Association of platelet distribution width with left ventricular remodeling in children with primary hypertension: a body mass index–matched case–control study

Background Platelet distribution width (PDW) is a routinely available marker of platelet size heterogeneity, but its association with cardiac remodeling in pediatric hypertension is unclear. We assessed the associations of PDW with left ventricular remodeling (LVR) and left ventricular mass index (LVMI), and its incremental discriminatory value. Methods This retrospective, single-center, BMI-matched case–control study included 126 treatment-naïve children with primary hypertension, comprising 63 children with LVR and 63 matched controls with normal left ventricular geometry. LVR was defined as any abnormal left ventricular geometry, including concentric remodeling, concentric hypertrophy, or eccentric hypertrophy. Controls were selected using 1:1 optimal matching without replacement, with exact matching on age and sex and minimization of BMI differences. Conditional logistic regression and linear regression were used to assess the associations of PDW with LVR and LVMI, respectively. Receiver operating characteristic curve analysis and the DeLong test were used to evaluate changes in discrimination after adding PDW to the clinical models. Results PDW was higher in children with LVR than in matched controls (12.22 ± 1.44 vs. 11.13 ± 1.18 fL; P < 0.001). Each 1-fL increase in PDW was associated with higher odds of LVR after adjustment for hypertension grade, obstructive sleep apnea syndrome, and visceral adiposity index (OR, 1.523; 95% CI, 1.105–2.100; P = 0.010) and after additional adjustment for erythrocyte sedimentation rate and parental history of hypertension (OR, 1.526; 95% CI, 1.094–2.129; P = 0.013). In the matched sample, higher PDW was associated with higher LVMI (β = 1.77 g/m 2.16 per 1-fL increase; P = 0.002). Adding PDW increased the AUC from 0.703 to 0.791 in the prespecified clinical model (DeLong P = 0.009) and from 0.730 to 0.803 in the extended model (DeLong P = 0.013). Conclusions Higher PDW was associated with LVR after multivariable adjustment and with higher LVMI in the matched sample. PDW also improved discrimination beyond the evaluated clinical models; however, these findings require prospective external validation.