Diagnostic Ramifications of Repeated Plasma Cholesterol and Triglyceride Measurements in Children: Regression Toward the Mean in a Pediatric Population

[摘要] To assess the diagnostic ramifications of the phenomenon of regression toward the mean for plasma cholesterol and triglyceride in a pediatric population group, two sequential plasma cholesterol and triglyceride measurements were made six weeks or more apart (visits 1 and 2), in children from the Cincinnati Lipid Research Clinic, Princeton school study group. In 326 children with plasma cholesterol ≥205 mg/dl (the 95th percentile) at visit 1, mean plasma cholesterol fell 10.5 mg/dl at visit 2 ( P <0.1). Only 58%, 188 of these 326 children with cholesterol ≥205 mg/dl at visit 1, had levels ≥205 at visit 2. In 180 children with plasma triglyceride ≥125 mg/dl at visit 1 (the 95th percentile), only 31% had levels ≥125 at visit 2. These marked shifts toward less extreme values for cholesterol and triglyceride at a second sampling of children initially at the extremes of the distribution reflect, in part, the phenomenon of regression toward the mean, and argue for at least a second and preferably a third resampling of all children initially above some "cutpoint" for cholesterol and triglyceride. Failure to repetitively retest before making diagnostic conclusions or contemplating therapeutic interventions could result in misdiagnosis, unnecessary treatment programs, and overestimation of treatment efficacy.

[学科分类] 儿科学