The use of serum total cholesterol measurement was evaluated as a screening tool to predict elevated levels of low-density lipoprotein cholesterol in 2857 children and adolescents, aged 5 to 17 years, examined in 1981 and 1982. Subjects were from the biracial community of Bogalusa, Louisiana. For selected serum total cholesterol values (150 to 210 mg/dL, 3.88 to 5.43 mmol/L), sensitivities were higher for blacks than whites and higher for females than males, whereas the positive predictive values were higher for whites than blacks and higher for males than females. With the age-, race-, and sex-specific 95th percentiles of serum total cholesterol levels as cutoff points, only 44% to 50% of subjects with elevated low-density lipoprotein cholesterol levels (greater than or equal to 95th percentile) were detected, and approximately 50% of those identified had false-positive results. Lowering the serum total cholesterol cutoff point increased the sensitivity, but decreased the specificity and positive predictive value. At the 75th percentiles of serum total cholesterol levels, sensitivities were 92% to 95% for females and 100% for males and specificities were 78% to 79%, but the false-positive results increased to 81% to 84%. The low cost and ease of obtaining serum total cholesterol measurements contribute to its appeal as a screening tool for hyperlipidemia. However, its poor test characteristics make serum total cholesterol measurement inefficient as a screening tool for detecting elevated levels of low-density lipoprotein cholesterol in children and adolescents.