Impaired global longitudinal strain predicted cardiovascular outcomes in incident peritoneal dialysis patients: a prospective study
K Kam, J Ng, V Kwong, C C Szeto, P W LeeAbstract
Introduction
Volume overload is common among peritoneal dialysis (PD) patients and is associated with cardiac dysfunction. Left ventricular global longitudinal strain (LV-GLS) is a more sensitive echocardiographic marker than ejection fraction (EF) in detecting early LV systolic dysfunction. The prognostic value of GLS in PD patients remains uncertain.
Purpose
To evaluate the relationship between volume overload and echocardiographic parameters, and the association of GLS with clinical outcomes in incident PD patients.
Methods
This was a prospective observational study which enrolled incident PD patients from a single dialysis center from December 2018 to March 2020. GLS was determined by two-dimensional speckle-tracking echocardiography. Threshold for defining impaired GLS was derived from the 95% percentile of age and sex matched control. Volume status was determined by multifrequency bioimpedance spectroscopy and expressed as relative hydration index (RHI). Primary outcome was a composite of composite of death from cardiovascular (CV) causes, nonfatal myocardial infarction, nonfatal stroke, or hospitalization for heart failure. Secondary outcome was all-cause mortality.
Results
We studied 95 incident PD patients (53.7% male, mean age 59.3 years). Mean GLS was -19.8 ± 3.0%. A total of 35 (36.8%) and 91 (95.8%) PD patients had impaired GLS (> -19.0%) and preserved EF (≥ 50%), while 65 (68.4%) had volume overload (RHI > 15%). RHI was significantly correlated with left ventricular mass index (r= 0.30, P= 0.003) and left atrial volume index (r= 0.37, P <0.001), but not GLS. After a median follow-up of 49.0 (IQR 25.9-60.8) months, primary outcome occurred in 25 (71.4%) patients with compromised GLS and 25 (41.6%) patients with normal GLS (log-rank test, P=0.003). Patients with Impaired GLS (> -19%) had approximately two-fold increase in risk to develop primary outcome (adjusted hazard ratio [HR]=1.96, P =0.021), after adjustment of demographics, comorbidities, EF and volume status. Each 1% increase in GLS (indicating worse LV systolic function) was independently associated with 13% increase in risk of primary outcome (adjusted HR= 1.13, P= 0.030). At 48 months, patient survival was 70.6% and 58.2% in patients with normal and impaired GLS, respectively (log-rank test =0.041). After adjustment of demographics, comorbidities, albumin and EF, GLS significantly predicted all-cause mortality (adjusted HR= 1.19, P =0.002). Sensitivity analysis by considering kidney transplant and death as competing events showed that GLS remained an important predictor for composite of CV outcomes (subdistribution HR= 1.13, P= 0.006).
Conclusions
Subclinical LV systolic dysfunction was present in considerable number of PD patients despite preserved EF. Hypervolemia was closely related to LV hypertrophy and diastolic dysfunction but not GLS. Nonetheless, patients with compromised GLS were associated with significant increase in risk of CV events and mortality.