Response by Chaker et al to Letter Regarding Article, "Thyroid Function and Sudden Cardiac Death: A Prospective Population-Based Cohort Study"

In Response:
We would like to thank Dr Cerit and colleagues for their letter to the editor in response to our recent publication. We fully agree that the association between free thyroxine and sudden cardiac death (SCD) is puzzling. In our recent study, we describe that higher free thyroxine levels are associated with a higher risk of SCD, in both the general population and euthyroid participants. In their letter, Cerit and colleagues suggest that high-density lipoprotein (HDL) cholesterol and lipoprotein (a) [Lp(a)] levels could be a possible explanation for the link between free thyroxine and SCD. Although we highly appreciate the suggested hypothesis for this association, we think this is unlikely.
As is shown in the literature cited by Cerit and colleagues, lower HDL-cholesterol values and higher Lp(a) are associated with a higher risk of SCD. However, for the hypothesis of Cerit and colleagues to hold true, higher thyroid hormone levels need to be associated with lower HDL-cholesterol and higher Lp(a) levels. In contrast, the association is precisely in the opposite direction. Thyroid hormone has several important effects on hepatic lipid metabolism, resulting in a net beneficial effect with a decrease of low-density lipoprotein-cholesterol and an increase of HDL-cholesterol. Also, thyroid hormone analogues have been shown to reduce low-density lipoprotein -cholesterol levels in animal and human studies. In a 12-week randomized placebo-controlled, double-blind trial, the thyromimetic drug eprotirome showed a reduction in low-density lipoprotein -cholesterol of ≤32% in statin-treated patients. As a secondary outcome, the trial also showed a significant and dose-dependent decrease in Lp(a) levels in eprotirome-treated patients with baseline values ˃70 mg/dL.
Lp(a) is not available in the Rotterdam Study, but we have information on HDL-cholesterol in ˃10 000 participants within our previously described study population. The risk estimates in euthyroid participants of the most adjusted model that we previously reported remained similar after adding HDL-cholesterol to the model with hazard ratios of 2.25 (1.05–4.82) and 2.36 (1.09–5.11), respectively. These results are in line with previous literature on the association of thyroid hormone with lipid profiles in general and with HDL-cholesterol in particular. It is also in agreement with the biological mechanisms of thyroid hormone, explained earlier.
In short, higher thyroid hormone levels are associated with a favorable lipid profile, and in turn a favorable lipid profile is associated with a lower risk of SCD. In our study, we find that higher free thyroxine is associated with a higher risk of SCD. Thus, HDL-cholesterol and Lp(a) are unlikely mediators of the association between higher thyroid hormone levels and SCD risk. Therefore, the mechanism underlying the association remains to be established and is pivotal for questions on thyroid dysfunction treatment and cardiovascular risk management.