Cardiovascular disease research should focus more on patients’ T3 levels than TSH

Science has already shown that at T3 is more important than TSH to the functional parameters of the cardiovascular system:

• At the molecular and tissue level, the T3 hormone directly affects the cardiovascular system in a myriad of important ways (see “Low T3’s effects on the cardiovascular system.”)

Why, then, when researchers shift their focus from the molecular level to the “epidemiological” and “clinical” level (the study of real populations of patients), do they shift their focus away from T3?

• When studying patients, research tends to use TSH to categorize patients.
• When studying patients, their T3 levels are not examined in relation to health factors.

The only major exception to this generalization is the study of “Low T3 syndrome” / “non-thyroidal illness” in cardiovascular syndromes, which has proven that low T3 can be very harmful to human health.

This creates a fundamental problem in research.

• If T3 is the main hormone that influences the cardiovascular system,
• If T3 deficiency can cause varying degrees of cardiovascular harm,
• Then T3 should also be the main hormone that is analyzed in relation to cardiovascular health, not TSH.

TSH and cardiovascular biomarkers

First, let’s look at the very broad associations between patients’ TSH and cardiovascular disease, focusing on the extremes between “normal” TSH and “high” TSH.

In Klein & Danzi, 2016 table 2, they give specific numbers to quantify the degree of hypothyroidism in the cardiovascular system, most likely defining hypothyroidism by TSH, since no other definition was specified:

• Systemic vascular resistance (dyne-cm/s-5)
  • NORMAL: 1500-1700. HYPO: 2100-2700
• Heart rate (beats per minute)
  • NORMAL: 72-84. HYPO: 60-80 (measure at rest)
• Ejection fraction (%)
  • NORMAL: 60%. HYPO: < 60%
• IVRT (isovolumic relaxation time) (milliseconds)
  • NORMAL: 60-80. HYPO: >80
• Cardiac output (liter / minute)
  • NORMAL 5.8. HYPO: 4.5
• Blood volume (% of normal)
  • NORMAL: 100%. HYPO: 84.5

How would these definitions and values look if used T3 was used rather than TSH levels?

Does a given category or level of TSH always mean the same degree of T3 deficit or excess?

Some of the values in the list above appear to depend on averages across a broad TSH-defined cohort. Therefore, the average “hypothyroid” level would depend on the characteristics of the population that was included in the study as well as the upper TSH cutoff, which varied over the years.

Another study, Martin et al, 2017, which examined over 11,000 patients, gave average values relative to supposedly “Euthyroid” patients.  Using TSH categories, they subcategorized patients into “mild” (MILD) and “moderate or severe” (SEVERE) hypothyroidism. These were the statistically significant (p = <0.001) items in their Table 2:

• LDL-C  (mg/dL) MILD +3.2, SEVERE +15.1
• Triglycerides (mg/dL) MILD +4.6, SEVERE +24.5
• Diastolic blood pressure (mm Hg) MILD +0.3, SEVERE +2.0
• Heart rate (beats per minute) MILD -0.7, SEVERE -1.7

Again, we do not know which T3 levels make a person “mildly” or “severely” hypothyroid.

Broad TSH categories are biologically irrelevant to the cardiovascular system, which is sensitive to subtle changes in blood T3 levels as well as tissue T3 levels.

As a third example, in a review article by Razvi, et al, 2018, TSH categories were broadly associated with cardiovascular risk factors.

• “Hypothyroidism is associated with a small but significant increase” in total cholesterol and LDL cholesterol levels (p. 1786), whereas hyperthyroidism showed the opposite trend.
• Both hyper- and hypothyroid patients can have increased carotid artery intimal-media thickness; in hypothyroidism it may be caused by carotid intima hyperplasia as well as by atherosclerosis.
• Both hyper- and hypothyroid patients can have “increased arterial stiffness” (Table 2).
• Both overt hyper- and hypothyroid patients can have “diastolic dysfunction” (Table 2)
• Both “TH [thyroid hormone] deficiency and excess can alter the coagulation pathway [blood clot formation], although the precise clinical relevance of this finding is unclear.”

This list includes several biomarkers that appear to be similar on both sides of the hyper- and hypo spectrum.

Because their review article so frequently blurred the distinction between thyroid hormone (T3) status and TSH status and presumed them to be largely interchangeable or equivalent, it is difficult to know which factors were likely caused by higher T3 or lower T3.

The review frequently mentioned that thyroid hormone replacement “resolves” cardiovascular disorders or risk factors, without saying to what degree it resolved them or whether it normalized T3 levels.

T3 and cardiovascular biomarkers

In contrast, let’s look at one recent study that actually analyzed patients’ T3 levels across the entire normal T3 reference range. It was a promising study, but it was disappointing in the end.

Martin et al, 2017 , in addition to categorizing patients by TSH (as shown above), analyzed cardiovascular risk parameters in relation to patients’ Total T3 levels (not Free T3).

They proudly highlighted their inclusion of T3 hormone levels as an uniquely important feature of their study:

Triiodothyronine (T3) has been unavailable in most epidemiologic studies to date.

T3 is considered the bioactive form of thyroid hormone that mediates peripheral effects and therefore of specific interest with respect to risk for CVD events.”

A single cohort study with standardized procedures, long-term follow-up, and a comprehensive thyroid panel (inclusive of T3), would be highly informative.

Only one such study has been done that examined thyroid function and adverse outcomes in 2843 adults enrolled in the Cardiovascular Health Study, with thyrotropin (TSH), free thyroxine (FT4), and T3 concentrations in the euthyroid range.  (p. 3307)

NOTE: There is an error in their article. “Only one such study” refers to item 6 in their reference list by Chaker et al, 2015. However, the study they refer to is actually Chaker et al, 2016, with a similar title.  In fact, NEITHER article by Chaker et al included an analysis of T3 hormone levels.

The T3 section of Miller et al’s study gave a much more nuanced picture of the relationship between the most important thyroid hormone and cardiovascular risk.

However, they excluded from the analysis all levels below or above the normal T3 reference range.

In addition, they excluded patients on drugs that would modify their risk factors (i.e. statin drugs, when analyzing cholesterol).

And because their entire study focused on endogenous T3 only, they also excluded all patients on thyroid medication.

In their study, they confirmed that two biomarkers had a very strong linear relationship to T3:

1. Heart rate. (Lower T3 patients had lower heart rate.)
2. Non-HDL cholesterol. (Lower T3 patients had higher LDL cholesterol, while higher T3 patients had lower cholesterol.)

In contrast, several other cardiovascular risk parameters showed not a linear pattern but rather a U-shaped or L-shaped pattern on a set of graphs (Figure 1).

• BOTH had higher HDL-cholesterol,
• BOTH had lower Systolic and Diastolic blood pressure,
• BOTH had higher Hemoglobin A1c, and
• BOTH had higher Blood Glucose.

Therefore, patients with Higher T3 and Lower T3 shared several “increased cardiovascular risk” factors compared to those with mid-range T3.

Unfortunately, the researchers concluded that there was “no significant interaction” between T3 and these health factors, since they were looking for a linear relationship in order to determine a hazard ratio (HR) risk calculation.

But why must the relationship be forced into a linear mold? We already know that hypothyroidism and hyperthyroidism share several disorders in common, and it’s a false oversimplification to say that they are complete opposites.

• hair loss
• heart failure
• osteoporosis
• fatigue
• cognitive difficulty
• muscle weakness

It appears that similar aberrations are involved in both extremes of higher and lower T3. Is that not instructive in any way?

Perhaps the body is attempting to compensate for both T3 aberrations in a complex manner and therefore the end result (the variable measured) is similar on both ends of the spectrum.

Interestingly, there was a common pattern across many of the graphs. It appeared that certain biomarkers were more focused on lower T3 than higher T3.  The shaded area representing the range of data was often more broad at the extreme edge of higher T3, showing that an indicator varied more widely among higher-T3 patients than it did among lower-T3 patients.

Unfortunately, this aspect of their study was confined to the normal T3 reference range. As a result, the reader could not see the data of patients with extremely low baseline T3 or levels above reference range. We cannot see the full direction of the trend lines and the spread of data points.

In contrast, the first part of their article included all TSH values below and above the reference range.

These T3 data are real human data, just like TSH data. If there was a concern about distorting the data visually because there were only a few abnormal outliers, then the researchers should have presented a scatterplot instead of a shaded area, and a table could have included more precise data than a graph.

T3 levels and long-term cardiovascular risk

Next, when Martin et al analyzed baseline T3 levels and incidence of myocardial infraction and stroke in the ensuing 22.5 years,

• Lower T3 and Higher T3 patients both had more incidents of myocardial infarction than the norm.
• Lower T3 patients appeared to have a normal risk of stroke, but higher T3 levels significantly reduced the risk of stroke.

Most importantly for the long-term outcomes of stroke or myocardial infarction,

• These risks were assessed by patients’ baseline (1990) T3 levels, not by later T3 levels measured throughout 22.5 years.
• In contrast with the other T3 data, this data set included thyroid patients on therapy. Among the participants who were mildly or severely hypothyroid in TSH, many had begun thyroid hormone replacement therapy, and 78% of them no longer had an elevated TSH by visit 5 (2011-2013).

Therefore, it is impossible for a reader to determine whether the risk of stroke or MI shown in their graphs and tables were wholly dependent on the “15-20%” of patients whose T3 was not normalized by the 23-year follow up point, or whether the risks were present even among those whose T3 had been normalized by therapy for many years or decades.

A faulty conclusion

In the end, it appears that Martin et al jumped to an unfounded and largely irrelevant conclusion in the final sentence of their study:

“Persons with various degrees of thyroid function, by clinical categories or concentrations of T3, had cardiovascular outcomes over two decades of followup that were similar to those with normal thyroid function.

These results support existing guideline recommendations and provide reassurance that clinical management of moderate or severe thyroid disease in the United States is effectively mitigating any long-term cardiovascular consequences.” (p. 3314)

These broad claims have insufficient basis in their own research study.

• Their study did, in fact, show effects of T3 levels that differed from the norm in both linear and non-linear ways. They could only be “similar” if the real differences were mathematically averaged away into statistical similarity.
• It is illogical to make a claim about thyroid therapy in a study that excluded patients on thyroid therapy from the T3 hormone effect analyses.
• Their study could have, but did not, relate patients’ TSH categories to patients’ T3 levels. The TSH section of the paper was separated from the T3 section of the paper and the study did not examine their interrelationships.

In conclusion, even though this study did something very important by analyzing T3 levels, unfortunately the researchers appeared to be biased by the TSH-based categories at the very foundation of their study.

The researchers appeared to cling to the belief that TSH category, not T3 level, determined risk, despite the fact that, as they had stated earlier, the T3 hormone, is the “bioactive” hormone.

More T3- focused research needed

It is time to bring T3 as a key measure into more clinical studies of cardiovascular diseases and their risk factors:

• Study patients on thyroid therapy with or without a diagnosis of cardiovascular disease.
• Study thyroid patients on all sorts of therapies, not just Levothyroxine monotherapy.
• Study patients without thyroid glands (less than 0.5 mL thyroid volume) so as to discern whether the T3 is thyroidally secreted or not.
• Study autoimmune thyroid disease patients, and categorize them by their thyroid gland health (do an ultrasound to determine gland size and echogenicity), again to determine the degree to which their risk is based on their own gland’s production of thyroid hormone.
• Divide patients by tertiles, quartiles or quintiles according to their T3 levels, and try to avoid being biased by the edges of the statistical normal reference range when setting up these divisions.
• Collect T3 data frequently over the long term and record any changes in ongoing health status.
• Avoid being trapped by dualistic thinking that presumes that high and low T3 must always have opposite effects. Try to understand all patterns in the data and not just linear relationships.
• Keep an open and curious mind rather than just trying to prove a hypothesis.

And most urgently,

• Stop excluding human beings from studies of T3 and cardiovascular disease just because they are taking thyroid medications.

No matter where the T3 supply comes from, the human body still needs sufficient T3.

Thyroid patients have been excluded from studies of “non-thyroidal illness” (Low T3 syndrome) for far too long. Many of us still suffer from low or suboptimal T3 under thyroid therapy. Thyroid patients may suffer more harm from smaller variations in T3 levels if our body is unable to adapt to them (see “Thyroid patients excluded from research“).