05–1.49]; Supplementary Fig. S3). In the sensitivity analysis to identify heterogeneity, the pooled RR barely changed, remaining statistically significant after omitting each study (Supplementary Fig. S4); the pooled RR range was 1.25–1.36.

FIG. 1.

Age- and sex-adjusted relative risk (RR) of combined fatal and non-fatal cardiovascular disease (CVD) associated with subclinical hypothyroidism (SCH) compared to euthyroidism according to CVD risk.

Twenty-nine studies with 532,332 individuals and 68,100 events analyzed all-cause mortality (14 –18,23,27 –32,35 –51). In all participants, a significantly higher risk for all-cause mortality was found in participants with SCH than in those who were euthyroid (RR = 1.20 [CI 1.07–1.34]), with significant heterogeneity (I ² = 84%; Fig. 2) among studies. The funnel plot showed no publication bias (Supplementary Fig. S3). In the sensitivity analysis to identify heterogeneity, the pooled RR barely changed, remaining statistically significant after omitting each study (Supplementary Fig. S4); the range was 1.16–1.22.

FIG. 2.

Age- and sex-adjusted RR of all-cause mortality associated with SCH compared to euthyroidism according to CVD risk.

When the subgroup analysis was performed according to location (Table 2), a significant association of SCH with CVD risk and all-cause mortality was found in Asia (RR for CVD = 1.90 [CI 1.50–2.40]; RR for all-cause mortality = 1.60 [CI 1.28–2.00]) and Europe (RR for CVD = 1.34 [CI 1.03–1.74]; RR for all-cause mortality = 1.15 [CI 1.00–1.32]), whereas studies from America failed to show significant results (RR for CVD = 1.02 [CI 0.92–1.12]; RR for all-cause mortality = 1.04 [CI 0.96–1.13]).

Table 2.

Subgroup Analyses of Pooled RR of CVD and All-Cause Mortality

				Sensitivity analysis
Subgroup	No. of studies	RRs [CI]	I²	No. of outlier studies	Corrected RRs [CI]	I²
Combined fatal and non-fatal CVD events
Age
Mean age <65 years	12	1.54 [1.21–1.96]	80%	1	1.32 [1.20–1.45]	49%
Mean age ≥65 years	17	1.07 [0.97–1.18]	49%	—	—	—
Location
Asia	5	1.90 [1.50–2.40]	32%	—	—	—
Europe	15	1.34 [1.03–1.74]	76%	1	1.25 [1.13–1.39]	38%
Oceania	1	1.78 [1.22–2.59]	—	—	—	—
America	6	1.02 [0.92–1.12]	0%	—	—	—
All-cause mortality
Age
Mean age <65 years	13	1.28 [1.10–1.48]	80%	—	—	—
Mean age ≥65 years	18	1.13 [0.97–1.33]	78%	1	1.06 [0.98–1.20]	25%
Location
Asia	6	1.60 [1.28–2.00]	64%	1	1.56 [1.46–1.66]	44%
Europe	17	1.15 [1.00–1.32]	67%	2	1.03 [0.98–1.08]	0%
Oceania	2	1.15 [0.96–1.38]	54%	—	—	—
America	4	1.04 [0.96–1.13]	1%	—	—	—

RR was adjusted for age and sex.

RR, relative risk; CVD, cardiovascular disease.

Effect of SCH on CVD risk and all-cause mortality according to CVD risk

Effects of SCH on the risk of CVD and all-cause mortality were analyzed according to the status of CVD risk. In the subgroup analysis with high CVD risk groups, SCH significantly increased risk of CVD and all-cause mortality (RR for CVD = 2.20 [CI 1.28–3.77]; I ² = 84%; RR for all-cause mortality = 1.66 [CI 1.41–1.94]; I ² = 45%; Figs. 1 and 2). In sensitivity analysis for CVD, the pooled RR barely changed and remained significant after omitting each study; the range was 1.63–2.32. However, in the subgroup analysis with low CVD risk groups, a modest association was observed with CVD (RR = 1.15 [CI 1.06–1.22]; I ² = 39%; Fig. 1), while there was no significant association with all-cause mortality (RR = 1.10 [CI 0.98–1.24]; I ² = 87%; Fig. 2). In sensitivity analysis for all-cause mortality, two outlier studies were found (14,15). When these studies were excluded, the pooled RR remained nonsignificant (RR = 1.03 [CI 0.99–1.07]; I ² = 9%).

Effect of SCH on the risk of CVD and all-cause mortality in aged individuals

It has been suggested that the clinical significance of SCH is different in aged populations (13). Therefore, a meta-analysis was performed after dividing the studies according to the mean age. Subgroup analysis of the studies in which participants had a mean age of <65 years showed a significant association between SCH and risk of CVD and all-cause mortality with significant heterogeneity (RR for CVD = 1.54 [CI 1.21–1.96]; I ² = 80%; RR for all-cause mortality = 1.28 [CI 1.10–1.48]; I ² = 80%; Table 2). Sensitivity analysis for CVD found one outlier study (24). The statistical significance remained after omitting the outlier study (RR = 1.32 [CI 1.20–1.45]; I ² = 49%). In sensitivity analyses for all-cause mortality, no outlier studies could be found, and the pooled RR value barely changed after omitting each study. However, a meta-analysis of the studies in which participants had a mean age of ≥65 years found no significant association between SCH and risk of CVD and all-cause mortality (RR for CVD = 1.07 [CI 0.97–1.18]; I ² = 49%; RR for all-cause mortality = 1.13 [CI 0.97–1.33]; I ² = 78%; Table 2). Sensitivity analysis found one outlier study (14). After omitting the outlier study, the pooled RRs of the remaining studies remained nonsignificant (RR = 1.06 [CI 0.98–1.20], I ² = 25%).

Effect of SCH on the risk of CVD and all-cause mortality based on CVD risk in aged individuals

To determine whether CVD risk factors affect the association between SCH and the risk of CVD and all-cause mortality in aged populations, a further subgroup was made according to CVD risk in the aged population. Among the studies in which participants had a mean age of ≥65 years, subgroup analysis with low CVD risk groups showed no significant association between SCH and the risk of CVD and all-cause mortality (RR for CVD = 1.02 [CI 0.92–1.13]; I ² = 2%; RR for all-cause mortality = 1.07 [CI 0.89–1.29]; I ² = 84%; Fig. 3). When one outlier study (14) was excluded from sensitivity analysis for all-cause mortality, the pooled RR remained nonsignificant (RR = 1.04 [CI 0.96–1.12]; I ² = 18.5%; Fig. 3). However, in contrast to studies with low CVD risk groups, the subgroup analysis with the high CVD risk group showed a significantly high risk of all-cause mortality in the SCH patients without significant heterogeneity (RR = 1.41 [CI 1.08–1.85]; I ² = 0%; Fig. 3). Although the results for CVD did not reach statistical significance (RR = 1.50 [CI 0.89–2.54]; I ² = 68%; Fig. 3), studies with high CVD risk groups showed higher risk than those with low CVD risk groups. The results of a subgroup analysis of the studies in which participants had a mean age of <65 years showed similar results (Supplementary Fig. S5).

FIG. 3.

Subgroup analyses of studies with mean age ≥65 years according to CVD risk. (A) Combined fatal and non-fatal CVD events. (B) All-cause mortality.

Discussion

Based on this large-scale meta-analysis with 555,530 participants from 35 cohort studies, it is concluded that SCH is associated with an increased risk of CVD and all-cause mortality, specifically among individuals with high CVD risk. The increased risk of all-cause mortality in participants with SCD and high CVD risk was also observed in aged populations.

In this study, SCH was associated with increased CVD events and all-cause mortality by 20% compared to euthyroid status, and this risk was >60% in the high CVD risk group. Although there was significant heterogeneity, this risk effect persisted after sensitivity analysis and correction of publication bias. This result is different from the previous meta-analysis by Rodondi et al. (10), which reported no association after a meta-analysis of 12 studies. The difference may be due to the recent large population-based studies included in the current analysis (14 –16). Since there was significant heterogeneity, subgroup analyses were conducted to find associations between SCH and the risk of CVD events and all-cause mortality in the different groups.

Interestingly, a significant association between SCH and risk of CVD events and all-cause mortality was found in studies from Asia and Europe, whereas studies from America failed to show significant results. This phenomenon might result from specific ethnic characteristics. However, it also might reflect possible confounding effects stemming from heterogeneity of the study population. All eight American studies were conducted with participants from the general population, whereas four of seven Asian studies and 7 of 18 European studies were performed with participants at high CVD risk. Therefore, further studies with high CVD risk groups in America are needed.

The American Thyroid Association (ATA) and the American Association of Clinical Endocrinologists (AACE) recommend consideration of treatment of SCH for subjects with TSH levels ≥10 mIU/L; symptoms suggestive of hypothyroidism; a positive test for antibodies to thyroid peroxidase; or evidence of atherosclerotic cardiovascular disease, heart failure, or CVD risk factors (13). Because of a lack of data, further analysis according to TSH levels could not be conducted. However, a subgroup analysis was performed stratified by underlying CVD or CVD risk factors. In a previous meta-analysis by Rodondi et al., the association between SCH and CVD events and all-cause mortality did not differ among participants with and without underlying CVD (10). However, several newer studies suggest that groups with SCH and CVD or high underlying CVD risk have a high risk of CVD events and all-cause mortality (17,36,52). The present meta-analysis, which includes the more recent data, demonstrates that SCH is associated with a higher risk of CVD and all-cause mortality in participants with a high CVD risk compared to that of participants with a low CVD risk. Considering the different results compared to the previous meta-analysis, it is assumed that the current study could provide a new perspective to find a subgroup of patients with SCH based on CVD risk that could benefit from treatment. Furthermore, this may provide important evidence to support the current clinical guidelines of the ATA and AACE (13).

The European Thyroid Association (ETA) suggested recommendations for SCH should vary according to patient age (53). In younger subjects (<65–70 years), ETA guidelines recommend treatment of SCH for those with TSH levels of ≥10 mIU/L or symptoms suggestive of hypothyroidism. In contrast, in aged individuals (>70 years), the ETA recommend treatment of SCH for those with TSH levels of ≥10 mIU/L only when they exhibit clear symptoms of hypothyroidism or high CVD risk. Advancing age is associated with increased prevalence of SCH, and around 10% of aged individuals have SCH (3,52,54 –59). SCH is a possible contributor to diverse problems in aged individuals, such as CVD, cognitive dysfunction, musculoskeletal problems, and deleterious effects on bone (10,55,60,61). However, because of a widening of the reference range for TSH with increasing age (57,62), the clinical significance of SCH in aged individuals is poorly understood. In addition, the association between SCH and mortality in elderly people is inconsistent (14,50,58). Albeit the previous meta-analysis failed to find any association between SCH and all-cause mortality in older people, recent large population-based studies reported significant RRs of all-cause mortality (10,14 –16). This updated meta-analysis with recent studies confirms that SCH has no substantial association with all-cause mortality in a general population of aged people, but it did have an association in those with high CVD risk, providing evidence to support the ETA guidelines (53). In contrast to effects in older individuals, the deleterious effect of SCH on CVD events and mortality is well established among young adults (13). The present study also found a deleterious effect of SCH on mortality in studies that investigated relatively younger subjects.

Notably, SCH subjects with high CVD risk persistently show a significantly increased risk of all-cause mortality in both younger and aged communities. Considering the present results, SCH is a clinical heterogeneous entity, and this might be one possible explanation for the contradictory results between previous studies (10,14,50,58). Importantly, the present study demonstrates that assessment of CVD risk might help to find a group in whom treatment of SCH is beneficial, regardless of age.

The considerable strengths of this study are that a number of cohort studies with large populations were included, and predefined subgroup analyses could be performed. However, the present study has some limitations. First, most of the included studies measured TSH and fT4 at a single time point at baseline, and different cutoff values of TSH were used to define SCH across the studies included in the meta-analysis. Second, although effect size from a multivariate model that adjusted for age and sex was predominantly used, several studies did not supply age, sex, and other confounding factors for adjusted hazard ratios. Third, subgroup analysis could not be performed with healthy participants, but this was possible using a general population. Considering that these studies included both healthy participants and preexisting CVD patients, there might be a confounding effect. However, in view of the fact that the SCH patients with high CVD risk had an increased risk of adverse clinical outcomes in the subgroup analysis, the confounding by preexisting CVD appears to have only a minor influence on the results.

In conclusion, the findings demonstrate that SCH is associated with a greater risk of CVD and all-cause mortality in subpopulations with underlying CVD or high CVD risk. Further well designed prospective cohort studies are necessary to find high-risk groups, and randomized controlled trials are needed to investigate possible benefits of a levothyroxine replacement therapy in these high-risk groups.

Footnotes

Author Disclosure Statement

The authors declare no conflict of interest.

References

Helfand

. 2004. Screening for subclinical thyroid dysfunction in nonpregnant adults: a summary of the evidence for the U.S. Preventive services task force. Ann Intern Med, 140:128–141.

Cooper

, Biondi

. 2012. Subclinical thyroid disease. Lancet, 379:1142–1154.

Kim

, Park

. 2014. Prevalence and risk factors of subclinical thyroid disease. Endocrinol Metab (Seoul), 29:20–29.

Leese

, Flynn

, Jung

, Macdonald

, Murphy

, Morris

. 2008. Increasing prevalence and incidence of thyroid disease in Tayside, Scotland: the Thyroid Epidemiology Audit and Research Study (TEARS). Clin Endocrinol (Oxf), 68:311–316.

Asvold

, Bjoro

, Nilsen

, Vatten

. 2007. Association between blood pressure and serum thyroid-stimulating hormone concentration within the reference range: a population-based study. J Clin Endocrinol Metab, 92:841–845.

Asvold

, Vatten

, Nilsen

, Bjoro

. 2007. The association between TSH within the reference range and serum lipid concentrations in a population-based study. The HUNT Study. Eur J Endocrinol, 156:181–186.

Biondi

, Klein

. 2004. Hypothyroidism as a risk factor for cardiovascular disease. Endocrine, 24:1–13.

Biondi

, Palmieri

, Lombardi

, Fazio

. 2002. Effects of subclinical thyroid dysfunction on the heart. Ann Intern Med, 137:904–914.

Biondi

, Palmieri

, Lombardi

, Fazio

. 2002. Subclinical hypothyroidism and cardiac function. Thyroid, 12:505–510.

10.

Rodondi

, den Elzen

, Bauer

, Cappola

, Razvi

, Walsh

, Asvold

, Iervasi

, Imaizumi

, Collet

, Bremner

, Maisonneuve

, Sgarbi

, Khaw

, Vanderpump

, Newman

, Cornuz

, Franklyn

, Westendorp

, Vittinghoff

, Gussekloo

. 2010. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA, 304:1365–1374.

11.

Gencer

, Collet

, Virgini

, Bauer

, Gussekloo

, Cappola

, Nanchen

, den Elzen

, Balmer

, Luben

, Iacoviello

, Triggiani

, Cornuz

, Newman

, Khaw

, Jukema

, Westendorp

, Vittinghoff

, Aujesky

, Rodondi

; 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.

12.

Chaker

, Baumgartner

, den Elzen

, Ikram

, Blum

, Collet

, Bakker

, Dehghan

, Drechsler

, Luben

, Hofman

, Portegies

, Medici

, Iervasi

, Stott

, Ford

, Bremner

, Wanner

, Ferrucci

, Newman

, Dullaart

, Sgarbi

, Ceresini

, Maciel

, Westendorp

, Jukema

, Imaizumi

, Franklyn

, Bauer

, Walsh

, Razvi

, Khaw

, Cappola

, Völzke

, Franco

, Gussekloo

, Rodondi

, Peeters

; Thyroid Studies Collaboration. 2015. Subclinical hypothyroidism and the risk of stroke events and fatal stroke: an individual participant data analysis. J Clin Endocrinol Metab, 100:2181–2191.

13.

Garber

, Cobin

, Gharib

, Hennessey

, Klein

, Mechanick

, Pessah-Pollack

, Singer

, Woeber

. 2012. Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Thyroid, 22:1200–1235.

14.

Grossman

, Weiss

, Koren-Morag

, Shimon

, Beloosesky

, Meyerovitch

. 2016. Subclinical thyroid disease and mortality in the elderly: a retrospective cohort study. Am J Med, 129:423–430.

15.

Kovar

, Fang

, Perkmann

, Haslacher

, Slavka

, Födinger

, Endler

, Wagner

. 2015. Subclinical hypothyroidism and mortality in a large Austrian cohort: a possible impact on treatment?. Wien Klin Wochenschr, 127:924–930.

16.

Wang

, Guan

, Gerdes

, Iervasi

, Yang

, Tang

. 2015. Thyroid status, cardiac function, and mortality in patients with idiopathic dilated cardiomyopathy. J Clin Endocrinol Metab, 100:3210–3218.

17.

Rhee

, Curhan

, Alexander

, Bhan

, Brunelli

. 2013. Subclinical hypothyroidism and survival: the effects of heart failure and race. J Clin Endocrinol Metab, 98:2326–2336.

18.

Iervasi

, Molinaro

, Landi

, Taddei

, Galli

, Mariani

, L'Abbate

, Pingitore

. 2007. Association between increased mortality and mild thyroid dysfunction in cardiac patients. Arch Intern Med, 167:1526–1532.

19.

Moher

, Liberati

, Tetzlaff

, Altman

; The PRISMA Group. 2009. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med, 6:e1000097.

20.

World Health Organization. Cardiovascular diseases fact sheet. Available at: www.who.int/mediacentre/factsheets/fs317/en (accessed December 3, 2017 ).

21.

Inker

, Astor

, Fox

, Isakova

, Lash

, Peralta

, Kurella Tamura

, Feldman

. 2014. KDOQI US commentary on the 2012 KDIGO clinical practice guideline for the evaluation and management of CKD. Am J Kidney Dis, 63:713–735.

22.

Tohidi

, Derakhshan

, Akbarpour

, Amouzegar

, Mehran

, Baghbani-Oskouei

, Azizi

, Hadaegh

. 2018. Thyroid dysfunction states and incident cardiovascular events: the Tehran Thyroid Study. Horm Metab Res, 50:37–43.

23.

Segna

, Méan

, Limacher

, Baumgartner

, Blum

, Beer

, Kucher

, Righini

, Matter

, Frauchiger

, Cornuz

, Aschwanden

, Banyai

, Osterwalder

, Husmann

, Egloff

, Staub

, Lämmle

, Angelillo-Scherrer

, Aujesky

, Rodondi

. 2016. Association between thyroid dysfunction and venous thromboembolism in the elderly: a prospective cohort study. J Thromb Haemost, 14:685–694.

24.

Afsar

, Yilmaz

, Siriopol

, Unal

, Saglam

, Karaman

, Gezer

, Sonmez

, Eyileten

, Aydin

, Hamcan

, Oguz

, Covic

, Kanbay

. 2016. Thyroid function and cardiovascular events in chronic kidney disease patients. J Nephrol, 30:235–242.

25.

Hayashi

, Hasegawa

, Kanzaki

, Funada

, Amaki

, Takahama

, Ohara

, Sugano

, Yasuda

, Ogawa

, Anzai

. 2016. Subclinical hypothyroidism is an independent predictor of adverse cardiovascular outcomes in patients with acute decompensated heart failure. ESC Heart Fail, 3:168–176.

26.

Martin

, Daya

, Lutsey

, Matsushita

, Fretz

, McEvoy

, Blumenthal

, Coresh

, Greenland

, Kottgen

, Selvin

. 2015. Thyroid function, cardiovascular risk factors, and incident atherosclerotic cardiovascular disease: the Atherosclerosis Risk in Communities (ARIC) study. J Clin Endocrinol Metab, 102:3306–3315.

27.

Laulund

, Nybo

, Brix

, Abrahamsen

, Jørgensen

, Hgedüs

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

28.

, Yang

, Wang

, Ding

, Wang

, Hua

. 2014. The prevalence and prognostic effects of subclinical thyroid dysfunction in dilated cardiomyopathy patients: a single-center cohort study. J Card Fail, 20:506–512.

29.

van de Ven

, Netea-Maier

, de Vegt

, Ross

, Sweep

, Kiemeney

, Smit

, Hermus

, den Heijer

. 2014. Associations between thyroid function and mortality: the influence of age. Eur J Endocrinol, 17:183–191.

30.

Drechsler

, Schneider

, Gutjahr-Lengsfeld

, Kroiss

, Carrero

, Krane

, Allolio

, Wanner

, Fassnacht

. 2014. Thyroid function, cardiovascular events, and mortality in diabetic hemodialysis patients. Am J Kidney Dis, 63:988–996.

31.

Yeap

, Alfonso

, Hankey

, Flicker

, Golledge

, Norman

, Chubb

. 2013. Higher free thyroxine levels are associated with all-cause mortality in euthyroid older men: the Health In Men Study. Eur J Endocrinol, 169:401–408.

32.

Ceresini

, Ceda

, Lauretani

, Maggio

, Usberti

, Marina

, Bandinelli

, Guralnik

, Valenti

, Ferrucci

. 2013. Thyroid status and 6-year mortality in elderly people living in a mildly iodine-deficient area: the aging in the Chianti Area Study. J Am Geriatr Soc, 61:868–874.

33.

Hyland

, Arnold

, Lee

, Cappola

. 2013. Persistent subclinical hypothyroidism and cardiovascular risk in the elderly: the cardiovascular health study. J Clin Endocrinol Metab, 98:533–540.

34.

LeGrys

, Funk

, Lorenz

, Giri

, Jackson

, Manson

, Schectman

, Edwards

, Heiss

, Hartmann

. 2013. Subclinical hypothyroidism and risk for incident myocardial infarction among postmenopausal women. J Clin Endocrinol Metab, 98:2308–2317.

35.

Tseng

, Lin

, Lee

, Li

, Sung

, Huang

. 2012. Subclinical hypothyroidism is associated with increased risk for all-cause and cardiovascular mortality in adults. J Am Coll Cardiol, 60:730–737.

36.

Molinaro

, Iervasi

, Lorenzoni

, Coceani

, Landi

, Srebot

, Mariani

, L'Abbate

, Pingitore

. 2012. Persistence of mortality risk in patients with acute cardiac diseases and mild thyroid dysfunction. Am J Med Sci, 343:65–70.

37.

Nanchen

, Gussekloo

, Westendorp

, Stott

, Jukema

, Trompet

, Ford

, Welsh

, Sattar

, Macfarlane

, Mooijaart

, Rodondi

, de Craen

. 2012. Subclinical thyroid dysfunction and the risk of heart failure in older persons at high cardiovascular risk. Clin Endocrinol Metab, 97:852–861.

38.

Waring

, Arnold

, Newman

, Bùzková

, Hirsch

, Cappola

. 2012. Longitudinal changes in thyroid function in the oldest old and survival: the cardiovascular health study all-stars study. J Clin Endocrinol Metab, 97:3944–3950.

39.

Frey

, Kroiss

, Berliner

, Seifert

, Allolio

, Güder

, Ertl

, Angermann

, Störk

, Fassnacht

. 2013. Prognostic impact of subclinical thyroid dysfunction in heart failure. Int J Cardiol, 168:300–305.

40.

de Jongh

, Lips

, van Schoor

, Rijs

, Deeg

, Comijs

, Kramer

, Vandenbroucke

, Dekkers

. 2011. Endogenous subclinical thyroid disorders, physical and cognitive function, depression, and mortality in older individuals. Eur J Endocrinol, 165:545–554.

41.

Schultz

, Kistorp

, Raymond

, Dimsits

, Tuxen

, Hildebrandt

, Faber

. 2011. Cardiovascular events in thyroid disease: a population based, prospective study. Horm Metab Res, 43:653–659.

42.

Sgarbi

, Matsumura

, Kasamatsu

, Ferreira

, Maciel

. 2010. Subclinical thyroid dysfunctions are independent risk factors for mortality in a 7.5-year follow-up. Eur J Endocrinol, 162:569–577.

43.

Razvi

, Weaver

, Vanderpump

, Pearce

. 2010. The incidence of ischemic heart disease and mortality in people with subclinical hypothyroidism: reanalysis of the Whickham Survey cohort. J Clin Endocrinol Metab, 95:1734–1740.

44.

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.

45.

Asvold

, Bjøro

, Nilsen

, Gunnell

, Vatten

. 2008. Thyrotropin levels and risk of fatal coronary heart disease. Arch Intern Med, 168:855–860.

46.

Chen

, Wu

, Jap

, Lu

, Wang

, Chen

, Lin

. 2007. Subclinical hypothyroidism is a risk factor for nephropathy and cardiovascular diseases in type 2 diabetic patients. Diabet Med, 24:1336–1344.

47.

Cappola

, Fried

, Arnold

, Danese

, Kuller

, Burke

, Tracy

, Ladenson

. 2006. Thyroid status, cardiovascular risk, and mortality in older adults. JAMA, 295:1033–1041.

48.

Rodondi

, Bauer

, Cappola

, Cornuz

, Robbins

, Fried

, Ladenson

, Vittinghoff

, Gottdiener

, Newman

. 2008. Subclinical thyroid dysfunction, cardiac function, and the risk of heart failure. J Am Coll Cardiol, 52:1152–1159.

49.

Walsh

, Bremner

, Bulsara

, O'Leary

, Leedman

, Feddema

, Michelangeli

. 2005. Subclinical thyroid dysfunction as a risk factor for cardiovascular disease. Arch Intern Med, 165:2467–2472.

50.

Gussekloo

, van Exel

, de Craen

, Meinders

, Frolich

, Westendorp

. 2004. Thyroid status, disability and cognitive function, and survival in old age. JAMA, 292:2591–2599.

51.

Parle

, Maisonneuve

, Sheppard

, Boyle

, Franklyn

. 2001. Prediction of all-cause and cardiovascular mortality in elderly people from one low serum thyrotropin result: a 10-year cohort study. Lancet, 358:861–865.

52.

McQuade

, Skugor

, Brennan

, Hoar

, Stevenson

, Hoogwerf

. 2011. Hypothyroidism and moderate subclinical hypothyroidism are associated with increased all-cause mortality independent of coronary heart disease risk factors: a PreCIS database study. Thyroid, 21:837–843.

53.

Pearce

, Brabant

, Duntas

, Monzani

, Peeters

, Razvi

, Wemeau

. 2013. 2013 ETA guideline: management of subclinical hypothyroidism. Eur Thyroid J, 2:215–228.

54.

Vanderpump

. 2011. The epidemiology of thyroid disease. Br Med Bull, 99:39–51.

55.

Somwaru

, Rariy

, Arnold

, Cappola

. 2012. The natural history of subclinical hypothyroidism in the elderly: the cardiovascular health study. J Clin Endocrinol Metab, 97:1962–1969.

56.

Hollowell

, Staehling

, Flanders

, Hannon

, Gunter

, Spencer

, Braverman

. 2002. Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab, 87:489–499.

57.

Atzmon

, Barzilai

, Hollowell

, Surks

, Gabriely

. 2009. Extreme longevity is associated with increased serum thyrotropin. J Clin Endocrinol Metab, 94:1251–1254.

58.

Pearce

, Razvi

, Yadegarfar

, Martin-Ruiz

, Kingston

, Collerton

, Visser

, Kirkwood

, Jagger

. 2016. Serum thyroid function, mortality and disability in advanced old age: the Newcastle 85+ Study. J Clin Endocrinol Metab, 101:4385–4394.

59.

Cho

, Choi

, Kim

, Lee

, Choi

, Lim

, Park

, Jang

, Cho

. 2010. Interaction between cigarette smoking and iodine intake and their impact on thyroid function. Clin Endocrinol (Oxf), 73:264–270.

60.

Surks

, Ortiz

, Daniels

, Sawin

, Col

, Cobin

, Franklyn

, Hershman

, Burman

, Denke

, Gorman

, Cooper

, Weissman

. 2004. Subclinical thyroid disease: scientific review and guidelines for diagnosis and management. JAMA, 291:228–238.

61.

Andersen

, Pedersen

, Bruun

, Laurberg

. 2002. Narrow individual variations in serum T(4) and T(3) in normal subjects: a clue to the understanding of subclinical thyroid disease. J Clin Endocrinol Metab, 87:1068–1072.

62.

Surks

, Hollowell

. 2007. Age-specific distribution of serum thyrotropin and antithyroid antibodies in the US population: implications for the prevalence of subclinical hypothyroidism. J Clin Endocrinol Metab, 92:4575–4582.