HAEMOGLOBIN reference curves may improve paediatric laboratory interpretation by providing continuous age specific estimates and identifying differences between traditional reference intervals, health based optimal curves, and current World Health Organization (WHO) thresholds.
Clinicians typically rely on reference intervals to interpret laboratory test results. However, in paediatric populations, estimating these intervals often requires dividing data by age, sex, and other relevant characteristics, reducing sample sizes within each group and limiting the precision of estimates. Investigators therefore developed sex specific haemoglobin reference curves that modelled haemoglobin as a continuous function of age, removing the need for age partitioning.
The cross-sectional study included healthy Canadian children aged 2 weeks–10 years who attended scheduled primary care visits in Toronto, Ontario, between 3 June 2008 and 26 February 2020. Blood samples were analysed for haemoglobin, ferritin, and C reactive protein concentrations. Parents also completed questionnaires to collect variables used as optimality criteria. Data were analysed between 16 October 2024 and 1 February 2026.
Haemoglobin Reference Curves Compared with Optimal Curves
Reference curves were developed using blood samples from 4,597 children, including 2,451 males, with a median age of 38 months. Optimal curves were estimated from a subgroup of 3,426 children, including 1,798 males, with a median age of 45 months, after excluding children with indicators of suboptimal iron status.
Among females, the lower optimal curve limits were slightly below the lower reference curve limits until age 2 years before becoming higher after age 6 years. At 6 months, the lower optimal limit was: 9.91 g/dL (90% CI:9.70–10.13) compared with 10.00 g/dL (90% CI:9.78–10.23) for the reference curve. Among males, lower optimal limits remained higher than lower reference limits until approximately 20 months of age before becoming similar thereafter. At 6 months, the lower optimal limit was: 9.74 g/dL (90% CI:9.46–10.02) compared with 9.28 g/dL (90% CI:8.94–9.63). Differences between upper limits were minimal for both sexes.
Haemoglobin Reference Curves Compared with WHO Thresholds
The investigators also examined the findings alongside current WHO haemoglobin thresholds. Across all ages, WHO thresholds remained higher than the lower limits of the optimal curves. However, they exceeded the fifth percentile curve only among children aged 5–10 years in both sexes. For example, among males aged 1 year, the lower optimal limit was: 10.06 g/dL (90% CI:9.92–10.21) compared with the WHO threshold of 10.5 g/dL.
The findings suggest that haemoglobin reference curves based on continuous age modelling may overcome limitations associated with age partitioning while providing clinicians with additional context through health based optimal curves. The investigators concluded that these findings also highlight potential misalignment between current WHO haemoglobin thresholds and age specific haemoglobin distributions, particularly during younger childhood.
Reference
Bijelić V et al. Pediatric Reference and optimal curves for hemoglobin. JAMA Netw Open. 2026;9;(6):e2620863.
- Author:
