Metabolomic biomarkers and metabolic remodelling in genetically characterised hypertrophic cardiomyopathy: a systematic review

Abstract

Introduction

Hypertrophic cardiomyopathy (HCM) is the most common inherited cardiomyopathy and a major cause of heart failure, malignant arrhythmias and sudden cardiac death. Although pathogenic variants in sarcomeric genes such as MYBPC3, MYH7, TNNT2, and TNNI3 are well-recognised contributors to the disease, the mechanisms linking these mutations to the heterogeneous phenotypes remain poorly understood. Increasing evidence suggests that HCM is fundamentally associated with disturbances in myocardial energetics and metabolic remodelling.

Objectives

This systematic review aimed to synthesise metabolomic evidence from genetically characterised HCM cohorts, with three specific objectives: (1) to characterise the metabolic pathways dysregulated in HCM, (2) to determine whether distinct metabolic signatures are associated with specific sarcomeric gene variants and (3) to evaluate the diagnostic, prognostic, and therapeutic relevance of identified metabolite biomarkers in HCM management. By explicitly examining whether genotype-specific metabolomic signatures exist in the current literature, this review addresses a question of direct relevance to precision medicine approaches in HCM.

Methods

A systematic search of PubMed, Scopus and Web of Science identified 14 eligible studies comprising 1511 participants.

Results

Most studies employed liquid chromatography-mass spectrometry (LCMS) based metabolomics using plasma or myocardial tissue samples. Across studies, recurrent metabolic alterations were observed in pathways related to fatty acid oxidation, branched-chain amino acid (BCAA) metabolism, ketone body utilisation, tricarboxylic acid (TCA) cycle intermediates and lipid metabolism. Circulating acyl carnitines, amino acids, and eicosanoid metabolites were among the most frequently reported biomarkers associated with disease presence or severity. Myocardial tissue analyses demonstrated impaired mitochondrial energetics, reductions in high-energy phosphate metabolites, and shifts toward alternative metabolic substrates.

Conclusion

No distinct genotype-specific metabolic signature was identified across studies; the metabolic remodelling pattern reflected phenotypic severity rather than the causal variants. While metabolomics shows potential in improving disease characterisation and biomarker discovery, further longitudinal and externally validated studies are required before clinical translation.

Introduction

Hypertrophic cardiomyopathy (HCM) is the most common inherited cardiomyopathy, characterised by otherwise unexplained left ventricular hypertrophy (LVH) in the absence of abnormal loading conditions. Patients present with substantially heterogeneous phenotypes, ranging from asymptomatic to manifestations of cardiac decompensation such as dyspnoea, heart failure, ventricular arrhythmias and even sudden cardiac death (SCD) (Maron and Maron, 2013). A landmark study on the prevalence of HCM, estimated that this disease affects approximately 1 in 200 individuals worldwide (Semsarian et al., 2015). The genetic mutations that cause HCM are well-studied, with variants reported in 60% of cases (Basit et al., 2024). Most frequently, HCM is associated with pathogenic variants in genes encoding sarcomeric proteins, such as the myosin binding protein C3 (MYBPC3) Arg502Trp and c.1928–2 A > G, myosin heavy chain 7 (MYH7) Arg403Gln, along with contributions from other genes involved in the contractile function such as cardiac troponin type 2 (TNNT2) and cardiac troponin I3 (TNNI3) (Popa-Fotea et al., 2019; Ho et al., 2015).

These mutations disrupt normal sarcomere and function, forming the molecular basis for the development of hypertrophic remodelling. A major consequence of this is the metabolic remodelling of the myocardium, and increasing evidence suggests that HCM is fundamentally a bioenergetic disorder rather than a mere structural pathology (Marques-Alves et al., 2025). Across the studies synthesised in this review, genotyping most commonly targeted the core sarcomeric genes MYBPC3 (Jørgenrud et al., 2015, Jansen et al., 2023a), MYH7 (Ranjbarvaziri et al., 2021), TNNT2, and TNNI3 (Wang et al., 2022), with a minority of cohorts additionally screening non-sarcomeric genes such as the cardiac sodium-channel gene SCN5A (Schuldt et al., 2021) and the desmosomal gene DSP (Zhang et al., 2025). Sarcomeric gene mutations destabilise the energy-conserving states of myosin, impair relaxation and promote hypercontractility, thereby reducing the efficiency of cardiomyocyte contraction and increasing ATP consumption (Anderson et al., 2018). Previous studies demonstrated reduced myocardial phosphocreatine (PCr) to ATP ratios in mutation carriers without overt hypertrophy, indicating that energy deficiency precedes structural remodelling (Toepfer et al., 2020). This energetic imbalance is thought to trigger a cascade of downstream cellular responses, including altered calcium handling, oxidative stress, and signalling pathways that promote cardiomyocyte hypertrophy (Frey et al., 2011; Ormerod et al., 2016).

Under physiological conditions, the human adult heart derives the majority of its ATP from mitochondrial oxidative phosphorylation, through fatty acid oxidation (Brown et al., 2017). In HCM, however, increased energy demand and impaired oxidative capacity drives an incomplete and inefficient substrate shift away from fatty acids toward greater dependence on glycolysis, ketone bodies, and amino acid-related pathways (Previs et al., 2022, Nollet et al., 2024). Human myocardial tissue studies have identified reductions in high-energy phosphate metabolites, decreased fatty-acid oxidation intermediates such as acyl carnitines, and changes in tricarboxylic acid (TCA) cycle metabolites, indicating impaired oxidative metabolism and structural mitochondrial abnormalities due to damaged cristae and disordered mitochondrial architecture, as well as reduced NAD(H)-linked respiratory capacity (Kirichenko et al., 2025; Nollet et al., 2023; Ranjbarvaziri et al., 2021). These findings support a model in which mitochondrial dysfunction is integral to disease progression, further exacerbating ATP deficiency, redox stress, and inefficient energy substrate switching.

Although genetic discoveries have improved our understanding of HCM, the relationship between genotype and phenotype remains highly variable. Individuals carrying similar pathogenic variants may exhibit different clinical manifestations. Some genotype-positive individuals may have no structural abnormalities, while other patients present with severe hypertrophy, left ventricular outflow tract (LVOT) obstruction, diastolic dysfunction and progressive heart failure. Current diagnostic approaches, including electrocardiography, echocardiography and circulating biomarkers such as natriuretic peptides, largely detect structural or functional abnormalities that occur relatively late in the disease course. As such, these methods have limited ability to capture early disease processes or to predict phenotypic progression among mutation carriers, which highlight the need for more novel predictors and biomarkers capable of delineating disease stage and severity (Ho et al., 2015).

Against this background, metabolomics offers a novel approach for studying HCM. By profiling small-molecule metabolites in biological samples such as plasma, serum, or myocardial tissue, metabolomics may provide an integrated view of cellular metabolic activity and reveal signatures associated with cardiovascular disease outcomes (Ussher et al., 2016). Alongside conventional imaging and biomarkers which reflect structural or haemodynamic abnormalities in HCM, studies have suggested that metabolomics may be translated clinically and enhance therapeutic approaches (Jørgenrud et al., 2015; Schuldt et al., 2021). However, these findings are based mostly on small-cohort studies. Additionally, narrative reviews on metabolic remodelling of HCM do not systematically synthesise the use of metabolomics, appraise consistency across analytical platforms, or evaluate its diagnostic and prognostic performance. Particularly, there is lack of systematic review focusing on metabolomic findings specifically in genetically characterised cohorts, in which genotype status is known. Hence, it remains unclear whether HCM has a reproducible, genotype-informed metabolic signature and whether this signature could support earlier diagnosis or risk stratification.

To address these gaps, the present systematic review synthesises metabolomic evidence from genetically characterised HCM cohorts, with three specific objectives: (1) to characterise the metabolic pathways consistently dysregulated in HCM regardless of genotype, (2) to determine whether distinct metabolic signatures are associated with specific sarcomeric gene variants, and (3) to evaluate the diagnostic, prognostic, and therapeutic relevance of the identified metabolite biomarkers.

Methodology

This systematic review was conducted to synthesise evidence on metabolomic alterations in genetically characterised HCM and their potential roles in biomarker discovery and pathogenesis. A full protocol is available with the PROSPERO registration number CRD420261355397.

Study inclusion and exclusion criteria

Documents were included if they fulfilled the following criteria:

• Studies involving human participants with HCM were eligible. Included studies must involve HCM cohorts in which genetic testing was performed at the individual or cohort level and genotype information must be reported sufficiently to classify the cohort as genetically characterised, or where genotype status was explicitly defined for study participants. Genetic testing and reporting of genetic findings were considered helpful in supporting the diagnosis of sarcomeric HCM, given that imaging-defined unexplained LVH may include phenocopies or secondary causes of LVH with distinct metabolic profiles. Studies which did not report individual genes have been indicated in Table 1.

• Studies published in the English language, up to the 9th of March 2026.

• Studies performing metabolomics, lipidomics or metabolomics within multi-omics studies.

• Studies that utilised spectroscopy/spectrometric techniques such as LC-MS, GC-MS or NMR spectroscopy on human biospecimen.

• Studies reporting findings related to at least one of the following outcomes were included: (1) diagnostic biomarkers, (2) risk stratification, (3) prognostic outcomes, (4) survival or adverse clinical outcomes or (5) altered metabolic pathways relevant to HCM pathophysiology.

Table 1.

A comprehensive table of data extracted from 14 studies, including the population, genetic variants reported, biospecimen used, metabolomics platform used, research approach, the metabolite and pathway findings and metabolite identification level as per the Metabolomics Standards Initiative (MSI)

Author	Population	Genes reported	Sample	Instrument/Vendor Platform	Approach	Metabolite findings	Metabolic pathways altered	MSI Identification level
Jørgenrud et al. (2015)	34 MYBPC3-Q1061X mutation carriers with LVH, 19 carriers without LVH and 20 non-carrier, non-hypertrophy controls	MYBPC3-Q1061X truncating mutation	Fasting plasma	Agilent 6890 N GC; Leco Pegasus 4D GCxGC-TOFMS	Untargeted metabolomics	In carriers vs. controls: ↑ BCAAs (leucine, isoleucine, valine), threonine, triglycerides, ether phospholipids; ↓ glutamate. Findings associated with LVH phenotype and systolic dysfunction.	Amino acid metabolism, phospholipid metabolism, fatty acid oxidation	2
Schuldt et al., 2021	31 asymptomatic carriers, 14 HOCM carriers, 9 healthy controls (family members)	MYBPC3, MYH7, TNNT2, TNNI3, SCN5A, SMN, MYL2	Fasting serum	Direct infusion HRMS (vendor not disclosed)	Untargeted metabolomics	In HOCM carriers: ↑ alanyl-leucine, stearic acid, glutarylcarnitine; tetracosanoic acid, monoacylglycerols, ↓ acyl carnitines	Fatty acid oxidation, amino acid metabolism and lipid metabolism	2
Wolter et al. (2021)	18 HOCM patients (15 G-, 3 G+) with LVOT obstruction undergoing SRT	MYBPC3, MYH7	Plasma collected pre and post-myectomy	Waters ACQUITY UPLC; Thermo Q-Exactive Orbitrap HRMS	Untargeted metabolomics	Post-myectomy: ↓, phosphatidylcholine, sphingomyelins, arginine metabolites, β-hydroxybutyrate, trigonelline. Also reported ↓ bilirubin, biliverdin, PFOS, and PFOA suggesting improved perfusion and excretory function post-myectomy.	Suggested improvement in heme metabolism, arginine metabolism, phospholipid metabolism.	2
Liu et al. (2021)	2 familial pedigrees (n = 9), 5 G+ with HCM and 4 G- healthy controls	Family 1: TNNT2-rs397516484 and MYH6-rs372446459 (n = 3) Family 2: MYBPC3-rs786204339 frameshift mutation (n = 2)	Plasma	Agilent UPLC; AB Sciex QTOF-MS	Untargeted metabolomics, WES and echocardiographic phenotyping	Individual metabolites not reported	Family 1 (TNNT2 + MYH6 carriers): Glycolysis/gluconeogenesis, pentose and glucuronate interconversions, purine metabolism, amino sugar and nucleotide sugar metabolism Family 2 (MYBPC3 carriers): Citrate cycle (TCA cycle), glycerolipid metabolism, fructose and mannose metabolism, glycerophospholipid metabolism	2
Ranjbarvaziri et al. (2021)	27 HOCM patients undergoing septal myectomy, 13 non-failing donor hearts; 2 mitral stenosis samples for respiration assays only Mixed cohort of G+, G- and genotype unknown	MYBPC3, MYH7, CSRP3, TNNT2, SCN5A	Left ventricle myocardial tissue	Thermo Q Exactive HF MS (HILIC/RPLC)	Multi-omics (untargeted and targeted metabolomics, transcriptomics and lipidomics)	↑ Free fatty acids, ceramides, sphingomyelins; ↓ acyl carnitines, carnitine, TCA metabolites, ATP/ADP/PCr, many amino acids	Fatty acid oxidation, glucose metabolism, pentose phosphate pathway, TCA cycle, nucleotide biosynthesis, oxidative phosphorylation, ROS pathways, mitophagy	1, 2
Ravi et al. (2022)	38 carriers of HCM variants (8 diagnosed with HCM, 28 not clinically diagnosed) vs. 180 age and BMI-matched healthy controls	MYH7 Arg1053Gln (n = 24) Arg941His (n = 4) MYBPC3 Gln1061Ter (n = 3), c.655–2 A > C (n = 4), Gly853fs (n = 1) TPM1 Asp175Asn (n = 2).	Fasting plasma	Waters ACQUITY UPLC; Thermo Q Exactive Orbitrap HRMS	Untargeted metabolomics	In HCM: ↑ vanillactate, des-Arg 9-bradykinin, dimethylglycine, ethyl beta-glucopyranoside. ↓ 2-hydroxypalmitate, (14 or 15)-methylpalmitate, EPA, arachidonate, DHA, multiple lysophosphatidylcholines, choline-lysoplasmalogens, lysophosphatidylinositol.	Amino acid metabolism, fatty acid oxidation, glycerophospholipid metabolism	2
Previs et al. (2022)	41 HCM myectomy samples (12 genotype negative, 29 variant carriers) vs. donor hearts	MYBPC3 (n = 19), MYH7 (n = 10)	Interventricular septum myocardial tissue	Dionex UltiMate 3000 UHPLC; Thermo Q Exactive MS	Multi-omics (proteomics and targeted metabolomics)	In HCM: ↓ acyl carnitines, ATP, PCr, acetyl-CoA, NAD(H), serine, methionine. ↑ β-hydroxybutyrate, lactate, BCAAs (leucine, isoleucine, valine). Found no differences between genotypes.	Fatty acid oxidation; ketone metabolism; BCAA metabolism; energy metabolism	1, 2
Wang et al. (2022)	349 HCM patients vs. 16 non-HCM controls (cardiac tissue); 143 HCM vs. 60 controls (plasma); Additional 46 DCM comparison	MYH7, MYBPC3, TNNT2, TNNI3, MYL2, MYL3, TPM1, ACTC1 (~ 58% mutation carriers; MYH7 and MYBPC3 most common)	Left ventricle myocardial tissue (septum) and plasma	Shimadzu Nexera×2 UPLC; Sciex Qtrap 6500 + MS	Multi-omics (targeted metabolomics, lipidomics, proteomics, WES, machine learning, clustering)	In HCM: ↓ acyl carnitines (impaired fatty acid oxidation) ↓ glycerophospholipids (LPE, LPC) ↑ sphingolipids (HexCer, SM), DAG ↓ TAGs (especially low double-bond species) ↑ gluconate (PPP), nucleotide metabolites ↑ oxidative stress markers (GSSG) Short-chain acyl carnitines ↓ in both plasma and tissue; long-chain acyl carnitines ↓ in tissue but ↑ in plasma	Fatty acid oxidation; glycolysis; pentose phosphate pathway; TCA cycle; purine and pyrimidine metabolism; glycerophospholipid metabolism; sphingolipid metabolism; oxidative phosphorylation; redox/oxidative stress pathways	1, 2
Jansen et al. (2023a)	30 carriers with severe HCM and 30 carriers with mild/no phenotype vs. 10 genotype-negative controls	MYBPC3 c.2373dupG, c.2827 C > T, c.2864_2865delCT, c.3776delA	Plasma	Direct-infusion HRMS (vendor not disclosed)	Untargeted metabolomics	Aminoadipic acid, 9,12-hexadecadienoylcarnitine, uric acid, 3-methylhistidine	Acyl carnitine, histidine, lysine, purine, steroid metabolism, proteolysis	2
Jansen et al. (2023b)	85 carriers (59 severe HCM, 26 mild HCM) vs. 39 non-carriers	MYBPC3 c.2373dupG, c.2827 C > T, c.2864_2865delCT, and c.3776delA.	Plasma	Waters Acquity UPLC-MS/MS	Targeted metabolomics (acyl carnitine panel)	In carriers with severe HCM: ↑ acyl carnitines C3, C4, C6-DC, C8:1, C16, C18, and C18:2. In mild HCM: ↑ C3, C6-DC, C8:1, and C18. C6-DC correlated with maximum wall thickness and LVEF, and C8:1 correlated with maximum wall thickness	Fatty acid oxidation intermediates	1, 2
Nollet et al. (2024)	47 carriers with HOCM (G+(n = 19), G- (n = 22) and genotype unknown (n = 6)undergoing septal myectomy, 20 non-failing donor hearts	Individual genes not reported	Myocardial tissue	Dionex Ultimate 3000 NanoLC; Thermo Orbitrap Elite/Fusion Lumos MS/MS	Multi-omics (transcriptomic, proteomic, untargeted metabolomics, lipidomics)	In G + HCM: ↓ acyl carnitines C16/C18; ↑ β -hydroxybutyrate, leucine/isoleucine/valine. G- HCM: ↑ lactate	Fatty acid oxidation; ketone metabolism; BCAA metabolism; glycolysis; mitochondrial energetics	2
Wu et al. (2025)	5 carriers with HCM vs. 10 healthy controls	MYBPC3 P459fs (c.1377delC)	Whole blood	Shimadzu UFLC; AB Sciex 6500 Q TRAP MS/MS	Multi-omics (untargeted metabolomics, WES, transcriptomics, iPSC-derived modelling)	In carriers with HCM: ↓ long-chain acyl carnitines (C16, C18), ATP. ↑β-hydroxybutyrate, BCAAs (leucine, isoleucine, valine) lactate.	Fatty acid oxidation; ketone metabolism; BCAA metabolism; glycolysis	2
Zhang et al. (2025)	73 HCM vs. 78 healthy controls Mixed G+, G- and genotype unknown cohort	DSP, FLNC, MYBPC3, MYH7, MYPN, TNNI3, TNNT2	Plasma	Waters Acquity UPLC; Sciex QTRAP-5500 MS	Targeted lipidomics	In HCM: ↓ eicosanoids (AA, 5,6-DHET, 12-HETE, LXA4, EPA) 12-HETE and EPA Did not report genotype differences	Arachidonic acid/eicosanoid metabolism	1, 2
Ikoma et al. (2026)	61 HOCM (SRT, n = 20, non-SRT, n = 41)	MYBPC3, TNNT2, TNNI3	Myocardial tissue	Waters Xevo G2-XS QTOF-MS (DESI-IMS)	Spatial metabolomics	In HCM, ↑ testosterone glucuronide	Steroid metabolism and androgen signalling pathways	2

AA, arachidonic acid; ADP, adenosine diphosphate; ATP, adenosine triphosphate; BCAAs, branched-chain amino acids; DAG, diacylglycerol; DCM, dilated cardiomyopathy; DESI-IMS, desorption electrospray ionization imaging mass spectrometry; DHA, docosahexaenoic acid; DHET, dihydroxyeicosatrienoic acid; EPA, eicosapentaenoic acid; ESI, electrospray ionization; genotype positive; G-, genotype negative; GC, gas chromatography; GCxGC-TOFMS, two-dimensional gas chromatography-time-of-flight mass spectrometry; GSSG, glutathione disulfide; HCM, hypertrophic cardiomyopathy; HESIII, heated electrospray ionization source; HETE, hydroxyeicosatetraenoic acid; HexCer, hexosylceramide; HILIC, hydrophilic interaction liquid chromatography; HOCM, hypertrophic obstructive cardiomyopathy; HRMS, high-resolution mass spectrometry; iPSC, induced pluripotent stem cell; LC-MS/MS, tandem liquid chromatography-mass spectrometry; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; LVEF, left ventricular ejection fraction; LVH, left ventricular hypertrophy; LVOT, left ventricular outflow tract; LXA4, lipoxin A4; MRM, multiple reaction monitoring; MS, mass spectrometry; NanoLC, nanoflow liquid chromatography; PCr, phosphocreatine; PFOA, perfluorooctanoic acid; PFOS, perfluorooctanesulfonic acid; PPP, pentose phosphate pathway; QTOF/Q-TOF, quadrupole time-of-flight; ROS, reactive oxygen species; RPLC, reversed-phase liquid chromatography; SM, sphingomyelin; SRT, septal reduction therapy; TAG, triacylglycerol; TCA, tricarboxylic acid; UFLC, ultra-fast liquid chromatography; UHPLC, ultra-high pressure liquid chromatography; UPLC, ultra-performance liquid chromatography; and WES, whole exome sequencing

Documents were excluded if they fulfilled the following criteria:

• Studies which did not investigate HCM, or where HCM data could not be extracted separately.

• Documents which are not original research articles (such as review articles, editorials, commentaries or conference proceedings).

• Studies reporting on phenocopies such as Fabry disease, mitochondrial diseases, amyloidosis, or secondary cardiac hypertrophy resulting from other diseases (such as hypertension or aortic stenosis).

• Studies conducted using imaging energetic techniques alone, such as phosphorus magnetic resonance spectroscopy (31P-MRS) and positron emission tomography (PET).

• Animal studies, in vitro studies or cell-line experiments.

• Studies without extractable metabolomic findings and pathways relevant to the review objectives.

Search methodology and data extraction

The review methodology followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020, Page et al., 2020) guidelines (Fig. 1). A comprehensive literature search was conducted on three major databases, namely PubMed, Scopus and Web of Science on the 9th of March 2026. The search strategy combined terms related to hypertrophic cardiomyopathy, metabolomics and its translational impacts, using a combination of medical search headings (MeSH) and free-text keywords. An example of the search string on PubMed is as follows:

Fig. 1.

PRISMA flowchart of the study selection process. PRISMA: preferred reporting items for systematic review and meta-analyses

(Cardiomyopathy, Hypertrophic[Mesh] OR hypertrophic cardiomyopathy*[tiab] OR hypertrophic obstructive cardiomyopath*[tiab] OR HCM[tiab] OR HOCM[tiab]) AND (Metabolomics [Mesh] OR Lipidomics[Mesh] OR metabolomic*[tiab] OR lipidomic*[tiab] OR metabonomic*[tiab] OR metabolome*[tiab] OR metabolite profil*[tiab] OR metabolic profil*[tiab] OR metabolic signature*[tiab] OR metabolite signature*[tiab] OR fatty acid*[tiab] OR ketone bod* [tiab] OR acylcarnitine*[tiab] OR amino acid*[tiab] OR glucose[tiab]) AND (biomarker*[tiab] OR marker*[tiab] OR diagnos*[tiab] OR prognos*[tiab] OR “risk stratification“[tiab] OR survival[tiab] OR outcome*[tiab] OR pathway*[tiab] OR pathophysiolog*[tiab])

In addition to the electronic database search, relevant articles were also manually searched via Google Scholar to identify additional eligible studies. One study that met the inclusion criteria but was not captured by the search strategy was identified through this process and included in the review. Our initial search retrieved a total number of 985 entries combined. Using a web-based platform (www.tera-tools.com), the initial dataset was deduplicated, resulting in a clean list of 413 records. Thereafter, 2 authors independently reviewed each title and abstract. In the initial screening, documents which were not in the review scope, such as review articles, non-HCM related, LVH phenocopies, studies conducted on animal models or cell lines, and imaging energetic modalities only (31P-MRT, PET CT) studies were excluded, leaving a preliminary total of 27 records. Studies which did not report genetically studied cardiomyopathy cases, no metabolite pathways, possible biomarkers or translational implications, were excluded. Discrepancies were resolved by consensus discussion, yielding a final total of 14 studies in the review.

To evaluate whether the class-specific free-text terms (fatty acid*, ketone bod*, acylcarnitine*, amino acid*, glucose) disproportionately influenced study retrieval, we cross-checked each of the 14 included studies against the individual components of the search string. Thirteen of the 14 studies (93%) were retrievable using the broad, generic omics terms alone (Metabolomics[Mesh], Lipidomics[Mesh], or free-text equivalents such as metabolomic*, lipidomic*, metabolome*, metabolite/metabolic profil*, metabolite/metabolic signature*), independent of any metabolite-class term. Only one study (Jansen et al. (2023b)) was retrieved solely via a class-specific term (acylcarnitine*). This indicates that the class-specific terms served their intended purpose as a narrow sensitivity safeguard for literature not indexed under general omics terminology, rather than functioning as the primary determinant of which studies or metabolite classes were captured in this review.

Data synthesis and reporting

We extracted data from each included study using a standardised framework developed prior to analysis. To address the expected heterogeneity in study design, metabolomics platforms, and reported metabolites across studies, we compiled and tabulated data across two domains to summarise our findings.

DOMAIN 1. General study characteristics and design

The characteristics of included studies were summarised, including name of authors, year of publication, geographical distribution, study designs, participant demographics, omics platforms used, and types of biospecimens analysed. Throughout this review, genotype-positive denotes participants identified as carriers of a pathogenic or likely-pathogenic variant in a sarcomere-encoding gene or associated non-sarcomeric gene, whereas genotype-negative denotes those who underwent comprehensive genetic testing in which no such causative variant was identified.

DOMAIN 2. Metabolites, metabolic pathways and translation impact reported

Metabolites associated with diagnostic discrimination, risk stratification, and prognostic outcomes in HCM were summarised. Where available, performance metrics such as area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, hazard ratios, and odds ratios were extracted. Metabolites reported as significantly altered in included studies were compiled into a comprehensive table, and those that consistently appeared across independent datasets were highlighted as potential metabolic signals associated with HCM. Based on the Metabolomics Standards Initiative (MSI), metabolite identification levels were assigned based on the confidence of metabolite annotation reported in the included studies (Sumner et al., 2007). MSI Level 1 identification corresponds to metabolites that were confirmed using authentic reference standards, typically through matching of retention time, mass spectra, or fragmentation patterns. MSI Level 2 represents putatively annotated metabolites, identified based on spectral similarity to entries in established metabolomics databases, such as the Human Metabolome Database (HMDB) and the Kyoto Encyclopedia of Genes and Genomes (KEGG).

Results

Study characteristics

The 14 included studies were published between 2015 and 2026 and comprised a total of 1511 participants aged 23 to 71 years. A comprehensive summary of the studies can be found in Table 1. Most studies included both male and female participants, with one study conducted in a male-only cohort (Ravi et al., 2022). Across studies, comparisons were commonly made between individuals with varying HCM severity and healthy controls or non-carrier groups, while fewer studies compared carriers with non-carriers without detailed stratification by disease stage (Jørgenrud et al., 2015; Schuldt et al., 2021; Zhang et al., 2025; Wu et al., 2025). Clinical classification was primarily performed using either echocardiography or cardiac magnetic resonance imaging (MRI), with parameters such as impaired left ventricular outflow tract (LVOT) gradient, reduced left ventricular ejection fraction (LVEF), increased maximal wall thickness (MWT), and increased left ventricular mass used to define disease severity and obstruction (Ikoma et al., 2026; Zhang et al., 2025; Wu et al., 2025; Nollet et al., 2024; Jansen et al., 2023b; Previs et al., 2022; Liu et al., 2021; Schuldt et al., 2021; Wolter et al., 2021). A number of studies also incorporated circulating biomarkers such as the brain natriuretic peptides (BNP) and N-terminal pro-B-type natriuretic peptides (NT-proBNP) (Ikoma et al., 2026; Zhang et al., 2025; Wu et al., 2025; Schuldt et al., 2021), all of which reported increased levels in obstructive HCM patients.

A range of biospecimens and analytical platforms were used, among which plasma was the most frequently analysed sample type (Jørgenrud et al., 2015; Wolter et al., 2021; Liu et al., 2021; Ravi et al., 2022; Jansen et al., 2023b; Zhang et al., 2025), followed by left ventricular or septal myocardial tissue (Ranjbarvaziri et al., 2021; Previs et al., 2022; Nollet et al., 2024; Ikoma et al., 2026), serum (Schuldt et al., 2021), whole blood (Wu et al., 2025), and combined plasma plus myocardial tissue analysis (Wang et al., 2022). LC-MS based platforms were the most commonly utilised (Ranjbarvaziri et al., 2021; Ravi et al., 2022; Previs et al., 2022; Wang et al., 2022; Nollet et al., 2024; Wu et al., 2025; Zhang et al., 2025; Wolter et al., 2021; Liu et al., 2021), although studies also used direct-infusion high-resolution mass spectrometry (HR-MS) (Schuldt et al., 2021; Jansen et al., 2023a), gas chromatography (Jørgenrud et al., 2015), and spatial metabolomics using desorption electrospray ionisation imaging mass spectrometry (DESI-IMS) (Ikoma et al., 2026). Untargeted metabolomics was the approach of choice in many studies (Jørgenrud et al., 2015; Schuldt et al., 2021; Wolter et al., 2021; Liu et al., 2021; Ravi et al., 2022; Jansen et al., 2023a; Wu et al., 2025), while a minority employed targeted approaches or broader multi-omics workflows integrating lipidomics, proteomics and transcriptomics. (Ranjbarvaziri et al., 2021; Previs et al., 2022; Wang et al., 2022; Nollet et al., 2024; Zhang et al., 2025).

Although all included studies involved genetically characterised HCM cohorts, there was substantial heterogeneity in individual genotype reporting and classification. Studies either included mixed populations of genotype-positive and genotype-negative individuals or did not directly compare metabolic profiles across specific genetic variants (Wang et al., 2022.; Previs et al., 2022; Nollet et al., 2024; Wolter et al., 2021; Zhang et al., 2025; Ikoma et al., 2026). Consequently, the findings presented reflect metabolomic alterations observed in genetically studied HCM populations rather than genotype-specific metabolic differences. The most frequently reported gene across studies was MYBPC3 (Jørgenrud et al., 2015, Jansen et al., 2023a; Jansen et al., 2023b) followed by other core sarcomeric genes including MYH7, TNNT2, TNNI3 (Schuldt et al., 2021; Wang et al., 2022, Zhang et al., 2025), MYH6 (Liu et al., 2021), MYL2, MYL3, ACTC1 (Wang et al., 2022), TPM1 (Ravi et al., 2022). A smaller number of studies additionally screened Z-disc and sarcomere-associated genes, namely CSRP3 (Ranjbarvaziri et al., 2021), MYPN and FLNC (Zhang et al., 2025), as well as non-sarcomeric genes included in broader inherited arrhythmia and cardiomyopathy panels, such as SCN5A (Ranjbarvaziri et al., 2021; Schuldt et al., 2021) and DSP (Zhang et al., 2025).

A subset of studies investigated metabolomic changes in patients undergoing septal reduction therapy (SRT) or septal myectomy (Wolter et al., 2021; Ranjbarvaziri et al., 2021; Nollet et al., 2024; Ikoma et al., 2026). These studies used either longitudinal plasma sampling or myocardial tissue obtained during surgery, providing additional insight into metabolic changes associated with severe obstructive disease and presented metabolomics data in the context of surgical intervention. Among the 14 studies included in this review, six did not attempt metabolic discrimination based on genotype, prioritising instead clinical disease staging, phenotypic severity, or comparisons involving a single specific mutation. Wolter et al., (2021) focused on the longitudinal metabolic effects of surgical myectomy, noting that their low prevalence of sarcomere mutations (17%) precluded balanced genotype stratification. Both Jørgenrud et al., (2015) and Wu et al., (2025) conducted investigations restricted to single specific founder variants (MYBPC3-Q1061X and MYBPC3-P459fs, respectively) comparing carriers to healthy relatives or wild-type controls rather than comparing different genetic backgrounds. Similarly, the exploratory studies by Jansen et al. (2023a, 2023b) utilised a genetically homogeneous population of MYBPC3 founder variant carriers to identify biomarkers that distinguished severe HCM from mild or preclinical phenotypes. Finally, while Schuldt et al., (2021) and Ravi et al., (2022) identified the genetic makeup of their participants, their metabolic modelling was designed to discriminate clinical disease stages (preclinical versus obstructive) or carriers of pathogenic variants generally from healthy controls, without attempting to group or differentiate patients by specific genotype.

Metabolites and metabolic pathways implicated in HCM

Across studies, metabolomic profiling revealed recurrent disturbances in pathways related to myocardial energy metabolism, despite heterogeneity in study design, biospecimen type, and analytical platforms. Both circulating and myocardial tissue analyses demonstrated disturbances in fatty acid oxidation, amino acid metabolism, ketone utilisation, glycolysis, and tricarboxylic acid (TCA) cycle metabolism.

Disturbances in fatty acid oxidation were among the most consistent metabolic findings across HCM cohorts, although the direction of change differed between myocardial tissue and circulating metabolites. Myocardial tissue studies consistently reported reduced long-chain acyl carnitines alongside depletion of high-energy phosphates and acetyl-CoA(Ranjbarvaziri et al., 2021; Previs et al., 2022; Wang et al., 2022; Nollet et al., 2024; Wu et al., 2025). These studies also reported evidence of mitochondrial damage and reduced citrate synthase activity (Ranjbarvaziri et al., 2021). In contrast, circulating studies reported elevated long-chain acyl carnitines in patients with more severe phenotypes (Jansen et al., 2023b). Several plasma studies additionally reported reductions in polyunsaturated fatty acids (PUFAs) and lipid-related species, including lysophosphatidylcholines and glycerophospholipids (Ravi et al., 2022; Wang et al., 2022). Targeted lipidomics further identified reductions in arachidonic acid, EPA, DHA, and 12-HETE (Zhang et al., 2025), while myocardial studies reported increased ceramides and sphingomyelins (Ranjbarvaziri et al., 2021; Wang et al., 2022).

Alterations in amino acid metabolism, particularly involving branched-chain amino acids (BCAAs), were also consistently reported across many HCM cohorts. Elevated leucine, isoleucine, and valine were observed in both plasma and myocardial tissue and were associated with greater hypertrophic disease severity (Jørgenrud et al., 2015; Previs et al., 2022; Nollet et al., 2024; Wu et al., 2025). Other amino acid-related metabolites, including aminoadipic acid and 3-methylhistidine, were associated with more severe phenotypes in an MYBPC3-confirmed cohort, while dimethylglycine was also reported as a circulating amino acid-related abnormality (Jansen et al., 2023a; Ravi et al., 2022). Vanillactate and des-Arg9-bradykinin were additionally identified in circulating metabolomic studies, although these represent distinct catabolic pathways related to tyrosine/catecholamine metabolism and the kinin-kallikrein system, respectively (Ravi et al., 2022).

Evidence of increased reliance on alternative energy substrates was observed in both tissue-based and circulating studies. Elevated levels of β-hydroxybutyrate were reported in hypertrophied myocardium, consistent with increased ketone utilisation in the setting of impaired oxidative metabolism (Previs et al., 2022; Nollet et al., 2024; Wu et al., 2025). Lactate accumulation was also reported in myocardial tissue and circulating analytes, indicating a shift toward glycolytic metabolism (Previs et al., 2022; Nollet et al., 2024; Wu et al., 2025). This shift toward alternative substrate use in myocardial tissue was accompanied by depletion of high-energy phosphate metabolites, including ATP, ADP, and phosphocreatine (Ranjbarvaziri et al., 2021; Previs et al., 2022; Wang et al., 2022; Wu et al., 2025). Interestingly, Wolter et al. provided longitudinal plasma data showing postoperative reductions in β-hydroxybutyrate, as well as in bilirubin, biliverdin, PFOS, PFOA, trigonelline, phosphatidylcholines, sphingomyelins, and arginine-related metabolites. As PFOS and PFOA are renally and hepatically cleared exogenous compounds, their reduction, together with that of bilirubin and biliverdin, is consistent with improved renal and hepatic excretory function following surgical relief of LVOT obstruction, in line with the study’s interpretation of improved organ perfusion after myectomy, and suggests that some systemic abnormalities in HCM extend beyond the myocardium and may be reversible with surgical treatment (Wolter et al., 2021).

Alterations in central carbon metabolism were also recurrently observed. Distinct carbohydrate-metabolism disturbances were reported in a plasma metabolomic study of two unrelated HCM family pedigrees (nine participants in total, five of whom carried pathogenic sarcomeric variants: three carriers of combined TNNT2 and MYH6 variants in one family, and two carriers of a MYBPC3 variant in the other). In the TNNT2/MYH6-affected family, altered metabolites were enriched in glycolysis and gluconeogenesis, pentose and glucuronate interconversions, purine metabolism, and amino sugar and nucleotide sugar metabolism, whereas in the MYBPC3-affected family, altered metabolites were enriched in the TCA cycle, fructose and mannose metabolism, and glycerophospholipid metabolism (Liu et al., 2021). Myocardial tissue studies showed reductions in TCA cycle intermediates such as citrate, malate, and succinate, consistent with impaired oxidative metabolism in the failing myocardium (Ranjbarvaziri et al., 2021). Other work identified disruption of the pentose phosphate pathway, purine and pyrimidine metabolism, and markers of oxidative stress or redox imbalance in association with disease severity. In the same study, galactose 1-phosphate was positively associated with worse cardiac function, whereas UDP-galactose, G6P-F6P, and GSSG showed inverse associations with NYHA class-based severity measures (Wang et al., 2022).

Genotype and metabolic phenotype correlations in HCM

A synthesis of data across the reviewed literature indicates that the overall metabolic profile of HCM is independent of genotype status once the clinical disease phenotype has manifested. Several studies explicitly reported no significant metabolomic differences between genotype-positive and genotype-negative HCM groups or stated that their principal metabolic findings were independent of genotype (Previs et al., 2022; Wang et al., 2022). Some studies reported that genotype-positive and genotype-negative patients do not cluster separately in principal component analyses (PCA), suggesting that metabolic remodelling represents a “final common pathway” of the disease (Ranjbarvaziri et al., 2021; Wang et al., 2022). Furthermore, at a group level, no individual metabolites have been identified with significantly different absolute abundances when comparing genotype-positive directly to genotype-negative cohorts (Nollet et al., 2024; Wang et al., 2022).

Similar findings were also observed when comparing different pathogenic sarcomere variants, such as MYH7 and MYBPC3, which displayed nearly identical metabolic signatures (Previs et al., 2022; Wang et al., 2022). This convergent remodelling is also reflected in circulating eicosanoid profiles, which are reduced across a variety of mutations, including FLNC, MYBPC3, MYH7, TNNI3, and TNNT2, compared to healthy controls (Zhang et al., 2025). Even in populations specifically enriched for founder MYBPC3 mutations, the metabolic signatures aligned more strongly with disease phenotype than with genotype (Jørgenrud et al., 2015; Jansen et al., 2023a, b). While absolute metabolite levels are similar between genotypes, some differences exist in how these molecules associate with remodelling; for instance, biosynthetic and anabolic precursors such as purines and pyrimidines display distinct correlation patterns with septal hypertrophy in genotype-positive versus genotype-negative patients (Nollet et al., 2024).

Biomarker performance and translational relevance

A targeted eicosanoid panel comprising 12-HETE and EPA exhibited strong diagnostic capability for distinguishing HCM cases, achieving an AUC of 0.884, which indicates good discriminatory power between affected individuals and controls (Zhang et al., 2025). Serum metabolite signatures identified by multivariate modelling discriminated preclinical and obstructive HCM carriers with very high accuracy and performed comparably with imaging-derived energetic parameters (Schuldt et al., 2021). Wang and colleagues achieved excellent discrimination between HCM and controls (AUC 0.976, AUPRC 0.991) using a plasma-based random forest model using five metabolites, namely 8:0-carnitine, hypoxanthine, creatine, phenylalanine, and tryptophan. They also proposed that metabolomic profiles could stratify HCM into distinct subtypes with differing clinical outcomes, with poorer prognosis associated with increased purine metabolism and reduced carnitines, and better survival linked to higher TCA intermediates such as isocitrate and fumarate (Wang et al., 2022). In addition, Jansen et al. reported that circulating acyl carnitines correlated with ventricular wall thickness and LVEF in MYBPC3-positive cohorts, suggesting the potential use of metabolomics for subtype classification (Jansen et al., 2023b). Collectively, these findings support the potential translational value of metabolomics for disease stratification and outcome prediction, although external validation remains limited.

Emerging metabolic pathways

In addition to the major pathways above, several studies identified less commonly reported pathways that may represent emerging aspects of HCM metabolism. These included steroid-related metabolites such as testosterone glucuronide (Ikoma et al., 2026), purine metabolites including uric acid and broader purine pathway activation (Jansen et al., 2023a; Wang et al., 2022), histidine-related metabolites and proteolysis-associated signals, and lipid mediators from eicosanoid metabolism (Zhang et al., 2025). Although these pathways were less consistently reported than fatty acid oxidation or amino acid metabolism, they suggest that HCM metabolic remodelling may extend beyond core bioenergetic pathways into broader endocrine, oxidative stress, and inflammatory signalling networks (Zhang et al., 2025; Wang et al., 2022; Ikoma et al., 2026).

Discussion

This systematic review gathers metabolomics evidence from genetically characterised HCM cohorts and provides a unified interpretation of metabolic remodelling in the disease. Across 14 studies, a consistent pattern emerges in which sarcomeric gene mutations result in a shared metabolic phenotype characterised by impaired myocardial energetics, mitochondrial dysfunction, and altered substrate utilisation. Rather than identifying genotype-specific metabolic signatures, the current evidence supports a model in which diverse genetic variants lead to a common energy deficit that leads to extensively variable clinical presentations.

Despite methodological differences across studies, including different sample types, analytical platforms, sample size and population, recurrent metabolic disturbances were observed in fatty acid oxidation, amino acid metabolism, ketone utilisation, and central carbon metabolism. These findings reinforce the concept that HCM is fundamentally a metabolic and energetic disorder, with structural remodelling representing a consequence rather than the primary driver.

Mechanistic interpretation of metabolic remodelling in HCM

Sarcomeric gene mutations in MYBPC3, MYH7, and TNNT2 disrupt cross-bridge cycling regulation, promoting hypercontractility and impaired relaxation, which increases ATP consumption and reduces contractile efficiency (Anderson et al., 2018). This energetic imbalance is evident even in genotype-positive individuals without overt hypertrophy (Toepfer et al., 2020), indicating that bioenergetic impairment is an early, potentially initiating event in disease development.

This energy demand is compounded by mitochondrial dysfunction; reduced oxidative phosphorylation capacity, impaired NAD(H)-linked respiration, and structural abnormalities (Nollet et al., 2024). This dysfunction diminishes ATP generation and increases ROS production, contributing to oxidative stress and myofilament dysregulation (Ranjbarvaziri et al., 2021; Cuello et al., 2018). Impaired calcium handling further elevates intracellular calcium burden, exacerbating this cycle (Viola and Hool, 2019), while coronary microvascular dysfunction and relative myocardial ischaemia limit oxygen delivery and oxidative metabolism at the tissue level (Karamitsos et al., 2013; Van der Velden et al., 2018). Together, these processes form a self-reinforcing cycle, in which sarcomere inefficiency raises energy demand while mitochondrial injury and microvascular ischaemia reduce supply, leading to downstream dysregulation of AMPK, mTOR, and PPAR-related signalling implicated in hypertrophic remodelling and fibrosis, though their precise causal role in human HCM remains incompletely defined (Skórka et al., 2025; Lumish et al., 2024).

This model is consistent with the findings reviewed here: hypertrophied myocardium shows reduced fatty acid oxidation capacity, depletion of high-energy phosphate metabolites, and increased reliance on glycolysis, ketone bodies and amino acids, reflecting a state of energetic inefficiency that likely underpins disease progression and phenotypic heterogeneity in HCM (Fig. 2).

Fig. 2.

Proposed model of metabolic remodelling in genetically characterised hypertrophic cardiomyopathy (HCM). This figure summarises how sarcomeric gene variants lead to energetic imbalance and mitochondrial dysfunction, driving widespread metabolic remodelling across multiple pathways and contributing to the diverse clinical manifestations observed in HCM. Key: ↑ increased/elevated; ↓ decreased/reduced; FAO, fatty acid oxidation; BCAA, branched-chain amino acids; TCA, tricarboxylic acid; PPP, pentose phosphate pathway; ROS, reactive oxygen species; ETC, electron transport chain; SCD, sudden cardiac death

Reduced myocardial acyl carnitines and depletion of energetic intermediates may reflect impaired mitochondrial fatty acid oxidation and diminished ATP-generating capacity, potentially related to mitochondrial damage and reduced citrate synthase activity (Ranjbarvaziri et al., 2021). Conversely, the accumulation of long-chain acyl carnitines in the circulation, particularly in more severe disease, may represent incomplete fatty acid oxidation and systemic metabolic spillover rather than enhanced fatty acid utilisation (Jansen et al., 2023b). The accompanying changes in PUFAs, eicosanoids, ceramides, and sphingomyelins further suggest a broader disruption of lipid homeostasis, including membrane remodelling and altered lipid signalling. Collectively, this pattern is consistent with impaired mitochondrial lipid utilisation and reduced metabolic flexibility, potentially contributing to energetic stress and greater reliance on alternative substrates in HCM, which are metabolically inefficient.

Consequential to impaired fatty acid oxidation, alternative substrate use is evident. Tissue and circulating studies consistently reported elevated β-hydroxybutyrate (Previs et al., 2022; Nollet et al., 2024; Wu et al., 2025), indicating increased ketone utilisation as an oxygen-efficient compensatory substrate. This shift appears at least partially reversible: β-hydroxybutyrate declined following relief of LVOT obstruction by septal myectomy (Wolter et al., 2021), suggesting dynamic rather than fixed substrate remodelling. Whether increased ketone metabolism is adaptive or simply marks underlying inefficiency remains unclear. Similarly, elevated lactate reflects increased, less-efficient glycolytic flux, potentially contributing to intracellular acidosis. Comparable substrate shifts occur in other forms of cardiac hypertrophy and heart failure, lending broader relevance to targeted metabolic therapies such as sodium-glucose cotransporter-2 (SGLT2) inhibitors (Kim and Jung, 2025; Greene et al., 2024), though their specific role in HCM requires further study.

The accumulation of BCAAs and other amino acid-related metabolites may reflect altered mitochondrial substrate utilisation and protein turnover under chronic energetic stress. BCAAs are catabolised predominantly through the mitochondrial branched-chain α-keto acid dehydrogenase (BCKDH) complex (White et al., 2018), hence their accumulation may be consistent with impaired mitochondrial oxidative capacity. BCAAs have also been proposed to modulate hypertrophic signalling through pathways such as mTOR (Walejko et al., 2021), although direct evidence for a causal role in human HCM remains limited. However, whether these changes represent adaptive responses, downstream consequences, or contributors to HCM pathophysiology remains uncertain. Additionally, circulating BCAA concentrations are also influenced by systemic factors, including insulin sensitivity and peripheral substrate utilisation, and therefore may not directly reflect myocardial metabolism (Lynch and Adams, 2014).

Beyond these core pathways, several emerging findings warrant cautious interpretation. Elevated myocardial testosterone glucuronide (Ikoma et al., 2026) may reflect broader endocrine and neurohormonal adaptation to chronic stress, given the known roles of steroid hormones in cardiac metabolism and fibrotic signalling (Montisci et al., 2012). Elevated purine metabolites, including uric acid, may indicate increased ATP turnover (Lopez-Schenk et al., 2023), while histidine- and proteolysis-related metabolites suggest possible disturbances in amino acid turnover and antioxidant defence. Given the limited consistency across studies, these findings should be regarded as preliminary, pending validation in larger cohorts.

Evidence indicates that HCM, irrespective of genotype status, is characterised by a recurring pattern of metabolic remodelling involving impaired fatty acid oxidation, altered amino acid metabolism, accumulation of alternative substrates, and broader disruption of mitochondrial energetics. Although metabolomic profiling may capture clinically relevant variation related to disease severity or phenotype, current evidence does not demonstrate consistent discrimination of genotype-specific metabolic profiles. Instead, the convergence of diverse sarcomeric mutations on similar metabolic alterations suggests that these signatures reflect shared downstream consequences of myocardial energetic stress rather than gene-specific biochemical effects. This may be explained by the common functional consequences of sarcomeric mutations, including impaired cross-bridge cycling, increased ATP demand, and reduced mechanical efficiency, which collectively drive a state of energetic insufficiency in the myocardium (Pioner et al., 2023). Therefore, the metabolic phenotype of HCM appears to represent the severity and consequences of myocardial energy imbalance more than the underlying genetic background.

Metabolomics and the different causes of left ventricular hypertrophy

A major clinical challenge in HCM management is distinguishing true sarcomeric HCM from phenocopies and other causes of LVH, including hypertensive heart disease, athlete’s heart, cardiac amyloidosis, and Fabry disease, each of which may present with overlapping structural features on imaging. This review, however, would not be able to directly address whether metabolomics is capable of making this distinction, as every comparator group across the 14 included studies consisted of healthy controls or genotype-negative relatives rather than patients with alternative causes of LVH. This follows from our exclusion criteria, which deliberately excluded phenocopy and secondary LVH studies to isolate genuine sarcomeric HCM biology and preserve the internal validity of the genotype-phenotype comparisons central to this review. Evidence from outside the scope of this systematic review suggests some LVH mimics have distinct metabolic signatures. For example, Fabry disease is characterised by accumulation of globotriaosylceramide (Lyso-Gb3), reflecting its underlying sphingolipid storage defect, mechanistically distinct from the fatty acid oxidation and BCAA disturbances described here for sarcomeric HCM (Alharbi et al., 2018). However, without direct head-to-head metabolomic comparisons between HCM and these other LVH aetiologies, the diagnostic specificity of any candidate HCM metabolite panel relative to these mimics remains unknown. Future studies designed explicitly to compare metabolomic profiles across LVH aetiologies, rather than against healthy controls alone, will be necessary before metabolomics can be considered for differential diagnosis in clinical practice.

Implications and emerging translational opportunities

The diagnosis of HCM relies predominantly on structural imaging and genetic testing, neither of which captures the metabolic state of disease activity. Current echocardiographic and cardiac MRI criteria identify hypertrophy only after structural remodelling has occurred, while genetic testing, although definitive when positive, leaves approximately 40% of cases without a pathogenic variant (Marian and Braunwald, 2017). Metabolomic profiling offers complementary information by reflecting the functional metabolic state of the myocardium rather than its structural or genetic makeup and may possibly address the diagnostic gaps of current tools.

The most promising diagnostic candidates emerge from machine learning approaches applied to multi-metabolite plasma panels. This convergence between the observed biomarker signal and the underlying pathobiology increases confidence that these panels are capturing disease-relevant metabolic dysregulation rather than an incidental variation. Similarly, the performance of an eicosanoid-based panel targeting 12-HETE and EPA suggests that inflammatory and lipid-mediator pathways may contribute an additional, orthogonal dimension to the disease characterisation (Zhang et al., 2025).

In terms of disease stratification, several studies reported clear associations between metabolite profiles and measures of disease severity. Circulating acyl carnitines, particularly medium and long-chain species, have been shown to correlate with left ventricular wall thickness and ejection fraction in MYBPC3-positive cohorts (Jansen et al., 2023b).Other metabolites, including aminoadipic acid, uric acid, and 3-methylhistidine, were consistently associated with more severe phenotypes, reflecting alterations in amino acid metabolism, purine turnover, and proteolysis. Furthermore, clustering approaches applied to metabolomic datasets have identified distinct metabolic subgroups of HCM patients with differing clinical outcomes, with poorer prognosis linked to increased purine metabolism and reduced carnitine-related metabolites, while relatively preserved TCA cycle intermediates were associated with more favourable profiles (Wang et al., 2022). These observations indicate that metabolomics may capture biologically meaningful heterogeneity that is not fully reflected by conventional imaging or clinical parameters.

Therapeutically, these metabolic disturbances point toward myocardial energetics as an intervention target. Reversibility is supported by longitudinal data showing β-hydroxybutyrate and lipid-related species decline after septal myectomy, consistent with haemodynamic and energetic recovery (Wolter et al., 2021). These findings support the concept that metabolic remodelling is dynamic rather than fixed. Pharmacological agents that modulate substrate utilisation may therefore have indirect benefits through improving myocardial energetic efficiency, although direct evidence in HCM remains limited (Coats et al., 2019). In parallel, cardiac myosin inhibitors such as mavacamten target the upstream sarcomeric dysfunction by reducing ATP consumption, thereby addressing the primary driver of energetic imbalance identified in metabolomic studies (Kim and Jung, 2025).

Despite these developments, metabolomics should currently be regarded as a tool that enhances mechanistic understanding and identifies candidate biomarkers, rather than a clinically established modality. Most proposed metabolite panels lack external validation, and their incremental value over established diagnostic and prognostic tools has not yet been determined. Future studies integrating metabolomic data with clinical parameters, imaging, findings and other omics disciplines will be prudent to define its role within precision medicine approaches for HCM.

Beyond these biological and mechanistic considerations, several structural barriers currently limit clinical translation. Nearly all included studies were single-centre, single-population cohorts with modest sample sizes, limiting statistical power and generalisability. Relatedly, proposed metabolite panels were derived and evaluated within the same discovery cohort in most studies, without replication in an independent external validation cohort, leaving their real-world diagnostic performance unconfirmed. Included studies also varied in fasting status, timing of sample collection relative to diagnosis or surgical intervention, and sample processing and storage pipelines, all of which are known to affect circulating metabolite measurements and complicate pooled or comparative interpretation. Large-scale, multi-centre studies with standardised collection protocols and independent external validation cohorts are therefore a prerequisite for clinical translation of any HCM metabolomic biomarker.

Analytical platform considerations and emerging approaches

The cross-study comparability of metabolomic findings in HCM is significantly influenced by the analytical platforms used, which vary in sensitivity, resolving power, and coverage. LC-MS is the dominant platform, with most studies employing UPLC or UHPLC systems from vendors such as Waters (ACQUITY), Thermo Fisher Scientific (Dionex UltiMate 3000), or Shimadzu (Nexera/UFLC). These are frequently coupled to HRMS instruments, particularly the Thermo Scientific Orbitrap series (Q-Exactive, Elite, or Fusion Lumos), which provide the high mass accuracy necessary for untargeted metabolite identification. Other researchers have utilised QTOF-MS platforms, such as those from AB Sciex or Waters, to achieve rapid, high-resolution scanning of the HCM metabolome. For targeted approaches, triple quadrupole systems like the Sciex Qtrap 6500 + or 5500 are preferred due to their superior sensitivity and dynamic range for quantifying specific panels, such as acyl carnitines and eicosanoids. LC-MS offers excellent sensitivity and adequately resolves low-abundance metabolites, such as acyl carnitines. However results remain dependent on chromatographic separation conditions, ionisation mode, and reference library coverage, all of which vary across laboratories and contribute to inconsistent metabolite annotation between studies.

While LC-MS predominates, GC remains a vital tool for specific applications. This two-dimensional GC approach offers enhanced resolving power for complex samples, though it typically requires metabolite derivatization, which may limit the detection of certain polar species compared to standard LC-MS methods. Furthermore, direct-infusion HRMS has emerged as a high-throughput alternative for untargeted profiling. While this method used by Schuldt et al. (2021) and Jansen et al. (2023a, 2023b), affords rapid analysis and high mass accuracy, however the absence of chromatographic separation makes it challenging to distinguish between isomeric or isobaric species without additional ion mobility capabilities.

Spatial metabolomics represents a significant emerging approach in the field, as demonstrated by the use of DESI-IMS. This technique is able to map the regional distribution of metabolites such as testosterone glucuronide directly within myocardial tissue sections. This provides a localised metabolic context that traditional solution-based platforms cannot capture. However, DESI-IMS remains largely semi-quantitative and possesses a narrower dynamic range than standard UPLC-MS/MS. Additionally, no included study employed ion mobility spectrometry, which may offer future value in resolving isomeric and isobaric metabolites not distinguishable using the direct-infusion methods described above.

Ultimately, these platform-dependent differences in sensitivity and resolution contribute to discrepancies in which specific metabolites reach statistical significance across studies. These technical variations, alongside reference libraries used by different vendors, must be carefully considered when interpreting the consistency of individual biomarker findings across this body of literature, as well as when synthesising a unified metabolic signature for HCM.

Critical appraisal of study designs

The overall quality of the included studies was moderate, with methodological limitations that warrant caution when interpreting the findings. Most used case-control designs with relatively small samples, limiting statistical power and increasing vulnerability to selection bias. The predominance of cross-sectional data also prevents assessment of temporal relationships, which limits evaluation of metabolite biomarkers for disease progression or prognosis.

There was substantial heterogeneity across a few components, namely cohort composition, genotype reporting, biospecimen type, analytical platform, sample handling, fasting status, and statistical modelling. Some cohorts combined genotype-positive and genotype-negative individuals without stratified analyses, which may have limited the identification of genotype-specific metabolic signatures. In one study, participants were identified as carriers of HCM-associated variants, although many did not have clinically diagnosed HCM (Ravi et al., 2022).

Although most studies matched or adjusted comparison groups for age and sex, important confounders including diabetes, renal function, BMI, and medication use, were inconsistently reported or accounted for. Study populations were also largely drawn from homogeneous single-population cohorts, including populations of Finnish, Chinese, Dutch, and North American descent. Race and ethnicity were rarely reported or analysed independently, which limits the generalisability of the findings across more diverse populations.

Additionally, most reported metabolites were assigned at MSI Level 2, based on putative spectral matching rather than confirmation with authentic standards. Differences in metabolite identification and reporting may further affect reproducibility of these studies, which could consequently lead to metabolite misclassification. Several studies also reported mainly statistically significant metabolites, creating a risk of selective reporting and limiting comparison across independent datasets.

Studies by Schuldt et al. (2021) and Jansen et al. (2023a) did not specify the platform used for their direct-infusion high-resolution mass spectrometry analysis. This missing information could limit the reproducibility of their findings and the assessment of cross-study comparability, as instrument-specific variations in sensitivity, ionisation efficiency, and reference library coverage can significantly influence which metabolites are annotated and deemed statistically significant.

Interpretation is additionally influenced by the biological source of samples. Circulating metabolite profiles reflect systemic factors such as diet, comorbidities, medications, and peripheral metabolism, whereas myocardial tissue offers a more direct view of cardiac metabolism but is generally available only from patients undergoing surgery. Circulating and tissue-based findings should therefore be regarded as complementary, rather than interchangeable.

Despite these limitations, we must acknowledge that the field appears to be moving from exploratory untargeted profiling toward more focused targeted and integrative multi-omics approaches.

Future directions

At present, metabolomics is best viewed as a hypothesis-generating tool that provides mechanistic insight into disease biology and identifies candidate biomarkers, rather than a clinically established diagnostic or prognostic modality. Future longitudinal studies following genetically confirmed cohorts are urgently needed, integrating metabolomic data with genetic, imaging, and clinical information to define biologically meaningful HCM subtypes and support the development of precision medicine approaches for this heterogeneous disorder.

A key priority is the establishment of prospective longitudinal cohorts in which genotype-positive individuals are followed serially from the pre-hypertrophic stage throughout disease progression. Such study designs would allow metabolomic changes to be characterised in relation to disease onset, enabling identification of metabolites that predict disease progression rather than merely reflecting diagnosis. Ideally, serial plasma and urine sampling should be integrated alongside standard clinical surveillance in HCM family screening programmes, which already provide access to pre-clinical mutation carriers. Moreover, future metabolomic studies should adopt MSI Level 1 metabolite identification, given that the majority of included studies reported only Level 2, such as suggested by Sumner and colleagues (Sumner et al., 2007).

Conclusions

Evidence synthesised in this systematic review suggests that HCM is characterised less by genotype-specific metabolic profiles and more by a shared pattern of metabolic remodelling reflecting myocardial energy deficiency. Across the included studies, recurring findings include impaired fatty acid oxidation, accumulation of BCAAs, increased reliance on ketone bodies and glycolysis, and disruption of central carbon metabolism, with additional contributions from purine-related pathways. These alterations appear to align more closely with phenotypic severity than with the underlying sarcomeric variant, suggesting a potential convergence toward a common downstream bioenergetic state. For biomarker development, these findings imply that metabolite panels reflecting energy impairment may offer greater clinical utility for disease characterisation than gene-specific signatures. Regarding therapy, this synthesis highlights cardiac energetics as a central potential target, where interventions aimed at improving metabolic efficiency or reducing energy demand may have broad applicability across the HCM spectrum. However, given the methodological heterogeneity and the predominance of Level 2 metabolite identification in current research, these signals should be viewed as preliminary. Future work must prioritise longitudinal, genotype-defined cohorts to determine whether these metabolic changes can reliably predict disease progression and be translated into validated clinical tools.

Study limitations

Several limitations arise from the design of this review itself. First, inclusion was deliberately restricted to genetically characterised cohorts. While this was essential to isolate sarcomere-associated disease from phenocopies and secondary hypertrophy, it necessarily narrowed the evidence base, and the findings should therefore not be generalised to undifferentiated or non-genotyped left ventricular hypertrophy, which represents a distinct question warranting a separate review. Second, and for the same reason, the exclusion of secondary hypertrophic phenocopies and other mimics means the evidence base contains no direct comparison between HCM and alternative causes of hypertrophy; the diagnostic specificity of the identified metabolites therefore cannot be established from the present synthesis, only their association with HCM relative to healthy or non-carrier controls. Third, eligibility was limited to English-language publications, introducing a potential language bias. Finally, the substantial methodological and analytical heterogeneity across the fourteen eligible studies precluded a quantitative meta-analysis, so the evidence was synthesised narratively; combined with the small number of eligible studies, this limits the strength of the conclusions that can be drawn. Taken together, these constraints indicate that the convergent metabolic signature described here should be regarded as a hypothesis-generating framework, requiring confirmation in larger, prospectively designed, genotype-defined cohorts with standardised protocols and independent external validation.

Author contributions

A.H.A.R. wrote the main manuscript text, conducted the literature search, performed study screening and data extraction, and prepared the tables and figures. N.A.L. and S.A.R. contributed to data interpretation and critically reviewed the manuscript. S.H.S.A.K. conceptualised the study, involved in study screening, provided supervision, and reviewed the manuscript. All authors reviewed and approved the final manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.