This randomized clinical trial at a tertiary referral center in the United Kingdom compares the effects of oral trimetazidine hydrochloride vs placebo on exercise capacity among patients with nonobstructive hypertrophic cardiomyopathy.
Key Points
Question
Does oral trimetazidine dihydrochloride therapy improve exercise capacity for patients with nonobstructive hypertrophic cardiomyopathy?
Findings
In this randomized clinical trial of 51 patients with drug-refractory symptoms, oxygen consumption of patients with nonobstructive hypertrophic cardiomyopathy at peak exercise was not improved after 3 months of oral trimetazidine therapy, a statistically significant finding.
Meaning
Metabolic modulators, such as trimetazidine, have limited application for patients with nonobstructive hypertrophic cardiomyopathy.
Abstract
Importance
Hypertrophic cardiomyopathy causes limiting symptoms in patients, mediated partly through inefficient myocardial energy use. There is conflicting evidence for therapy with inhibitors of myocardial fatty acid metabolism in patients with nonobstructive hypertrophic cardiomyopathy.
Objective
To determine the effect of oral therapy with trimetazidine, a direct inhibitor of fatty acid β-oxidation, on exercise capacity in patients with symptomatic nonobstructive hypertrophic cardiomyopathy.
Design, Setting, and Participants
This randomized, placebo-controlled, double-blind clinical trial at The Heart Hospital, University College London Hospitals, London, United Kingdom was performed between May 31, 2012, and September 8, 2014. The trial included 51 drug-refractory symptomatic (New York Heart Association class ≥2) patients aged 24 to 74 years with a maximum left ventricular outflow tract gradient 50 mm Hg or lower and a peak oxygen consumption during exercise of 80% or less predicted value for age and sex. Statistical analysis was performed from March 1, 2016 through July 4, 2018.
Interventions
Participants were randomly assigned to trimetazidine, 20 mg, 3 times daily (n = 27) or placebo (n = 24) for 3 months.
Main Outcomes and Measures
The primary end point was peak oxygen consumption during upright bicycle ergometry. Secondary end points were 6-minute walk distance, quality of life (Minnesota Living with Heart Failure questionnaire), frequency of ventricular ectopic beats, diastolic function, serum N-terminal pro–brain natriuretic peptide level, and troponin T level.
Results
Of 49 participants who received trimetazidine (n = 26) or placebo (n = 23) and completed the study, 34 (70%) were male; the mean (SD) age was 50 (13) years. Trimetazidine therapy did not improve exercise capacity, with patients in the trimetazidine group walking 38.4 m (95% CI, 5.13 to 71.70 m) less than patients in the placebo group at 3 months after adjustment for their baseline walking distance measurements. After adjustment for baseline values, peak oxygen consumption was 1.35 mL/kg per minute lower (95% CI, −2.58 to −0.11 mL/kg per minute; P = .03) in the intervention group after 3 months.
Conclusions and Relevance
In symptomatic patients with nonobstructive hypertrophic cardiomyopathy, trimetazidine therapy does not improve exercise capacity. Pharmacologic therapy for this disease remains limited.
Trial Registration
ClinicalTrials.gov identifier: NCT01696370
Introduction
Hypertrophic cardiomyopathy (HCM) is defined clinically by the presence of left ventricular hypertrophy unexplained by loading conditions; it predisposes to sudden cardiac death and progressive heart failure.1 In most patients, it is an autosomal dominant genetic trait caused by mutations of cardiac sarcomere protein genes.2 Individuals with exertional symptoms caused by left ventricular outflow tract obstruction can be treated with drugs or septal reduction. Similar symptoms occur in patients with the absence of left ventricular outflow tract obstruction but medical treatment is often ineffective.1
Many of the gene mutations that cause HCM increase the energetic cost of cardiomyocyte contraction and relaxation.3 Conventional therapies, such as β-blockers and nondihydropyridine calcium antagonists, reduce myocardial energy demands by decreasing heart rate and blood pressure, but their use is limited by adverse effects or lack of clinical efficacy. An alternative approach is to stimulate glucose oxidation and reduce fatty acid oxidation through inhibition of fatty acid uptake into the mitochondrion and direct inhibition of β-oxidation.4 Compared with placebo, perhexiline maleate (a carnitine palmitoyl transferase–1 inhibitor) was shown to improve symptoms and exercise performance in patients with HCM in association with improved myocardial energetics and diastolic filling.5 However, perhexiline is limited by its narrow therapeutic index and potential neurotoxic and hepatotoxic effects when patients are exposed to sustained high plasma levels of the drug. We hypothesized that trimetazidine dihydrochloride, a safe and well-tolerated direct inhibitor of β-oxidation, improves symptoms and exercise capacity in medically refractory symptomatic patients with nonobstructive HCM.
Methods
Trial Design
The noncommercial, investigator-led, single center, randomized, double-blind, placebo-controlled, parallel-group trial was performed during 3 months, from May 31, 2012, through September 8, 2014. The trial was approved by the National Research Ethics Committee, Glasgow, United Kingdom, and participants provided written informed consent before enrollment. Participants were allocated to either the placebo group or intervention group in a ratio of 1:1 (the full trial protocol is available in Supplement 1). The study was conducted according to the principles of the Declaration of Helsinki6 and monitored by an independent data and safety monitor.
Participants
Eligible participants were 18 years or older with a diagnosis of nonobstructive HCM and receiving stable medical therapy. Participants were symptomatic (New York Heart Association class ≥2), with reduced exercise capacity during symptom-limited bicycle ergometry defined by peak oxygen consumption (V̇O2) 80% or less of predicted values for age and sex. Individuals with diabetes, renal impairment (estimated glomerular filtration rate <60 mL/min), or liver impairment were excluded. After trial commencement, eligibility criteria were amended to include patients with left ventricular outflow tract obstruction <50 mm Hg (instead of <30 mm Hg) and individuals with permanent atrial fibrillation and a ventricular rate (<90 beats/min). The criteria were amended to improve recruitment of additional symptomatic patients who were refractory to current drug treatment.
Interventions
After baseline evaluation, patients were randomized in a double-blind fashion to receive either trimetazidine, 20 mg (n = 27), or placebo (n = 24) 3 times daily. No dose adjustments were made, and concomitant medications were continued for the duration of the trial.
Outcomes
The primary end point was peak V̇O2. Secondary end points were (1) symptom status assessed by the Minnesota Living with Heart Failure questionnaire7; (2) exercise capacity assessed by 6-minute walk distance and the submaximal cardiopulmonary exercise measure (ventilation/carbon dioxide production slope); (3) biomarkers N-terminal pro–brain natriuretic peptide, troponin T, insulin to glucose ratio, and homeostatic model assessment of insulin resistance; (4) systolic and diastolic function using echocardiography; and (5) ventricular ectopic beats recorded using 24-hour ambulatory monitoring. Investigations were performed at baseline and at 3 months. The study design is shown in eFigure in Supplement 2. No changes were made to the primary or secondary end points after the trial commenced. Adverse events were assessed during the 4-week telephone call and the 3-month follow-up visit.
Sample Size
A sample size of 72 patients (36 in each group) was required to detect a change in peak V̇O2 of 2 mL/kg per minute (power of 80% and significance level of 5%) using a 2-sample t test. An estimated SD of 3 mL/kg per minute was used in the calculation.5
Randomization
A masked internet randomization service supplied by Sealed Envelope8 was used with a linked randomization list for administrator unmasking. Participants, researchers, and clinicians were masked to the type of treatment received.
Statistical Analysis
Data were analyzed with Stata, version 14 (Stata Corp). Analyses were carried out by treatment allocated, using all available data (complete case) with intention to treat principles. Continuous variables were summarized using mean (SD) or median (interquartile range). Categorical variables were presented as frequencies and percentages. A linear regression model adjusting for baseline peak V̇O2 was used to estimate the treatment effect on the primary outcome. Appropriate regression models were used for the secondary outcomes, adjusted for baselines values. A 2-tailed P < .05 indicated statistical significance.
Results
Recruitment
All participants were recruited from cardiomyopathy clinics at The Heart Hospital, University College London Hospitals, London, United Kingdom, between May 31, 2012, and September 8, 2014. Of 51 patients randomized, 27 were in the intervention group and 24 were in the control group. One patient withdrew from the study, and another was excluded from the final analysis because of poor compliance. Figure 1 summarizes participant flow during the study. Recruitment was incomplete at the end of the prespecified enrollment period despite a 6-month extension and protocol amendments. With agreement from the sponsor and an independent advisor, an interim statistical analysis between March 1, 2016, and July 4, 2018 was performed. The trial was terminated on April 14, 2015. Reasons for screening failure are summarized in eTable 1 in Supplement 2.
Figure 1. Flow Diagram.
Baseline Data
Of 49 participants who received trimetazidine (n = 26) or placebo (n = 23) and completed the study, 34 (70%) were male; the mean (SD) age was 50 (13) years. Patient demographics and baseline clinical variables are shown in Table 1. The trimetazidine and placebo groups were well matched.
Table 1. Descriptive Statistics for Baseline Variables in 51 Patientsa.
| Baseline Variable | Trimetazidine (n = 27) | Placebo (n = 24) |
|---|---|---|
| Male sex | 18 (67) | 18 (75) |
| White race/ethnicity | 17 (63) | 19 (79) |
| Age, mean (SD) [range], y | 49 (13) [25-69] | 51 (14) [24-74] |
| Weight, mean (SD) [range], kg | 87 (23) [47-132] | 87 (18) [49-133] |
| Height, mean (SD) [range], cm | 171 (10) [155-191] | 175 (10) [153-192] |
| Body mass index, mean (SD) [range]b | 29 (6) [18-42] | 28 (5) [21-38] |
| Pulse, mean (SD) [range], beats/min | 69 (11) [50-88] | 69 (10) [51-85] |
| Blood pressure, mean (SD) [range], mm Hg | ||
| Systolic | 118 (15) [90-151] | 120 (19) [92-173] |
| Diastolic | 73 (9) [56-87] | 72 (6) [62-89] |
| Creatinine level, μmol/L | 90 (18) [50-131] | 83 (13) [60-105] |
| Hemoglobin level, median (interquartile range), g/L | 14.7 (13.7-15.4) | 14.7 (13.9-15.5) |
| Current or former smoker | 19 (70) | 14 (58) |
| Alcohol use, mean (SD) [range], units/wk | 5 (8) [0-30] | 8 (9) [0-40] |
| Electrocardiogram | ||
| Sinus rhythm | 19 (70) | 17 (71) |
| Atrial fibrillation | 2 (8) | 4 (17) |
| Paced rhythm | 6 (22) | 3 (12) |
| QRS interval, mean (SD) [range], ms | 122 (34) [82-212] | 126 (35) [80-194] |
| Medical history | ||
| Hypertension | 5 (19) | 0 |
| Paroxysmal atrial fibrillation | 1 (4) | 1 (4) |
| Implantable cardioverter defibrillator | 11 (41) | 12 (50) |
| Septal myectomy | 5 (19) | 8 (33) |
| Alcohol septal ablation | 1 (4) | 2 (8) |
| Cardiac resynchronization therapy device | 1 (4) | 0 |
| Concomitant medication | ||
| Bisoprolol | 12 (44) | 10 (42) |
| Verapamil | 9 (33) | 5 (21) |
| Disopyramide | 1 (4) | 3 (13) |
| Aspirin | 12 (44) | 5 (21) |
| Clopidogrel | 1 (4) | 1 (4) |
| Warfarin | 10 (37) | 13 (54) |
| Spironolactone | 8 (30) | 11 (46) |
| Simvastatin | 5 (19) | 3 (13) |
Data are presented as number (percentage) of patients unless otherwise indicated.
Calculated as weight in kilograms divided by height in meters squared.
Outcomes and Estimation
The effects of placebo and trimetazidine on the primary and secondary outcomes are shown in eTables 2 and 3 in Supplement 2. The mean peak V̇O2 increased from 17.36 (3.59) mL/kg per minute to 19.01 (4.68) mL/kg per minute in the placebo group and from 17.35 (3.89) mL/kg per minute to 17.66 (3.53) mL/kg per minute in the trimetazidine group. After adjustment for baseline peak V̇O2, the trimetazidine group had a lower peak V̇O2 by 1.35 mL/kg per minute (95% CI, –2.58 to –0.11; P = .03) (Table 2). Inspection of the distribution of residuals and residuals vs fitted values did not indicate a violation of the assumptions of linear regression. Figure 2 shows individual changes in peak V̇O2.
Table 2. Mean Difference in Outcome at 3 Months Comparing Trimetazidine Dihydrochloride and Placebo.
| Primary Outcome | Mean Difference (95% CI) |
|---|---|
| Peak V̇O2, mL/kg per minutea | −1.35 (−2.58 to −0.11) |
| 6-min walking distance, m | −38.4 (−71.70 to −5.13) |
| Ventricular ectopic beats, >500 per 24 h | 0.49 (0.04 to 6.96)b |
| Minnesota Living With Heart Failure score | −0.77 (−9.22 to 7.68) |
| Grade of diastolic function | 0.89 (0.15 to 5.49)c |
| Left ventricular ejection fraction, % | 0.72 (−2.45 to 3.88) |
| Left atrial area, cm2 | −0.95 (−3.16 to 1.26) |
| Global left ventricular systolic strain | −0.07 (−1.40 to 1.26) |
| Left atrial volume index, mL/m2 | −0.909 (−7.45 to 5.64) |
| Log N-terminal pro–brain natriuretic peptide, pmol/L | −0.07 (−0.28 to 0.14) |
| Troponin T, ng/L | 0.001 (−0.013 to 0.016) |
| Insulin to glucose ratio | 0.14 (−0.77 to 1.05)d |
| VE/V̇O2 slope | −0.74 (−3.76 to 2.28) |
| V̇O2 work slope, ML/min/W | 0.009 (−1.09 to 1.11) |
| VE/V̇CO2 slope | 0.13 (−1.52 to 1.78)d |
Abbreviations: VE, minute ventilation; V̇CO2, carbon dioxide production; V̇O2, oxygen production.
P = .03.
Odds ratio comparing trimetazidine group with placebo (based on logistic regression).
Odds ratio comparing trimetazidine group with placebo (based on ordinal regression; (patients in sinus rhythm only).
Median difference after adjusting for baseline values (based on quantile regression).
Figure 2. Primary Outcome Results.
The effects of placebo and trimetazidine on exercise capacity. The mean (SD) peak V̇O2 increased from 17.36 (3.59) to 19.01 (4.68) mL/kg per minute in the placebo group and from 17.35 (3.89) to 17.65 (3.53) mL/kg per minute in the trimetazidine group.
Exercise Capacity and Symptom Status
Patients in the trimetazidine group walked 38.4 m (95% CI, 5.13 to 71.70 m) less than patients in the placebo group at 3 months, after adjustment for their baseline walking distance measurements. The median ventilation/carbon dioxide production slope was higher by 0.13 (95% CI, –1.52 to 1.78). At 3 months, the adjusted Minnesota Living with Heart Failure questionnaire score, after adjustment for the baseline values, was lower in the trimetazidine group by 0.77 (95% CI, –9.22 to 7.68). (Table 2).
Echocardiography and Arrhythmia
After adjustment for the baseline values, there was no change in adjusted grade of diastolic function (mean difference, 0.89 [95% CI, 0.15 to 5.49]), left ventricular ejection fraction (mean difference, 0.72% [95% CI, −2.45% to 3.88%]), left atrial area (mean difference, −0.95 cm2 [95% CI, −3.16 to 1.26 cm2]), or global left ventricular longitudinal systolic strain (mean difference, −0.07 [95% CI, −1.40 to 1.26]). The odds ratio of more than 500 ventricular ectopic beats at 3 months in the trimetazidine group compared with the placebo group was 0.49 (95% CI, 0.04 to 6.96).
Biomarkers
After adjustment for the baseline values, there was no change in adjusted log N-terminal pro–brain natriuretic peptide level (mean difference, −0.07 pmol/L [95% CI, −0.28 to 0.14 pmol/L]) and troponin T level at 3 months (mean difference, 0.001 ng/L [95% CI, −0.013 to 0.016 ng/L]) in the trimetazidine group. The insulin to glucose ratio (mean difference, 0.14 [95% CI, −0.77 to 1.05]) and homeostatic model assessment of insulin resistance (−0.039 [95% CI, −1.76 to 1.68]) decreased in the trimetazidine group indicating improved insulin sensitivity consistent with the known effects of trimetazidine (eTable 2 and eTable 3 in Supplement 2).
Adverse Events
Chest pain and respiratory tract infection requiring hospital assessment were recorded as 2 serious adverse events affecting a single patient in the placebo group. Nonserious adverse events are summarized in eTable 4 in Supplement 2.
Discussion
The myocardium depends on oxygen for high-energy phosphate (adenosine triphosphate) production by oxidative phosphorylation. In the normal heart, adenosine triphosphate is produced primarily by the metabolism of free fatty acids and carbohydrates, with free fatty acids accounting for approximately 70% of adenosine triphosphate production in the fasting state.4 In health, free fatty acid oxidation is directly related to its plasma concentration, whereas glucose and lactate uptake are inversely related to plasma free fatty acid levels via the Randle effect. Free fatty acids are less efficient as a source of myocardial energy because they require approximately 10% more oxygen than glucose to produce an equivalent amount of adenosine triphosphate.4 The HCM is characterized by a reduction in the concentration of high-energy phosphates in the myocardium4,9 possibly because of myocardial ischemia or an energy wasting effect of sarcomere protein gene mutations.10,11
Fatty acid oxidation is regulated by the concentration of plasma free fatty acids, activity of carnitine palmitoyl transferase–1, and β-oxidation in the mitochondria. Drugs that inhibit cardiac fatty acid oxidation act in 1 of 3 ways: suppression of fatty acid release from adipocytes (eg, β-blockers), inhibition of carnitine palmitoyl transferase–1 inhibitor and fatty acid uptake into the mitochondria (eg, perhexiline), and direct inhibition of β-oxidation (eg, trimetazidine and ranolazine).
Trimetazidine, a reversible competitive inhibitor of 3-ketoacyl-coenzyme A thiolase has a good safety and tolerability profile and in placebo-controlled trials has been shown to improve exercise performance in patients with stable angina and ischemic cardiomyopathy.12,13,14 Trimetazidine appears to reduce free radical production and prevents accumulation of protons, sodium, and calcium in the myocyte.15
In this study, we showed no beneficial effect of trimetazidine on exercise capacity in patients with HCM. A negative result was also reported with ranolazine in HCM.16 This may reflect the weaker inhibition of fatty acid metabolism compared with carnitine palmitoyl transferase–1 inhibitors or that the 3-month duration of therapy was insufficient to improve symptoms. The fact that there was a 1.65 mL/kg per minute increase in the placebo group could potentially represent a harmful effect of the drug, but the change was within the 95% CI of the study design.
Limitations
The study was stopped before the end of planned recruitment because of unanticipated reluctance of patients to participate (travel distance and work commitments). The sample size in both study arms was small; thus, the power to examine the association between baseline characteristics and outcomes was limited. There was no measure of blood trimetazidine concentration; thus, compliance was assessed by medical diary and pill counting.
Conclusions
Trimetazidine therapy does not improve exercise capacity in symptomatic patients with nonobstructive HCM. Randomized therapeutic clinical trials are feasible in nonobstructive HCM, but a therapeutic role for metabolic modulators is not confirmed.
Study Protocol
eFigure. Schematic of Study Design
eTable 1. Reasons for Screen Failure
eTable 2. Descriptive Statistics for the Primary and Secondary Outcomes at Baseline
eTable 3. Descriptive Statistics for the Primary and Secondary Outcomes at 3 Months
eTable 4. Summary of Nonserious Adverse Events
Data Sharing Statement
References
- 1.Elliott PM, Anastasakis A, Borger MA, et al. ; Authors/Task Force members . 2014 ESC Guidelines on diagnosis and management of hypertrophic cardiomyopathy: the Task Force for the Diagnosis and Management of Hypertrophic Cardiomyopathy of the European Society of Cardiology (ESC). Eur Heart J. 2014;35(39):2733-2779. doi: 10.1093/eurheartj/ehu284 [DOI] [PubMed] [Google Scholar]
- 2.Marian AJ, Braunwald E. Hypertrophic cardiomyopathy: genetics, pathogenesis, clinical manifestations, diagnosis, and therapy. Circ Res. 2017;121(7):749-770. doi: 10.1161/CIRCRESAHA.117.311059 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ashrafian H, Watkins H. Myocardial dysfunction in hypertrophic cardiomyopathy. Circulation. 2001;104(25):E165. doi: 10.1161/circ.104.25.e165 [DOI] [PubMed] [Google Scholar]
- 4.Stanley WC, Recchia FA, Lopaschuk GD. Myocardial substrate metabolism in the normal and failing heart. Physiol Rev. 2005;85(3):1093-1129. doi: 10.1152/physrev.00006.2004 [DOI] [PubMed] [Google Scholar]
- 5.Abozguia K, Elliott P, McKenna W, et al. Metabolic modulator perhexiline corrects energy deficiency and improves exercise capacity in symptomatic hypertrophic cardiomyopathy. Circulation. 2010;122(16):1562-1569. doi: 10.1161/CIRCULATIONAHA.109.934059 [DOI] [PubMed] [Google Scholar]
- 6.World Medical Association World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191-2194. doi: 10.1001/jama.2013.281053 [DOI] [PubMed] [Google Scholar]
- 7.Rector TS, Cohn JN; Pimobendan Multicenter Research Group . Assessment of patient outcome with the Minnesota Living with Heart Failure questionnaire: reliability and validity during a randomized, double-blind, placebo-controlled trial of pimobendan. Am Heart J. 1992;124(4):1017-1025. doi: 10.1016/0002-8703(92)90986-6 [DOI] [PubMed] [Google Scholar]
- 8.Envelope S. https://www.sealedenvelope.com/. Accessed December 29, 2018.
- 9.Crilley JG, Boehm EA, Blair E, et al. Hypertrophic cardiomyopathy due to sarcomeric gene mutations is characterized by impaired energy metabolism irrespective of the degree of hypertrophy. J Am Coll Cardiol. 2003;41(10):1776-1782. doi: 10.1016/S0735-1097(02)03009-7 [DOI] [PubMed] [Google Scholar]
- 10.Olivotto I, Girolami F, Sciagrà R, et al. Microvascular function is selectively impaired in patients with hypertrophic cardiomyopathy and sarcomere myofilament gene mutations. J Am Coll Cardiol. 2011;58(8):839-848. doi: 10.1016/j.jacc.2011.05.018 [DOI] [PubMed] [Google Scholar]
- 11.Timmer SA, Knaapen P. Coronary microvascular function, myocardial metabolism, and energetics in hypertrophic cardiomyopathy: insights from positron emission tomography. Eur Heart J Cardiovasc Imaging. 2013;14(2):95-101. doi: 10.1093/ehjci/jes242 [DOI] [PubMed] [Google Scholar]
- 12.Fragasso G, Palloshi A, Puccetti P, et al. A randomized clinical trial of trimetazidine, a partial free fatty acid oxidation inhibitor, in patients with heart failure. J Am Coll Cardiol. 2006;48(5):992-998. doi: 10.1016/j.jacc.2006.03.060 [DOI] [PubMed] [Google Scholar]
- 13.Vitale C, Wajngaten M, Sposato B, et al. Trimetazidine improves left ventricular function and quality of life in elderly patients with coronary artery disease. Eur Heart J. 2004;25(20):1814-1821. doi: 10.1016/j.ehj.2004.06.034 [DOI] [PubMed] [Google Scholar]
- 14.Ciapponi A, Pizarro R, Harrison J. Trimetazidine for stable angina. Cochrane Database Syst Rev. 2005;(4):CD003614. [DOI] [PubMed] [Google Scholar]
- 15.Stanley WC, Marzilli M. Metabolic therapy in the treatment of ischaemic heart disease: the pharmacology of trimetazidine. Fundam Clin Pharmacol. 2003;17(2):133-145. doi: 10.1046/j.1472-8206.2003.00154.x [DOI] [PubMed] [Google Scholar]
- 16.Olivotto I, Camici PG, Merlini PA, et al. Efficacy of ranolazine in patients with symptomatic hypertrophic cardiomyopathy: the RESTYLE-HCM randomized, double-blind, placebo-controlled study. Circ Heart Fail. 2018;11(1):e004124. doi: 10.1161/CIRCHEARTFAILURE.117.004124 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Study Protocol
eFigure. Schematic of Study Design
eTable 1. Reasons for Screen Failure
eTable 2. Descriptive Statistics for the Primary and Secondary Outcomes at Baseline
eTable 3. Descriptive Statistics for the Primary and Secondary Outcomes at 3 Months
eTable 4. Summary of Nonserious Adverse Events
Data Sharing Statement