Obesity is the main intermediate phenotype of primary hypertension (PH), and increased fat mass is directly related to target organ damage (TOD) and metabolic syndrome (MS). The aim of the study was to assess the sensitivity and specificity of body mass index (BMI), percentile-based, definitions of obesity [BMI > 95th percentile (pc)], and overweight (BMI > 85th pc), and BMI thresholds for cardiovascular (cv) complications (BMIcv) described by Katzmarzyk et al. (Pediatrics 114:198-205, 2004) in predicting risk of TOD and MS in 122 adolescents with PH. Our results indicated that the prevalence of left ventricular hypertrophy (LVH) and carotid intima-media thickness (cIMT) above 2 standard deviations (SDS) was the same, irrespective of the criteria used. BMIcv was more sensitive as a marker of LVH than were the cut-off values of the 85th pc and 95th pc of BMI (87.5%, 75%, 62.5%, respectively; P < 0.0001). BMIcv thresholds and cut-off values of the 85th pc of BMI were of the same sensitivity in predicting the presence of MS (95.8% and 95.8%, respectively) and were more sensitive than the cut-off values of the BMI 95th pc (87.5%; P = 0.02). Metabolic abnormalities, including insulin resistance, were more marked in patients with greater BMI, irrespective of cut-off value. However, only when a stratification system using the 85th pc of BMI was used, were the differences significant for a homoeostasis model assessment for insulin resistance (HOMA-IR) and for serum concentrations of high-density lipoprotein (HDL)-cholesterol, triglycerides and adiponectin. We concluded that BMIcv is more sensitive for diagnosing the presence of LVH and that the cut-off value of the 85th pc of BMI is more sensitive for predicting presence of MS in children with PH.