Effects of levothyroxine in subclinical hypothyroidism and heart failure with reduced ejection fraction: An open-label randomized trial

Summary

We report a randomized, multicenter, open-label trial (ClinicalTrials.gov: NCT03096613) to investigate the clinical benefits of levothyroxine (L-T4) administration in subclinical hypothyroidism (SCH) patients with heart failure with reduced ejection fraction (HFrEF). Overall, 117 patients were enrolled and received L-T4 plus standard HFrEF treatment (experimental group, N = 57) or standard HFrEF therapy alone (control group, N = 60). The change of 6-min walk test distance in the experimental group was significantly higher than that in the control group at 24 weeks (70.08 ± 85.76 m vs. 27.73 ± 82.00 m, mean difference [95% confidence interval (CI)] 46.90 [12.90, 80.90], p < 0.001). Improvements in New York Heart Association (NYHA) classification (p = 0.033) and thyroid function were significant. Adverse event incidence was similar between groups (risk ratio [95% CI]: 0.942 1.053 (0.424, 2.616); p = 0.628). L-T4 addition to HFrEF treatment improved activity tolerance, NYHA class, and thyroid function within 6 months, suggesting its potential for combined therapy in HFrEF patients with SCH. Future double-blind, placebo-controlled trials should be performed to confirm these results.

Graphical abstract

Highlights

• •Subclinical hypothyroidism is prevalent in HFrEF population
• •Levothyroxine can enhance exercise tolerance in HFrEF with subclinical hypothyroidism
• •A low-dose levothyroxine strategy is safe and effective in HFrEF patients

In a multicenter, open-label, randomized study (ThyroHeart-CHF), Wang et al. demonstrate that levothyroxine addition to standard heart failure treatment improved activity tolerance, cardiac function classification, and thyroid function. Levothyroxine replacement may be considered in combination therapy for patients with subclinical hypothyroidism and heart failure with reduced ejection fraction.

Introduction

Heart failure (HF) is characterized by cardiac structural and functional abnormalities, confirmed by elevated levels of circulating natriuretic peptide or the presence of associated signs and symptoms.^1,2 Approximately 64 million people suffer from HF worldwide, with an estimated prevalence of 1%–3% among adults.³ Despite improvements in medical management of the disease,⁴ total mortality from HF has shown an increase.⁵ Endocrine disturbances, including thyroid dysfunction, are also frequently observed in HF patients.⁶ Current guidelines recommend evaluating thyroid hormones in HF patients to prevent possible adverse events,⁷ underscoring the importance of timely recognition and management of thyroid dysfunction in HF patients.

Subclinical hypothyroidism (SCH) is a prevalent condition observed in HF and is characterized by elevated serum thyroid-stimulating hormone (TSH) levels with free thyroxine (FT4) values within the normal range.⁸ It represents an early stage of hypothyroidism. Current guidelines recommend thyroid hormone replacement in cases of overt hypothyroidism; however, routine therapy for SCH is not advised.⁸ This stance is primarily grounded in the absence of observed benefits from thyroid hormone replacement in the broader SCH population, regardless of the presence of heart conditions.⁹ For SCH patients combined with HF, the guidelines give no clear treatment recommendations,⁸ yet emerging evidence has shown that the prevalence of SCH in the HF population could reach 5%–12%,¹⁰ and there is a strong association between SCH and cardiovascular diseases.^11,12 SCH can alter cardiovascular morphology and function, exacerbating the progression of HF.¹³ Accumulating evidence indicates that hypothyroidism plays a crucial role in the occurrence and development of HF and is associated with poor prognosis and even death.^14,15 The absence of clinical evidence poses challenges for patients with HF combined with SCH.¹⁶ We aim to provide initial evidence to furnish data for larger, appropriately designed randomized controlled trials (RCTs) in this area.

We focused on patient HF with reduced ejection fraction (HFrEF) for our study due to the lack of consensus on defining HF with preserved ejection fraction (HFpEF) during project conception. Besides, there are significant differences in the pathophysiological mechanisms and standard therapy strategies between HFrEF and HFpEF. Our primary endpoint was the classic surrogate endpoint, the 6-min walk test (6MWT), chosen to validate treatment efficacy without posing potential patient harm.

We hypothesized that levothyroxine (L-T4) replacement could enhance exercise tolerance in HFrEF patients with SCH. Addressing the absence of randomized trials establishing L-T4’s role in treating HF patients with SCH, we initiated a multicenter, open-label, randomized trial. The primary objective of this study was to evaluate the efficacy and safety of L-T4 replacement therapy in HFrEF patients with SCH. By conducting this comprehensive investigation, we shed light on potential benefits and risks for this specific patient group.

Results

Patients

Between March 2017 and December 2018, we screened 272 patients diagnosed with HFrEF combined with SCH across 5 Chinese centers. Among them, 124 patients met the inclusion criteria and were randomly assigned at a 1:1 ratio to either receive thyroxine replacement therapy along with standard HFrEF treatment (experimental group) or standard HFrEF therapy alone (control group). 117 patients were included in the full analysis set (FAS) (experimental group, n = 57; control group, n = 60), as depicted in Figure 1.

Figure 1.

Patient flow diagram

The mean age of enrolled patients was 58.54 ± 13.73 years in the experimental group and 57.77 ± 15.12 years in the control group; of these, 73.5% were male. Although there were significant differences in the history of hypertension between the two groups, it did not impact the subsequent efficacy or safety analysis due to the lack of significant differences in systolic and diastolic blood pressure. This is likely due to the fact that the hypertensive patients were using medications. The distribution of the other demographic and clinical characteristics between the two groups were well balanced (Table 1).

Table 1.

Baseline characteristics of patients

Characteristic	Experimental group (n = 57)	Control group (n = 60)	p Value
Male (n [%])	39 (68.42%)	47 (78.33%)	0.225
Age (years)	58.54 ± 13.73	57.77 ± 15.12	0.772
Height (cm)	167.21 ± 14.30	170.34 ± 6.94	0.135
Weight (kg)	73.86 ± 21.11	72.84 ± 15.22	0.767
Respiratory rate (times/min)	19.16 ± 2.73	19.03 ± 4.48	0.858
Systolic blood pressure (mm Hg)	114.58 ± 20.01	119.34 ± 17.79	0.178
Diastolic blood pressure (mm Hg)	74.12 ± 11.01	77.42 ± 13.43	0.151
Heart rate (beats/min)	80.67 ± 20.45	80.39 ± 15.74	0.935
Diabetes mellitus (n [%])	14 (24.56%)	17 (28.81%)	0.605
History of hypertension (n [%])	21 (36.84%)	33 (55.00%)	0.049
Hyperlipidemia (n [%])	18 (31.58%)	17 (28.33%)	0.702
Smoking (n [%])	27 (47.37%)	29 (48.33%)	0.917
Angina or myocardial ischemia (n [%])	1 (1.75%)	6 (10.00%)	0.115
Ventricular dysfunction (n [%])	56 (98.25%)	59 (98.33%)	1.000
Family history of coronary heart disease (n [%])	6 (10.53%)	4 (6.67%)	0.678
Previous PCI (n [%])	1 (1.75%)	1 (1.67%)	1.000
Previous CABG (n [%])	1 (1.75%)	0	0.487
Drugs for HFa (n [%])
β blockers	45 (83.33%)	48 (90.57%)	0.267
ACEI drugs	36 (66.67%)	34 (64.15%)	0.784
Aldosterone receptor antagonists	47 (88.68%)	45 (84.91%)	0.566
Diuretics – loop diuretics	48 (88.89%)	49 (92.45%)	0.763
Positive inotropic drugs	30 (55.56%)	22 (41.51%)	0.146
Aspirin	19 (35.19%)	19 (35.85%)	0.943

PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft; ACEI, angiotensin converting enzyme inhibitor.

^aAt the initiation of L-T4 intervention.

The doses and application details of L-T4 in the experimental group are shown in Table S1 and Figure S1. In this trial, the median dose of L-T4 used in the experimental group was 12.50 μg, and 26.32% participants received no L-T4 treatment at week 24.

As shown in Table 2, at the 24-week visit, the experimental group showed significantly greater improvement in TSH levels compared with the control group (4.07 mIU/L vs. 5.23 mIU/L, p = 0.008), with free triiodothyronine (FT3; 2.99 pg/mL vs. 2.80 pg/mL, p = 0.014) and thyroxine (TT4; 8.20 μg/dL vs. 7.35 μg/dL, p = 0.024) levels also being significantly higher. However, there was no significant difference in FT4 or triiodothyronine (TT3) between the two groups (all p > 0.05). Furthermore, a higher proportion of patients in the treatment group achieved TSH levels within the normal range at the 24-week follow-up, with significantly more patients in the experimental group having TSH levels lower than 4.78 mIU/L compared with the control group (73.9% vs. 45.8%, risk ratio [95% confidence interval (CI)]: 1.613 [1.134, 2.294], p = 0.006), while there was no significant difference in the number of patients with TSH levels greater than 10 mIU/L between the two groups (0% vs. 2.1%, p = 0.325).

Table 2.

Comparison of thyroid indicators between the two groups

Thyroid indicators	Experimental group (n = 57)			Control group (n = 60)			p Value
	Baseline	24 weeks	P	Baseline	24 weeks	p Value
TSH (mIU/L)	6.23 (5.30, 7.92)	4.07 (2.56, 5.12)	<0.001	6.72 (5.78, 7.63)	5.23 (3.26, 7.14)	<0.001	0.008
FT3 (pg/mL)	2.64 (2.31, 3.02)	2.99 (2.73, 3.81)	<0.001	2.47 (2.12, 3.13)	2.80 (2.51, 3.04)	0.039	0.014
FT4 (ng/dL)	1.20 (1.06, 1.34)	1.20 (1.09, 1.37)	0.404	1.13 (1.00, 1.30)	1.19 (1.02, 1.28)	0.845	0.163
TT3 (ng/mL)	0.92 (0.75, 1.12)	1.00 (0.91, 1.25)	0.005	0.85 (0.72, 1.03)	0.92 (0.82, 1.15)	0.010	0.093
TT4 (μg/dL)	7.90 (6.43, 8.68)	8.20 (7.30, 9.10)	0.109	6.80 (5.20,7.95)	7.35 (6.53, 8.58)	0.106	0.024
Grouping of TSH level (n [%])
24-week TSH < 4.78 (mIU/L)	34 (73.9%)			22 (45.8%)			0.006
Risk ratio (95% CI)	–			1.613 (1.134, 2.294)			–
24-week TSH > 10 mIU/L	0			1 (2.1%)			0.325

TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxine; TT3, triiodothyronine; TT4, thyroxine.

Primary endpoint

Both groups exhibited significant improvements after 24 weeks of therapies (Table 3). At the 24-week visit, significant differences were observed between the two groups. The 6MWT distance improvement from baseline to the 24-week visit in the experimental group was significantly greater than that in the control group (70.08 ± 85.76 m vs. 27.73 ± 82.00 m, p < 0.001; Table 3). These findings indicate that the combination of L-T4 and standard HF treatment led to more substantial improvements in 6MWT distance. Because of the differences in some of the variables at baseline, we further used a multifactor correction to compare the differences in the primary endpoint. After adjustment for baseline age, gender, left ventricular ejection fraction (LVEF), TSH, and history of hypertension, the administration of L-T4 significantly altered the distance difference of the 6MWT, with a mean difference of 47.40 m (95% CI, 11.10–83.71; p = 0.011; Table 3).

Table 3.

Primary endpoint and secondary endpoints of the two groups

	Experimental group (n = 57)	Control group (n = 60)	p Value
Primary endpoint
6MWT at baseline (m)
Mean ± SD	214.75 ± 87.84	204.62 ± 84.14	0.507
Median (Q1, Q3)	220 (142.50, 280)	200 (160, 238.75)
Mean difference (95% CI)	10.14 (−21.36, 41.63)	–
6MWT at 24-week visit (m)
Mean ± SD	292.40 ± 92.15	237.16 ± 125.21	0.003
Median (Q1, Q3)	310 (202.50, 357.25)	200 (165, 291)
Mean difference (95% CI)	55.24 (12.01, 98.47)	–
Distance difference of the 6MWT (m)
Mean ± SD	70.08 ± 85.76	27.73 ± 82.00	<0.001
Median (Q1, Q3)	60 (31.25, 113.75)	−5 (−20, 68.50)
Mean difference (95% CI)	46.90 (12.90, 80.90)	–
Adjusted distance difference of the 6MWT (m)a
Mean difference (95%CI)	47.40 (11.10, 83.71)	–	0.011
Secondary endpoints
MLHFQ score
Baseline	45 (32, 59.50)	47.50 (33.50)	0.639
24-week visit	21 (11, 27)b	29.50 (8, 54.75)b	0.128
NT-proBNP (ng/L)
Baseline	1,911 (981.25, 5,363.80)	4,275.90 (1,184, 8,252)	0.124
24-week visit	1,360 (589.80, 3,564)b	1,654 (774.45, 4,621.50)b	0.461
EF (%)
Baseline	34 (28, 39)	34.50 (28, 40)	0.746
24-week visit	40 (35.50, 49.50)b	41 (32, 51.50)b	0.908
LA (mm)
Baseline	46 (41.50, 50)	46 (41, 51.75)	0.675
24-week visit	42 (38.50, 47.25)	46 (40.25, 53)	0.066
LV (mm)
Baseline	62 (57, 72)	63 (57.33, 68)	0.501
24-week visit	56.50 (48, 63.25)b	56 (50, 65.50)	0.705
AO (mm)
Baseline	32.00 (30.00, 37.00)	32.00 (27.50, 35.50)	0.302
24-week visit	32.00 (28.00, 36.00)	32.00 (29.00, 35.00)	0.572
IVS (mm)
Baseline	9.00 (8.00, 11.00)	9.00 (8.00, 10.00)	0.148
24-week visit	9.00 (8.00, 10.00)	9.00 (8.00, 10.00)	0.313
LVPW (mm)
Baseline	9.00 (8.00, 10.00)	9.00 (8.00, 10.00)	0.884
24-week visit	9.00 (8.00, 10.00)	9.00 (8.00, 10.00)	0.845

6MWT, 6-min walk test; MLHFQ, Minnesota Living with Heart Failure Questionnaire; NT-proBNP, N-terminal pro-B-type natriuretic peptide; EF, ejection fraction; LV, left ventricular; LA, left atrium; AO, aortic diameter; IVS, interventricular septum; LVPW, left ventricular posterior wall.

^aThe difference of the 6MWT between the experimental group and control groups was adjusted for baseline age, gender, EF, TSH, and history of hypertension.

^bIntragroup comparison showed statistically significant differences between baseline and 24-week visit. The p value is presented as a non-parametric test result.

Secondary endpoints

The comparison between two groups revealed no significant difference in Minnesota Living with Heart Failure Questionnaire (MLHFQ) score, N-terminal pro-B-type natriuretic peptide (NT-proBNP), or echocardiographic measures (ejection fraction [EF], left atrium [LA], left ventricular [LV], aortic diameter [AO], interventricular septum [IVS], and left ventricular posterior wall [LVPW]) at both baseline and the 24-week visit (Table 3). The New York Heart Association (NYHA) functional classification was assessed at each visit. As shown in Table S2 and Figure 2, there was no significant difference in NYHA functional classification between the experimental group and the control group at baseline (p > 0.999). However, after treatment with L-T4 plus standard HF treatment, the frequency of NYHA class I patients gradually increased, while the frequency of NYHA class III–IV patients gradually decreased in the experimental group. Notably, the analysis indicated that L-T4 plus standard HF treatment led to superior improvements in NYHA functional classification at the 8-week visit compared with the control group (p = 0.033). Improvement of NYHA functional classification was observed at 8 weeks, and there was no difference at 6 months, suggesting that L-T4 contributes to short-term cardiac function improvement.

Figure 2.

NYHA classification results of the two groups

(A and B) Changes of NYHA classification during the research period in (A) experimental group and (B) control group.

Subgroup analyses

We conducted subgroup analyses to investigate the potential impact of baseline characteristics on patient outcomes when administering additional L-T4 supplementation in conjunction with HF therapy. Subgroup analyses suggested that the result of the distance difference of the 6MWT obtained was not significantly influenced by age, gender, baseline TSH, FT3, FT4, NT-proBNP level, or EF (all p > 0.05; Figure 3). It was observed that the efficacy of L-T4 supplementation was relatively more pronounced in the following subgroups: males, patients under the age of 65 years, TSH above 6.52 mIU/L, EF below 34%, and FT3/FT4 below median. The findings in different age subgroups align with the established guidelines for patients with SCH in the absence of HF.

Figure 3.

Effect of levothyroxine (L-T4) compared with placebo on the distance difference of the 6MWT according to the prespecified subgroups

Safety

The occurrence of adverse events (AEs) was observed on 8 occasions (14.04%) in the experimental group versus 8 (13.33%) in the control group, and there was no significant difference between the two groups (risk ratio [95% CI]: 1.053 -0.424, 2.616]; p = 0.628; Table 4). The incidence rate of AEs in patients over 65 years had no significant difference between the two groups (p = 0.326), which might suggest that it is safe to take the current strategy with a small dose of L-T4 in this population. At the 24-week visit, there was no significant difference in the incidence of atrial fibrillation between the two groups at 24-week visit (5 [8.77%] vs. 6 [10.00%]; risk ratio [95% CI]: 0.877 [0.283, 2.716]; p = 0.820). During follow-up, no statistically significant difference was observed regarding death (5 [8.77%] vs. 7 [11.67%]; risk ratio [95% CI]: 0.752 [0.253, 2.234]; p = 0.510).

Table 4.

Summary of adverse events (AEs)

N (%)	Experimental group (n = 57)	Control group (n = 60)	Risk ratio (95% CI)	p Value
Total events	8 (14.04%)	8 (13.33%)	1.053 (0.424, 2.616)	0.628
Readmission for HF	6 (10.52%)	5 (8.33%)	–	–
Angina pectoris	1 (1.75%)	0	–	–
Heart transplantation	0	1 (1.67%)	–	–
Ventricular arrhythmia	1 (1.75%)	2 (3.34%)	–	–
AEs in patients > 65 years old	1 (1.75%)	3 (5.00%)	–	0.326
Atrial fibrillation at 24-week visita	5 (8.77%)	6 (10.00%)	0.877 (0.283, 2.716)	0.820
Death	5 (8.77%)	7 (11.67%)	0.752 (0.253, 2.234)	0.510
Cardiac death	4 (7.02%)	7 (11.67%)	–	–
Non-cardiac death	1 (1.75%)	0	–	–

^aAtrial fibrillation was detected by 24-h Holter test.

Discussion

In this multicenter, open-label, randomized trial, treatment with L-T4 for 24 weeks could improve activity tolerance in HFrEF patients combined with SCH. Besides, L-T4 replacement therapy demonstrated notable improvements in NYHA functional class and thyroid function within the 6-month treatment period with a favorable safety profile. Previously, studies in the realm of HF combined with hypothyroidism mainly consisted of non-randomized observational investigations with modest sample sizes ranging from 10 to 40 patients.^{17,18,19,20,21} The largest clinical RCT contained only 86 patients.²² In contrast, our present study, serving as an exploratory venture, enrolled the largest sample size within this domain thus far. A key implication of this RCT is to establish a foundation and provide valuable insights for the future design and implementation of larger-scale clinical trials.

In some previous studies of thyroid hormone replacement therapy for HF, both conventional thyroid hormone doses and high-dose thyroid hormone analogs²² improved cardiac function indices. However, these approaches also correlated with a notable rise in adverse effects, including hyperthyroid-like reactions such as accelerated heart rate and weight loss. When formulating the present study, we devised a strategy that commenced with a small dose and gradually increased it. The intention was to mitigate the incidence of adverse effects. Consequently, the treatment group primarily received thyroid hormone at very low doses, ranging from 12.5 to 37.5 μg per day. The results of the present study confirm that a low-dose L-T4 strategy is safe and effective in patients with HFrEF.

In addition, high-dose thyroid hormone analog studies were halted due to severe adverse reactions, prompting our choice of a low-dose approach. A previous study assessed high-dose thyroid hormone analog supplementation for HF but was discontinued due to severe side effects.²² Nonetheless, this result does not clarify whether supplemental hormone therapy potentially benefits patients with concurrent hypothyroidism. Moreover, the dosing strategy and final dosage are vital for therapeutic efficacy. For patients with HF combined with SCH, we advocate a gradual, low-dose initiation strategy as a more rational and safe option.

Interestingly, our findings in patients with HFrEF complicated by SCH do not contradict the guidelines for general SCH patients without the HF condition. The 2013 SCH management guidelines from the European Thyroid Association recommend that young patients with serum TSH levels greater than 10 mIU/L (age < 65 years) and even those with lower serum TSH levels should receive L-T4 replacement.¹¹ SCH leads to insufficient thyroid hormone secretion and compensatory elevated TSH secretion, resulting in dyslipidemia, characterized by increased total cholesterol and low-density lipoprotein cholesterol levels.²³ In this study, continuous L-T4 replacement therapy for 6 months restored the baseline thyroid function of HFrEF patients with SCH. This was evidenced by notable reductions in TSH along with a significant increase in FT3 levels.

How did our findings impact HFrEF therapy guidelines? Unlike previous studies, our research implemented a rigorous RCT within the population of individuals concurrently dealing with SCH and HFrEF. SCH increases the risk of HF and all-cause death.¹⁶ Hypothyroidism could aggravate myocardial damage and result in cardiac dysfunction, affecting the short-term and long-term prognoses of HF patients. In the general population, L-T4 replacement has shown potential to lower cardiovascular disease incidence in patients with elevated TSH levels.^24,25 Badran et al.²⁶ have reported safe and effective short-term use of L-T4 in idiopathic dilated cardiomyopathy patients, suggesting that synthetic thyroid hormones might have a role in HF treatment when standard drugs fall short.

To the best of our knowledge, the ThyroHeart-CHF study was a prospective, multicenter, randomized, controlled clinical trial assessing the efficacy and safety of L-T4 replacement in patients with HFrEF complicated by SCH. Our subgroup analysis yielded additional insights. For example, Zijlstra et al.²⁷ found that L-T4 replacement treatment did not change the risk of all cardiovascular outcomes in older adults with SCH.²⁷ The analysis of data from several clinical trials indicated that L-T4 treatment did not significantly alleviate the hypothyroid symptoms in older patients with SCH.^9,28 L-T4 treatment might only benefit certain subgroups, such as younger patients or those at higher cardiovascular risk.²⁹ In our study, L-T4 replacement combined with standard HFrEF treatment demonstrated efficacy and safety, particularly within subgroups of “baseline TSH level ≥ 6.52 mIU/L″ and “age below 65 years.” These findings, together with previous reports, indicate that the benefit of thyroid hormone replacement in HFrEF patients depend on their baseline characteristics. Younger patients and those with higher TSH levels are more likely to benefit from L-T4 treatment.

In a study published in 2020,³⁰ researchers enrolled 95 patients with acute myocardial infarction combined with SCH. They investigated the impact of thyroid hormone replacement therapy on enhancing cardiac function. Notably, the control group also displayed a tendency of TSH recovery, aligning with our study’s findings. While the overall endpoints yielded negative outcomes, certain subgroups exhibited treatment benefits, particularly those characterized by “baseline TSH > 5.7 mU/L” and “decreased EF.” The baseline characteristics that benefitted from hormone replacement therapy align with our subgroup characteristics that exhibited the most benefit, even though the patients had different heart conditions. This suggests that thyroid hormone replacement therapy can broadly contribute to improving heart conditions.

As this study is one of the RCTs in this area, we chose the classic surrogate endpoint (6MWT) as the primary endpoint to validate the efficacy of the study while avoiding potential harm to patients. In the present study, exercise tolerance was assessed using the 6MWT, which is widely used in many different populations due to its simplicity and reliability. The 6MWT has important prognostic value in HFrEF patients, and it has been recommended to monitor the disease course and to assess the efficacy of intervention.³¹ In a previous study examining the physical performance of patients with SCH, the 6MWT has been reported to be decreased in SCH patients compared with healthy controls.³² In our study, the distance change of the 6MWT in experimental group was significantly higher than that in control group, indicating that L-T4 plus standard HFrEF treatment exhibited greater improvements in exercise tolerance. Meanwhile, we found that the L-T4 plus standard HFrEF treatment resulted in superior improvements in NYHA functional classification at the 8-week visit compared with the control group. However, improvement of NYHA classification showed no statistical difference at later follow-up between the two groups. This indicates that benefits of L-T4 supplement to clinical symptoms may appear at a relatively early stage, which helps transition to a more stable treatment phase. During a longer term of treatment, cooperative treatment effects of drugs for HF and thyroid dysfunction may narrow down the difference between the two groups.

In our study, improvement in the 6MWT results was observed, while LV function or quality of life has no difference between the two groups. The following three reasons might help explain this finding. First, low-dose L-T4 improved cardiomyocyte energy metabolism, which, although it may not make a difference in heart structure or function during the 6-month follow-up period, is enough to improve exercise tolerance. Second, considering that the 6-month follow-up period is relatively short, significant improvements in LV function or QoL may been observed over a longer period. Third, thyroid hormone also improves skeletal muscle system function and peripheral circulation tension, which helps enhance exercise tolerance.³³

In conclusion, in patients with HFrEF complicated by SCH, the addition of L-T4 replacement therapy to standard HFrEF treatment demonstrated notable improvements in activity tolerance, NYHA functional class, and thyroid function within the 6-month treatment period, with a favorable safety profile. As a result, L-T4 replacement may be considered as combination therapy for HFrEF patients with SCH. Future double-blind, placebo-controlled trials should be performed to confirm these results.

Limitations of the study

First, while the total sample size met our predetermined target with sufficient power for primary endpoint evaluation, it unavoidably restricted the precision of the subgroup analysis. Second, this trial only included HFrEF patients due to the lack of universally established definitions for HFpEF at the time of trial design. Consequently, further trials targeting individuals with HFpEF are necessary in the future to validate therapeutic effectiveness. Third, the relatively short follow-up period limited the identification of treatment-associated adverse effects over the long term. Fourth, a placebo control was lacking in this trial. Fifth, lack of blinding could mean that participants and assessors could be biased in their reporting and assessments. The main reason for the lack of blinding is that our study requires multiple dosage adjustments of L-T4 during the research process. Therefore, blinding required additional packaging of drugs (additional containers) to avoid doctors and patients being aware of changes in drug dosage. During the ethics review, the ethics committee suggested that the container required for blinding involved changes in drug properties and needed further permission from the China Food and Drug Administration. In addition, the implementation of blinding required a separate group of physicians to adjust the medication. These processes were very difficult to achieve for our exploratory study, and considering feasibility, we ultimately did not choose blinding or placebo control.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Levothyroxine (L-T4)	Merck Serono China Co. Ltd.	N/A
Critical commercial assays
Immulite 2000	Siemens Healthcare Diagnostics Inc., CA, USA	N/A
NT-proBNP assays	Roche Diagnostics, Basel, Switzerland	N/A
Software and algorithms
SPSS software version 26.0	SPSS Inc., Chicago, IL, USA	www.spss.com

Resource availability

Lead contact

Further information and requests for recourses and reagents should be directed to the lead contact, Yi-Da Tang (tangyida@bjmu.edu.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •All data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

Study design

In this prospective, multi-center, open-label, randomized study, patients were recruited from five study centers in China from March 2017 to December 2018. The study protocol was approved by all the institutional review board (Fuwai Hospital, No.2016-798; Beijing Luhe Hospital, No.2017-LHYW-01; the Second Hospital of Tianjin Medical University, No.KY2017K037; The First Hospital of Hebei Medical University, No. 2017001; Henan Provincial People’s Hospital, No.2017–12). The study was conducted in accordance with Declaration of Helsinki and International Conference on Harmonisation Good Clinical Practice guidelines. All patients signed the written informed consent.

Patient eligibility

Patients aged 18–80 years were eligible if they had a clinical diagnosis of HFrEF and SCH. All individuals had thyroid function checked within 24 h of admission to the hospital. Patients received 2-week standard HF treatment and one more thyroid function test 10–14 days after the diagnosis of HFrEF. Eligible patients met the following inclusion criteria: (1) had an New York Heart Association (NYHA) functional class of II-III; (2) left ventricular ejection fraction (LVEF) ≤ 40% by echocardiography during screening and randomization; (3) SCH diagnosed according to two thyroid function tests (serum TSH: upper limits of normal to 10 mIU/L, and FT4 level in the normal range); (4) had received optimal medical treatment with target dose or maximum tolerable dose for HFrEF. The major exclusion were as follows: (1) had acute HF or acute exacerbation of HFrEF within the past 2 weeks; (2) scheduled cardiac resynchronization therapy or heart transplantation; (3) had history of malignant tumor or life expectancy under 12 months; (4) had history of thyroid dysfunction, or taken medications that might affect thyroid function (LT4, carbimazole, propylthiouracil, amiodarone, lithium); (5) pregnancy and lactation period; (6) participated in another clinical trial within the past 30 days; (7) contraindication or intolerance to evidence-based therapy for HFrEF, such as beta-blocker, angiotensin-converting enzyme inhibitor, or angiotensin receptor blocker; (8) hypersensitivity to the trial treatments; (9) severe renal dysfunction (estimated glomerular filtration rate ≤30 mL/min/1.73 m²) or significant hepatic impairment (serum GPT >120 U/L); (10) had any disorder, which might jeopardize patients’ safety or compliance. The complete eligibility criteria were listed in the published protocol.³⁴

Method details

Study protocol

Eligible patients were randomly assigned in a 1:1 manner using an electronic centralized randomization system to achieve randomization and allocation concealment across study centers to receive L-T4 replacement therapy plus standard HFrEF treatment (experimental group) or only standard HFrEF therapy (control group). The standard treatment for HFrEF adhered to the protocol outlined in a previously published report.³⁵ L-T4 replacement therapy was initiated at an initial daily dose of 12.5 μg for all patients in the experimental group. To ensure appropriate dosage adjustments, thyroid function assessments were conducted at 4, 8, and 12 weeks. The dosage regulation of L-T4 was primarily guided by monitoring the levels of thyroid-stimulating hormone (TSH), as specified in the published protocol.³⁴ The aim of dose adjustments was to maintain TSH levels within the reference range. Specifically, if TSH levels remained above the upper limit of the normal range, an additional 12.5 μg/day of L-T4 was administered. When TSH levels returned to the normal range, the dosage of L-T4 was maintained. In cases where TSH levels fell below the lower limit of the normal range, the prescription of L-T4 was discontinued for those patients. These carefully managed dosage adjustments and evaluations of thyroid function were essential in ensuring the safety and effectiveness of the L-T4 replacement therapy in this study.

Biochemical analyses

Twelve-hour-fasting blood samples were drawn and the serum levels of thyroid hormones and TSH were measured using radioimmunoassay (Immulite 2000; Siemens, Germany). The reference intervals of TH and TSH in our laboratory were as follows: TSH = 0.55–4.78 mIU/L; fT3 = 1.79–4.09 pg/mL; fT4 = 0.8–1.88 ng/dL; TT3 = 0.65–1.91 ng/mL; TT4 = 4.29–12.47 μg/dL. Other biochemical indexes (blood lipid, creatinine, transaminases, NT-proBNP, etc) were analyzed using the Roche assays (Roche Diagnostics, Switzerland).

Endpoints and assessments

Patients attended follow-up appointments at 4-, 8-, 12-, and 24-week of treatment. The primary endpoint was the distance difference of the 6MWT between 24-week visit and baseline. Secondary endpoints included: (1) change in NYHA classification within 24-week visit; (2) difference in plasma N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels between 24-week visit and baseline; (3) quality of life assessment, that was, the difference in Minnesota Living With Heart Failure Questionnaire (MLHFQ) scores between 24-week visit and baseline; (4) difference in echocardiographic and cardiac magnetic resonance imaging (MRI) measures between 24-week visit and baseline; (5) percentage of patients experiencing cardiovascular death, re-hospitalization for HF, cardiovascular disease, severe arrhythmia, stroke and others. Meanwhile, the thyroid function indexes at baseline and 24-week visit were also recorded. At each visit, patients were asked about the occurrence of any clinical event or adverse effect.

Quantification and statistical analysis

Sample size

The sample size and power calculations for this study were performed with careful consideration of the following assumptions: a two-sided log rank test, an α level of 0.05, and a randomization ratio of 1:1. The calculation of the sample size was based on the anticipated change in the 6MWT distance at the 24-week visit. Previous research reported a mean difference of 40 m between groups in terms of the change in 6MWT distance from baseline, with a standard deviation (SD) of 70 m.³¹ There are also trials with similar intervention, outcome and target population.^2,3 Based on these assumptions, a total sample size of 124 patients, with 62 patients in each group, was determined to be necessary. This sample size would enable the detection of a mean treatment effect of at least 40 m at the 24-week visit, with 80% power, at a two-sided significance level of 5%. Furthermore, this calculation accounted for a potential dropout rate of 20%, ensuring the robustness and reliability of the study findings.

Statistical analysis

All statistical analyses were performed with the statistical package SPSS software (Version 26.0, SPSS Inc., Chicago, IL, USA). The efficacy and safety analysis were performed on the full analysis set (FAS) and safety analysis set (SS), respectively. As the missing rate is very low in the present study, the missing values are not included in the FAS analysis. Continuous variables were expressed as mean ± standard deviation. Categorical variables were described as the frequency or ratio. Kolmogorov-Smirnov test was used to assess whether the data conform to a normal distribution. The normally distributed continuous variables were compared by using two-sided Student’s t test, and continuous variables that were not normally distributed were compared using Mann-Whitney U test. Categorical variables were compared using the two-sided likelihood ratio chi-squared test or Fisher exact test. The primary endpoint was adjusted using analysis of covariance (ANCOVA), and the adjusted variables included age, gender, baseline EF, baseline TSH, and baseline history of hypertension. Prespecified subgroup analyses was performed to explore subgroups by treatment interactions to investigate potential differential estimated treatment effects. ANCOVA was used to calculate the P-value for the interaction in the subgroup analysis for age, gender, baseline TSH, baseline FT3, baseline FT4, baseline NT-proBNP and baseline EF. Statistical significance was set at p < 0.05.

Additional resources

This trial is registered with ClinicalTrials.gov, NCT03096613.

Published: March 26, 2024