Advanced lipid testing in children: a call for pediatric reference intervals

Abstract Introduction Abnormal serum lipid levels in childhood are associated with dyslipidemia and increased risk of cardiovascular disease in adulthood. High prevalence of childhood obesity in the US has increased concerns about the adverse effects of abnormal lipids in children, prompting recommendations for universal lipid screening in adolescents. While specialized analysis of lipoproteins by nuclear magnetic resonance (NMR) has been predominately used in adults with risk factors for CVD, our laboratory also receives pediatric samples for NMR testing, most commonly associated with diagnostic codes for hyperlipidemia and/or abnormal weight gain. Currently adult-based reference ranges for the NMR lipoprotein measurements are used in children. The objective of this pilot study was to compare lipoprotein profiles of healthy adults and children and to assess the impact of reference intervals for children on clinical interpretation of pediatric results. Methods Serum samples were collected from 60 healthy children, ages 1 – 17 years, and 148 healthy adults, ages 19 – 64 years old. NMR lipoprotein analysis was performed using the Axinon lipoFIT by NMR assay on a 600 MHz NMR (Bruker). Statistical analysis was conducted in Excel and GraphPad Prism. Reference intervals (RIs) were determined in EP Evaluator as the central 95% interval. Results The levels of atherogenic low-density lipoprotein particle number (LDL-P) and small LDL-P (sLDL-P) were significantly lower in children compared to adults: 783±222 vs 1212±412 nmol/L and 302±142 vs 498±306 nmol/L, respectively (p<0.00001). Additionally, LDL size (LDL-s) was significantly larger in children than adults (21.23±0.29 vs 21.06±0.44 nm, p=0.006), reflecting prevalence of less dense LDL-P in the former group. Pediatric samples also had lower number of anti-atherogenic high-density lipoprotein particles (HDL-P) compared to adults: 29.4±4.6 vs 33.5±4.6 µmol/L, p<0.00001. The significant differences in the number and size of lipoprotein particles between adults and children implies that currently used adult reference intervals might not be appropriate for evaluating pediatric NMR samples. Therefore, we estimated RIs for our heathy pediatric and adult cohorts and applied these RIs to our retrospective data of 618 samples from children 1-17 years old, which were submitted for NMR lipoprotein testing in recent years. The analysis showed that using the adult RIs, 37% and 27% of pediatric samples would not flag for abnormally elevated LDL-P and sLDL-P, respectively; but they would be flagged based on the RI derived from our pediatric cohort. Moreover, 12% pediatric samples flagged for low HDL-P would have normal HDL-P levels if pediatric RIs are used instead of adult RIs. Conclusions Use of adult reference intervals in children may result in misclassification of lipid status and underestimation or overestimation of dyslipidemia and CVD risk in children. This pilot study highlights the need for pediatric reference intervals to guide interpretation of advanced lipid testing in children.