Changes in the cardiovascular risk profile in children approaching kidney replacement therapy

Abstract

Background: Despite significant cardiovascular (CV) morbidity in children on dialysis and after kidney transplantation, data on the evolution of CV damage in children with chronic kidney disease (CKD) approaching kidney replacement therapy (KRT) is unknown.

Methods: The burden, progression, and predictors of CV damage before KRT onset were explored in two prospective multicenter cohorts from Europe and Canada: Cardiovascular Comorbidity in Children with CKD (4C) and Haemodiafiltration, Heart and Height (3H) studies, conducted from 2009-19 and 2013-16, respectively. CV damage and risk factors were evaluated (i) cross sectionally at KRT-start (n = 248), and (ii) longitudinally over the 2-years preceding KRT start (n = 157; 331 patient-visits). Longitudinal analyses with mixed-effects models estimated associations of modifiable CV risk factors with change in carotid intima-media thickness (cIMT) standard deviation score (SDS), pulse wave velocity (PWV-SDS), left ventricular (LV) mass and systolic dysfunction.

Findings: 248 patients, age 14.3 (12.2, 16.2) years were evaluated at median 35 (28-114) days before KRT start. Elevated cIMT-SDS and PWV-SDS were present in 43% and 25%, and LV hypertrophy and systolic dysfunction in 49% and 33%. Aortic stiffness and LV hypertrophy significantly increased, especially in the year before KRT start (adjusted odds ratio, OR 0.33, P = 0.002 and OR 0.54, P = 0.01, respectively). 79% of children had >3 modifiable CV risk factors at KRT onset. Diastolic BP and BMI were strongly associated with a linear increase in all CV measures. After controlling for CV risk factors, the time to KRT onset no longer predicted the burden of CV damage.

Interpretation: This comprehensive CV evaluation shows the progressive accrual of modifiable risk factors and a high burden of CV damage in the years preceding KRT onset. CV damage in the pre-KRT period is preventable.

Funding: Supported by EU4Health Programme (101085068) and Kidney Research UK (RP39/2013).

Keywords: Carotid intima-media thickness (cIMT); Kidney replacement therapy (KRT); Left ventricular mass index (LVMI); Mid-wall fractional shortening; Pulse wave velocity (PWV).

Fig. 1

Flow of study participants included from the prospective 4C and 3H study cohorts Parameters at initiation of kidney replacement therapy (KRT) were analysed cross-sectionally in 248 patients (T₀). Of these, 157 patients with at least one visit before KRT initiation were included for longitudinal analyses. Data at visits 1-year (T_-1) and 2-year (T_-2) prior to KRT start were available in 110 and 66 patients respectively. CKD chronic kidney disease, eGFR estimated glomerular filtration rate.

Fig. 2

Scatter plots showing (A) carotid intima-media thickness standard deviation score (SDS), (B) pulse wave velocity SDS, (C) left ventricular mass index, and (D) mid-wall fractional shortening in the 2-years prior to kidney replacement therapy (KRT) onset. Lines show predicted mean trajectory, grey areas are 95% confidence intervals. Slopes (± standard errors) represent one unit (SDS, g/m^2.16 or %) change per year and are regression coefficients obtained in a linear mixed-effect model with patient-level random intercepts, controlling for patients' visit age, gender and country.

Fig. 3

Prevalence of cardiovascular damage in the form of elevated carotid intima-media thickness (>2 standard deviation score (SDS)), elevated pulse wave velocity (>2 SDS), left ventricular hypertrophy (left ventricular mass index >45 g/m^2.16) and systolic dysfunction (mid-wall fractional shortening <15.7%) at onset, 1-yr and 2-yr before onset of kidney replacement therapy (KRT). P values represent change in prevalence per year, obtained from a logistic mixed-effect model with patient-level random intercepts, controlling for patients' visit age, gender, and country. Prevalence of elevated elevated carotid intima-media thickness SDS was higher than elevated pulse wave velocity SDS (Chi squared P < 0.001, at all time points), prevalence of left ventricular hypertrophy was higher than systolic dysfunction (Chi squared P < 0.001, P = 0.002 and P = 0.01, at onset, 1-yr and 2-yr before KRT start, respectively).