Relief with a catch: hemodilution in chronic mountain sickness sparks an erythropoietin surge Open Access

TO THE EDITOR:

Life at high altitude poses a formidable physiological challenge, to which the body responds through acclimatization.¹ However, some highlanders develop a debilitating lifelong condition called chronic mountain sickness (CMS) or Monge disease characterized by excessive erythrocytosis (EE) and symptoms including headache, breathlessness, palpitations, cyanosis, tinnitus, and disturbed sleep.^2,3 The notion that red cells are produced excessively is borne from Monge’s original observation linking high hematocrit with CMS symptoms,⁴ as well as the fact that phlebotomy brings symptomatic relief.^5,6 However, this relief is only transient, which suggests a persistent underlying erythropoietic drive. A proposed refinement to conventional phlebotomy is isovolumetric hemodilution (IVHD), whereby the volume of removed blood is replaced with fluid.⁵ IVHD has gained traction because early studies have shown longer-lasting benefits, albeit in small cohorts and based on anecdotal testimony.⁷ However, other studies have reported a worsening of symptoms—notably apnea-hypopnea index—after IVHD.⁸ Additionally, hemodilution raises concerns about iron depletion, which may trigger an increase in pulmonary artery pressure and lead to, or exacerbate, pulmonary hypertension.⁹ 

An important consideration in designing more effective CMS treatments is their impact on erythropoietin (EPO), the renal hormone that stimulates red blood cell production. To that end, it is important to understand how the kidney detects whole-body oxygen availability. The oxygen partial pressure (PO₂) probed by peritubular interstitial fibroblasts of the inner cortex and outer medulla is set by the balance between arterial oxygen supply and tissue oxygen consumption. Although kidney cells sense PO₂ locally, this is more closely related to arterial oxygen content (CaO₂), which is the sum of free and hemoglobin-bound oxygen, than to arterial PO₂. The reason is that the amount of oxygen diffusing into peritubular interstitial fibroblasts depends on how much oxygen is available in the blood.^10,11 This principle was elegantly demonstrated in human studies that reduced CaO₂ with or without altering arterial PO₂.¹² At steady state, oxygen delivery is pitched against renal respiration set by the energetic requirements of tubular transport.¹³ Indeed, this coupling explains why the kidney is uniquely suited as an oxygen sensor. In most other tissues, local PO₂ is not a reliable proxy for overall oxygen availability because changes in tissue perfusion affect oxygenation, even at constant CaO₂ (eg, hyperemia raising tissue PO₂). In contrast, renal blood flow is linked to glomerular filtration rate (GFR), hence tubular transport workload. Consequently, an increase in renal blood flow raises both oxygen delivery and oxygen consumption, imposing a stoichiometric coupling that stabilizes renal PO₂.¹⁴ Circumstances that reduce renal PO₂ and stimulate release of EPO include anemia,¹⁵ where oxygen delivery is reduced and tubular oxygen demand is increased due to greater filtration of plasma-enriched blood. Accordingly, ascent to high altitude acutely stimulates EPO secretion due to reduced arterial oxygen delivery;¹⁶ however, this response becomes ablated¹⁷ with longer hypoxic exposure because plasma volume contracts and is replaced with red cells,¹⁸ ostensibly decreasing GFR and renal work.

A salient prediction of this model is that IVHD should (1) reduce arterial oxygen delivery in proportion to the reduction in hematocrit and (2) increase oxygen demand because of the higher GFR from the additional fluid volume. In contrast, a conventional phlebotomy would reduce arterial oxygen content but not necessarily increase renal oxygen demand because GFR is not expected to increase. In summary, IVHD is predicted to increase renal workload and trigger an EPO surge. Stimulated EPO release would put into question the benefit of IVHD as a refinement. We set out to address this conundrum by measuring serum EPO, alongside blood and alveolar gases, in CMS-affected highlanders before and after IVHD to a target hematocrit matching highlanders without CMS.

The study was approved by the Institutional Ethics Committee of Universidad Peruana Cayetano Heredia (CIEH-UPCH 081-03-17; SIDISI number 59285) and was conducted in accordance with the Declaration of Helsinki. We recruited 14 consenting males (aged 24-64 years), who were born at >4000 m above sea level, remained lifelong residents of Cerro de Pasco, Peru (4340 m), and self-reported as being of Quechua ancestry. Of these, 6 were diagnosed with CMS. Only male highlanders were enrolled because CMS has vanishingly low prevalence among premenopausal women.¹⁹ Exclusion criteria were history of pulmonary, cardiovascular or renal disease, a current smoker, work in mining, blood transfusions or underwent phlebotomies in the previous 6 months, travel to lower altitudes (<3000 m) for >7 days in the previous 6 months, or abnormal electrocardiogram or spirometry during screening. Participants with hematocrit >63% (hemoglobin >21 g/dL) were classified as having excessive erythrocytosis. General health and Qinghai CMS score questionnaires were applied.²⁰ To reduce hematocrit in participants with CMS by 20%, the necessary volume of blood was estimated by the method of Gross.²¹ Up to 4 units (450 mL each) of blood were removed in supine position from the cephalic vein over 2 consecutive days and replaced with an equal volume of colloid plasma (Polygelyne 3.5%; Hisocel). Polygeline has a half-life of 3 to 6 hours, and its hemodynamic stabilization effects last for 24 hours.^22,23 Hematocrit was measured before each IVHD session to titrate the need for additional blood volume removal. Each hemodilution session was conducted slowly (>3 hours) to avoid acute hemodynamic changes.

At 48 hours after IVHD, participants with CMS had significantly reduced hematocrit, total blood volume, and red cell volume, and raised plasma volume to match levels in non-CMS highlanders (Figure 1A-D). There was no significant effect on alveolar PO₂, arterial PO₂, or oxygen saturation (Figure 1E-G). Consistent with reduced oxygen-carrying capacity, CaO₂ decreased after IVHD by 18% to levels in non-CMS controls (Figure 1H). Significantly, IVHD improved CMS score (Figure 1I). Whole-body respiratory rate and cardiac output were unchanged by IVHD and remained similar to non-CMS participants (Figure 1J-K). At constant cardiac output, the reduced hematocrit and expanded plasma volume after IVHD are expected to increase GFR⁵, hence workload on tubular transport, alongside a decrease in arterial oxygen delivery. This combination is predicted to decrease renal PO₂ and inadvertently trigger EPO. Indeed, our measurements of serum EPO revealed a substantial approximate threefold increase to the range ∼70 mIU/mL (Figure 1L). The magnitude of this effect is substantial, implicating a synergy between decreased oxygen supply and increased oxygen demand (Figure 2). The reduction in CaO₂, at constant cardiac output is likely too small to explain the spike in EPO and highlights the critical importance of considering GFR and tubular transport. A plausible contributor to the EPO surge is the inevitable but transient (1 hour/unit) reduction in blood volume during IVHD; however, testing this effect by autologous transfusion would likely be declined by participants with CMS and carries significant safety risks.

Our findings corroborate the beneficial effect of IVHD on CMS score—at least acutely—and reveal a potentially problematic response in the form of an EPO surge. This renal response may affect hematocrit recovery (eventually curtailing the effect of IVHD) and should be considered in a cost-benefit analysis of treatment options. Our findings highlight the importance of considering renal function in patients with CMS for a more complete appreciation of their treatment responses. Furthermore, we argue that kidney disease should feature in treatment guidelines. Assessment of renal function may take the form of direct GFR measurement or an estimate from creatinine, provided that the formula is developed specifically for highlanders. A final observation from the results is that, despite normalizing hematocrit to levels seen in non-CMS residents at high altitude, a strong erythropoietic drive resurfaces after IVHD in participants with CMS. This indicates that non-CMS and post-IVHD CMS groups have different signaling thresholds for EPO release. One plausible explanation is impaired oxygen unloading from red cells, consistent with the notion that patients with CMS experience tissue hypoxia. Indeed, our human kidney perfusion experiments using stored blood that releases oxygen slowly revealed a decrease in diffusive capacity for oxygen and renal oxygen extraction.²⁴ Dysfunctional oxygen release from red cells may be a new aspect of CMS that could explain our present findings and help direct better treatments.

Acknowledgment: The study was funded by Wellcome Trust grant 107544/Z/15/Z (F.C.V.).

Contribution: F.C.V. designed and performed research, contributed vital new reagents or analytical tools, analyzed data, and wrote the manuscript; and P.S. contributed analytical tools, analyzed data, and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Francisco C. Villafuerte; Universidad Peruana Cayetano Heredia, Av. Honorio Delgado 430, Lima 15102, Perú; email: francisco.villafuerte@upch.pe.