TREATMENT OF HEART FAILURE WITH NORMAL EJECTION FRACTION

The invention relates to treatment of heart failure with normal left ventricular (LV) ejection fraction syndrome (HFnEF).

Background of the Invention

Significant advances in therapy for heart failure (HF) with impaired systolic function have improved quality of life, and increased survival. However up to 50% of patients who have clinical evidence of HF are found to have a normal left ventricular ejection fraction (HF with normal left ventricular (LV) ejection fraction syndrome (HFnEF), also referred to as HF with preserved left ventricular ejection fraction syndrome (HFpEF). Patients with HFnEF represent a rapidly increasing proportion of patients hospitalised and suffering mortality from heart failure.

Despite a normal EF, HFnEF patients manifest subtle systolic dysfunction but the principal abnormality in most is a disorder of active relaxation and/or passive filling of the LV. However resting measures of active relaxation and filling relate poorly to symptoms and exercise capacity therefore no 'gold standard' diagnostic echocardiographic test exists for HFnEF. Effective ventricular filling results from a highly energy dependent active relaxation process and from passive filling which is dependent on loading conditions as well as the intrinsic (passive) properties of the LV. Since both these parameters change markedly during exercise due to sympathetic activation, it is not surprising that these resting parameters are so poorly predictive of exercise capacity and symptoms.

Perhexiline (2-(2,2-dicyclohexylethyl) piperidine) is a known anti-anginal agent that operates principally by virtue of its ability to shift metabolism in the heart from free fatty acid metabolism to glucose, which is more energy efficient.

WO-A-2005/087233 discloses the use of perhexiline for the treatment of chronic heart failure (CHF) where the CHF is a result of an initial inciting influence of ischaemia or where the CHF is a result of an initial non-ischaemic inciting influence.

Summary of the Invention

According to a first aspect of the present invention, there is provided a method of

treating HFnEF, which comprises administering to an animal in need thereof an effective amount of perhexiline, or a pharmaceutically acceptable salt thereof, to treat said HFnEF. The animal is preferably a mammal and most preferably a human.

According to another aspect of the present invention, perhexiline, or a pharmaceutically acceptable salt thereof, is provided for use in the treatment of HfnEF.

According to a further aspect of the invention there is provided a treatment programme for treating HFnEF, which involves the co-use or co-administration of perhexiline or pharmaceutically acceptable salt thereof with one or more other compounds that are advantageous in treating HFnEF or the symptoms thereof, for example a diuretic, an angiotensin receptor blocker or a calcium channel blocker.

Brief Description of the Drawings

Figure 1 displays variables correlating with Aerobic Exercise Capacity (Vθ2max).

Figure 2 shows MR images of a patient with HFpEF lying prone over a ³¹ P surface coil (Panel A) and the corresponding localized ³¹ P MR spectra from the left ventricle (Panel B). Panel C is Individual PCr/ γ~ ATP ratio in Patients with HfpEF and Controls.

Detailed description of the invention

In aspects of the present invention, the perhexiline exists in the form of a salt of perhexiline, preferably the maleate salt. The perhexiline may be used at doses titrated to achieve therapeutic but non-toxic plasma perhexiline levels (Kennedy JA,

Kiosoglous AJ, Murphy GA, PeIIe MA, Horowitz JD. "Effect of perhexiline and oxfenicine on myocardial function and metabolism during low-flow ischemia/reperfusion in the isolated rat heart", J Cardiovasc Pharmacol 2000; 36(6):794-801). Typical doses for a normal patient would be 100mg to 300mg daily, although smaller doses may be appropriate for patients who are slow metabolisers of perhexiline.

Physiologically acceptable formulations, such as salts, of the compound perhexiline, may be used in the invention. Additionally, a medicament may be formulated for

administration in any convenient way and the invention therefore also includes within its scope use of the medicament in a conventional manner in a mixture with one or more physiologically acceptable carriers or excipients. Preferably, the carriers should be "acceptable" in the sense of being compatible with the other ingredients of the formulation and not deleterious to the recipient thereof. The medicament may be formulated for oral, buccal, parental, intravenous or rectal administration. Additionally, or alternatively, the medicament may be formulated in a more conventional form such as a tablet, capsule, syrup, elixir or any other known oral dosage form.

The invention is illustrated by the following non-limiting examples.

Example 1

The role of exercise related changes was evaluated in left ventricular (LV) relaxation and of vasculo-ventricular coupling as the mechanism of exercise limitation in patients with heart failure with normal (or preserved) LV ejection fraction (HFnEF) and whether cardiac energetic impairment may underlie these abnormalities.

The study involved 37 patients with HFpEF and 20 matched controls. Vasculo- ventricular coupling (WC) and Time to Peak LV Filling (a measure of LV active relaxation) (nTTPF) were assessed at rest and on exercise by Multiple Uptake Gated Acquisition scanning. Cardiac energetic status (PCr/ATP ratio) was assessed by ³¹ P Magnetic Resonance Spectroscopy. At rest nTTPF and WC were similar in patients and controls. Cardiac PCr/ATP ratio was reduced in patients vs. controls (1.57±0.52 vs.2.14±0.62; P=0.003). VU2max was lower in patients vs. controls (19+4 vs.36±8 ml/kg/min; P<0.001). During maximal exercise the heart rate increased less in patients vs. controls (52±16 vs.81±14 bpm; pθ.001) and the relative changes in stroke volume and cardiac output during submaximal exercise were lower in patients vs. controls (0.99±0.34 vs. 1.25±0.47, P=0.04; 1.36±0.45 vs. 2.13±0.72, P<0.001). nTTPF fell during exercise in controls, but increased in patients (-0.03±12 sec vs. +0.07±0.11; P=0.005). WC decreased on exercise in controls but was unchanged in patients (-0.01 ±0.15 vs. -0.25±0.19; pθ.001). Heart rate, WC and nTTPF were independent predictors of Vθ2max.

Methods

Patients

The study involved 37 HFpEF patients prospectively recruited from heart failure clinics. Also studied were twenty age-gender-matched healthy controls with no cardiac history or diabetes mellitus. Study participants had clinical examination, 12- lead electrocardiogram, pulmonary function test, echocardiogram, metabolic exercise test, MUGA studies and a subgroup underwent cardiac ³¹ P MRS studies to assess cardiac energetic status. All controls had a normal cardiovascular examination, 12- Iead electrocardiogram and echocardiogram. HFpEF patients were defined in accordance with ACC/AHA recommendation (1): i) symptoms and signs of heart failure, ii) ejection fraction >50%, iii) no valvular abnormalities. In addition it was stipulated that patients should have iv) VCtemax <80% of age and gender predicted with a pattern of gas exchange on metabolic exercise testing indicating a cardiac cause for limitation, v) absence of objective evidence of lung disease on formal lung function testing and/or absence of arterial desaturation during exercise and with a ventilatory reserve at peak exercise ≥15L. Patients with rhythm other than sinus were excluded.

Echocardiography

Echocardiography was performed with participants in the left lateral decubitus position with a Vivid 7 echocardiographic machine using a 2.5-MHz transducer. Cardiac quantifications were determined in accordance with European Association of Echocardiography. (2) Left ventricular hypertrophy was defined as a left ventricular mass indexed to body surface area that exceeded 88 g/rri2 for women and 102 g/nri2 for men. (2) LV end-systolic elastance (Ees) was determined using the non-invasive single-beat technique. (3) Arterial elastance (Ea) was calculated as the ratio of LV end-systolic pressure/stroke volume. Studies were stored digitally and analyzed offline.

³¹ P Cardiac Magnetic Resonance Spectroscopy (MRS)

³¹ P cardiac magnetic resonance spectroscopy was performed using a Phillips

Achieva 3τ scanner and a linearly polarized transmitter and receiver ³¹ P coil with a diameter of 14 cm. Localization was achieved by ISIS volume selection (4).

Localized iterative 1st order shimming was performed including the entire heart using

the unsuppressed water signal acquired with the body coil. A short axis cine scan was acquired to optimise shimming and Fo determination. A 3-D voxel of acquisition was planned to include the left ventricle. The repetition time was 10000 ms with 136 averages and 512 samples. Acquisition was ECG gated and the trigger delay was set to acquire in diastole. Total scan time was 23 minutes. Java magnetic resonance user interface v3.0 (jMRUI) was used for analysis. (5) Using the preprocessing operations within jMRUI, signals were fourier transformed, phase corrected and apodized by 15Hz. AMARES (advanced method of accurate, robust and efficient spectroscopic fitting) algorithm was used to fit for phosphorcreatine (PCr), adenosine triphosphate (ATP), phosphodiesters and 2, 3- bisphosphoglycerate with prior knowledge of all peaks incorporated into the processing algorithm. PCr and γ-ATP was used to determine the PCr/ATP ratio which is a measure of cardiac energetic state (6). Patients with diabetes, ischemic heart disease or had contraindication were excluded from the MRS study (N=12). One patient's spectra was excluded from the analysis due to poor quality. Three controls had contraindication to MRS study. Data were analysed separately by an investigator unaware of participants' clinical status. To test the reproducibility a single subject was scanned on eight separate occasions. The mean PCr/ATP ratio was 2.11+0.25 and Bland Altman plots demonstrated a variance of 12% in the measurement of PCr/ATP ratios. As a measure of the quality of the data the Cramer- Rao lower bounds (CRLBs) (7) was calculated for the PCr peak and γ-ATP, the results were 6% and 10%, respectively.

Multiple Uptake Gated Acquisition Scan (MUGA)

LV ejection fraction and diastolic filling were assessed by equilibrium R-wave gated blood pool scintigraphy at rest and during graded semi erect exercise on a cycle ergometer as previously described. (8,9) Three minutes of data were acquired at rest and during exercise after a 30-second period for stabilisation of heart rate at the commencement of each stage. Exercise was performed at 50% workloads of heart rate reserve. Data were analysed using LinkMedical MAPS software, Sun Microsystems (Hampshire, UK). Peak left ventricular filling rate in terms of end- diastolic count per second (EDC/s) and time to peak filling normalised for R-R interval (nTTPF) in milliseconds after end systole were calculated from the first derivative of the diastolic activity-time curve. Venous blood samples were obtained

for weighing and for counting of blood gamma activity during each scan in order to correct for physical and physiological decay as well as for determination of relative volume changes. (10) The validity of these radionuclide measures of diastolic filling at high heart rates has been established previously. (11)

All gated blood pool scan-derived volumes were normalized to body surface area, yielding their respective indexes: end-diastolic volume index (EDVI), end-systolic volume index (ESVI), stroke volume index (SVI), and cardiac index. The following indexes were calculated: a) arterial elastance index (EaI) = ESP/SVI; b) LV end- systolic elastance index (ELVI) = ESP/ESVI and c) vasculo-ventricular coupling ratio (WC) = Eal/ELvl = (1 /EF)-I . (12)

Metabolic Exercise Test

All participants underwent a symptom-limited erect treadmill exercise using a standard ramp protocol with simultaneous respiratory gas analysis. (13)

Statistics

Continuous variables are expressed as means±SD. Unpaired Student's t-test (2-tail) was used to assess differences between mean values. Categorical variables were compared with Pearson Chi-Square test. All reported P values were calculated on the basis of two sided tests and a P value of <0.05 was considered to indicate statistical significance. Variances of data sets were determined using F-test. Pearson correlation coefficient (r) was used to describe the relationship between variables. All subjects were included into the model. Variables of interest that were found to correlate with the dependent variable on univariate analysis were included in a stepwise linear regression analysis to identify independent variables. SPSS (v15.0) was used to perform the statistical operations.

Results

The results obtained are set forth in Tables 1-3 below and in Figures 1 and 2.

In Figure 1 variables correlating with Aerobic Exercise Capacity (Vθ2max) are shown. Panel A: Vθ2max correlated negatively with Exercise-induced Changes in nTTPF. Panel B: VChmax correlated negatively with Exercise-induced Changes in

Vasculo- Ventricular Coupling Ratio. Panel C: Vθ2tnax correlated directly with Exercise-induced Changes in Heart Rate. Black circles indicate patients with HFpEF, and Open circles represents healthy controls. When patients on beta blockers were excluded from analysis, the level of significance were similar.

In Figure 2, Panel A shows an MR images of a patient with HFpEF lying prone over a ³¹ P surface coil and the corresponding localized ³¹ P MR spectra from the left ventricle is shown in panel B. The resonances derive from PCr and the v-, α-, and β- phosphate Resonances of the ATP. Panel C Individual PCr/ Y- ATP ratio in Patients with HfpEF and Controls. The PCr/ y- ATP ratio was significantly reduced in patients with HfpEF compared to healthy controls, P= 0.003

Plus-minus values are means ± SD. When patients on beta blockers were excluded from analysis, the level of significance were similar apart from resting HR (P=O.14). NYHA denotes New York Heart Association, ACE angiotensin-converting enzyme,

ARB angiotensin Il receptor blockers, BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, MABP mean arterial blood pressure, LA left atrium, E/E' mitral E-wave velocity-E' tissue velocity (PW-TDI) at basal inferoseptum ratio, Ees denotes Left Ventricular End-Systolic Elastance and Ea is Arterial elastance. The bodymass index is the weight in kilograms divided by the square of the height in meters.

Plus-minus values are means ± SD. When patients on beta blockers were excluded from analysis, the level of significance were similar apart from peak filling rates during exercise (P=0.08). EDC end diastolic count. SBP systolic blood pressure, DBP diastolic blood pressure, MABP mean arterial blood pressure. Relative δ Stroke Volume Index is SVi EXERCISE/ SVi REST, Relative δ Cardiac Output Index is COi EXERCISE / COi REST. Relative δ ELVI is ELVIEXERCISE / ELVIREST. Relative δ EaI is EalEXERcisE / EalREST. δ Vasculoventricular coupling ratio is (Eal/E_vl)EXERcιsE - (Eal/ELvl)RESτ. δ WC -0.01 ±0.15 -0.25±0.19 <0.001

^* Predictors: δHR, fPredictors: δHR and δ WC coupling ratio, φPredictors: δHR, δ WC coupling ratio and age, §Predictors: δHR, δ WC coupling ratio, age and δTTPF. Multivariate analysis was adjusted for the variable that some patients were on betablockers.

Characteristics of the Patients

HFpEF Patients were generally females, overweight, aged 67±9 years old with a history of hypertension, however blood pressure was well treated (systolic BP 138±19mmHg vs. 131±23mmHg; p=0.23, in patients vs. controls) (see Table 1 below). The tissue Doppler E/E' at the basal infero-septum (a measure of left ventricular end-diastolic pressure) (14), was significantly higher in patients than controls. There was also a trend (non-significant) to higher Ees in patients than in the control group. HFpEF patients also had significantly reduced Vθ2max and reduced peak HR on metabolic exercise testing. There was a positive correlation between VO∑max and δHR (HREXERCISE - HRREST) (r=0.7, P<0.001) (see Figure 1). During semi-erect cycle exercise the relative stroke volume (SVi EXERCISE /SVi REST) was lower in patients compared to controls (0.99±0.34 vs. 1.25±0.47; P=0.04), and

relative cardiac output (COiEXERCisE/COϊREST) was also lower (1.36±0.45 vs 2.13±0.72; p<0.001). (see Table 2)

Left Ventricular Active Relaxation

nTTPF is determined by the rate of active relaxation (15) and by transmural pressure gradient at the time of mitral valve opening. nTTPF was similar at rest in HFpEF patients and controls. During exercise it shortened in controls, but lengthened in patients (Table 2). There was a negative correlation between VCtemax and δnTTPF

(πTTPFEXERCISE - πTTPFREST) (r=-0.4, P=0.005) (see Figure 1). Furthermore, during exercise other MUGA diastolic filling variables such as peak filling rates as well as systolic function parameters e.g. EF and peak emptying rates, were significantly reduced in patients compared to controls, (see Table 2)

In vivo Myocardial Energetic state

At rest, cardiac PCr/ATP ratio in HFpEF patients (N=24) was significantly reduced compared to healthy controls (N=17), 1.57±0.52 and 2.14±0.63, respectively, P=0.003 (see Figure 2).

Vasculo-Ventricular Coupling

WC was similar at rest in HFpEF patients and controls. During exercise, WC decreased in controls because of a marked increase in LV end-systolic elastance accompanied by only a modest increase in arterial elastance. In contrast WC was essentially unchanged in patients on exercise and was significantly higher than in controls. This was because of a significantly smaller increase in LV end-systolic elastance and a trend to greater increase in arterial elastance in patients vs. controls (Table 2). There was a negative correlation between Vθ2inax and δ WC on exercise (r=-0.6, P<0.001) (see Figure 1).

Independent Predictors of Aerobic Exercise Capacity

In the multivariate analysis, a linear-regression model was used to examine Vθ2max as the dependent variable and found that exercise-induced changes in HR, WC and nTTPF were independent predictors of VChmax. (see Table 3)

Discussion

This study aims to investigate the dynamic intrinsic and systemic pathophysiology of

HFpEF. The principal findings are: a) HFpEF patients manifest a significant reduction in PCr/ATP ratio at rest, indicating a resting defect in myocardial energetics, b) As a corollary, during exercise, the energetically demanding active relaxation stage of diastole lengthens in patients (vs. a shortening in controls) and there is also a failure of contractile function to increase in patients. These abnormalities result in a lower stroke volume on exercise, c) Consistent with previous studies, HFpEF patients demonstrate chronotropic incompetence on exercise. (16- 18)

The pathophysiology of HFpEF has been the subject of considerable controversy. These patients are typically hypertensive and exhibit impaired LV active relaxation and/or increased passive left ventricular diastolic stiffness at rest. (19) This has led many to conclude that exercise limitation is primarily a result of impaired LV diastolic filling and to the use of the term 'diastolic heart failure' by some. (20) However, diastolic dysfunction is also a common finding at rest in healthy elderly subjects. (21) Furthermore, 'subtle' abnormalities of systolic function, in particular long axis systolic function, are also almost universally observed in HFpEF patients despite normal LV ejection fraction. (22) This has led others to propose that HFpEF is predominantly a disorder of contractile function. (23) Little attention has been directed to assessing changes in systolic and diastolic function during dynamic exercise, which is after all when the majority of patients experience their symptoms. In one study, ten patients with HFpEF were assessed with invasive pressure volume loops and compared with age-matched controls. (24) The former had increased arterial elastance (a measure of the stiffness of the entire arterial tree), and increased LV end-systolic elastance (a measure of the stiffness of the ventricle during systole, and the relatively load independent measure of the contractile state of the left ventricle. (25) Whilst diastolic abnormalities were not universally present in patients at rest, marked differences appeared during handgrip exercise. The rate of LV active relaxation increased in healthy subjects but it slowed in patients. (24) Another study from the same group, exercise-related symptoms in Afro-Caribbean hypertensive patients appeared to be strongly associated with chronotropic incompetence and an inadequate vasodilator reserve on exercise. (16)

The present study examined the pathophysiological mechanisms and predictors of exercise limitation in a substantially larger series of patients during a much more physiologically relevant form of exercise (dynamic leg exercise). There were marked dynamic abnormalities in both contractile and diastolic function of the left ventricle, and a lower peak exercise HR in patients. The independent predictors of impaired exercise capacity were abnormal ventricular-arterial coupling on exercise, a reduced HR response on exercise and a 'paradoxical' slowing of the rate of LV active relaxation on exercise (manifest as a prolongation of nTTPF). On the face of it, the independent role of an impaired chronotropic response in predicting exercise capacity in this setting appears counter-intuitive. On the one hand, Vθ2max is largely determined by cardiac output on exercise and the latter is simply the product of HR and SV. On this basis, a detrimental effect of an impaired HR might appear logical. However, in the setting of a profound slowing of active relaxation and increased LV passive diastolic stiffness, a larger diastolic filling period might be expected to be beneficial, by increasing SV. Indeed, this is in part the basis of therapy with β- blockers in hypertrophic cardiomyopathy, a classic paradigm for HFpEF. (26) However in this study, despite a longer diastolic filling time, the relative change in SV was lower in patients during sub-maximal exercise. Whilst an impaired HR response was independently predictive of reduced Vθ2max this relationship may or may not be causal. An alternative explanation is that it may be a consequence of the heart failure as an impaired chronotropic response (associated with slow HR recovery following exercise) (27) is typically present in systolic heart failure and is in part a manifestation of impaired vagal tone; (28) or it may be adaptation to improve diastolic filling. The latter seems at least plausible, since increasing heart rate by atrial pacing has been shown to reduce supine resting stroke volume and cardiac output in patients with HFpEF. (18) Clearly it will be important to undertake further studies to assess whether heart rate plays a causal role in exercise limitation in HFpEF, because if so this may be amenable to rate responsive pacing.

The patients in this study had a history of hypertension but were well treated with antihypertensives (in most cases including vasodilators) therefore resting blood pressure and arterial elastance were not significantly higher than in the control group. Consistent with prior studies (24), at rest, LV end-systolic elastance (a measure of contractility or systolic stiffness) tended to be higher in patients although this did not

reach significance. The increase in arterial elastance during exercise tended to be greater in patients vs. controls (presumably reflecting increased large artery stiffness). However, whilst left ventricular end-systolic elastance almost doubled during exercise in controls, the increase was only 35% in patients; hence WC reduced by 33% during exercise in controls but was unchanged in patients. These findings indicate a blunting of the physiological increase in the contractile state of the left ventricle on exercise.

The physiological increase in the rate of LV active relaxation during exercise is a consequence of sympathetic activation, via cAMP-dependent protein kinase (PKA) mediated phosphorylation of key proteins including Troponin I, Sarco/Endoplasmic Reticulum Ca ²⁺ -ATPase (SERCA) and Titin. (29-31) In experimental models, large acute increases in afterload resulted in an acute impairment of LV active relaxation (32), with the threshold for this phenomenon being lower in the diseased heart, leading to the concept of 'relative load' as a determinant of afterload related impairment of LV active relaxation. (33) In the study of HFpEF patients described earlier (5), handgrip exercise was associated with a substantially greater increase in LV end-systolic pressure than in controls, potentially explaining the observed slowing of LV active relaxation. A key coupler of this load dependent LV relaxation is Troponin I - Protein Kinase A (TnI-PKA) phosphorylation (34). It is known that the energy dependent process of phosphorylation of Troponin I by PKA decreases myofibrillar calcium sensitivity (35) and increases the rate at which calcium dissociates from Troponin C (36) which can lead to increase rate of LV relaxation. Indeed, in a study involving transgenic mice in which PKA phosphorylation sites on Troponin I were constitutively active, acute aortic constriction led to a lengthening of Tau (an invasive measure of active relaxation) in the wild type mice but not in the transgenic mice. (34)

In summary, in the present study, exercise led to both a paradoxical slowing of the rate of active relaxation of the LV and an attenuated increase in end-systolic elastance (=contractility) compared to controls. Resting cardiac energetic status was substantially reduced in HFpEF patients. It is proposed that energetic impairment may underlie these pathophysiological changes. Patients with ischemic heart disease and diabetes were excluded from the MRS studies because these conditions are associated with impaired cardiac energetics. (8,39). The causes for this resting

energy deficit remain; elusive they probably relate to insulin resistance (40), impaired mitochondrial function as a result of ageing (41), neuroendocrine activation and aberrant substrate metabolism of heart failure. (42,43) These observations indicate the potential for therapeutic benefit from 'metabolic agents' that increase cardiac energetic status by altering cardiac substrate use (44). These agents have shown promise in patients with systolic heart failure. (45,46)

Study limitations

The radionuclide exercise protocol involved asking subjects to maintain a HR which was 50% of HR reserve above their resting HR. Since this HR reserve was calibrated to peak HR rate on metabolic exercise testing, the absolute workload was lower in patients. Nonetheless, most of the changes in SV occur in the first part of exercise with subsequent increases in cardiac output being principally due to increases in HR. (47) A small proportion of patients were on β-blockers which may have affected their cardiovascular response to exercise, however, except where stated, the level of significance remained similar when these patients were excluded from the analysis. MRS and Radionuclide studies also require a regular rhythm, thus patients with atrial fibrillation were excluded from the study.

Conclusion

HFpEF Patients have abnormal cardiac energetic status which is expected to contribute to the abnormal active relaxation on exercise and to a failure of LV end- systolic elastance to increase. In addition chronotropic response was markedly impaired on exercise in patients. The independent predictors of exercise capacity in patients with HFpEF are exercise-induced changes in active relaxation, heart rate and ventricular-arterial coupling.

Example 2

A double-blind, randomised, placebo-controlled study is undertaken in order to investigate the effects of perhexiline on HFnEF.

70 patients who meet the selection criteria are recruited from HF clinics. The inclusion criteria are:

HFnEF is defined as:-

1. Clinical features consistent with HF.

2. LVEF > 50%, with no evidence of significant valvular disease, no hypertrophic cardiomyopathy, and no evidence of pericardial constriction.

3. A Peak VO2 < 80% predicted, with RER>1 and with a pattern of gas exchange on metabolic exercise testing indicating a cardiac cause for limitation. This ensures a strict definition of patients.

In addition all patients recruited are in sinus rhythm.

Exclusion criteria are:

1. BMI >35.

2. Objective evidence of lung disease on formal lung function testing.

3. Reversible myocardial ischaemia on contrast-enhanced myocardial stress Echocardiography, and no evidence of exercise-induced mitral regurgitation (>2+).

4. Impaired hepatic function; known hypersensitivity to perhexiline

The following are performed prior to and at the end of the 3 months treatment:

Metabolic Exercise Test

Metabolic exercise testing is performed during treadmill exercise testing to measure peak VO2, VO2 at anaerobic threshold and VE/VCO2 slope.

Assessment of Resting and Exercise Diastolic Function

LVEF and diastolic filling are assessed by equilibrium R-wave gated blood pool scintigraphy at rest and during graded semi erect exercise on a cycle ergometer. The rest and exercise gated blood pool scintigraphs are analysed by a single operator blinded to the patient's data. Peak LV filling rate in terms of end-diastolic volumes per second (EDV/s) and time to peak filling (TTPF) are calculated from the second

derivative of the diastolic activity-time curve. The δTTPF/δHR is calculated at peak and submaximal exercise.

Echocardiography

Standard echocardiographic views are obtained on held expiration. Transmitral flow profiles, pulmonary vein flows and tissue Doppler (TDI) assessment of mitral annulus velocities (as a measure of long axis systolic and diastolic function) are performed. The E/Ea ratio is calculated as an indirect measure of LVEDP. TDI measurements are also performed in multiple segments to assess radial and longitudinal systolic and diastolic function in multiple segments.

³¹ P Magnetic Resonance Spectroscopy - Cardiac Muscle

Cardiac high-energy phosphate metabolism is measured using ³¹ P MRS on a 3-Tesla Philips whole-body magnet using a linearly polarized transmit and receive ³¹ P coil with a diameter of 14 cm. The participants are positioned supine with the coil directly over the precordium. The ³¹ P spectrum is acquired with a repetition time of 10000 ms and 512 averages. PCr/ATP ratio is determined after correcting the ATP peak for blood contamination.

Intervention

After the investigations are performed, the subjects are randomised to receive either 100 mg of perhexiline a day or placebo. Dose titration (and dummy dose titration in the placebo group) are performed (based on plasma levels) by an unblended independent observer according to a standard algorithm.

References

1. Hunt SA, Abraham WT, Chin MH et al. ACC/AHA 2005 Guideline Update for the Diagnosis and Management of Chronic Heart Failure in the Adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure): developed in collaboration with the American College of Chest Physicians and the International Society for Heart and Page I 18 Lung Transplantation: endorsed by the Heart Rhythm Society. Circulation. 2005; 112:e154-e235.

2. Lang RM, Bierig M, Devereux RB et al. Recommendations for chamber quantification: a report from the American Society of Echocardiography's Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr. 2005; 18: 1440-1463.

3. Chen CH, Fetics B, Nevo E et al. Noninvasive single-beat determination of left ventricular end-systolic elastance in humans. J Am Coll Cardiol. 2001 ;38:2028- 2034.

4. Ordidge RJ, Van de Vyver FL. Re: Separate water and fat MR images. Radiology. 1985; 157:551 -553.

5. Naressi A, Couturier C, Castang I et al. Java-based graphical user interface for MRUI, a software package for quantitation of in vivo/medical magnetic resonance spectroscopy signals. Comput Biol Med. 2001 ;31 :269-286.

6. Neubauer S, Krahe T, Schindler R et al. 31 P magnetic resonance spectroscopy in dilated cardiomyopathy and coronary artery disease. Altered cardiac high- energy phosphate metabolism in heart failure. Circulation. 1992;86:1810-1818.

7. Cavassila Si Deval S, Huegen C et al. Cramer-Rao bounds: an evaluation tool for quantitation. NMR Biomed. 2001; 14:278-283. Page 1 19

8. LeIe SS, Thomson HL, Seo H et al. Exercise capacity in hypertrophic cardiomyopathy. Role of stroke volume limitation, heart rate, and diastolic filling characteristics. Circulation. 1995;92:2886-2894.

9. LeIe SS, Macfarlane D, Morrison S et al. Determinants of exercise capacity in patients with coronary artery disease and mild to moderate systolic dysfunction.

Role of heart rate and diastolic filling abnormalities. Eur Heart J. 1996; 17:204- 212.

10. Atherton JJ, Moore TD, LeIe SS et al. Diastolic ventricular interaction in chronic heart failure. Lancet. 1997;349: 1720-1724.

11. Bacharach SL GMBJSESE. Left ventricular peak ejection rate, filling rate and ejection fraction-frame requirements at rest and exercise: concise communication. Journal of Nuclear Medicine 20, 189-193. 1979.

12. Najjar SS, Schulman SP, Gerstenblith G et al. Age and gender affect ventricularvascular coupling during aerobic exercise. J Am Coll Cardiol. 2004;44:611-617.

13. Davies NJ, Denison DM. The measurement of metabolic gas exchange and minute volume by mass spectrometry alone. Respir Physiol. 1979;36:261-267.

14. Ommen SR, Nishimura RA, Appleton CP et al. Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: A comparative simultaneous Doppler-catheterization study. Circulation. 2000; 102: 1788-1794.

15. Magorien DJ, Shaffer P, Bush C et al. Hemodynamic correlates for timing intervals, ejection rate and filling rate derived from the radionuclide angiographic volume curve. Am J Cardiol. 1984;53:567-571. Page 1 20

16. Borlaug BA, Melenovsky V, Russell SD et al. Impaired chronotropic and vasodilator reserves limit exercise capacity in patients with heart failure and a preserved ejection fraction. Circulation. 2006; 114:2138-2147.

17. Brubaker PH, Joo KC, Stewart KP et al. Chronotropic incompetence and its contribution to exercise intolerance in older heart failure patients. J Cardiopulm Rehabil. 2006;26:86-89.

18. Westermann D, Kasner M, Steendijk P et al. Role of left ventricular stiffness in heart failure with normal ejection fraction. Circulation. 2008; 117:2051-2060.

19. ZiIe MR, Baicu CF, Gaasch WH. Diastolic heart failure-abnormalities in activerelaxation and passive stiffness of the left ventricle. N Engl J Med. 2004;350: 1953-1959.

20. ZiIe MR, Brutsaert DL. New concepts in diastolic dysfunction and diastolic heart failure: Part I: diagnosis, prognosis, and measurements of diastolic function. Circulation. 2002; 105: 1387-1393.

21. Mantero A, Gentile F ₁ Gualtierotti C et al. Left ventricular diastolic parameters in 288 normal subjects from 20 to 80 years old. Eur Heart J. 1995; 16:94-105.

22. Yu CM, Lin H, Yang H et al. Progression of systolic abnormalities in patients with "isolated" diastolic heart failure and diastolic dysfunction. Circulation. 2002; 105: 1195-1201.

23. Burkhoff D, Maurer MS, Packer M. Heart failure with a normal ejection fraction: is it really a disorder of diastolic function? Circulation. 2003; 107:656-658.

24. Kawaguchi M, Hay I, Fetics B et al. Combined ventricular systolic and arterial stiffening in patients with heart failure and preserved ejection fraction: implications for systolic and diastolic reserve limitations. Circulation. 2003;107:714-720.

25. Grossman W, Braunwald E, Mann T et al. Contractile state of the left ventricle in man as evaluated from end-systolic pressure-volume relations. Circulation. 1977;56:845-852. Page | 21

26. Nihoyannopoulos P, Karatasakis G, Frenneaux M et al. Diastolic function in hypertrophic cardiomyopathy: relation to exercise capacity. J Am Coll Cardiol. 1992; 19:536-540.

27. Borlaug BA, Melenovsky V, Russell SD et al. Impaired chronotropic and vasodilator reserves limit exercise capacity in patients with heart failure and a preserved ejection fraction. Circulation. 2006; 114:2138-2147.

28. Eckberg DL, Drabinsky M, Braunwald E. Defective cardiac parasympathetic control in patients with heart disease. N Engl J Med. 1971 ;285:877-883.

29. Pena JR, Wolska BM. Troponin I phosphorylation plays an important role in the relaxant effect of beta-adrenergic stimulation in mouse hearts. Cardiovasc Res. 2004;61 :756-763.

30. Fukuda N, Wu Y, Nair P et al. Phosphorylation of titin modulates passive stiffness of cardiac muscle in a titin isoform-dependent manner. J Gen Physiol. 2005; 125:257-271.

31. Kiriazis H, Kranias EG. Genetically engineered models with alterations in cardiac membrane calcium-handling proteins. Annu Rev Physiol. 2000;62:321-351.

32. Leite-Moreira AF, Correia-Pinto J, Gillebert TC. Afterload induced changes in myocardial relaxation: a mechanism for diastolic dysfunction. Cardiovasc Res. 1999;43:344-353.

33. Gillebert TC, Leite-Moreira AF, De Hert SG. Relaxation-systolic pressure relation. A load-independent assessment of left ventricular contractility. Circulation. 1997;95:745-752. Page | 22

34. Takimoto E, Soergel DG, Janssen PM et al. Frequency- and afterload-dependent cardiac modulation in vivo by troponin I with constitutively active protein kinase A phosphorylation sites. Circ Res. 2004;94:496-504.

35. Zhang R, Zhao J, Mandveno A et al. Cardiac troponin I phosphorylation increases the rate of cardiac muscle relaxation. Circ Res. 1995;76: 1028-1035.

36. Robertson SP, Johnson JD, Holroyde MJ et al. The effect of troponin I phosphorylation on the Ca2+-binding properties of the Ca2+-regulatory site of bovine cardiac troponin. J Biol Chem. 1982;257:260-263.

37. Lamb HJ, Beyerbacht HP, van der LA et al. Diastolic dysfunction in hypertensive heart disease is associated with altered myocardial metabolism. Circulation. 1999;99:2261-2267.

38. Smith CS, Bottomley PA, Schulman SP et al. Altered creatine kinase adenosine triphosphate kinetics in failing hypertrophied human myocardium. Circulation. 2006; 114:1151-1158.

39. Scheuermann-Freestone M, Madsen PL, Manners D et al. Abnormal cardiac and skeletal muscle energy metabolism in patients with type 2 diabetes. Circulation. 2003; 107:3040-3046.

40. Scheuermann-Freestone M NSCK. Abnormal cardiac muscle function in heart failure is related to insulin resistance. Cardiovasc J S Afr. 15, s12.

41. Szibor M, Holtz J. Mitochondrial ageing. Basic Res Cardiol. 2003;98:210-218.

42. Ashrafian H, Frenneaux MP, Opie LH. Metabolic mechanisms in heart failure. Circulation. 2007; 116:434-448. Page 1 23

43. Ingwall JS, Weiss RG. Is the failing heart energy starved? On using chemical energy to support cardiac function. Circ Res. 2004;95: 135-145.

44. Abozguia K, Clarke K, Lee L et al. Modification of myocardial substrate use as a therapy for heart failure. Nat CHn Pract Cardiovasc Med. 2006;3:490-498.

45. Fragasso G, Perseghin G, De Cobelli F et al. Effects of metabolic modulation by trimetazidine on left ventricular function and phosphocreatine/adenosine triphosphate ratio in patients with heart failure. Eur Heart J. 2006;27:942-948.

46. Lee L, Campbell R, Scheuermann-Freestone M et al. Metabolic modulation with perhexiline in chronic heart failure: a randomized, controlled trial of short-term use of a novel treatment. Circulation. 2005; 112:3280-3288.

47. Higginbotham MB, Morris KG, Williams RS et al. Regulation of stroke volume during submaximal and maximal upright exercise in normal man. Circ Res. 1986;58:281-291.