Pediatric ionized calcium reference intervals from archived radiometer data.

Background: Measurement of ionized calcium (iCa) reflects bioavailable calcium and has significant utility in children. However, robust pediatric iCa reference intervals (RI) have not been well-established. In this study, we retrospectively calculated RI for iCa in a pediatric population by accessing archived data acquired on Radiometer instruments and applying stringent exclusion criteria.

Methods: Data saved on 4 Radiometer ABL800 FLEX blood gas analyzers were queried. Exclusion criteria were applied based on information available from these instruments. iCa results were plotted and inflection points were visually identified. Following outlier removal and partition verification, age-specific RI were calculated using a nonparametric rank order approach from > 5,000 individuals. Finally, the stringency of the exclusion criteria was assessed by calculating RI from additional results in the dataset and comparing to existing in-house ranges.

Results: Six age-specific iCa partitions were established from 0 to 19 years. Relative to adults, wider ranges for the central 95th percentile were observed early in life that progressively narrowed with increasing age and approached adult concentrations by 2.5 years. Analysis of concurrent data for sodium and creatinine in the dataset suggest the applied exclusion criteria reduced the likelihood of including results from acutely-ill children.

Conclusions: Normal concentrations of iCa in children are more variable than adults. Observed differences may reflect the transition from maternally supplied calcium to nutritional sources, the maturation of calcium homeostatic mechanisms, and/or the need for calcium for growth/development. This study also demonstrates the feasibility and advantages of using data archived on Radiometer analyzers to establish pediatric RI. This approach enables rapid, cost-effective evaluation of large datasets and may be a feasible option when prospective or detailed retrospective analyses are not possible.