Ethical approval was granted by the ethics committee of the National University of Colombia at Bogotá (reference: ID 06/2015). The study conformed to the standards set by the Declaration of Helsinki. Written informed consent was obtained from all of the children and their parents. The subjects volunteered to participate in the study and were free to withdraw at any time without any need to provide a reason.

In total, 475 healthy children and adolescents (girls n = 217, boys n = 258; untrained group n = 171, endurance trained group n = 304) living at two different altitudes (∼1,000 m, n = 204, ∼2,600 m, n = 271) and aged from 9 to 18 years old participated in the study. Data on Hbmass and BV in relation to serum testosterone and erythropoietin concentrations were already determined from 313 participants in a recent publication (Mancera-Soto et al., 2021). Hematological data were recorded for all 475 children, and maximum oxygen uptake (VO₂max) was determined in 404 participants. For anthropometrical data and the exact distribution of the test subjects to the individual subgroups, see Tables 1, 2.

All of the trained children and adolescents practiced endurance sports, i.e., medium- and long-distance running, cycling, speed skating, race walking and triathlon. The precondition for participation in the trained group was a training history of at least 1 year, a training volume of at least 6 h per week and a training frequency of at least 3 times per week (for detailed information, see Table 2). The participants in the untrained group did not practice any other sport apart from school sports. The subjects were recruited in the Colombian regions around Bogotá (Central East Andean region, ∼2,600 m above sea level) and Tuluá/Cali (West Andean region, ∼1,000 m). Ethnicity of the participants corresponded to the regional distribution, which differed only slightly between both regions [Central East/West Andes: African origin 5.1%/14.0%, European origin 58.9%/55.4%, Native American origin: 36.0%/30.5% (Ossa et al., 2016)]. With the exception of 17 participants from the Central-East group who had lived in Bogotá for at least 5 years, they had lived their entire lives at their respective altitudes and had not changed their altitude for more than a week over the past year.

None of the female participants indicated the use of hormonal contraceptives, and none used iron, folic acid or any supplements that could affect the parameters determined in this study.

This cross sectional study with an independent groups design was conducted in Bogotá and in Tuluá on two consecutive days. On the first day, a medical examination was performed, followed by an anthropometric evaluation. Hemoglobin mass was determined using the CO-rebreathing method according to Schmidt and Prommer, (2005) and Prommer and Schmidt, (2007) and a cubital venous blood sample; i.e., 4 ml of heparinized blood were obtained for the determination of basic hematological parameters. On the second day, an incremental step test on a cycle ergometer or treadmill was performed to assess VO₂max. The cyclists, speed skaters, race walkers and triathletes, as well as the untrained children, were tested on a cycle ergometer, and the runners were tested on a treadmill.

Before the study, the children and their parents were advised not to practice exhaustive sports before both test days, to sleep at least 8 h, and to eat carbohydrate-rich meals. Health status was checked by medical examination, including electrocardiography, under resting conditions. The stage of biological maturation was evaluated using the method of coevaluation by a physician and self-report, according to Tanner, (1962), and the children were classified in stages I to V according to their external primary and secondary sexual characteristics.

For the anthropometrical evaluation, skinfold caliper measurements were performed by an identical scientist using the triceps, subscapular, supra-iliac, abdominal, thigh and medial calf (Stewart and Marfell-Jones, 2011). Percentage body fat and absolute lean body mass were estimated using a correction for the age and gender of the participants (Slaughter et al., 1988).

The whole procedure, including blood sampling, sample transport, sample storage and sample analyses, was performed under standardized conditions oriented to current WADA guidelines (WADA, 2019). Four milliliters of blood were drawn from a cubital vein using EDTA vacutainers after the subject was left in a sitting position for at least 15 min. The samples were stored in a +4°C refrigerator until they were transported within 24 h under controlled, cool conditions to the WADA-related Institute for Doping Analysis in Bogotá, Colombia.

The hematological measurements were obtained at the Institute of Doping Analysis in Bogotá. A Sysmex XT 2000i hematological analyzer (Sysmex, Norderstedt, Germany) was used for the determination of hemoglobin concentrations ([Hb]) and hematocrit (Hct).

Hemoglobin mass (Hbmass) was determined using the CO-rebreathing method as described by Schmidt and Prommer, (2005) and modified by Prommer and Schmidt, (2007). Briefly, a bolus of 99.97% carbon monoxide (boys and girls < 14 years old—untrained 0.7 ml/kg body mass, trained 0.8 ml/kg; boys > 14 years old—untrained 1.0 ml/kg, trained 1.2 ml/kg; girls > 14 years old—untrained 0.8 ml/kg, trained 1.0 ml/kg) was administered to subjects and rebreathed along with 2–3 L of 100% O₂ for 2 min. The COHb concentration was measured in sextuplicate before and 7 min after starting the CO-rebreathing in capillarized blood from an earlobe using the CO-oximeter OSM3 (Radiometer, Copenhagen, Denmark). The total Hbmass was then calculated from the difference in the COHb concentration before and after the rebreathing maneuver. A more detailed description was previously provided by Mancera-Soto et al., (2021). During the whole study, identical equipment was used by the same staff. The typical error (TE) of this method obtained in our laboratory was 2.2% and was in accordance with the TE published by Gore et al., (2005). BV, red cell volume (RCV) and plasma volume (PV) were calculated using Hb mass, venous [Hb] and venous hematocrit values (obtained from the Sysmex XT 2000i system) as described by Schmidt and Prommer, (2005).

VO₂max was determined by an incremental step test until subjective exhaustion. Except for the runners, all of the trained and untrained children were tested on the identical cycle ergometer (Monarc Ergomedic 839, Monark Exercise AB, Vansbro, Sweden). The reason why the untrained children were tested on the bicycle ergometer was that they could be graded much more precisely than on the treadmill. Among the trained children, cyclists, triathletes and speed skaters were sufficiently familiar with bicycles to reach their discipline-specific maxima, while runners were tested on a treadmill (HP Cosmos Quasar, h/p/cosmos sports and medical GmbH, Nussdorf—Traunstein, Germany) to obtain their discipline-specific maximum values. Since in untrained subjects and in moderately trained endurance athletes, the VO₂max on the treadmill is 5–10% higher than on the bicycle ergometer because of the greater muscle mass involved (Millet et al., 2009), the values of the runners were corrected by 7% (Hermansen and Saltin, 1969). Before starting the testing, all of the children were familiarized with the ergometer or treadmill. In all of the tests, a 3-min warm-up was provided with the initial setting, and the load was increased differently in the individual groups according to the scheme presented in the supplemental material. The test was ended when the participant could not maintain the number of revolutions (cycle ergometer) or speed (treadmill). Ninety-seven percent reached at least two of the following criteria: 1) occurrence of a plateau of VO₂ despite an increase in exercise intensity; 2) heart rate greater than 180 beats/min; and 3) respiratory exchange ratio (RER) > 1.1. Measurement of VO₂max was always performed using the identical portable spirometry system (COSMED Model Quark CPET, COSMED, Rome, Italy). For VO₂max, values were averaged over a period of 30 s before exhaustion.

For statistical analysis, IBM SPSS software, version 28, was used. Data refer to the stage of biological maturation and are presented as the means and standard deviations for the respective stage of puberty. To increase the statistical power, we pooled the data on Tanner stages I (pre puberty) and II (early puberty) and named them pre/early puberty, as well as stages IV and V (late puberty), and compared them with Tanner stage III (mid puberty). These pooled data are presented in this paper, and the data divided into the five Tanner subgroups are shown in the supplemental material.

The following methods were used for the statistical analysis.

1) In the first step, changes in the hematological variables and VO₂max over the course of pubertal development were examined. For the hematological parameters, a two-way ANOVA with puberty and sex as the independent variables was used. For VO₂max, training status was added as an additional independent variable, so a three-way ANOVA was used here.

In both sexes, Hbmass was elevated in the later puberty stages compared to pre/early puberty; this was, however, much more pronounced in the boys (from 408 ± 88 g to 788 ± 133 g) than in the girls (from 367 ± 86 g to 531 ± 107 g; Table 3). Significantly higher absolute Hbmass and blood volumes (RCV, BV, PV) were observed in girls during mid puberty compared to pre/early puberty, and in boys during late puberty compared to pre/early and mid puberty. When Hbmass, RCV, BV and PV were normalized to LBM, no difference or only small differences were detected between girls and boys at pre/early puberty, and no significant difference existed between the female puberty groups. In boys, however, considerably higher values for Hbmass, RCV and BV were detected in the mid puberty and late puberty groups compared to pre/early puberty (Table 3). The PV did not differ between the sex and puberty groups with the exception of higher values in mid-puberty boys (Table 3).

Hemoglobin concentrations ([Hb]) in the boys was higher at late puberty (15.8 ± 1.2 g ^. dl⁻¹) than at pre/early puberty (14.6 ± 1.0 g ^. dl⁻¹) while in the girls, no difference was observed among the different puberty groups (Table 3). At pre/early puberty, no sex-related differences in [Hb] and Hct were detected, while at mid puberty and late puberty, all of the hematological values were considerably higher in the boys (Table 3).

Multifactorial ANOVA (see #3 in the section “statistics”) with absolute Hbmass and with normalized Hbmass (g kg⁻¹ and g kgLBM⁻¹) as the dependent variables yielded significant main effects of phase of puberty, sex, training status and place of residence, as well as of the interaction of phase of puberty with sex and training status (Table 4). No interaction was found for the phase of puberty with place of residence.

When the whole group was divided and the pre/early-, mid- and late pubertal phases were separately analyzed (see #4 in the section “statistics”), only a slight, significant effect of sex and no effect of training status were observed in the pre/early pubertal group for absolute and normalized Hbmass, while both effects were highly significant in the later stages. (Figure 1B shows Hbmass normalized to kg LBM). Moderate altitudes had significant, positive effects in all puberty stages (Figure 1B).

Multifactorial ANOVA (see #3 in the section “statistics”) for BV as a dependent variable showed significant, positive main effects of sex and training status, as well as interactions for sex and training status, with the phase of puberty but no effect of altitude (Table 4). The main effects of sex and training status (see #4 in the section “statistics”) were highly pronounced in the mid- and late pubertal stages but did not exist pre/early puberty (Figure 1C). The effects of sex and training status on PV were similar to those on BV, while moderate altitude exerted a negative effect on PV (Table 4; Figure 1D).

The absolute VO₂max showed a well-known sex- and training-specific increase with increasing sexual maturity (Table 5). When VO₂max was normalized to LBM, there was still a clear, positive influence of sex and training status but not of the puberty stage.

The linear regression analysis between Hbmass and VO₂max showed a clear difference between the trained and untrained groups, with a steeper slope of the regression line of the trained group (3.63 vs. 2.59. The relationship between the two variables became weaker when the values related to LBM were used (trained, absolute vs. normalized: r = 0.93 vs r = 0.54, untrained: r = 0.83 vs r = 0.30), but the difference between the groups persisted (Figure 2).

Multifactorial ANOVA (see #3 in the section “statistics”) with the absolute VO₂max of the entire group as the dependent variable showed significant main effects for puberty status, sex, training status, and Hbmass and interactions between training status and Hbmass (Table 4). For normalized VO₂max as the dependent variable, also a small positive effect was found for altitude.

In all stages of puberty, there were significant main effects (see #4 in the section “statistics”) of the confounding factors of sex, training, and Hbmass. Moderate altitude also tended to have a positive effect on VO₂max. For the magnitude of the main effects, see Figure 3.

The development of Hbmass, BV and PV during maturation corresponds to the data we recently published (Mancera-Soto et al., 2021). In girls, there was an equal percentage increase in RCV or Hbmass and PV by approximately 45% for both parameters from pre/early to late puberty, reflected in unchanged [Hb] and Hct levels. However, in boys, Hbmass increased by 95%, while PV increased by only 65%, resulting in an increase in [Hb] from 14.6 g dl⁻¹–15.8 g dl⁻¹. These data clearly show that [Hb] is not a good indicator of the absolute amount of hemoglobin in the body and does not represent the quantitative development of hemoglobin in childhood and adolescence.

The influence of training has only been investigated in a few studies with a small number of test subjects. Prommer et al., (2018) described a small training effect of 25 g (7%) over the course of 2.5 years in children aged 8–11 years old at the start of the study, whereas other studies did not find any specific training effects in 12–15-year-old boys and girls (Landgraff and Hallén, 2020) or in 15–17-year-old boys (Ulrich et al., 2011).

As demonstrated in Figure 1, at pre/early puberty, there were no associations between training status and Hbmass or blood volumes. In the mid- and late pubertal states, however, there were significant effects ranging between 10 and 12% occurring independently on LBM in both sexes (Figures 1, 4). Possible reasons for these delayed training adaptations could be the training status of the different age groups, although the youngest participants already had a training history of at least 1 year (mean 2.7 years) and a training volume of at least 6 h/week (mean 11 h/week) completed. In contrast, the mean training history and the weekly training volume in the end-pubertal group were up to twice as high (see Table 2).

Conversely, the relatively small increase in Hbmass at later developmental ages, which is only slightly greater than in adults who begin intensive endurance training (Schmidt and Prommer, 2008), demonstrates only slight erythropoietic stimulation from exercise. Since there was no interaction between training status and sex, the increase in Hbmass in the mid- and late pubertal training groups was not achieved through training-related testosterone stimulation. This outcome is in line with our recently published data, which showed no association between training status and serum testosterone levels during puberty (Wagner, 2000).

BV shows a very similar association with training status as Hbmass, which could indicate that the increase in Hbmass occurs less via primary erythropoietic stimulation but rather as a consequence of the training-related increase in BV with subsequent normalization of [Hb] and Hct via a so-called critmeter (Montero and Lundby, 2018). This outcome would mean that the undoubtedly higher and pronounced Hbmass values in elite endurance athletes are mainly genetically determined and can only be modulated to a relatively small extent by training. However, data on the genetic mechanisms are scarce. Studies with endurance athletes have identified only NFIA-AS2 polymorphisms as a mediator of erythropoiesis, with carriers of the variant genotype having higher hemoglobin mass, BV, and PV (Malczewska-Lenczowska et al., 2020; Kim et al., 2022). The view of the involvement of the kidney as a critmeter is also supported by the behavior of the PV, which was significantly increased in the mid- and late puberty training groups and thus, together with the larger erythrocyte volume, contributes to the increase in BV.

The influence of moderate altitude on the [Hb] is approximately 1.0 g/dl in all stages of puberty and does not show sex-specific characteristics, corresponding to data from Böning et al., (2004); Schmidt et al., (2002) in adults from the same regions, as well as data from Gassmann et al., (2019), who showed significantly greater [Hb] in inhabitants of the Andes compared to inhabitants of similar high altitudes from Tibet and East Africa. The influence of altitude on Hbmass and BV has not yet been investigated in children. For the first time, we showed here that an equal erythropoietic adaptation as in adults also occurs in childhood and adolescence. There are data available from adults from the same regions where this study was conducted showing almost the same absolute Hbmass values and altitude-related differences in women as we do here in Bogotá for late pubertal girls (Böning et al., 2004). In men, however, trained and untrained subjects from Bogotá show significantly higher values than adolescents during late puberty (Schmidt et al., 2002). This outcome demonstrates that, in contrast to the girls, the development of Hbmass and the associated BV is not yet complete at the end of puberty in boys, which is in good agreement with the data of Steiner and Wehrlin, (2011), who observed an increase in Hbmass up to the age of 21 years old.

In our recently published study, we did not find any altitude-related effects on serum erythropoietin and testosterone concentrations in children and adolescents (Wagner, 2000) and can therefore exclude a simple increase in these hormones as the cause of the higher Hbmass at higher altitudes. However, it seems possible that, at higher altitudes, there occurs a decrease in the soluble EPO receptor, improving the binding of EPO to its membrane receptor (Villafuerte et al., 2014) and thereby augmenting the erythropoietic effectiveness at identical serum EPO concentrations.

The whole BV did not differ between the groups from low and moderate altitudes since the higher erythrocyte volume at moderate altitudes is compensated for by a lower PV (Table 3; Figure 1), which can be attributed to a hypoxia-related change in the concentration of the volume-regulating hormones aldosterone and ANP (Schmidt et al., 1999). In adults, this behavior of the PV has not only been described for short-term stays at different altitudes (Beidleman et al., 2017) but also for chronic stays at the identical moderate altitudes to those in this study (Böning et al., 2004).

In this study, we did not detect any interactions between the training effects and the altitude effects on Hbmass but only the main effects, which are demonstrated in Figure 1. This outcome suggests independent regulatory mechanisms, which at moderate altitudes are likely to be hypoxic renal stimulation and, with regard to the training effects, compensation for the expansion of the PV (Schmidt et al., 1988).

The increase in the absolute VO₂max with age, as well as the sex differences, correspond to the model developed by Armstrong and Welsman, (2019). Since LBM is the main influencing factor on VO₂max (Armstrong and Welsman, 2019), the normalized VO₂max should be constant during the maturation and growth process of children and adolescents. Landgraff et al., (2021), however, demonstrated a slight decrease over the course from 12 to 15 years of age in endurance-trained boys. Our data do not show any significant differences in the normalized values during the growth and maturation process but a clear trend toward decreasing values with increasing age in the untrained children and increasing values in the trained children. Because in our study, the normalized data differed greatly between the trained and untrained children, factors other than LBM also influenced VO₂max and even made variation in the normalized values during maturation likely.

The improvements through training shown by Bar-Or for children and adolescents (Bar-Or and Rowland, 2004) were less than 5% in pre/early pubescent children and more than 15% in mid- and late pubertal adolescents. This different behavior in the age groups could in principle also be observed in this study. However, the differences between the trained and untrained groups in all stages of development were greater than those in Bar-Or’s review and amount to ∼20% in the pre/early pubertal stage and ∼40% in the mid- and late pubertal stages for both sexes (Table 5). While Bar-Or does not provide any definitive causes for better trainability at older ages, the finding in our study could also be due to a longer training history and possibly more intensive training in the older children.

To date, VO₂max has been examined several times in children and adolescents, but the importance of both Hbmass and BV on VO₂max has rarely been quantified. In statistical analysis, Hbmass and BV have almost the same effect on VO₂max. Both parameters are associated with the maximum stroke volume (Schierbauer et al., 2021), and a high Hbmass guarantees an advantageous [Hb] and thus O₂ transport capacity. In children, both maximum SV and a − v ¯   O 2   diff are mainly determined by the magnitude of LBM (Armstrong and Welsman, 2020). The effects of endurance training on VO₂max, in contrast, are—like in adults (Lundby et al., 2017)—mainly brought about by increasing the SV and less so by increasing the a − v ¯   O 2   diff (Obert et al., 2003). As a reason for this change, cardiac remodeling through the development of cardiac morphology and function in youth athletes has been reported (Nottin et al., 2002). Recently, Bjerring et al., (2020) demonstrated distinct phases in the development of the young athletic heart with a tendency toward concentric remodeling in endurance athletes at the age of 12, with increased wall thickness and cardiac mass compared to sedentary peers. In later ages, this development changed to eccentric remodeling. Cardiac changes during adolescence appear to be associated with the development of BV and Hbmass, as Perkins et al., (2022) showed a simultaneous increase in left ventricular mass and expansion of Hbmass and BV in trained children in the later stages of puberty.

In this study, the effect of the training status on VO₂max, already present in the pre/early pubescent stage but even more pronounced in the mid- and late pubertal stages, was remarkable and indicates—because LBM was similar in the trained and untrained groups—LBM-independent training adjustments of the O₂ supply, likely via an increased stroke volume, as mentioned above.

In the pre/early pubertal stage, there is merely a tendency toward an expansion of the BV, and the hypothesized higher SV should be obtained by functional improvements within the circulation. In the mid- and late pubertal stages, however, the increase in BV of ∼400–600 ml in both sexes should lead to an increase in the maximum SV of ∼12 ml (Schierbauer et al., 2021). This increase, in turn, should lead to a higher maximum cardiac output (Q_max) of ∼2.3 L  min⁻¹ and thus a higher VO₂max of ∼200 ml  min⁻¹ (Schierbauer et al., 2021). Conversely, although this adjustment indicates an important improvement, it can only explain part of the difference in VO₂max between the trained and untrained groups. In the later stages of development, both the functional improvements as in the pre/early pubertal stage and the BV increase-related optimizations should be mainly responsible for the improved VO₂max.

A similar conclusion emerges when one examines the relationship between Hbmass and VO₂max, as demonstrated in Figure 2. With the same Hbmass of 13.1 g kg LBM⁻¹ (average value of boys in the mid pubertal stage), untrained children had a VO₂max of 49.1 ml  min^−1. kg LBM⁻¹, but trained children had a VO₂max of 61.3 ml/min/kg LBM⁻¹, indicating 25% more effectiveness in O₂ transport and/or O₂ metabolism. Since the trained groups of both sexes in the mid- and late pubertal stages also possess ∼1.5 g kg LBM⁻¹ more Hbmass (see Figure 1B), their VO₂max increases by an additional 3.5 ml  min^−1. kg LBM⁻¹, i.e., by 7% more. The adaptation processes via increased erythropoiesis thus constitute ∼22% of the total training adaptations.

Immediately after ascent to a moderate altitude, VO₂max decreases by ∼0.65% per 100 m, i.e., by 17% from sea level to 2,600 m (Clark et al., 2007). During longer stays in altitude training camps at ∼2,300 m, the VO₂max improves, but after 3 weeks, it is still ∼5% less than at sea level (Schuler et al., 2007), and the performance, measured as competition pace, is still decreased by ∼3% (Chapman and Levine, 2007). In the case of chronically altitude-adapted cyclists at the national level from Bogotá, however, comparable values (69 ml kg^−1. min⁻¹) were measured at moderate altitudes to those of European cyclists with the same performance level at sea level (68 ml kg^−1. min⁻¹) (Schmidt et al., 2002). To date, however, there are no data available comparing adolescent athletes born and living at different altitudes.

In the present study, we did not detect any negative, but even a tendency to a positive effect of altitude on VO₂max (Figure 3). This result cannot be attributed to methodological causes since, when comparing the cycle ergometer and treadmill tests, the discipline-specific performance determined as power and speed at 2,600 m did not differ from that at 1,000 m. Rather, the following mechanisms could contribute to the complete compensation for the hypoxic surroundings at 2,600 m. 1) There is an increase in Hbmass, which is ∼7% in all stages of development; 2) The Hb-O₂ binding properties are adapted to the altitude in the form of a steeper Hb-O₂ binding curve, resulting in a left shift in the upper part and a right shift in the lower part (Schmidt et al., 1990). This process enables a higher arterial O₂ saturation in the lungs and a more effective O₂ release in the muscle tissue. In that study, we were able to show that, with a capillary pO₂ of 15 mm Hg, this mechanism delivers ∼4% more O₂ to the tissue. 3) Adequate ventilation and improved diffusion properties in the lungs (Wagner et al., 2002) lead to higher arterial O₂ saturation at rest (Böning et al., 2004), especially during exercise, which is much more pronounced than in altitude-adapted lowlanders (Favier et al., 1995).

Based on these considerations, the importance of a high Hbmass for VO₂max can be emphasized. Since the effects of the training-related increase in Hbmass are almost nonexistent at pre/early pubescent ages and relatively small at mid- and late pubertal ages, genetic predispositions for a high Hbmass and thus high VO₂max in adulthood are absolute prerequisites. Screening Hbmass as one biological marker at a young age was therefore recommended by Wehrlin and Steiner, (2021) to characterize possible talent for endurance sports. In this study, for example, after correction for the training and altitude effects, 9% of all of the values of the children and adolescents from the untrained groups were greater than the 90th percentile, and 3% were greater than the 95th percentile of the sex- and puberty stage-specific Hbmass values, while the numbers for the trained children were 12 and 5%, respectively. This outcome indicates that a large number of children who have not yet practiced any sports might still have the potential to become successful endurance athletes. These subjects could be those individuals who respond to endurance training with particularly large increases in VO₂max, as described by Bouchard et al., (1999) in their genetic studies.

Living at a moderate altitude from childhood fully compensates for the decreased ambient oxygen pressure so that VO₂max is not restricted. Athletes from moderate altitudes, therefore, have a clear advantage in competitions at altitude over athletes from sea level even if the latter have adapted to high-altitude training camps. This advantage should also exist at lower altitudes and could be one reason for the success of elite Colombian cyclists, who mainly grew up in the area around Bogotá that we examined in this study. However, it is still unclear to what extent and in what time frame this advantage over athletes from sea level will be reduced over time when competing at low altitudes.

The study was conducted in Colombia, whose residents are of different ethnic origins. Although the ethnic groups in the two regions from which the children studied here derived are fairly homogeneous, it must be noted that the majority of residents of these regions are mestizos with more or less European or Native American origin. In order to be able to assess possible ethnic influences on the parameters determined here, it would therefore be desirable to carry out a genetic typing, which, however, could not be done in this study.

In the present cross-sectional study, relationships between Hbmass or BV and important confounding factors were derived by means of regression analysis. Highly significant correlations with Hbmass and other influencing factors were also established for VO₂max. However, the correlations calculated here do not prove cause and effect but only associations between the respective variables. Cause and effect can only be estimated through longitudinal studies over several years. However, since it is extremely difficult to observe training effects from the untrained state to the competitive athlete over the course of child and adolescent development on a statistically sufficiently large number of test subjects, cross-sectional studies such as these could be valuable. Therefore, it cannot be excluded that children with a genetically predetermined high Hbmass and BV facilitating a higher VO₂max are more likely to be found in the trained group and thus influence the conclusions about the trainability of Hbmass and BV.

In boys, the parameters related to body mass and LBM agree very well. In the case of girls, however, the data diverge notably over the course of puberty due to the increasing body fat percentage. Therefore, the hematological variables are presented as normalized to LBM to better compare boys and girls. The LBM was not measured directly here but was estimated by skin fold measurements, and based on these measurements, the body fat percentage was determined. Although this method correlates sufficiently with the gold standard methods [dual-energy X-ray absorptiometry (Hofsteenge et al., 2015), three-component method (Aguirre et al., 2015)], the results should be interpreted with caution.

The VO₂max of the runners was corrected by 7% because it was determined on the treadmill and not on the bicycle ergometer, as was the case with the other participants. This adjustment was necessary because very well-trained cyclists show almost identical values on both devices, but runners are significantly less efficient on the ergometer. Since there are interindividual differences when comparing the VO₂max on the treadmill and on the bicycle ergometer, inaccuracies due to the 7% correction cannot be excluded. However, since the same number of runners were tested at both altitudes, this possible inaccuracy should not affect the validity of this work.