In Lundby et al. (2007), we demonstrated that recombinant human erythropoietin (rHuEpo) may increase arterial oxygen content by two means, (1) the well known increase in red cell mass and (2) a new observation: a concomitant decrease of plasma volume. It seems they are equally important in enhancing arterial oxygen content (Lundby et al. 2007). To us this is a potentially very important observation, and we are of course happy to clarify the methodological and physiological points raised in the letter.
The study was conducted in Copenhagen with Danish subjects. They were medical or physiology students and none of them participated in organized sports. Like most students in Denmark, however, they commuted by bicycle to the university and were active in recreational sport activities such as jogging 1–2 times per week. This was not controlled for, but subjects were asked not to deviate from their normal lifestyle, and this was well accepted by all. Mean pre-rHuEpo was a modest 3.95 ± 0.4 l min−1 (Thomsen et al. 2007). To ensure physical activity did not influence the collected data, all CO re-breathing was completed in the fasting state between 07.00 and 09.00 h after 15 min of resting in a chair. The care taken when completing the experiments is reflected in the low CV for red cell volume, plasma volume, and blood volume (2.7, 3.7, and 2.6%, respectively). As pointed out in the letter, these volumes are derived from the measured total haemoglobin mass and it is suggested that total haemoglobin mass should be reported also. Total haemoglobin mass increased from 14.86 ± 2.04 mmol before rHuEpo to 16.28 ± 1.81 (P= 0.011), 15.81 ± 2.82 (P= 0.025) and 16.09 ± 2.87 mmol (P= 0.011) after 5, 11 and 13 weeks, respectively. Since haematocrit of the venous blood sample was corrected neither for trapped plasma, nor for peripheral blood sampling (Fcell ratio = 1) (Harrison, 1985), it is suggested in the letter that this may have interfered with our results. However, by using femoral arterial haematocrit as an estimate of whole body haematocrit in our subjects, we calculated a mean Fcell ratio of 0.99 ± 0.02 before rHuEpo and of 1.00 ± 0.02 after 14 weeks of treatment. This suggests that, in the present study, peripheral venous haematocrit did not overestimate whole body haematocrit, either during baseline or after rHuEpo treatment.
In the letter it is argued that an increase in the filling state of the heart could be the main cause of the reported reduction in renin and aldesterone. Although appealing, we are of different opinion. Renin and aldesterone were reduced by approximately 50% already 96 h after the first injection. At this early stage it seems very unlikely that rHuEpo should have increased red cell mass, and concomitantly total blood volume, and hence ultimately the filling state of the heart. It is argued that a tendency for increases in SV and in our study would support an increased filling state. The P values for the suggested tendency, however, are 0.33 and 0.22, respectively, leaving this rather speculative. As mentioned in our paper Epo may cause vasoconstriction. Heidenreich et al. (1991) were the first to demonstrate the vasopressor effects of Epo in renal resistance vessels, and subsequently it was shown that this may be caused by an Epo induced endothelin release and shift in prostaglandin balance (Bode-Böger et al. 1992, 1996). In our experiments, however, systemic vascular conductance remained unchanged with rHuEpo treatment (5.1 ml min−1 mmHg−1 increase, P= 0.43). The difference in results between our study and those demonstrating a vasopressor effect may very well be the result of a low versus high Epo concentration. During the majority of our study we injected 5000 IU per week, whereas up to 200 U ml−1 have been used to achieve a maximum vasoconstrictor response in renal resistance vessels.
The last point raised in the letter, i.e. that oxygen transport may only be increased with rHuEpo treatment if pre-haematocrit is below an optimal value, is well taken, but also seems a bit speculative since all previous studies performed with Epo treatment in humans have shown an increase in . It should be noted that although viscosity is increased as the haematocrit is increased (Pirofsky, 1953), this does not necessarily affect since the increase in has been reported to be of similar magnitude over a wide ranges of haematocrit (Ekblom & Berglund, 1991). At most, the point indicates that the optimal haematocrit for type of exercise in humans must be over 50%, but it should of course be kept in mind that the crucial factor in this regard is red cell mass. We have also performed (unpublished) maximal exercise trials in the same subjects including determination of cardiac output with the ICG method, and in these trials pre- and post-rHuEpo was unaltered. It should also be obvious that does not necessarily change when the haematocrit is altered – after all the haematocrit may easily be changed without changing haemoglobin mass, and it has been demonstrated that plasma volume expansion does not do much to , at least in trained subjects.
As we have stated in our paper, we are not the first to observe decreases in plasma volume with rHuEpo treatment. In chronic heart failure patients, that treatment with recombinant human Epo (rHuEpo) increased red blood cell volume by 250 ml and simultaneously decreased plasma volume by more than 1 litre (Mancini et al. 2003), and when calculating plasma volumes from Hb mass or blood volume and haematocrits reported in studies where rHuEpo has been injected in healthy subjects, plasma volumes are indeed also decreased in these studies (Ekblom & Berglund, 1991; Parisotto et al. 2001). In a newly performed study we have mimicked our original injection regime in 16 subjects, and the preliminary results demonstrate a decrease in plasma volume over time in all subjects, demonstrating a high reproducibility of our previous results.