The Increase in Hemoglobin Concentration With Altitude Differs Between World Regions and Is Less in Children Than in Adults

INTRODUCTION

The hemoglobin concentrations [Hb] of high-altitude residents increases with altitude.^{1, 2} Approximately 5% of the world's population lives at altitudes above 1500 m,³ and thus is at risk to suffer from anemia because highland residents rely on adequate iron and vitamin intake⁴ to foster erythropoiesis for maintaining this chronically elevated [Hb]. In turn, anemia adversely affects growth and development, and impairs physical performance. Moreover, abnormally high [Hb] is often related to chronic mountain sickness.⁵

Increasing [Hb] is an effective means to maintain a normal arterial blood O₂ content despite decreased oxygen partial pressure (PaO₂) in inspiratory air at high altitude that otherwise compromises tissue oxygen supply by decreasing arterial oxygen saturation (SaO₂).⁶ Mechanisms of adjustment are best observed in lowlanders ascending to high altitude, where the increase in [Hb] is achieved by a decrease in plasma volume⁷ and by stimulation of erythropoiesis via an HIF-2α dependent increase in blood erythropoietin levels⁴ inducing a slow elevation in total Hb-mass.^{7, 8} The increase in [Hb] in high altitude residents is caused by converse changes of plasma volume and total Hb-mass,⁹ too. Interestingly, at comparable total blood volume, the higher [Hb] of Andeans compared with Sherpas seems to be caused by a reduced plasma volume in the Andeans⁹ indicating that besides erythropoiesis also adjustments of plasma volume significantly affect blood oxygen transport.

Because hypoxia increases [Hb] in high-altitude residents it is a challenge to define whether an obtained [Hb] value is abnormal. The World Health Organization (WHO) provides cutoff values defining anemia¹⁰ and polycythemia¹¹ for sea-level populations and values for adjustment accounting for [Hb] increments above the sea level.¹² This adjustment is intended for worldwide use but appears inappropriate because of strong evidence for differences in the magnitude of [Hb] increase among world regions.¹ As an example, Tibetan and Ethiopian altitude-populations have considerably lower mean Hb concentrations than Andean populations.^{1, 13} It is therefore questionable whether fixed correction factors apply worldwide or whether region-specific adjustments are better suited for accurate diagnosis of abnormal [Hb]. Seeking to address this issue, we analyzed datasets from The Demographic and Health Surveys (DHS) Program from low- and middle-income countries where >4% of the population lives at altitudes above 1000 m.

MATERIALS AND METHODS

Groups

Sex

The surveys contained data on females and males. However, DHS-target populations are female and male children under age 5 and women age >15. Data on boys and girls were pooled because there is no difference in [Hb] at this age.^16-18 Analyses shown in the article are mostly among females because of incomplete data on the males (Table 1).

Age

Given the age-related increase in [Hb],¹⁶ the data were subgrouped into children age 6–24 and 25–59 months. Adults were subgrouped into age 15–20 and >20 years. No data were available on females and males age 5–14 years.

Pregnancy

Separate subgroups are reported for pregnant women age 15–20 and age >20 years. We did not distinguish among pregnancy-trimesters as this information was not collected.

Wealth index

Micronutrient status was not recorded in DHS surveys. The DHS wealth index, a composite measure of a household's living standard, categorizes households into 5 wealth quintiles, allowing to see how health differs between the poor and wealthy (dhsprogram.com). Therefore, the wealth index is a surrogate for the socio-economic status, which also includes access to goods, services, and assets.¹⁹ A value of 1 indicates the lowest and 5 the highest wealth status. Undernutrition is correlated with lower wealth status and so we used low wealth status as a surrogate for poor health.

Altitude (meters)

Altitude of residence was obtained from GPS readings at the sampling site. Altitude is categorized into the near sea level (−500 to 500 m), then into ranges of 500 m stepwise up to 5000 m, and into the group >5000 m. Table 1 shows the maximal altitude for each country, where the samples had been obtained.

Hemoglobin concentration [Hb]

Measurements

The DHS Program follows a standard protocol for [Hb] measurement. Trained data collectors obtained a capillary blood sample (finger or heel). [Hb] was measured in a HemoCue 201+ or HemoCue 301 analyzer (Radiometer Group, Ängelholm, Sweden).

Removal of outliers

Because some [Hb] were very low or high for apparently healthy, nonhospitalized individuals, thresholds suggested for a healthy low-altitude population were applied,²⁰ and outliers were not used for further analyses. All [Hb] values at the sea level between 7.0 and 18.0 g/dL for nonpregnant women age >15 were included. The inclusion range for pregnant women age >15 and for children of both sexes age <5 was 6.0–17.0 g/dL; it was 9.0–20.0 g/dL for men age >20.²⁰ Those upper and lower limits were adjusted for altitude with 2 different methods:

1. Upper and lower sea-level limits [Hb] were altitude-corrected based on suggestions by the WHO¹² using the formula

for all [Hb] regardless of world region, sex, age, and pregnancy. Suppl. Figures S1A and B show the application of this method for SSEA and for SAm, respectively. Values included for analysis are shown in black, outliers in orange, indicating that many more samples were removed in SSEA than in SAm. Numbers of outliers identified with this method are shown in Suppl. Table S2.

• 2. Upper and lower limits [Hb] were also altitude-corrected applying regression coefficients (Δ[Hb]/km) in Table 2. Suppl. Figures S2A and B show the application of this method for SSEA and for SAm, respectively, where values used for analysis are shown in black and outliers in orange.

Table 2. Intercepts and Regression Coefficients From Regression Analysis Describing the Change in [Hb] With the Residential Altitude, Sea-level to Maximal Altitude, in Children and in Nonpregnant and Pregnant Women From All World Regions

A. Change in Hb Concentration With Altitude (Slope; ΔHb/Altitude; g/dL/km)
World Region			CAm			SAm			WCAfr			SAfr			EAfr			ME			CWA			SSEA
Age	Sex	Pr	Slope	SE	R2	Slope	SE	R2	Slope	SE	R2	Slope	SE	R2	Slope	SE	R2	Slope	SE	R2	Slope	SE	R2	Slope	SE	R2
6–24 mo	bg	–	0.405	0.019	0.048	0.611	0.012	0.326	0.673	0.030	0.054	0.567	0.050	0.052	0.713	0.015	0.102	−0.251	0.043	0.008	0.279	0.036	0.015	0.443	0.013	0.022
25–59 mo	bg	–	0.534	0.013	0.096	0.642	0.008	0.397	0.716	0.021	0.063	0.348	0.033	0.023	0.703	0.009	0.102	−0.285	0.029	0.009	0.381	0.025	0.032	0.493	0.009	0.030
15–20 y	f	np	0.742	0.015	0.155	0.807	0.009	0.489	0.890	0.042	0.070	0.721	0.042	0.074	1.201	0.014	0.230	0.576	0.406	0.022	0.548	0.038	0.046	0.560	0.009	0.032
>20 y	f	np	0.648	0.008	0.100	0.769	0.004	0.418	0.752	0.023	0.044	0.649	0.022	0.053	1.121	0.008	0.184	0.529	0.047	0.016	0.636	0.017	0.051	0.558	0.004	0.032
15–20 y	f	pr	0.676	0.062	0.119	0.824	0.047	0.432	0.719	0.151	0.028	0.478	0.180	0.035	1.173	0.067	0.145	1.004	0.782	0.043	0.087	0.268	0.001	0.585	0.059	0.025
>20 y	f	pr	0.635	0.035	0.107	0.702	0.021	0.366	0.832	0.061	0.051	0.376	0.097	0.023	1.098	0.026	0.166	0.448	0.130	0.015	0.592	0.069	0.048	0.503	0.018	0.024
B. Intercept of regression line representing the Hb concentration at the sea level (g/dL)
World Region			CAm			SAm			WCAfr			SAfr			EAfr			ME			CWA			SSEA
Age	Sex	Prg		Int	SE		Int	SE		Int	SE		Int	SE		Int	SE		INT	SE		Int	SE		Int	SE
6–24 mo	bg	–		10.64	0.02		10.47	0.03		9.66	0.03		10.06	0.07		9.78	0.02		10.67	0.05		10.69	0.04		10.12	0.01
25–59 mo	bg	–		11.44	0.02		11.49	0.02		10.26	0.02		11.12	0.05		10.63	0.01		11.44	0.03		11.35	0.03		10.71	0.01
15–20 y	F	np		12.84	0.02		12.67	0.02		11.65	0.03		12.32	0.06		11.52	0.02		12.24	0.30		12.24	0.05		11.64	0.01
>20 y	F	np		12.87	0.01		12.74	0.01		11.72	0.02		12.36	0.03		11.57	0.01		12.04	0.04		12.05	0.02		11.66	0.01
15–20 y	F	pr		11.66	0.07		11.24	0.10		10.41	0.12		11.62	0.26		10.04	0.08		11.05	0.58		11.39	0.30		10.76	0.03
>20 y	F	Pr		11.63	0.04		11.60	0.04		10.53	0.05		11.54	0.13		10.35	0.03		11.14	0.11		11.01	0.08		10.80	0.01

• A: The regression coefficients (slope; g/dl/km) ± standard error (SE), and the coefficient of determination, R², which indicates the change in [Hb] with altitude of residency. B: The intercepts (int; g/dL) ± SE, which represent the calculated Hb concentration at the sea level. Both were calculated from all data obtained from sea-level up to the extreme altitudes within each region. Details on the regression statistics are provided in the Supplement. [Hb] values from boys and girls (bg) were combined because there is no sex-difference in [Hb] up to an age of ~15 years.^16-18 Age is shown in months (mo) for the children and in years (y) for the adult, not-pregnant (np) or pregnant (pr) females (f). P values for multiple comparisons of intercepts and slopes are shown is the supplement in Suppl. Table S8 (difference between age/pregnancy subgroups within world regions) and in Suppl. Table S9 (difference between regions within age/pregnancy subgroups).
• CAm = Central America; CWA = Central and Western Asia; EAfr = Eastern Africa; ME = Middle East; SAfr = Southern Africa; SAm = South America; SSEA = South and South-East Asia; WCAfr = West/Central Africa.

RESULTS

Dependency of [Hb] on altitude of residence and world region

[Hb] near sea level

Our data show that the mean [Hb] at altitudes <500 m (Suppl. Table S4) is all within a range of approximately 1.0 to 1.5 g/dL in the world regions within age-matched and pregnancy subgroups in females (Figure 2A) and in males (Figure 2B). ANOVA revealed that within each age group mean sea level [Hb] was significantly higher in CAm and SAm than in the other regions (not shown). The lower [Hb] in adult, not-pregnant females compared to adult males was also clearly evident (Figure 2A,B; Suppl. Table S4). Suppl. Table S4 also shows that the mean sea level [Hb] increased with age in females and males confirming earlier reports.¹⁶ Likewise, the [Hb] was lower in pregnant than in nonpregnant women of the respective age group.²⁵ These differences are also indicated by the intercepts of the regression analysis performed in females (Table 2).

Increase in [Hb] with altitude

Figure 2 indicates that, in most world regions, the mean [Hb] in females (Figure 2A) and males (Figure 2B) increased with the altitude of residence. This is additionally shown in look-up tables (Suppl. Table S4), which also provide indicators of variation for different altitude ranges from <500 to >5000 m in 500 m increments, grouped by world region, sex, age, and pregnancy. In Suppl. Table S7, we show that in children and not-pregnant females from CAm and SAm, a statistically significant difference in mean [Hb] was found between most low- and high-altitude increments. In WCAfr, ME, and CWA changes in mean [Hb] were irregular and sometimes [Hb] was lower at higher than at lower altitudes; patterns differ between age groups (Figure 2; Suppl. Table S4). No statistically significant difference between mean [Hb] values of the different altitude increments was found in children and 15 to 20 years old pregnant females from WCAfr, SSEA, whereas the Hb-differences between low and very high-altitude regions were statistically significant in the not-pregnant females of both age groups.

We also compared [Hb] values from the altitude range between 3500 and 4000 m, where data were only available from SAm, EAfr, SSEA, and CWA. Also, data from all not-pregnant women age >15 were combined. Numbers were too low to compare children and pregnant women. Figure 3 shows that mean [Hb] was significantly higher in SAm than in SSEA and CWA (P < 0.001), but was not different from EAfr (P = 1.0). There was also a statistically significant difference between SSEA and CWA.

Regression analysis, all altitudes

Table 2 shows the results of the linear regression analysis performed to quantify the increase in [Hb] with altitude. We performed two kinds of calculations: For Table 2, regressions were calculated using Hb-values obtained from all altitude ranges, that is, from the sea level to the region-specific highest altitude. Regression coefficients indicate the increase in [Hb] with altitude (ΔHb/km of altitude; g/dL/km), the intercepts indicate modeled [Hb] values for the sea level. Table 2 indicates several important aspects: (1) the ΔHb/km of children (boys and girls combined) and of females was highest in EAfr followed by CAm and SAm and was much lower in most other world regions, and even negative in children from ME. (2) Within each of the world regions, the Δ[Hb]/km was smallest in children age 6 to 24 months and increased with age. The Δ[Hb]/km was 20% to 50% lower in the children than in not-pregnant and pregnant women. It is important to note that, although correlations were statistically highly significant, R² was highest in samples from SAm (>0.32) followed by EAfr (R²: ~0.1 to ~0.33), whereas in the other world regions R² was well below 0.1 and the standard error of Δ[Hb]/km was high indicating a weak dependency of [Hb] on residential altitude. Suppl. Tables S8 and S9 show the respective P values for the difference between regressions within age groups and within regions, respectively.

Regression analysis, altitudes > 1000 m

Because the WHO in the current guidelines for Hb-adjustments for altitude does not suggest adjustment of [Hb] for altitudes below 1000 m,¹² we also calculated regression coefficients only from samples obtained at altitudes >1000 m. Of note, it has been shown recently that the [Hb] increases significantly every 300 m of altitude between 200 and 2000 m above sea-level in Switzerland.²⁶ Correlation analyses are summarized in Suppl. Table S3. Interestingly, in this calculation Δ[Hb]/km is considerably higher in SAm and EAfr compared with analyzing the sea level to maximal altitude (Table 2). In contrast, in SSEA the Δ[Hb]/km was much lower at altitudes >1000 m than with the lower altitudes included, and the correlation lost statistical significance in several subgroups (Suppl. Table S3). In ME, the Δ[Hb]/km became even more negative in the preschool children.

Subanalyses

We next evaluated whether the lack of [Hb] elevation with altitude in certain world regions was associated with the wealth status. Low wealth status is associated with undernutrition, which ultimately might impair erythropoiesis and thus affect [Hb]. Suppl. Figures S3A–H indicate that the increase—or the lack of increase—in [Hb] appeared not related to this parameter. Another subanalysis concerned the relatively small (compared with SAm) increase in [Hb] with altitude in SSEA that might be caused by specific subpopulations because of genetic similarities of Himalayan Sherpas and Tibetan highlanders.^{22, 24} Unexpectedly, the pattern of [Hb]-changes with altitude was similar in residents of the Himalayan regions compared to the rest of SSEA (Suppl. Figure S4).

Prevalence of anemia and polycythemia

The prevalence of anemia and polycythemia was calculated using the sea-level-thresholds and their adjustment for altitude as suggested by the WHO¹² that is set for use in all world regions, and by adjustment for altitude using our world region- and age- and pregnancy-specific Δ[Hb]/km (Table 2). For comparison, the lower and upper limits of the 95% RI calculated from data after removal of abnormally low and high [Hb] values were applied to the complete dataset. Because in some groups the number of samples was small, particularly in the high-altitude ranges, children of both age groups as well as females >15 years were pooled resulting in the groups of children age 6 to 59 months, not-pregnant women, and pregnant women age > 15 years. Figure 5 compares the prevalence of anemia, and Suppl. Table S5 provides absolute numbers and group-sizes.

Prevalence of anemia near sea level <500 m

Based on the WHO-thresholds, in all world regions, anemia-prevalence was higher in the children than in pregnant and nonpregnant females. In all age- and pregnancy groups, the highest prevalence of anemia was found in WCAfr, EAfr, ME, and SSEA.

Prevalence of anemia at altitudes >500 m

Based on the WHO model, there was a pronounced increase in anemia-prevalence with increasing altitude in most world regions (Figure 5, Suppl. Table S5). Prevalence was >80% in children from SAm, WCAfr, ME, and SSEA, and in not-pregnant females from WCAfr and SSEA in the extreme-altitude ranges. Anemia prevalence in pregnant females was lower than in the not-pregnant females in most regions. Using WHO-sea-level-thresholds and altitude adjustment with our linear regressions shows a somewhat increased anemia prevalence at the more moderate altitudes but also a lower prevalence at high-altitude ranges in nearly all world regions. The altitude, where the two curves intersect, is in the range of approximately 2500 m. Using the sea-level 95% CI-lower limit as anemia-threshold and linear altitude adjustment indicates a much lower prevalence, but also a slight increase with increasing residential altitude.

Prevalence of polycythemia at low and high altitudes

The prevalence of polycythemia (Suppl. Table S6) is, in general, much lower than that for anemia. Using literature-values for polycythemia-thresholds from the sea level^{16, 27} and WHO- and regression-derived altitude adjustments results in the prevalence of polycythemia below 1%, even at altitudes >2000 m in most world regions. Only in children and adult, nonpregnant women the prevalence reaches a prevalence up to 5% at altitudes >4000 m in SAm (Suppl. Table S6). In contrast, using the 95% CI and regressions-derived altitude adjustment, the prevalence for polycythemia is somewhat higher (1% to 2.5% at altitudes up to ~2500 m), and increases with increasing altitude reaching values up to ~11% in not-pregnant females living at altitudes >4000 m (Suppl. Table S6).

DISCUSSION

Our analysis of DHS data shows statistically significant correlation between residential altitude and [Hb] values (Δ[Hb]/km), but also pronounced differences in its magnitude between world regions, confirming earlier results¹ and literature comparing specific altitude regions of individual countries.^{13, 28} The highest values for Δ[Hb]/km were found in adult females from the 3 African and the 2 American regions (1.2 to 0.7 g/dL/km), whereas the Δ[Hb]/km was below 0.6 g/dL/km in the ME and the 2 Asian regions (Table 2). These values compare well with previous data from a meta-analysis.¹ We report here for the first time that in all world regions the Δ[Hb]/km was approximately 20%–25% smaller in the children than in the adult females. Importantly, in the children from ME, [Hb] even decreased with increasing altitude. It is of note that, although all correlations were statistically significant, in some age-subgroups from ME, WCAfr, and CWA, mean [Hb] values sometimes increased with altitude but then decreased at higher altitudes or there was no consistent change (Figure 2, Suppl. Table S4). [Hb] of girls and boys up to age of 5 years were similar at comparable altitude of residency, which is consistent with reports on low-altitude [Hb]¹⁶ indicating that in preschool children no sex-specific adjustment is required.

These findings have important clinical implications because applying a single set of values for adjusting [Hb] for altitude for worldwide and age-independent use as currently suggested by the WHO¹² may result in erroneous diagnosis of abnormal [Hb], particularly in children from all world-regions and for adult females in the Asian regions. The situation is complicated by the observation that, when only [Hb] values from altitudes >1000 m were used for calculation, the Δ[Hb]/km decreased considerably in the Asian regions and in WCAfr, and the correlation lost statistical significance in some subgroups. In contrast, the Δ[Hb]/km became even somewhat higher in CAm, SAm, SAfr, and EAfr (Suppl. Table S3) compared to the analysis including low-altitude data shown in Tables 2. Because of the divergent values of Δ[Hb]/km the diagnosis of anemia and polycythemia in high-altitude residents remains challenging.

Most studies have not examined the Δ[Hb]/km across countries. Rather, studies have focused on [Hb] values at high altitudes from different countries for comparison.¹³ Such comparisons are limited by the fact that altitude regions are often not the same. Altitudes studied in Ethiopia were considerably lower than those in South America or in the Tibet. Given the magnitudes of Δ[Hb]/km in the various regions (Table 2), this might result in erroneous conclusions. The DHS dataset contains sufficient data at altitudes >3500 m only from SAm, EAfr, CWA, and SSEA from adult, not-pregnant women age >15 years but not from the other subgroups. Comparison (Figure 3) shows that [Hb] from EAfr was not different from SAm, but that [Hb] from SSEA and CWA was lower than in the other regions.

Genomic diversity might explain the differences in the Δ[Hb]/altitude and in mean [Hb] values at certain altitudes between world regions. Such genomic variations have been found in Tibetans, where polymorphisms in EPAS1 and EGLIN1 genes had been associated with low [Hb] pointing to a role of erythropoietin (EPO) and erythropoiesis.^{24, 29, 30}

A different aspect had been put forward by Stembridge and colleagues,⁹ who showed that a smaller plasma volume at comparable total blood volume in South American Quechuas accounts for the difference in [Hb] to Nepalese Sherpas. On a first look, this result seems to speak against a major role of erythropoiesis in adjusting [Hb] in high altitude residents, unless the prolyl-hydroxylase/HIF-system and/or erythropoietin also control plasma volume. There is in fact some evidence that such interaction might exist: Treatment with EPO has been shown to decrease plasma volume,³¹ probably by affecting proximal tubular reabsorption.³² On the other hand, head-down tilt increased central blood volume and also elevated plasma EPO-levels.³³ Therefore, Tibetans and genetic relatives having a hyporesponsive HIF-system^{29, 34} due to a gain of function mutation of PHD2,³⁵ have a lower or even missing EPO response to hypoxia, as seen by similarly low EPO-levels in high-altitude Sherpas and non-Sherpa,³⁶ which would result in an increased plasma volume compared with Andeans carrying the fully responsive HIF-system and elevated EPO.³⁷ If this link between EPO-levels and plasma volume, which are the main determinants of [Hb], is based on these known genetic variations, and if those variations existed throughout the regions we have analyzed, one might speculate that populations carrying a hyperresponsive HIF-system have migrated from Africa through Asia and on different routes to the Americas,³⁸ where the high responsive HIF-system was preserved. A Denisovan-related introgression³⁹ resulted in a hyporesponsive HIF-system restricted to Asian regions. We cannot define which of these two [Hb] and HIF-systems provides a selective advantage, and secondary adaptations might also have improved survival in the specific world regions. Such likely HIF-independent effects might explain the lower ΔHb/km in children, and the lower [Hb] in pregnant women caused mainly by a disproportionate increase in plasma volume despite increased total erythrocyte mass.⁴⁰ Targeted genetic and functional analysis of the oxygen transport system are required to test the assumption on migration and adaptive advantages.

[Hb] is also affected by nutritional iron- and vitamin-supply essential for erythropoiesis.⁴ This is of particular importance and thus, the DHS focusses on low and middle-income countries, where malnutrition is likely to occur. It is well established that iron supplementation increases [Hb] in iron-deficient children and child-bearing women living at high altitude^{37, 41, 42} and prevents a [Hb]-decrease in non-anemic, SAm school children.⁴³ Our dataset did not contain such information. Therefore, we analyzed the wealth status as a surrogate for general health and well-being, which might include those parameters. Unexpectedly, we did not observe a systematic effect of the wealth index on the magnitude of change in mean [Hb] values with altitude (Suppl. Figure S3). Comparison of our results with other studies on [Hb] in healthy high-altitude residents revealed comparable results on Δ[Hb]/altitude.^{17, 18} In summary, these results indicate that the nutritional status seems not to be a major determinant of [Hb] in healthy high-altitude residents. Severe nutritional deficits, however, will certainly affect erythropoiesis and cause anemia.

The health status is another confounder affecting [Hb]. Chronic inflammation causes hypoferremia and anemia by upregulation of hepcidin.^{44, 45} The health status of the studied children is quite well characterized, but records on adults are mostly based on interviews, and chronic inflammation was not recorded. Individuals with reported health problems, amongst others Malaria, were excluded from analysis, but diseases such as thalassemia, sickle cell disease, hookworm infections, and chronic mountain sickness were not surveyed.

Along with the wealth status, cooking and heating systems might differ. As such, people might be affected by exposure to elevated levels of carbon monoxide (CO) by open fire in closed environments. Also, tobacco smoking increases CO-exposure. CO binds to hemoglobin with high affinity and prevents binding of oxygen.⁶ CO-exposure is thus another confounder affecting [Hb] in an altitude-independent manner,⁴⁶ and CO and altitude effects on [Hb] might be additive because elevated CO stimulates erythropoiesis⁴⁷ and decreases plasma volume.³¹ Both kinds of exposure had been recorded in the DHS-surveys, but were not accounted for in our calculations due to low sample numbers, particularly in the very high-altitude regions.

SUMMARY AND CONCLUSION

Results from this study—complemented by other reports^{1, 13}—clearly demonstrate world region-specific differences in Δ[Hb]/altitude with highest value found in Africans and South Americans, and the smallest in Asians. Here, we show for the first time that in all world regions the Δ[Hb]/altitude is less in children than in adult women. Because of this diversity we generated look-up tables defining a normal range of Hb for each studied world-region and the respective subgroups.

This divergence in Δ[Hb]/altitude indicates profound problems in the diagnosis of anemia and polycythemia in high-altitude residents because until now their diagnosis was based on a single algorithm for altitude adjustment of [Hb] as presented by the WHO for world region-unspecific, and age-independent use.¹² These discrepancies need to be overcome by re-defining altitude-dependent normal [Hb] ranges. The look-up tables and the regression coefficients might be useful tools. In that, the results of the present study contribute to current efforts of the WHO in generating improved tools to define thresholds for the diagnosis of anemia but also for polycythemia in high altitude residents. Furthermore, a more accurate definition of anemia might improve deciding on the necessity of costly iron and vitamin supplementation in different regions of the world.

AUTHOR CONTRIBUTIONS

MUM, MG, RB, and HM made substantial contributions to the conception or design of the work. MUM, MG, RB, SK, SS, and HM gave final approval for the manuscript to be published and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. SK, SS, and HM analyzed, MUM, MG, RB, SK, SS, and HM interpreted data for the article. MUM, MG, RB, SK, SS, and HM critically revised the manuscript for important intellectual content. HM drafted the manuscript and prepared the final draft. HM (corresponding author) had full access to all data and had the final responsibility for submission.

DATA AVAILABILITY

Data are available in the manuscript and supplementary digital content. No additional datasets were created as a result of this study.

DISCLOSURES

MUM is a HemaSphere editor. The authors have no conflicts of interest to disclose.