Individual patient meta-analyses of restrictive diastolic filling pattern and mortality in patients post acute myocardial infarction and in patients with chronic heart failure

Maelstrom Research guidelines for rigorous retrospective data harmonization

Background

It is widely accepted and acknowledged that data harmonization is crucial: in its absence, the co-analysis of major tranches of high quality extant data is liable to inefficiency or error. However, despite its widespread practice, no formalized/systematic guidelines exist to ensure high quality retrospective data harmonization.

Methods

To better understand real-world harmonization practices and facilitate development of formal guidelines, three interrelated initiatives were undertaken between 2006 and 2015. They included a phone survey with 34 major international research initiatives, a series of workshops with experts, and case studies applying the proposed guidelines.

Results

A wide range of projects use retrospective harmonization to support their research activities but even when appropriate approaches are used, the terminologies, procedures, technologies and methods adopted vary markedly. The generic guidelines outlined in this article delineate the essentials required and describe an interdependent step-by-step approach to harmonization: 0) define the research question, objectives and protocol; 1) assemble pre-existing knowledge and select studies; 2) define targeted variables and evaluate harmonization potential; 3) process data; 4) estimate quality of the harmonized dataset(s) generated; and 5) disseminate and preserve final harmonization products.

Conclusions

This manuscript provides guidelines aiming to encourage rigorous and effective approaches to harmonization which are comprehensively and transparently documented and straightforward to interpret and implement. This can be seen as a key step towards implementing guiding principles analogous to those that are well recognised as being essential in securing the foundational underpinning of systematic reviews and the meta-analysis of clinical trials.

The Effect of Metformin on Diastolic Function in Patients Presenting with ST-Elevation Myocardial Infarction

Introduction

Diastolic dysfunction is an important predictor of poor outcome after myocardial infarction. Metformin treatment improved diastolic function in animal models and patients with diabetes. Whether metformin improves diastolic function in patients presenting with ST-segment elevation myocardial infarction (STEMI) is unknown.

Methods

The GIPS-III trial randomized STEMI patients, without known diabetes, to metformin or placebo initiated directly after PCI. The previously reported primary endpoint was left ventricular ejection fraction at 4 months, which was unaffected by metformin treatment. This is a predefined substudy to determine an effect of metformin on diastolic function. For this substudy trans-thoracic echocardiography was performed during hospitalization and after 4 months. Diastolic dysfunction was defined as having the combination of a functional alteration (i.e. decreased tissue velocity: mean of septal e’ and lateral e’) and a structural alteration (i.e. increased left atrial volume index (LAVI)). In addition, left ventricular mass index and transmitral flow velocity (E) to mean e' ratio (E/e’) were measured to determine an effect of metformin on individual echocardiographic markers of diastolic function.

Results

In 237 (63%) patients included in the GIPS-III trial diastolic function was measured during hospitalization as well as at 4 months. Diastolic dysfunction was present in 11 (9%) of patients on metformin and 11 (9%) patients on placebo treatment (P = 0.98) during hospitalization. After 4 months 22 (19%) of patients with metformin and 18 (15%) patients with placebo (P = 0.47) had diastolic dysfunction. In addition, metformin did not improve any of the individual echocardiographic markers of diastolic function.

Conclusions

In contrast to experimental and observational data, our randomized placebo controlled trial did not suggest a beneficial effect of short-term metformin treatment on diastolic function in STEMI patients.

Adverse diastolic remodeling after reperfused ST-elevation myocardial infarction: An important prognostic indicator

OBJECTIVES

We sought to determine the relationship of adverse diastolic remodeling (ie, worsening diastolic or persistent restrictive filling) with infarct scar characteristics, and to evaluate its prognostic value after ST-segment elevation myocardial infarction (STEMI).

BACKGROUND

Severe diastolic dysfunction (restrictive filling) has known prognostic value post STEMI. However, ongoing left ventricular (LV) remodeling post STEMI may alter diastolic function even if less severe.

METHODS AND RESULTS

There were 218 prospectively recruited STEMI patients with serial echocardiograms (transthoracic echocardiography) and cardiac magnetic resonance imaging (CMR) performed, at a median of 4 days (early) and 55 days (follow-up). LV ejection fraction and infarct characteristics were assessed by CMR, and comprehensive diastolic function assessment including a diastolic grade was evaluated on transthoracic echocardiography. 'Adverse diastolic remodeling' occurred if diastolic function grade either worsened (≥1 grade) between early and follow-up imaging, or remained as persistent restrictive filling at follow-up. Follow-up infarct scar size (IS) predicted adverse diastolic remodeling (area under the curve 0.86) and persistent restrictive filling (area under the curve 0.89). The primary endpoint of major adverse cardiovascular events (MACE) occurred in 48 patients during follow-up (mean, 710±79 days). Kaplan-Meier analysis showed that adverse diastolic remodeling (n=50) and persistent restrictive filling alone (n=33) were significant predictors of MACE (both P<.001). Multivariate Cox analysis, when adjusted for TIMI risk score and CMR IS, microvascular obstruction, and LV ejection fraction, showed adverse diastolic remodeling (HR 3.79, P<.001) was an independent predictor of MACE, as was persistent restrictive filling alone (HR 2.61, P=.019).

CONCLUSIONS

Larger IS is associated with adverse diastolic remodeling. Following STEMI, adverse diastolic remodeling is a powerful prognostic marker, and identifies a larger group of 'at-risk' patients, than does persistent restrictive filling alone.

The prognostic impact of diastolic dysfunction in patients with chronic heart failure and post-acute myocardial infarction: Can age-stratified E/A ratio alone predict survival?

OBJECTIVE

To determine the prognostic implications of diastolic filling grades and identify whether age-stratified E/A ratio alone can identify patients at high risk of death post-AMI and HF. We hypothesized that in response to ageing and pathology, a normal E/A (>1) could be considered abnormal in patients post-AMI older than 65years, and that in patients with symptomatic HF, a normal E/A always represents advanced diastolic dysfunction.

METHODS AND RESULTS

This is a sub-analysis of the Meta-analysis Research Group in Echocardiography (MeRGE) which combined individual patient data from 30 prospective studies and demonstrated that restrictive filling was an important and independent predictor of all-cause mortality. This sub-analysis is restricted to those studies in which continuous E/A data were available (20 studies) and includes a total of 3082 AMI and 2321 HF patients. Patients were classified at the time of echocardiography into four filling patterns: normal, abnormal relaxation, pseudonormal, and restrictive filling. Post-AMI patients were divided into four groups on the basis of age and E/A, while patients with HF were classified into three groups, based on only E/A. Mortality across groups was compared using Kaplan-Meier survival analysis and Cox proportional hazards. In multivariable analyses in the AMI patients, age-stratified E/A was an independent predictor of outcome (HR 1.43 (95% CI: 1.31-1.56)), and in the HF cohort, E/A was confirmed as an independent predictor of mortality (HR 1.12 (95% CI 1.09-1.16)) alongside age and ejection fraction.

CONCLUSIONS

Age-stratified E/A is an independent predictor of mortality after AMI and in HF patients, regardless of left ventricular ejection fraction, age and gender. E/A ratio could be a first step echocardiographic risk stratification, which could precede and indicate the need for more advanced diagnostic and prognostic considerations in high-risk AMI and HF patients.

Independent and incremental prognostic value of Doppler-derived left ventricular total isovolumic time in patients with systolic heart failure

BACKGROUND

A prolonged total isovolumic time (T-IVT) has been shown to be associated with worsening survival in patients submitted to coronary artery surgery. However, it is not known whether it has prognostic significance in patients with chronic systolic heart failure (HF).

AIM

To determine the prognostic value of T-IVT in comparison with other clinical, biochemical and echocardiographic variables in patients with chronic systolic HF.

METHODS

Patients (n=107; age 68±12 years, 25% women) with chronic systolic HF, left ventricular ejection fraction (EF)<45%, and sinus rhythm, underwent a complete Doppler echocardiographic study, that included tissue Doppler long axis velocities and total isovolumic time (T-IVT), determined as [60-(total ejection time+total filling time)]. Plasma N-terminal pro-B natriuretic peptide (NT-pro-BNP) was also measured. The associations of dichotomous variables selected according to the Receiver Operator Characteristic analysis were assessed using the Cox proportional hazard model.

RESULTS

Follow-up period was 37±18 months. Multivariate predictors of events were T-IVT≥12.3% s/min, mean E/Em ratio≥10, log NT-pro-BNP levels≥2.47 pg/ml and LV EF≤32.5%. On Kaplan-Meier analysis, patients with prolonged T-IVT, high mean E/Em ratio, increased NT-pro-BNP levels and decreased LV EF had a worse outcome compared with those without. The addition of T-IVT and NT-pro-BNP to conventional clinical and echocardiographic variables significantly improved the chi-square for the prediction of the outcome from 33.1 to 38.0, (P<0.001).

CONCLUSIONS

Prolonged T-IVT added to the prognostic stratification of patients with systolic HF.

Independent relationship of left atrial size and mortality in patients with heart failure: an individual patient meta-analysis of longitudinal data (MeRGE Heart Failure)

AIMS

Left atrial (LA) size is considered a marker of poor prognosis in heart failure (HF) patients. Prior studies have recruited relatively few subjects limiting their power to adequately analyse the interaction between LA size, left ventricular (LV) systolic and diastolic function, and prognosis.

METHOD AND RESULTS

The MeRGE collaboration combines prospective data from 18 studies in HF patients. In this analysis of data from 1157 patients, the primary endpoint was death or hospitalization for worsening HF. In multivariate analysis (Cox proportion hazard model), LA area was associated with prognosis (HR 1.03 per cm(2), 95% CI 1.02, 1.05; P < 0.0001) independently of age, NYHA class, LV ejection fraction, and restrictive filling pattern (RFP). When LA area was used as a categorical variable, the HR associated with larger LA area (above median) was 1.4 (95% CI 1.13, 1.74) and when LA area index was used, the HR was 2.36 (95% CI 1.80, 3.08). When the patients with and without RFP were divided on the basis of either LA area or LA area index, significantly higher event rates were observed in those with larger LA area.

CONCLUSION

Left atrial area is a powerful predictor of outcome among HF patients with predominantly impaired systolic function, and is independent of, and provides additional prognostic information beyond LV systolic and diastolic function.

Prognostic value of brain natriuretic peptide and enhanced ventilatory response to exercise in patients with chronic heart failure

Whether brain natriuretic peptide (BNP), combined with a cardiopulmonary exercise test (CPx) parameters or echocardiography improves prognostic stratification in mild-to-moderate systolic heart failure (HF) is currently unclear. In 156 consecutive stable outpatients with mild to moderate HF and left ventricular ejection fraction (LVEF) <40%, we assessed the impact of BNP assay, Doppler echocardiography and CPx on survival. Median BNP plasma levels were 207 [90-520] pg/mL. Mean LVEF was 33 +/- 7%. Left bundle branch block (LBBB) was present in 52 patients (33%) and a restrictive filling pattern in 35 (22%). The slope of the relation between minute ventilation and carbon dioxide production (VE/VCO(2) slope) averaged 35 +/- 8; an enhanced ventilatory response (EVR) to exercise (VE/VCO(2) slope >35) was found in 67 patients (43%). During 759 +/- 346 days of follow-up, 24 patients died. By multivariate analysis, the strongest independent predictors of all-cause death among clinical, echocardiographic variables and BNP were LBBB and beta-blocker treatment. When CPx variables were added, the best predictors of mortality were LBBB, beta-blockade and VE/VCO(2) slope. This study highlights the value of a sequential approach, based on clinical, laboratory and functional data to identify high-risk HF patients. BNP assay might constitute a simple alternative tool for patients with an inability or with clinical contraindications to exercise, advanced physical deconditioning and unreliable CPx results. However, whenever feasible, CPx with assessment of EVR is recommended for a more accurate prediction of prognosis.

Prognostic value of N-terminal pro-type-B natriuretic peptide and Doppler left ventricular diastolic variables in patients with chronic systolic heart failure stabilized by therapy

Prognostication of patients with chronic heart failure (HF) stabilized by therapy may be difficult. Therefore, the aim was to evaluate whether combined assessment of plasma N-terminal pro-B natriuretic peptide (NT-pro-BNP) and Doppler left ventricular (LV) diastolic variables was relevant to the prognosis of patients with stable HF. Outpatients with LV systolic HF (ejection fraction < or =45%), classified using clinical criteria as decompensated (n = 94) and stable HF (n = 219), underwent a complete Doppler echocardiographic study. NT-pro-BNP was measured together with mitral wave velocities, E wave deceleration time, and tissue Doppler early septal annular velocity. Median follow-up was 22 months. Freedom from all-cause mortality or HF hospitalization at 24 months was worst (44%) in patients with decompensated HF, intermediate (58%) in patients with stable HF with NT-pro-BNP higher than the median (>1,129 pg/ml), and best (92%) in patients with lower NT-pro-BNP (log-rank p <0.0001). In patients with stable HF, NT-pro-BNP >1,129 pg/ml (hazard ratio [HR] 2.84, p = 0.003), E wave deceleration time <150 ms (HR 2.31, p = 0.004), and tissue Doppler early septal annular velocity <8 cm/s (HR 2.18, p = 0.01) were predictors of the end point at multivariate analysis. The addition of Doppler LV diastolic variables and NT-pro-BNP significantly improved the chi-square test for outcome prediction (from 14.4 to 46.4). In conclusion, NT-pro-BNP and spectral and tissue Doppler variables of LV diastolic dysfunction added independent and incremental contributions to prognostic stratification of patients with stable HF.

Independence of restrictive filling pattern and LV ejection fraction with mortality in heart failure: an individual patient meta-analysis

BACKGROUND

The Doppler echocardiographic restrictive mitral filling pattern (RFP) is an important prognostic indicator in patients with heart failure (HF), but the interaction between RFP, left ventricular ejection fraction (LVEF) and filling pattern remains uncertain.

AIMS

To determine whether the RFP is predictive of mortality independently of LVEF in patients with HF.

METHODS

Online databases were searched to identify studies assessing the relationship between prognosis and LV filling pattern in patients with HF. Individual patient data from 18 studies (3540 patients) were extracted and collated at the MeRGE Coordinating Centre (The University of Auckland).

RESULTS

Overall, RFP was associated with higher all-cause mortality than the non-restrictive filling pattern: hazard ratio 2.42 (95% CI 2.06, 2.83). In multivariable analysis the RFP, LVEF, NYHA class and age were independent predictors of mortality. The prevalence of the RFP was inversely related to LVEF but remained a predictor of mortality even in those patients with preserved LVEF.

CONCLUSIONS

The restrictive mitral filling pattern is a powerful predictor of mortality, independent of LVEF and age, in patients with HF. Doppler-derived LV filling patterns are an accessible marker from echocardiography that can readily be incorporated in risk stratification of all patients with HF.

Independent prognostic importance of a restrictive left ventricular filling pattern after myocardial infarction: an individual patient meta-analysis: Meta-Analysis Research Group in Echocardiography acute myocardial infarction

BACKGROUND

Restrictive mitral filling pattern (RFP), the most severe form of diastolic dysfunction, is a predictor of outcome after acute myocardial infarction (AMI). Low power has precluded a definite conclusion on the independent importance of RFP, especially when overall systolic function is preserved. We undertook an individual patient meta-analysis to determine whether RFP is predictive of mortality independently of LV ejection fraction (LVEF), end-systolic volume index, and Killip class in patients after AMI.

METHODS AND RESULTS

Twelve prospective studies (3396 patients) assessing the relationship between prognosis and Doppler echocardiographic LV filling pattern in patients after AMI were included. Individual patient data from each study were extracted and collated into a single database for analysis. RFP was associated with higher all-cause mortality (hazard ratio, 2.67; 95% CI, 2.23 to 3.20; P<0.001) and remained an independent predictor in multivariate analysis with age, gender, and LVEF. The overall prevalence of RFP was 20% but was highest (36%) in the quartile of patients with lowest LVEF (<39%) and lowest (9%) in patients with the highest LVEF (>53%; P<0.0001). RFP remained significant within each quartile of LVEF, and no interaction was found for RFP and LVEF (P=0.42). RFP also predicted mortality in patients with above- and below-median end-systolic volume index (1575 patients) and in different Killip classes (1746 patients). Importantly, when diabetes, current medication, and prior AMI were included in the model, RFP remained an independent predictor of outcome.

CONCLUSIONS

Restrictive filling is an important independent predictor of mortality after AMI regardless of LVEF, end-systolic volume index, and Killip class.