Predicting Survival in Patients With Heart Failure With an Implantable Cardioverter Defibrillator: The Heart Failure Meta-Score

Prophylactic ICD Survival Benefit Prediction: Review and Comparison between Main Scores

Current guidelines advocate for the use of prophylactic implantable cardioverter defibrillators (ICDs) for all patients with symptomatic heart failure (HF) with low ejection fraction (EF). As many patients will never use their device and some are prone to device-related complications, scoring systems for delineating subgroups with differential ICD survival benefits are crucial to maximize ICD benefit and mitigate complications. This review summarizes the main scores, including MADIT trial-based Risk Stratification Score (MRSS) and Seattle Heart Failure Model (SHFM), which are based on randomized trials with a control group (HF medication only) and validated on large cohorts of 'real-world' HF patients. Recent studies using cardiac MRI (CMR) to predict ventricular arrhythmia (VA) are mentioned as well. The review shows that most scores could not delineate sustained VA incidence, but rather mortality without prior appropriate ICD therapies. Multiple scores could identify high-risk subgroups with extremely high probability of early mortality after ICD implant. On the other hand, low-risk subgroups were defined, in whom a high ratio of appropriate ICD therapy versus death without prior appropriate ICD therapy was found, suggesting significant ICD survival benefit. Moreover, MRSS and SHFM proved actual ICD survival benefit in low- and medium-risk subgroups when compared with control patients, and no benefit in high-risk subgroups, consisting of 16-20% of all ICD candidates. CMR reliably identified areas of myocardial scar and 'channels', significantly associated with VA. We conclude that as for today, multiple scoring models could delineate patient subgroups that would benefit differently from prophylactic ICD. Due to their modest-moderate predictability, these scores are still not ready to be implemented into clinical guidelines, but could aid decision regarding prophylactic ICD in borderline cases, as elderly patients and those with multiple co-morbidities. CMR is a promising technique which might help delineate patients with a low- versus high-risk for future VA, beyond EF alone. Lastly, genetic analysis could identify specific mutations in a non-negligible percent of patients, and a few of these mutations were found to predict an increased arrhythmic risk.

Predicting heart failure outcomes by integrating breath-by-breath measurements from cardiopulmonary exercise testing and clinical data through a deep learning survival neural network

Aims

Mathematical models previously developed to predict outcomes in patients with heart failure (HF) generally have limited performance and have yet to integrate complex data derived from cardiopulmonary exercise testing (CPET), including breath-by-breath data. We aimed to develop and validate a time-to-event prediction model using a deep learning framework using the DeepSurv algorithm to predict outcomes of HF.

Methods and results

Inception cohort of 2490 adult patients with high-risk cardiac conditions or HF underwent CPET with breath-by-breath measurements. Potential predictive features included known clinical indicators, standard summary statistics from CPETs, and mathematical features extracted from the breath-by-breath time series of 13 measurements. The primary outcome was a composite of death, heart transplant, or mechanical circulatory support treated as a time-to-event outcomes. Predictive features ranked as most important included many of the features engineered from the breath-by-breath data in addition to traditional clinical risk factors. The prediction model showed excellent performance in predicting the composite outcome with an area under the curve of 0.93 in the training and 0.87 in the validation data sets. Both the predicted vs. actual freedom from the composite outcome and the calibration of the prediction model were excellent. Model performance remained stable in multiple subgroups of patients.

Conclusion

Using a combined deep learning and survival algorithm, integrating breath-by-breath data from CPETs resulted in improved predictive accuracy for long-term (up to 10 years) outcomes in HF. DeepSurv opens the door for future prediction models that are both highly performing and can more fully use the large and complex quantity of data generated during the care of patients with HF.

Factors Impacting Physician Prognostic Accuracy in Heart Failure Patients With Reduced Left Ventricular Ejection Fraction

BACKGROUND

A recent study showed that the accuracy of heart failure (HF) cardiologists and family doctors to predict mortality in outpatients with HF proved suboptimal, performing less well than models.

OBJECTIVES

The authors sought to evaluate patient and physician factors associated with physician accuracy.

METHODS

The authors included outpatients with HF from 11 HF clinics. Family doctors and HF cardiologists estimated patient 1-year mortality. They calculated predicted mortality using the Seattle HF Model and followed patients for 1 year to record mortality (or urgent heart transplant or ventricular assist device implant as mortality-equivalent events). Using multivariable logistic regression, the authors evaluated associations among physician experience and confidence in estimates, duration of patient-physician relationship, patient-physician sex concordance, patient race, and predicted risk, with concordant results between physician and model predictions.

RESULTS

Among 1,643 patients, 1-year event rate was 10% (95% CI: 8%-12%). One-half of the estimates showed discrepant results between model and physician predictions, mainly owing to physician risk overestimation. Discrepancies were more frequent with increasing patient risk from 38% in low-risk to ∼75% in high-risk patients. When making predictions on male patients, female HF cardiologists were 26% more likely to have discrepant predictions (OR: 0.74; 95% CI: 0.58-0.94). HF cardiologist estimates in Black patients were 33% more likely to be discrepant (OR: 0.67; 95% CI: 0.45-0.99). Low confidence in predictions was associated with discrepancy. Analyses restricted to high-confidence estimates showed inferior calibration to the model, with risk overestimation across risk groups.

CONCLUSIONS

Discrepant physician and model predictions were more frequent in cases with perceived increased risk. Model predictions outperform physicians even when they are confident in their predictions. (Predicted Prognosis in Heart Failure [INTUITION]; NCT04009798).

Predicting long-term survival after de novo cardioverter-defibrillator implantation for primary prevention: A population based study

Background

Implantable cardioverter-defibrillators (ICDs) reduce the risk of sudden cardiac death in patients with left ventricular dysfunction. While short-term mortality benefit of ICD insertion has been established in landmark randomized controlled trials, little is known about the long-term outcomes of patients with ICDs in clinical practice. In this paper, we describe the long-term survival of patients following de novo ICD implantation for primary prevention in clinical practice and determine the factors which help predict survival after ICD implant.

Methods

Retrospective population-based study of all patients receiving a de novo ICD for primary prevention in Ontario, Canada from 2007 to 2011 using the Ontario ICD Database housed within ICES. Simple random selection was used to split the population into a derivation and internal validation cohort in a ratio of 2:1. Cox proportional hazards regression was used to determine predictors of interest and predict 10-year survival, model performance was assessed using calibration and validation.

Results

In the derivation cohort (n = 3399), mean age was 65.3 years (standard deviation [SD] = 11.0), 664 patients were female (19.5 %) and 2344 patients (69.0 %) had ischemic cardiomyopathy. Ten year survival was 45.7 % (95 % confidence interval [CI] 44.0 %-47.4 %). The final prediction model included age, sex, disease factors (ischemic vs nonischemic cardiomyopathy, left ventricular ejection fraction) and patient factors (symptoms, comorbidities), and biomarkers at the time of ICD assessment. This model had good discrimination and calibration in derivation (0.79, 95 % CI 0.77, 0.81) and validation samples (0.78, 95 % CI 0.76, 0.79).

Conclusions

A combination of demographic and clinical factors determined at baseline can be used to predict 10-year survival in patients with implantable cardioverter-defibrillators with good accuracy. Our findings help to identify individuals at risk of long-term mortality and may be useful in targeting future prevention strategies to enhance longevity in this high-risk population.

Performance of current risk stratification models for predicting mortality in patients with heart failure: a systematic review and meta-analysis

AIMS

There are several risk scores designed to predict mortality in patients with heart failure (HF). This study aimed to assess performance of risk scores validated for mortality prediction in patients with acute HF (AHF) and chronic HF.

METHODS AND RESULTS

MEDLINE and Scopus were searched from January 2015 to January 2021 for studies which internally or externally validated risk models for predicting all-cause mortality in patients with AHF and chronic HF. Discrimination data were analysed using C-statistics, and pooled using generic inverse-variance random-effects model. Nineteen studies (n = 494 156 patients; AHF: 24 762; chronic HF mid-term mortality: 62 000; chronic HF long-term mortality: 452 097) and 11 risk scores were included. Overall, discrimination of risk scores was good across the three subgroups: AHF mortality [C-statistic: 0.76 (0.68-0.83)], chronic HF mid-term mortality [1 year; C-statistic: 0.74 (0.68-0.79)], and chronic HF long-term mortality [≥2 years; C-statistic: 0.71 (0.69-0.73)]. MEESSI-AHF [C-statistic: 0.81 (0.80-0.83)] and MARKER-HF [C-statistic: 0.85 (0.80-0.89)] had an excellent discrimination for AHF and chronic HF mid-term mortality, respectively, whereas MECKI had good discrimination [C-statistic: 0.78 (0.73-0.83)] for chronic HF long-term mortality relative to other models. Overall, risk scores predicting short-term mortality in patients with AHF did not have evidence of poor calibration (Hosmer-Lemeshow P > 0.05). However, risk models predicting mid-term and long-term mortality in patients with chronic HF varied in calibration performance.

CONCLUSIONS

The majority of recently validated risk scores showed good discrimination for mortality in patients with HF. MEESSI-AHF demonstrated excellent discrimination in patients with AHF, and MARKER-HF and MECKI displayed an excellent discrimination in patients with chronic HF. However, modest reporting of calibration and lack of head-to-head comparisons in same populations warrant future studies.

Prognostic value of natriuretic peptides in heart failure: systematic review and meta-analysis

Risk models, informing optimal long-term medical management, seldom use natriuretic peptides (NP) in ascertaining the absolute risk of outcomes for HF patients. Individual studies evaluating the prognostic value of NPs in HF patients have reported varying effects, arriving at best estimates requires a systematic review. We systematically summarized the best evidence regarding the prognostic value of brain natriuretic peptide (BNP) and NT-proBNP in predicting mortality and hospitalizations in ambulatory heart failure (HF) patients. We searched bibliographic databases from 2005 to 2018 and included studies evaluating the association of BNP or NT-proBNP with mortality or hospitalization using multivariable Cox proportional hazard models. We pooled hazard ratios using random-effect models, explored heterogeneity using pre-specified subgroup analyses, and evaluated the certainty of evidence using the Grading of Recommendations and Development Evaluation framework. We identified 67 eligible studies reporting on 76,178 ambulatory HF patients with a median BNP of 407 pg/mL (261-574 pg/mL). Moderate to high-quality evidence showed that a 100-pg/mL increase in BNP was associated with a 14% increased hazard of mortality (HR 1.14, 95% CI 1.06-1.22); a 1-log-unit increase was associated with a 51% increased hazard of mortality (HR 1.51, 95% CI 1.41-1.61) and 48% increased hazard of mortality or hospitalization (HR 1.48, 95% CI 1.29-1.69). With moderate to high certainty, we observed a 14% independent relative increase in mortality, translating to a clinically meaningful increase in absolute risk even for low-risk patients. The observed associations may help in developing more accurate risk models that incorporate NPs and accurately prognosticate HF patients.

Application of the heart failure meta-score to predict prognosis in patients with cardiac resynchronization defibrillators

BACKGROUND

The Heart Failure (HF) Meta-score may be useful in predicting prognosis in patients with primary prevention cardiac resynchronization defibrillators (CRT-D) considering the competing risk of appropriate defibrillator shock versus mortality.

METHODS

Data from 648 consecutive patients from two centers were used for the evaluation of the performance of the HF Meta-score. The primary endpoint was mortality and the secondary endpoint was time to first appropriate implantable cardioverter-defibrillator (ICD) shock or death without prior appropriate ICD shock. Fine-Gray model was used for competing risk regression analysis.

RESULTS

In the entire cohort, 237 patients died over a median follow-up of 5.2 years. Five-year cumulative incidence of mortality ranged from 12% to 53%, for quintiles 1 through 5 of the HF Meta-score, respectively (log-rank P < 0.001). Compared with the lowest quintile, mortality risk was higher in the highest quintile (HR 6.9; 95%CI 3.7-12.8). The HF Meta-score had excellent calibration, accuracy, and good discrimination in predicting mortality (C-statistic 0.76 at 1-year and 0.71 at 5-year). The risk of death without appropriate ICD shock was higher in risk quintile 5 compared to quintile 1 (sub HR 5.8; 95%CI 3.1-11.0, P < 0.001).

CONCLUSIONS

Our study demonstrated a good ability of the HF Meta-score to predict survival in HF patients treated with CRT-D as primary prevention. The HF Meta-score proved to be useful in identifying a subgroup with a significantly poor prognosis despite a CRT-D.

Hyponatremia in Heart Failure: Pathogenesis and Management

Hyponatremia is a very common electrolyte abnormality, associated with poor short- and long-term outcomes in patients with heart failure (HF). Two opposite processes can result in hyponatremia in this setting: Volume overload with dilutional hypervolemic hyponatremia from congestion, and hypovolemic hyponatremia from excessive use of natriuretics. These two conditions require different therapeutic approaches. While sodium in the form of normal saline can be lifesaving in the second case, the same treatment would exacerbate hyponatremia in the first case. Hypervolemic hyponatremia in HF patients is multifactorial and occurs mainly due to the persistent release of arginine vasopressin (AVP) in the setting of ineffective renal perfusion secondary to low cardiac output. Fluid restriction and loop diuretics remain mainstay treatments for hypervolemic/dilutional hyponatremia in patients with HF. In recent years, a few strategies, such as AVP antagonists (Tolvaptan, Conivaptan, and Lixivaptan), and hypertonic saline in addition to loop diuretics, have been proposed as potentially promising treatment options for this condition. This review aimed to summarize the current literature on pathogenesis and management of hyponatremia in patients with HF.

Risk Factors for 1-Year Graft Loss After Kidney Transplantation: Systematic Review and Meta-Analysis

BACKGROUND AND OBJECTIVES

With expansion of the pool of kidney grafts, through the use of higher-risk donors, and increased attention to donor management strategies, the 1-year graft survival rate is subject to change. It is, therefore, useful to elucidate 1-year graft survival rates by dissecting the characteristics of the low-risk and high-risk kidney transplant cases. The objective of our study was to evaluate factors purported to influence the risk of 1-year graft loss in kidney transplant recipients.

DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS

We searched bibliographic databases from 2000 to 2017 and included observational studies that measured the association between donor, recipient, the transplant operation, or early postoperative complications, and 1-year death-censored graft loss.

RESULTS

We identified 35 eligible primary studies, with 20 risk factors amenable to meta-analysis. Six factors were associated with graft loss, with moderate to high degree of certainty: donor age (hazard ratio [HR], 1.11 per 10-year increase; 95% confidence interval [95% CI], 1.04 to 1.18), extended criteria donors (HR, 1.35; 95% CI, 1.28 to 1.42), deceased donors (HR, 1.54; 95% CI, 1.32 to 1.82), number of HLA mismatches (HR, 1.08 per one mismatch increase; 95% CI, 1.07 to 1.09), recipient age (HR, 1.17 per 10-year increase; 95% CI, 1.09 to 1.25), and delayed graft function (HR, 1.89; 95% CI, 1.46 to 2.47) as risk factors for 1-year graft loss. Pooled analyses also excluded, with a high degree of certainty, any associations of cold ischemia time, recipient race, pretransplant body mass index, diabetes, and hypertension with 1-year graft loss.

CONCLUSIONS

Recipient age, donor age, standard versus extended criteria donor, living versus deceased donor, HLA mismatch, and delayed graft function all predicted 1-year graft survival. The effect of each risk factor is small.

Predicting Early Mortality Among Implantable Defibrillator Patients Treated With Cardiac Resynchronization Therapy

BACKGROUND

The beneficial effects of a cardiac resynchronization defibrillator (CRT-D) in patients with heart failure, low left ventricular ejection fraction (LVEF), and wide QRS have clearly been established. Nevertheless, mortality remains high in some patients. The aim of this study was to develop and validate a risk score to identify patients at high risk for early mortality who are implanted with a CRT-D.

METHODS AND RESULTS

For predictive modelling, 1282 consecutive patients from 5 centers (74% male; median age 66 years; median LVEF 25%; New York Heart Association class III-IV 60%; median QRS-width 160 ms) were randomly divided into a derivation and validation cohort. The primary endpoint is mortality at 3 years. Model development was performed using multivariate logistic regression by checking log likelihood, Akaike information criterion, and Bayesian information criterion. Model performance was validated using C statistics and calibration plots. The risk score included 7 independent mortality predictors, including myocardial infarction, LVEF, QRS duration, chronic obstructive pulmonary disease, chronic kidney disease, hyponatremia, and anemia. Calibration-in-the-large was suboptimal, reflected by a lower observed mortality (44%) than predicted (50%). The validated C statistic was 0.71 indicating modest performance.

CONCLUSION

A risk score based on routine, readily available clinical variables can assist in identifying patients at high risk for early mortality within 3 years after CRT-D implantation.

Performance of Prognostic Risk Scores in Heart Failure Patients: Do Sex Differences Exist?

BACKGROUND

Sex differences in the performance of prognostic risk scores in heart failure (HF) patients have not previously been investigated. We examined the performance of 2 commonly used scores in predicting mortality and a composite end point consisting of ventricular assist device, heart transplantation, or mortality in women vs men with HF.

METHODS

This was a retrospective study of 1,136 (25% women) consecutive ambulatory adult HF patients with reduced left ventricular ejection fraction (≤ 40%) followed at a single institution from 2000 to 2012. Discrimination, calibration, and absolute risk reclassification of the Seattle Heart Failure Model (SHFM) and the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score to predict 1- and 3-year outcomes were compared between women and men.

RESULTS

At 1- and 3-year follow-ups, 116 (22% women) and 231 (21% women) patients died, respectively. Survival was equal between sexes (P = 0.41). The SHFM and the MAGGIC score showed similar discriminatory capacity in women (c-statistics 0.84, 95% CI 0.77-0.92, and 0.74, 95% CI 0.64-0.83) and men (c-statistics 0.74, 95% CI 0.69-0.79, and 0.70, 95% CI 0.64-0.75). There was no difference in the predicted and observed 1-year mortality by the scores in both sexes. Compared with the SHFM, the MAGGIC score better reclassified 10% (95% CI 7%-14%) of women and 18% (95% CI 15%-20%) of men. At 3-year follow-up, similar results were seen for discrimination, whereas both scores overestimated mortality with more marked overestimation in women. The results were reproducible for the composite end point, with improved calibration at 3-year follow-up in both scores.

CONCLUSIONS

Our findings support the use of the MAGGIC score in both women and men owing to better risk classification.

Physician Judgement vs Model-Predicted Prognosis in Patients With Heart Failure

BACKGROUND

Previous evidence suggests that cardiologists and family doctors have limited accuracy in predicting patient prognosis. Predictive models with satisfactory accuracy for estimating mortality in patients with heart failure (HF) exist; physicians, however, seldom use these models. We evaluated the relative accuracy of physician vs model prediction to estimate 1-year survival in ambulatory patients with HF.

METHODS

We conducted a single-centre cross-sectional study involving 150 consecutive ambulatory patients with HF >18 years of age with a left ventricular ejection fraction ≤40%. Each patient's cardiologist and family doctor provided their predicted 1-year survival, and predicted survival scores were calculated using 3 models: HF Meta-Score, Seattle Heart Failure Model (SHFM), and Meta-Analysis Global Group in Chronic HF (MAGGIC) score. We compared accuracy between physician and model predictions using intraclass correlation (ICC).

RESULTS

Median predicted survival by HF cardiologists was lower (median 80%, interquartile range [IQR]: 61%-90%) than that predicted by family physicians (median 90%, IQR 70%-99%, P = 0.08). One-year median survival calculated by the HF Meta-Score (94.6%), SHFM (95.4%), and MAGGIC (88.9%,) proved as high or higher than physician estimates. Agreement among HF cardiologists (ICC 0.28-0.41) and family physicians (ICC 0.43-0.47) when compared with 1-year model-predicted survival scores proved limited, whereas the 3 models agreed well (ICC > 0.65).

CONCLUSIONS

HF cardiologists underestimated survival in comparison with family physicians, whereas both physician estimates were lower than calculated model estimates. Our results provide additional evidence of potential inaccuracy of physician survival predictions in ambulatory patients with HF. These results should be validated in longitudinal studies collecting actual survival.

An Appraisal of Biomarker-Based Risk-Scoring Models in Chronic Heart Failure: Which One Is Best?

PURPOSE OF REVIEW

While prediction models incorporating biomarkers are used in heart failure, these have shown wide-ranging discrimination and calibration. This review will discuss externally validated biomarker-based risk models in chronic heart failure patients assessing their quality and relevance to clinical practice.

RECENT FINDINGS

Biomarkers may help in determining prognosis in chronic heart failure patients as they reflect early pathologic processes, even before symptoms or worsening disease. We present the characteristics and describe the performance of 10 externally validated prediction models including at least one biomarker among their predictive factors. Very few models report adequate discrimination and calibration. Some studies evaluated the additional predictive value of adding a biomarker to a model. However, these have not been routinely assessed in subsequent validation studies. New and existing prediction models should include biomarkers, which improve model performance. Ongoing research is needed to assess the performance of models in contemporary patients.