Objectives : Current WHO hemoglobin (Hb) cutoffs (< 110g/L, 6-59mo; < 120g/L, 15-49y non-pregnant women) to define anemia were established in 1967 and validated in 1980s among iron replete USA survey participants. Our objective was to examine the consistency of Hb cutoffs from multiple countries and age-variations in Hb concentrations among apparently healthy people using contemporary data.
Methods : We analyzed altitude and smoking (women only)-adjusted Hb data from 25 countries for preschool children (PSC, 6-59mo, 24 surveys, n= 35,088) and women of reproductive age (WRA, 15-49y, 20 surveys, n=50,846) from the Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project. Healthy was defined as: [ferritin >12µg/L (< 5y), >15µg/L ( >5y)], RBP or retinol > 0.7µmol/L, CRP ≤5mg/L or AGP ≤g/L, and no malaria (if measured)]. We examined Hb levels, cut-offs at specified percentiles (%ile) and country variations with age-adjusted generalized mixed models using country as random intercept.
Results : The final analytic sample was 41,314; the healthy inclusion criteria resulted in 17-88% data loss of the original data depending on country. Age and country-adjusted mean (SE) of Hb was 116.9 (1.44)g/L for PSC, and 129.1 (1.49)g/L for WRA and was significantly (p< 0.001) associated with age in PSC but not in WRA, and varied by country (pheterogeniety < .001, in each target group). Country explained 22% of the total Hb variance for each group. The pooled countries’ Hb at the 5%ile was 94g/L for PSC [range: 79g/L (Pakistan) - 113g/L (USA], and 110g/L for WRA [92g/L (Nigeria) - 121g/L (USA)].
The high heterogeneity in Hb cutoffs among healthy people from different countries suggests that a single Hb cutoff for anemia may not work for every country. This multinational analysis suggests that different approaches may be warranted to establish Hb cutoffs.
Funding Sources : Bill & Melinda Gates Foundation