WHO hemoglobin thresholds for altitude increase the prevalence of anemia among Ethiopian highlanders

Gonzales et al., in “Correcting the cut-off point of hemoglobin at high altitude favors misclassification of anemia, erythrocytosis and excessive erythrocytosis”, report on the consequences of applying the current WHO hemoglobin concentration cut-off values for anemia, as adjusted by age, sex, and altitude of residence, to high-altitude Peruvian samples.1 Because the World Health Organization (WHO) has begun a 4-year project to review its global guidelines for hemoglobin thresholds used to define anemia at the individual and population level,2 we write to extend the findings of Gonzales et al.1 to East African highlanders. We show that the extent of misclassification depends on the degree of the hematological response to high-altitude hypoxia of each population. Two closely related ethnic groups reside at high altitude in Ethiopia.3 Our samples of Amhara residing at 3700 m and Oromo at 4000 m have been described in detail4. Here we report on a subsample of those data including only those who were iron replete as determined by measures of body iron stores above zero.5 Aural temperatures were below 38 °C (100.4 °F) apart from 5.1% of Amhara men. Hemoglobin analysis by high-pressure liquid chromatography found no evidence of hemoglobinopathies or beta-thalassemia; all Hb A2 and F concentrations were within the reference range. Both samples described here are healthy both by self-report and lack of physician-detected illness. Amhara highlanders have a dampened hemoglobin response to altitude compared with Oromo.6 Amhara and Oromo men have average hemoglobin concentrations of 16.7 ± 1.1 (n = 99) and 18.7 ± 1.7 gm/dL (n = 66), respectively (t = −8.7, P < .001). Similarly, Amhara and Oromo women average 15.5 ± 1.2 (n = 33) and 16.8 ± 1.6 gm/dL (n = 35), respectively at those altitudes (t = −3.8, P < .001). Figure 1 shows the distribution of hemoglobin concentration for these subsamples. The distribution of hemoglobin concentration among Amhara men at 3700 m (top left panel), Oromo men at 4000 m (bottom left panel), Amhara women (top right panel), and Oromo women (bottom right panel). Solid vertical lines indicate the WHO age-sex adjusted thresholds for anemia; the dashed vertical lines indicate the WHO age-sex-altitude adjusted thresholds Applying the current WHO hemoglobin threshold values, adjusted only for age and sex, no one was classified as anemic. In contrast, applying the current WHO hemoglobin threshold values adjusted for age, sex, and residence altitude, the apparent prevalence of anemia rose dramatically to 28.3% for Amhara men (28 men < 16.2 g/dL) and 9.1% among Oromo men (6 men < 16.5 g/dL). Similarly, the apparent prevalence of anemia increased steeply to 48.5% among Amhara women (16 women < 15.2 g/dL) and 14.3% among Oromo women (5 women < 15.5 g/dL). In the absence of acute and chronic inflammation, hemoglobinopathies and other inherited red blood cell disorders, the most likely cause of anemia in these healthy individuals would be iron deficiency. These East African highlanders are iron replete, suggesting that many or most have been misclassified as anemic using current adjusted WHO hemoglobin threshold values. Furthermore, the extent of apparent misclassification was larger among the Amhara with the lower average hemoglobin concentration, indicating the need for different altitude adjustments in the 2 ethnic groups. The WHO suggests modified thresholds for anemia for several ethnicities at low altitude. Taken together, the data presented here and by Gonzales et al. suggest that it also would be appropriate for the WHO to use modified thresholds for anemia for several ethnicities among high altitude populations. Several factors may contribute to the inapplicability of the current WHO-suggested high-altitude threshold values. The WHO states that the high-altitude hemoglobin cut-offs were adapted from Hurtado et al.,7 although it does not specify the methods used. Hurtado's review of the literature includes 21 publications describing just 589 Peruvians residing across the broad altitude range of 1520-5340 m. His paper added 52 men. Perhaps insufficient evidence underpinning the altitude adjustments explains the extensive misclassification of anemia resulting from their application. Hurtado was unaware of ethnic population differences in hematological responses to high-altitude hypoxia such as the dampened responses of Amhara and Tibetans compared with the responses of Oromo and Andean highlanders.8 In summary, evidence from highlanders in 3 populations on 2 continents demonstrate the over-estimation of anemia based on the application of current WHO age–sex-altitude adjusted hemoglobin concentration thresholds. To avoid misdiagnosis and mistreatment and to responsibly use medical and public health resources, we encourage the WHO to consider these and related data in its review of global guidelines for hemoglobin thresholds used to define anemia at the individual and population level. We are grateful to the study participants and communities for their hospitality and cooperation with our field work. Medical personnel from the Addis Ababa University Faculty of Medicine helped to collect data. We also thank the officials of the Ethiopian Science and Technology Commission, the Amhara Regional Government, and the staff of the Semien Mountains National Park who gave permissions and facilitated local arrangements.CMB received funding by the NSF Award NSF0452326. The authors declare that they have no conflicts of interest with the contents of this article.