Subclinical Hypothyroidism Tied to Increased Risk of Heart Disease Events, Deaths

Subclinical hypothyroidism appears to increase the risk of coronary heart events by up to 89% and coronary heart death by up to 58%, adding more fuel to the debate about whether to treat low-level thyroid dysfunction.

But although the findings clarify the level of heart disease risk associated with subclinical hypothyroidism, only a treatment trial can fully answer the question of who – and when – to treat, said Dr. Nicolas Rodondi, the study’s primary investigator.

“Now that we have clearly shown the increased risk and what we should test for, we need to know if we can decrease this risk by treating,” he said in an interview. “Unfortunately, our paper does not totally solve this issue. For that, we need a randomized, controlled interventional trial.”

Dr. Rodondi, an epidemiologist and internist at the University of Lausanne, Switzerland, presented the results Sept. 14 at the International Thyroid Congress in Paris. The full paper appeared in the Sept. 21 issue of JAMA (2010;304:1365-74).

The meta-analysis, comprising more than 55,000 patients, was intended to clarify an issue which until now has had no clear-cut answers, Dr. Rodondi said. “Before this study, there were very conflicting data as to whether there was any risk at all. Current guidelines mention this ‘magic’ cut-off number of 10 mIU/L [thyroid stimulating hormone level] as the level to have concern of increased risk for heart problems. But this number has never been supported by a lot of data. It’s been based mostly on expert opinion.”

He and his colleagues reviewed 11 prospective studies with nearly 543,000 person-years of follow-up. Because the studies used varying cut-off levels for subclinical hypothyroidism, the group defined the condition as described in the Cardiovascular Health Study: a serum thyroid stimulating hormone (TSH) of 4.5 mIU/L to 19.9 mIU/L with a normal free thyroxine (T4) level. Euthyroidism was defined as a TSH of 0.5 mIU/L to 4.49 mIU/L. Coronary heart disease events, coronary heart disease death, and overall mortality were the primary end points.

Of the 55,287 adults included in the meta-analysis, 3,450 (6%) had subclinical hypothyroidism; the rest were euthyroid. The rate of thyroid hormone replacement therapy at baseline varied among the studies, from 0% to 8%. In some studies, up to 12% of patients were taking thyroid hormone during the follow-up period. All 11 of the papers reported total and coronary heart disease mortality, while 7 also reported coronary heart disease events.

During the follow-up periods, which ranged from 2.5 to 20 years, 9,664 patients died; 2,168 of those from coronary heart disease (CHD). There were also 4,470 CHD events in the studies that examined this end point.

In an overall analysis of subclinical hypothyroidism vs. euthryoidism, adjusted for age and sex, there were no significant differences in the risk of CHD events or death, or total mortality.

However, significant differences appeared when the investigators examined these risks according to the degree of subclinical hypothyroidism: TSH 4.5-6.9 mIU/L; 7-9.9 mIU/L; and 10-19.9 mIU/L.

There were no significant between-group differences in the risk of overall mortality in any of the three TSH levels. “The finding of no increased risk of CHD among the high proportions of adults with minimal TSH elevations is also important because many patients with minimal TSH elevations are currently treated in clinical practice,” Dr. Rodondi and his colleagues noted.

However, those with a TSH of 7.0-9.9 mIU/L were a significant 42% more likely to die from CHD than were euthyroid patients.

The biggest differences emerged in the group with the highest TSH levels. The risk for CHD events was significantly higher than in euthyroid patients (hazard ratio 1.89). In this group of 235 patients, there were 70 events, for a rate of 38/1,000 person-years, compared with 20/1,000 person-years in euthyroid patients.

The highest TSH group also had a significantly increased risk of death from CHD (HR 1.58). There were 28 such deaths in that group of 333 patients, for an overall rate of 8/1,000 person-years, compared with 5 in the euthyroid patients. Although younger patients with elevated TSH appeared to have slightly higher risks, there were no significant overall associations with age.

The findings were essentially unchanged in a variety of sensitivity analyses that took into account such factors as excluding patients taking thyroid medication, adjustment for cardiovascular risk factors and drugs to manage those risks, and studies that included only cardiac patients.

The mechanism by which subclinical hypothyroidism could increase the risk of heart disease is not fully understood, Dr. Rodondi and his colleagues wrote. “Increased systemic vascular resistance, arterial stiffness, altered endothelial function, increased atherosclerosis, and altered coagulability have been reported to be associated with subclinical hypothyroidism and may accelerate CHD,” they noted. “The fact that adjustments for traditional cardiovascular risk factors did not alter risks could favor this hypothesis.”

Now that firmer data are established, the questions of screening and treatment remain to be answered, Dr. Rodondi said in the interview. Population-based screening is not warranted, but screening might someday be useful in specific groups – older patients, for example.

“Subclinical hypothyroidism is a very common finding, especially among older patients, with a prevalence of about 5% at age 50 [years] and 10% by age 65. But if we do screen these patients, and find abnormal levels, what do we do? At this point, we still don’t know,” he noted.

He also pointed out that there is no institutional support for any screening test without evidence that treatment can improve outcomes. “This is why we need an intervention trial,” he said. In fact, he and his colleagues are planning such a trial. Coauthor Dr. Douglas Bauer of the University of California, San Francisco, has secured a National Institutes of Health grant to begin its planning.

“It is likely to randomize older adults with subclinical hypothyroidism [to treatment or none] and examine not only cardiovascular disease, but musculoskeletal and cognitive outcomes as well,” Dr. Rodondi said.

In the meantime, physicians will still be left to weigh the existing evidence and apply it to each individual patient in making treatment decisions. “If a patient has a TSH above 10 mIU/L, we probably should be concerned that this person is at increased risk,” Dr. Rodondi said. “Our meta-analysis did not prove that it’s useful to treat that patient, but it is now clear that the person is likely to have a worse outcome than someone with a normal or near-normal TSH.”

Commenting on the study, Dr. Hossein Gharib said that it “confirms what is already known and applied by clinical endocrinologists – that a TSH of more than 10 mIU/L is clearly bad for the heart.” Furthermore, “it emphasizes that a borderline TSH between 5 and 10 mIU/L can also be bad, especially in a patient with other risk factors, such as antithyroid antibodies, goiter, hyperlipidemia, or pregnancy. Most clinical endocrinologists in the U.S. and in Europe would choose to treat these patients with thyroxine.”

Dr. Gharib’s personal practice “has been, and will continue to be, to favor treatment for these patients over no treatment.” Most clinical endocrinologists would agree in favor of treating a patient with a TSH of 6-10 mIU/L, he added, “once it has been confirmed on two separate occasions and in the presence of these other risk factors.”

A treatment trial such as the one Dr. Rodondi describes “will answer some of the concerns that still persist and is certainly desirable,” said Dr. Gharib, professor of medicine at the Mayo Clinic, Rochester, Minn. But until then, individual clinical judgment should override any blanket recommendations.

The study was sponsored by the National Institutes of Health. None of the authors reported any financial conflicts.

Subclinical hypothyroidism appears to increase the risk of coronary heart events by up to 89% and coronary heart death by up to 58%, adding more fuel to the debate about whether to treat low-level thyroid dysfunction.

But although the findings clarify the level of heart disease risk associated with subclinical hypothyroidism, only a treatment trial can fully answer the question of who – and when – to treat, said Dr. Nicolas Rodondi, the study’s primary investigator.

“Now that we have clearly shown the increased risk and what we should test for, we need to know if we can decrease this risk by treating,” he said in an interview. “Unfortunately, our paper does not totally solve this issue. For that, we need a randomized, controlled interventional trial.”

Dr. Rodondi, an epidemiologist and internist at the University of Lausanne, Switzerland, presented the results Sept. 14 at the International Thyroid Congress in Paris. The full paper appeared in the Sept. 21 issue of JAMA (2010;304:1365-74).

The meta-analysis, comprising more than 55,000 patients, was intended to clarify an issue which until now has had no clear-cut answers, Dr. Rodondi said. “Before this study, there were very conflicting data as to whether there was any risk at all. Current guidelines mention this ‘magic’ cut-off number of 10 mIU/L [thyroid stimulating hormone level] as the level to have concern of increased risk for heart problems. But this number has never been supported by a lot of data. It’s been based mostly on expert opinion.”

He and his colleagues reviewed 11 prospective studies with nearly 543,000 person-years of follow-up. Because the studies used varying cut-off levels for subclinical hypothyroidism, the group defined the condition as described in the Cardiovascular Health Study: a serum thyroid stimulating hormone (TSH) of 4.5 mIU/L to 19.9 mIU/L with a normal free thyroxine (T4) level. Euthyroidism was defined as a TSH of 0.5 mIU/L to 4.49 mIU/L. Coronary heart disease events, coronary heart disease death, and overall mortality were the primary end points.

Of the 55,287 adults included in the meta-analysis, 3,450 (6%) had subclinical hypothyroidism; the rest were euthyroid. The rate of thyroid hormone replacement therapy at baseline varied among the studies, from 0% to 8%. In some studies, up to 12% of patients were taking thyroid hormone during the follow-up period. All 11 of the papers reported total and coronary heart disease mortality, while 7 also reported coronary heart disease events.

During the follow-up periods, which ranged from 2.5 to 20 years, 9,664 patients died; 2,168 of those from coronary heart disease (CHD). There were also 4,470 CHD events in the studies that examined this end point.

In an overall analysis of subclinical hypothyroidism vs. euthryoidism, adjusted for age and sex, there were no significant differences in the risk of CHD events or death, or total mortality.

However, significant differences appeared when the investigators examined these risks according to the degree of subclinical hypothyroidism: TSH 4.5-6.9 mIU/L; 7-9.9 mIU/L; and 10-19.9 mIU/L.

There were no significant between-group differences in the risk of overall mortality in any of the three TSH levels. “The finding of no increased risk of CHD among the high proportions of adults with minimal TSH elevations is also important because many patients with minimal TSH elevations are currently treated in clinical practice,” Dr. Rodondi and his colleagues noted.

However, those with a TSH of 7.0-9.9 mIU/L were a significant 42% more likely to die from CHD than were euthyroid patients.

The biggest differences emerged in the group with the highest TSH levels. The risk for CHD events was significantly higher than in euthyroid patients (hazard ratio 1.89). In this group of 235 patients, there were 70 events, for a rate of 38/1,000 person-years, compared with 20/1,000 person-years in euthyroid patients.

The highest TSH group also had a significantly increased risk of death from CHD (HR 1.58). There were 28 such deaths in that group of 333 patients, for an overall rate of 8/1,000 person-years, compared with 5 in the euthyroid patients. Although younger patients with elevated TSH appeared to have slightly higher risks, there were no significant overall associations with age.

The findings were essentially unchanged in a variety of sensitivity analyses that took into account such factors as excluding patients taking thyroid medication, adjustment for cardiovascular risk factors and drugs to manage those risks, and studies that included only cardiac patients.

The mechanism by which subclinical hypothyroidism could increase the risk of heart disease is not fully understood, Dr. Rodondi and his colleagues wrote. “Increased systemic vascular resistance, arterial stiffness, altered endothelial function, increased atherosclerosis, and altered coagulability have been reported to be associated with subclinical hypothyroidism and may accelerate CHD,” they noted. “The fact that adjustments for traditional cardiovascular risk factors did not alter risks could favor this hypothesis.”

Now that firmer data are established, the questions of screening and treatment remain to be answered, Dr. Rodondi said in the interview. Population-based screening is not warranted, but screening might someday be useful in specific groups – older patients, for example.

“Subclinical hypothyroidism is a very common finding, especially among older patients, with a prevalence of about 5% at age 50 [years] and 10% by age 65. But if we do screen these patients, and find abnormal levels, what do we do? At this point, we still don’t know,” he noted.

He also pointed out that there is no institutional support for any screening test without evidence that treatment can improve outcomes. “This is why we need an intervention trial,” he said. In fact, he and his colleagues are planning such a trial. Coauthor Dr. Douglas Bauer of the University of California, San Francisco, has secured a National Institutes of Health grant to begin its planning.

“It is likely to randomize older adults with subclinical hypothyroidism [to treatment or none] and examine not only cardiovascular disease, but musculoskeletal and cognitive outcomes as well,” Dr. Rodondi said.

In the meantime, physicians will still be left to weigh the existing evidence and apply it to each individual patient in making treatment decisions. “If a patient has a TSH above 10 mIU/L, we probably should be concerned that this person is at increased risk,” Dr. Rodondi said. “Our meta-analysis did not prove that it’s useful to treat that patient, but it is now clear that the person is likely to have a worse outcome than someone with a normal or near-normal TSH.”

Commenting on the study, Dr. Hossein Gharib said that it “confirms what is already known and applied by clinical endocrinologists – that a TSH of more than 10 mIU/L is clearly bad for the heart.” Furthermore, “it emphasizes that a borderline TSH between 5 and 10 mIU/L can also be bad, especially in a patient with other risk factors, such as antithyroid antibodies, goiter, hyperlipidemia, or pregnancy. Most clinical endocrinologists in the U.S. and in Europe would choose to treat these patients with thyroxine.”

Dr. Gharib’s personal practice “has been, and will continue to be, to favor treatment for these patients over no treatment.” Most clinical endocrinologists would agree in favor of treating a patient with a TSH of 6-10 mIU/L, he added, “once it has been confirmed on two separate occasions and in the presence of these other risk factors.”

A treatment trial such as the one Dr. Rodondi describes “will answer some of the concerns that still persist and is certainly desirable,” said Dr. Gharib, professor of medicine at the Mayo Clinic, Rochester, Minn. But until then, individual clinical judgment should override any blanket recommendations.

The study was sponsored by the National Institutes of Health. None of the authors reported any financial conflicts.

Subclinical hypothyroidism appears to increase the risk of coronary heart events by up to 89% and coronary heart death by up to 58%, adding more fuel to the debate about whether to treat low-level thyroid dysfunction.

But although the findings clarify the level of heart disease risk associated with subclinical hypothyroidism, only a treatment trial can fully answer the question of who – and when – to treat, said Dr. Nicolas Rodondi, the study’s primary investigator.

“Now that we have clearly shown the increased risk and what we should test for, we need to know if we can decrease this risk by treating,” he said in an interview. “Unfortunately, our paper does not totally solve this issue. For that, we need a randomized, controlled interventional trial.”

Dr. Rodondi, an epidemiologist and internist at the University of Lausanne, Switzerland, presented the results Sept. 14 at the International Thyroid Congress in Paris. The full paper appeared in the Sept. 21 issue of JAMA (2010;304:1365-74).

The meta-analysis, comprising more than 55,000 patients, was intended to clarify an issue which until now has had no clear-cut answers, Dr. Rodondi said. “Before this study, there were very conflicting data as to whether there was any risk at all. Current guidelines mention this ‘magic’ cut-off number of 10 mIU/L [thyroid stimulating hormone level] as the level to have concern of increased risk for heart problems. But this number has never been supported by a lot of data. It’s been based mostly on expert opinion.”

He and his colleagues reviewed 11 prospective studies with nearly 543,000 person-years of follow-up. Because the studies used varying cut-off levels for subclinical hypothyroidism, the group defined the condition as described in the Cardiovascular Health Study: a serum thyroid stimulating hormone (TSH) of 4.5 mIU/L to 19.9 mIU/L with a normal free thyroxine (T4) level. Euthyroidism was defined as a TSH of 0.5 mIU/L to 4.49 mIU/L. Coronary heart disease events, coronary heart disease death, and overall mortality were the primary end points.

Of the 55,287 adults included in the meta-analysis, 3,450 (6%) had subclinical hypothyroidism; the rest were euthyroid. The rate of thyroid hormone replacement therapy at baseline varied among the studies, from 0% to 8%. In some studies, up to 12% of patients were taking thyroid hormone during the follow-up period. All 11 of the papers reported total and coronary heart disease mortality, while 7 also reported coronary heart disease events.

During the follow-up periods, which ranged from 2.5 to 20 years, 9,664 patients died; 2,168 of those from coronary heart disease (CHD). There were also 4,470 CHD events in the studies that examined this end point.

In an overall analysis of subclinical hypothyroidism vs. euthryoidism, adjusted for age and sex, there were no significant differences in the risk of CHD events or death, or total mortality.

However, significant differences appeared when the investigators examined these risks according to the degree of subclinical hypothyroidism: TSH 4.5-6.9 mIU/L; 7-9.9 mIU/L; and 10-19.9 mIU/L.

There were no significant between-group differences in the risk of overall mortality in any of the three TSH levels. “The finding of no increased risk of CHD among the high proportions of adults with minimal TSH elevations is also important because many patients with minimal TSH elevations are currently treated in clinical practice,” Dr. Rodondi and his colleagues noted.

However, those with a TSH of 7.0-9.9 mIU/L were a significant 42% more likely to die from CHD than were euthyroid patients.

The biggest differences emerged in the group with the highest TSH levels. The risk for CHD events was significantly higher than in euthyroid patients (hazard ratio 1.89). In this group of 235 patients, there were 70 events, for a rate of 38/1,000 person-years, compared with 20/1,000 person-years in euthyroid patients.

The highest TSH group also had a significantly increased risk of death from CHD (HR 1.58). There were 28 such deaths in that group of 333 patients, for an overall rate of 8/1,000 person-years, compared with 5 in the euthyroid patients. Although younger patients with elevated TSH appeared to have slightly higher risks, there were no significant overall associations with age.

The findings were essentially unchanged in a variety of sensitivity analyses that took into account such factors as excluding patients taking thyroid medication, adjustment for cardiovascular risk factors and drugs to manage those risks, and studies that included only cardiac patients.

The mechanism by which subclinical hypothyroidism could increase the risk of heart disease is not fully understood, Dr. Rodondi and his colleagues wrote. “Increased systemic vascular resistance, arterial stiffness, altered endothelial function, increased atherosclerosis, and altered coagulability have been reported to be associated with subclinical hypothyroidism and may accelerate CHD,” they noted. “The fact that adjustments for traditional cardiovascular risk factors did not alter risks could favor this hypothesis.”

Now that firmer data are established, the questions of screening and treatment remain to be answered, Dr. Rodondi said in the interview. Population-based screening is not warranted, but screening might someday be useful in specific groups – older patients, for example.

“Subclinical hypothyroidism is a very common finding, especially among older patients, with a prevalence of about 5% at age 50 [years] and 10% by age 65. But if we do screen these patients, and find abnormal levels, what do we do? At this point, we still don’t know,” he noted.

He also pointed out that there is no institutional support for any screening test without evidence that treatment can improve outcomes. “This is why we need an intervention trial,” he said. In fact, he and his colleagues are planning such a trial. Coauthor Dr. Douglas Bauer of the University of California, San Francisco, has secured a National Institutes of Health grant to begin its planning.

“It is likely to randomize older adults with subclinical hypothyroidism [to treatment or none] and examine not only cardiovascular disease, but musculoskeletal and cognitive outcomes as well,” Dr. Rodondi said.

In the meantime, physicians will still be left to weigh the existing evidence and apply it to each individual patient in making treatment decisions. “If a patient has a TSH above 10 mIU/L, we probably should be concerned that this person is at increased risk,” Dr. Rodondi said. “Our meta-analysis did not prove that it’s useful to treat that patient, but it is now clear that the person is likely to have a worse outcome than someone with a normal or near-normal TSH.”

Commenting on the study, Dr. Hossein Gharib said that it “confirms what is already known and applied by clinical endocrinologists – that a TSH of more than 10 mIU/L is clearly bad for the heart.” Furthermore, “it emphasizes that a borderline TSH between 5 and 10 mIU/L can also be bad, especially in a patient with other risk factors, such as antithyroid antibodies, goiter, hyperlipidemia, or pregnancy. Most clinical endocrinologists in the U.S. and in Europe would choose to treat these patients with thyroxine.”

Dr. Gharib’s personal practice “has been, and will continue to be, to favor treatment for these patients over no treatment.” Most clinical endocrinologists would agree in favor of treating a patient with a TSH of 6-10 mIU/L, he added, “once it has been confirmed on two separate occasions and in the presence of these other risk factors.”

A treatment trial such as the one Dr. Rodondi describes “will answer some of the concerns that still persist and is certainly desirable,” said Dr. Gharib, professor of medicine at the Mayo Clinic, Rochester, Minn. But until then, individual clinical judgment should override any blanket recommendations.

The study was sponsored by the National Institutes of Health. None of the authors reported any financial conflicts.