Repeated Echocardiograms Do Not Provide Incremental Prognostic Value to Single Echocardiographic Assessment in Minimally Symptomatic Patients with Chronic Heart Failure: Results of the Bio-SHiFT Study

Abstract

We aimed to compare the prognostic value of a single "baseline" echocardiogram with repeated echocardiography in stable chronic heart failure (CHF) patients. We hypothesized that repeated echocardiograms would contain incremental prognostic information. In the prospective Bio-SHiFT study, we performed 332 echocardiograms in 106 patients during a median follow-up of 2.3years. The endpoint comprised HF hospitalization, left ventricular (LV) assist device implantation, heart transplantation, and cardiovascular death. We compared hazard ratios (HRs; adjusted for N-terminal pro-brain natriuretic peptide) from Cox models for the first available measurement with HRs from joint models, which model individual trajectories based on the repeated measurements and link these to the time-to-event data. The mean age of the patients was 58.1years; 78.3% were male, 12.6% had New York Heart Association class>II, all had reduced ejection fraction, and the most common HF etiologies were cardiomyopathies (51%) and ischemia (40%). The endpoint occurred in 25 patients. Both the single measurements and the temporal trajectories were significantly associated with the endpoint (adjusted HR Cox model [95% CI] vs adjusted HR joint model [95% CI]): LV ejection fraction, 1.47 (0.93-2.31) vs 1.77 (1.13-2.93); diastolic LV diameter, 1.64 (1.09-2.47) vs 1.68 (1.12-2.57); systolic LV diameter, 1.72 (1.10-2.69) vs 1.68 (1.13-2.63); systolic left atrial diameter, 1.88 (1.18-3.00) vs 2.60 (1.48-4.97); E/A ratio, 2.73 (1.42-5.26) vs 3.87 (1.75-10.13); and E/e' ratio, 2.30 (1.38-3.84) vs 2.99 (1.68-6.19). None of the trajectories from the investigated parameters showed worsening prior to events. Although single baseline or repeatedly measured echocardiographic parameters were associated with the endpoint, all parameters remained on average stable during the 2.3years of follow-up in this largely minimally symptomatic CHF cohort. Thus, regular echocardiographic monitoring of systolic or diastolic LV function within this time frame does not carry incremental prognostic information over a single baseline measurement.