OBJECTIVE To determine the characteristics of obese children presenting at a tertiary care hospital and the frequency of metabolic syndrome (MS) in them using two paediatric definitions. STUDY DESIGN Cross-sectional study. PLACE AND DURATION OF STUDY The Endocrine Clinic of National Institute of Child Health, Karachi, from November 2005 till May 2008. METHODOLOGY A total of 262 obese children aged 4-16 years, with BMI greater than 95th percentile were included. Children having obesity due to syndromes, medications causing weight gain, chronic illness and developmental disability were excluded. Blood pressure, waist circumference, fasting triglycerides, HDL, insulin and glucose levels were obtained. Obesity was defined as BMI > 95th percentile for age and gender according to the UK growth reference charts. The prevalence of metabolic syndrome was estimated using to the De Ferrantis and Lambert definitions. RESULTS The frequency of MS varied between 16% and 52% depending on whether insulin levels were included in the definition. There was a significant positive correlation(r) when the metabolic parameters were correlated with waist circumference and insulin levels, except HDL which was negatively correlated. All the metabolic parameters like waist circumference, triglycerides, high density lipoprotein cholesterol and systolic blood pressure increased considerably across the insulin quartile (p < 0.05). The most noteworthy anthropometric and metabolic abnormality were the waist circumference (46.5%) and insulin levels (58%) respectively. CONCLUSION There was a marked difference in the frequency of metabolic syndrome according to the definition used. The waist circumference and hyperinsulinemia are significant correlates of MS in obese children. There is a need for establishing normal insulin ranges according to age, gender and pubertal status. The clinical examination and investigations ought to include waist circumference and insulin levels together as a part of the definition of MS, for early detection and intervention of childhood obesity.