Heart failure with improved ejection fraction

Article written and reviewed by Dr Luca Moretti

Heart failure with improved ejection fraction, abbreviated HFimpEF, identifies a patient who previously had a reduced left ventricular ejection fraction and who, after treatment or removal of the cause, shows a significant increase in systolic function. The updated definition requires previous HFrEF, an increase of at least 10 percentage points, and a subsequent measurement above 40%; the documented history is therefore part of the diagnosis and cannot be replaced by a single echocardiogram. The term describes a biological trajectory, not a new heart disease or a certain cure. Volumes, filling pressure, strain, contractile reserve, scar, and the molecular profile may remain abnormal even when the percentage enters the normal range; symptoms, natriuretic peptides, and arrhythmic risk do not necessarily follow the same course. The distinction from de novo HFrEF or de novo HFpEF is clinically decisive. A current ejection fraction of 50% in a patient with previous HFrEF retains different etiological, therapeutic, and prognostic implications from the same measurement in someone who has never had systolic dysfunction.

Improvement often results from reverse remodeling produced by neurohormonal therapy, resynchronization, or correction of ischemia, valvular disease, tachycardia, toxicity, or inflammation. The probability is higher in nonischemic forms of shorter duration, with less dilation and less fibrosis, but no characteristic guarantees permanent stability. Remission may last for years or be interrupted after infection, pregnancy, arrhythmia, hypertension, cardiotoxic exposure, or treatment reduction. TRED-HF demonstrated a high frequency of relapse after gradual drug withdrawal in recovered dilated cardiomyopathy, underpinning the recommendation to continue disease-modifying therapy. The goal is not only to maintain ejection fraction. Etiology, rhythm, biomarkers, volumes, valves, right heart function, functional capacity, and complications must be monitored because residual risk persists even in an asymptomatic patient.

Etiology, pathogenesis, and pathophysiology

Causes with greater reversibility include tachycardia-induced cardiomyopathy, dysfunction due to a high ectopic beat burden, resolved myocarditis, peripartum cardiomyopathy, alcohol or drug toxicity, and corrected dyssynchrony. Recovery after arrhythmia control may be rapid, but latent arrhythmia-induced cardiomyopathy or scar may explain relapses despite stable rhythm. In genetic dilated cardiomyopathy, reversibility depends on genotype and context. TTN variants may be associated with marked recovery, while laminopathies, desmosomal diseases, and FLNC or RBM20 variants maintain arrhythmic risk disproportionate to ejection fraction; the genetic result modifies surveillance and defibrillator decisions. In ischemic heart disease, stunned or hibernating myocardium may recover after revascularization, while transmural scar limits remodeling; global ejection fraction may improve through compensation by viable segments without eliminating the substrate for ventricular tachycardia.

ARNI, ACE inhibitors, or ARBs reduce afterload and profibrotic signaling; beta-blockers attenuate catecholamine toxicity and oxygen consumption; mineralocorticoid receptor antagonists and SGLT2 inhibitors act on the kidneys, volume, inflammation, and metabolism. The combined effect reduces end-systolic volumes and wall stress, promoting regression of remodeling beyond a simple inotropic increase. Resynchronization corrects heterogeneous activation, increases efficiency, and may produce an exceptional response in patients with left bundle branch block, particularly women and those with a nonischemic etiology and less scar. A super-responder nevertheless retains the substrate that made CRT necessary and remains dependent on effective pacing; correction of primary or secondary valvular regurgitation reduces overload, but postoperative ejection fraction depends on loading conditions. In mitral regurgitation, an initial decrease may reflect loss of the low-impedance ejection pathway rather than new injury.

Apparent recovery may be influenced by measurement variability. Image quality, Simpson method, contrast, rhythm, blood pressure, and volume status produce differences of several points; a change from 38% to 42% does not necessarily equate to biological change. Serial measurements should use the same technique and, when necessary, three-dimensional echocardiography or magnetic resonance imaging; ejection fraction is the ratio of ejected volume to end-diastolic volume. It may increase because the ventricle becomes smaller, because stroke volume increases, or because preload and afterload change; indexed volumes, cardiac output, strain, and valvular regurgitation must therefore be compared. A patient may reach a normalized ejection fraction but retain reduced longitudinal strain, reduced exercise reserve, and elevated filling pressure; the clinical phenotype may then remain symptomatic while meeting the HFimpEF definition.

The biology of remission includes reduction in mechanical stress, neurohormones, and fetal gene transcription, but not always normalization of the extracellular matrix, energetics, and immunity. Interstitial and replacement fibrosis constitute a memory of injury that predisposes to renewed dilation when stress returns. In dilated cardiomyopathy, an early increase in heart rate and blood pressure after drug withdrawal may precede the rise in natriuretic peptides and fall in ejection fraction. Relapse is therefore a dynamic process in which clinical and imaging signals may be timelier than BNP. Persistence of elevated galectin, troponin, or peptides does not define relapse on its own but signals risk; the combination of scar on magnetic resonance imaging, genotype, residual volumes, and biomarker trends provides a more robust assessment.

The trajectory is not unidirectional. Some patients move from HFrEF to HFimpEF and return to HFrEF; others fluctuate in the 40-50% range in relation to rhythm and loading conditions. Every substantial change requires distinguishing true relapse, technical error, and a transient change in hemodynamic conditions; atrial fibrillation may reduce ejection fraction with a rapid heart rate and make its measurement unstable from beat to beat. An increasing burden, even without continuous tachycardia, may signal loss of remission; devices and prolonged monitoring help in selected cases. Pregnancy, systemic infections, severe anemia, thyroid dysfunction, and cardiotoxic drugs may exceed limited reserve; previous improvement does not eliminate the need to prevent and treat these second hits.

Arrhythmic risk decreases with reverse remodeling but does not disappear. Ischemic or nonischemic scar, high-risk variants, previous tachycardias, and persistent inducibility may maintain an indication for protection; current ejection fraction alone is insufficient to decide on ICD explantation or nonreplacement. The right ventricle may improve in parallel with reduced pulmonary pressures, but advanced damage or tricuspid valve disease may persist; biventricular discordance explains edema and limitation despite left ventricular recovery. The kidneys, muscle, and vascular system may retain consequences of the chronic phase. Frailty and deconditioning require specific rehabilitation and do not regress automatically with echocardiographic improvement.

Clinical manifestations

Many patients become asymptomatic or move to a better NYHA class; disappearance of orthopnea and edema reflects reduced pressures, while recovery of capacity also depends on muscle, chronotropic function, and comorbidities. Clinical remission must be documented with a functional history and not inferred from ejection fraction alone. Residual dyspnea may result from congestion, chronotropic incompetence, ischemia, pulmonary disease, anemia, obesity, or overlapping HFpEF; automatic interpretation as “heart failure still active” exposes the patient to excessive diuresis; the opposite risks overlooking elevated pressures. The walk test and cardiopulmonary exercise testing quantify the limitation and partly separate cardiac, ventilatory, and peripheral components. Reduced peak VO2 remains prognostic even with recovered ejection fraction.

Signs of relapse include declining exercise tolerance, increased resting heart rate, rising blood pressure, palpitations, orthopnea, weight gain, and edema; recurrence may be gradual and precede hospitalization by weeks; a self-monitoring plan facilitates recognition. A new arrhythmia or a decrease in biventricular pacing may cause rapid deterioration. Palpitations, presyncope, syncope, or appropriate ICD shocks require urgent assessment of the electrical substrate, electrolytes, and ischemia. Chest pain in ischemic heart disease or after myocarditis must not be attributed to remission. Ischemia, inflammatory recurrence, and noncardiac causes follow distinct diagnostic pathways.

Asymptomatic low blood pressure during therapy may be compatible with good perfusion and does not require automatic treatment reduction. Dizziness, orthostatic symptoms, syncope, oliguria, and cold skin instead indicate hemodynamic intolerance or another cause requiring correction, starting with volume status and nonessential drugs. Bradycardia may result from beta-blockade, sinus node disease, or pacing; persistent tachycardia suggests infection, anemia, hyperthyroidism, arrhythmia, or drug withdrawal; the heart rate trend is often an early indicator of loss of balance. Examination assesses jugular veins, a third heart sound, murmurs, edema, and perfusion; a normal examination at rest does not exclude pathological pressures during activity.

Quality of life may remain impaired by fear of relapse, ICD shocks, work limitations, and uncertainty about the word “recovery.” Appropriate communication uses the concept of controlled remission: it acknowledges improvement without creating an expectation of stopping follow-up and therapy. Depression and anxiety influence adherence and activity. Rehabilitation, psychological support, and shared decision-making help patients safely resume exercise, sexual activity, and work. In women with peripartum cardiomyopathy, a subsequent pregnancy may entail relapse even after normalization; preconception counseling integrates ejection fraction, strain, reserve, history, and preferences, without treating a normal measurement as a guarantee.

Absence of congestion often allows a reduction in the diuretic dose, but does not imply withdrawal of disease-modifying drugs. Cramps, thirst, and hypotension may signal iatrogenic hypovolemia; the goal is the minimum diuretic dose that maintains euvolemia. Recurrence of ascites or isolated edema may depend on right heart dysfunction, kidney disease, liver disease, or venous insufficiency; the distribution of signs and venous and cardiac ultrasound prevent indiscriminate escalation. Respiratory infections, NSAID use, and failure to take medications are frequent precipitants and must be sought before attributing every symptom to structural progression.

Investigations and diagnosis

Diagnosis requires at least two reliable ejection fraction assessments and documentation of previous HFrEF. Reports, original images, and the treatment timeline must be retrieved; the phrase “previously reduced function” without a value and method does not always allow formal criteria to be applied. Echocardiography measures ejection fraction, volumes, strain, atria, right heart function, pulmonary pressure, and valves; serial comparison must consider rhythm, blood pressure, and volume, preferably with side-by-side image review. Contrast or magnetic resonance imaging is indicated when the endocardial border is inadequately visualized or the decision depends on a few percentage points; magnetic resonance imaging adds scar, edema, infiltration, and thrombi.

BNP or NT-proBNP may normalize, remain elevated because of the atrium, kidney function, or pressures, or increase at relapse; an isolated value has limited specificity; an individual trend associated with weight, heart rate, and imaging is more informative. Persistent troponin elevation suggests ongoing injury or a high-risk substrate, but requires interpretation with kidney function, ischemia, and cardiomyopathy. Complete blood count, creatinine, electrolytes, liver function, TSH, iron status, and glucose identify reversible causes and treatment safety; ECG documents QRS, conduction blocks, fibrillation, ectopy, and signs of scar. Ambulatory monitoring is useful for palpitations, suspected tachycardia-induced cardiomyopathy, ectopic beats, or genetic stratification.

The original cause should be reassessed if it remained idiopathic. Magnetic resonance imaging, coronary anatomy, family history, genetic testing, and investigation of exposures may transform prognosis; genetic cardiomyopathy is not excluded by recovery. Genetic testing is particularly indicated with a family history, conduction disorders, arrhythmias, sudden death, muscle involvement, or an unexplained dilated phenotype; the result requires counseling and appropriate cascade screening. In myocarditis, persistent late enhancement may maintain risk even with normal function. Biopsy is not a routine test in remission, but is reserved for scenarios in which histology changes treatment.

Longitudinal strain may remain reduced and predict less stability, but depends on software and loading conditions. A new, reproducible reduction may precede a fall in ejection fraction, particularly during cardiotoxicity; however, it must not generate a diagnosis from a change smaller than the limits of reproducibility. Cardiopulmonary exercise testing assesses integrated reserve, ventilatory inefficiency, and chronotropic function. It is useful with discordant symptoms, in occupational or sports assessment, and when determining whether the patient remains in an advanced phase despite recovery of the percentage. Exercise echocardiography or invasive hemodynamics may demonstrate an abnormal rise in pressures; this physiology may represent residual cardiomyopathy or overlapping HFpEF.

Surveillance is personalized according to cause and risk. After treatment changes or events, clinical status, blood pressure, kidney function, and potassium are checked; imaging is repeated when the result changes management, not according to a rigid schedule identical for everyone. In ICD or CRT recipients, arrhythmias, pacing percentage, threshold, battery, and signs of congestion are checked; loss of biventricular capture may precede worsening and requires correction of arrhythmia, programming, or leads. Diagnosis of relapse integrates a fall in ejection fraction, increased volumes, doubling of peptides, symptoms, or treatment requirements. Thresholds from studies such as TRED-HF help monitoring, but clinical decisions consider the whole picture.

ICD reassessment at generator replacement is complex. Previous appropriate therapies, scar, genotype, age, comorbidities, and duration of recovery weigh more than ejection fraction alone; shared decision-making also considers complications and preferences. CRT is not discontinued in a responder because electrical correction may be the cause of recovery. Turning off pacing may recreate dyssynchrony and deterioration. Dedicated assessments are needed for sports activity, pregnancy, and future cardiotoxicity; remission does not automatically authorize high-load exposures without stratification.

Treatment and prognosis

The central rule is to maintain HFrEF prognosis-modifying therapy even after improvement, unless true contraindications exist. ARNI or other renin-angiotensin system blockade, a beta-blocker, a mineralocorticoid receptor antagonist, and an SGLT2 inhibitor are continued at tolerated doses; remission is largely a result of therapy itself. TRED-HF showed early relapses after withdrawal in selected patients with apparently recovered dilated cardiomyopathy. Extended follow-up confirms that risk persists over time, supporting sustained monitoring and caution even when a patient has passed the initial period. If hypotension, hyperkalemia, or kidney dysfunction limits treatment, hypovolemia and drugs without benefit are addressed first, doses are spaced apart, and selective reductions are made; global deprescribing exposes the patient to loss of combined protection.

Diuretics are not mandatory if the patient is consistently euvolemic. The dose may be reduced with weight and symptom monitoring, while a rapid increase must be managed according to an agreed plan; this distinction avoids confusing decongestive therapy with disease therapy. Etiological treatment continues: rhythm control in tachycardia-induced cardiomyopathy, abstinence from alcohol or toxic substances, thyroid management, correction of ischemia, and valvular follow-up; removal of the trigger must be sustained because re-exposure may produce a faster second episode of dysfunction. In cancer therapy-related cardiotoxicity, cardio-oncology collaboration balances anticancer therapy, function, and biomarkers; recovery sometimes allows a protected rechallenge, not an automatic decision.

An SGLT2 inhibitor retains rationale and benefit even at increased ejection fractions. In the prespecified DELIVER analysis, dapagliflozin reduced events in patients with a previously reduced ejection fraction that subsequently improved, providing direct evidence beyond the rationale for maintenance. There is no ejection fraction threshold at which ARNI, a beta-blocker, or MRA suddenly becomes useless in a patient with previous HFrEF; the longitudinal history takes precedence over cross-sectional reclassification. Target doses remain goals when tolerated, but adherence to all classes may be more important than insisting on a single high dose. Blood pressure, heart rate, kidney function, potassium, and preferences guide optimization.

Aerobic and resistance exercise improves capacity and muscle mass. The program is started during stability and adapted to arrhythmias, ICD, ischemia, and frailty; recovery of ejection fraction does not make chronic rest necessary. Blood pressure, diabetes, obesity, sleep apnea, smoking, and lipids must be addressed; cardiovascular prevention reduces new insults that might reactivate remodeling. Vaccinations, medication reconciliation, and instructions on NSAIDs, infections, and dehydration complete prevention; self-management does not mean independently changing prognosis-modifying therapies.

ICD and CRT follow individual assessment; an implanted ICD is not removed because ejection fraction improves; arrhythmic risk and expected benefit are reexamined at generator replacement. Previous ventricular tachycardia or extensive fibrosis supports protection, while comorbidities and a prolonged absence of events may modify the choice; resynchronization continues in responders. If the battery or leads require a procedure, the causal role of CRT in recovery must be explicitly considered. Ablation of fibrillation or ectopic beats may preserve remission in arrhythmia-related phenotypes. Procedural success does not eliminate monitoring and treatment of the substrate.

Prognosis is better on average than in HFrEF with persistently reduced ejection fraction, with fewer hospitalizations and lower mortality, but worse than in people without heart failure. Outcomes depend on cause, duration, scar, genotype, right heart function, biomarkers, arrhythmias, and treatment completeness; predictors of relapse include greater residual dilation, incompletely normalized ejection fraction, abnormal strain, elevated peptides, fibrillation, drug withdrawal, and new triggers. No model allows generalizable safe withdrawal; a proportion progress to advanced heart failure despite transient recoveries. Repeated hospitalizations, drug intolerance, hypotension, or low capacity require timely referral, without being reassured by an isolated favorable measurement.

Complications

The characteristic complication is systolic relapse, which may initially be asymptomatic and then evolve into congestion or low cardiac output. Treatment consists of investigating triggers, reestablishing all tolerated classes, and correcting rhythm, ischemia, blood pressure, or toxicity, avoiding treating it as merely an inevitable failure. Repeated relapses increase fibrosis and reduce the probability of another complete recovery; timely recognition is therefore a prognostic component. A mild but confirmed fall may be clinically significant if accompanied by increased volumes or biomarkers; management need not wait for the 40% threshold.

Ventricular arrhythmias and sudden death remain possible because of scar or genotype. Appropriate ICD therapies may occur even after normalization, although absolute risk is often lower; electrolytes, ischemia, and adherence are modifiable factors. Atrial fibrillation, flutter, and ectopic beats may be both cause and consequence of deterioration; arrhythmic burden is more informative than an occasional ECG and may require device or prolonged monitoring. Conduction blocks and bradycardia are relevant in laminopathies, sarcoidosis, and after drug treatment. Syncope requires urgent assessment even with a normal ejection fraction.

Congestion, cardiorenal syndrome, and electrolyte disturbances may recur during infections or treatment changes. Rising creatinine must be interpreted with perfusion and volume status; simultaneously stopping RAASi, MRA, and SGLT2 without a plan may promote a spiral of treatment de-optimization. Hyperkalemia requires review of diet, supplements, NSAIDs, kidney function, and potential binders, when appropriate, to preserve useful therapies. Diuretic-induced hypokalemia and hypomagnesemia increase arrhythmic risk; hypovolemia is the opposite complication, particularly after a reduction in diuretic requirements. Orthostatic symptoms and kidney injury require targeted adjustment of decongestion.

Mitral or tricuspid regurgitation may persist or reappear with dilation; severity must be measured under comparable conditions because a temporary reduction with diuresis does not equate to structural resolution. Right heart dysfunction and pulmonary hypertension may become the dominant phenotype, causing edema, ascites, and hepatic congestion despite good left ventricular recovery; residual right ventricular impairment is an important determinant of outcome. Ventricular thrombus and embolism are less likely after recovery, but anticoagulation follows specific indications and is not discontinued without reassessing thrombus, fibrillation, or other causes.

Pregnancy may cause relapse in peripartum cardiomyopathy, and some drugs are fetotoxic. Planning requires specialist counseling, advance modification of therapy, and monitoring during pregnancy and the puerperium; preconception normalization reduces but does not eliminate risk. New chemotherapy, alcohol or stimulant abuse, and myocardial infections may reactivate injury; the history of HFimpEF must always appear in the documentation to prevent the patient from being treated as though cardiac function had never been abnormal. Anxiety, reduced adherence, and loss to follow-up are indirect complications of communicating a “cure.” Education and a realistic surveillance plan protect the result achieved.