More than 80 million people worldwide live permanently in high-altitude areas (i.e., > 2500 m) and are exposed to chronic hypobaric hypoxia.1 Mechanisms of adaptation to this hypoxic stress include increasing levels of hemoglobin concentration [Hb], especially in native Andean populations.2,3 However, among highlanders, excessive erythrocytosis (EE, defined as [Hb] ≥ 21 g·dL−1 in men and ≥ 19 g·dL−1 in women4) can occur and lead to severe clinical symptoms characterizing the chronic mountain sickness (CMS) syndrome. CMS is highly prevalent in Andean area,2–4 and is associated with systemic vascular dysfunction5–7 and/or pulmonary hypertension8,9 and may reduce life expectancy.2,3

While several pharmacological or non-pharmacological interventions have been tested for EE and CMS, the optimal treatment for these conditions remains to be identified.10,11 Among previously tested molecules, a 250 mg once-daily dose of acetazolamide (ACZ) has been shown to be the minimal effective dose to increase ventilatory response to hypoxia and to reduce both hypoxemia and hematocrit (Hct) up to 6 months in CMS highlanders at ∼4300 m.12–15 However, the exact mechanisms by which ACZ decreased Hct remain unclear, some results in highlanders suggesting a decrease in the hypoxia-induced erythropoiesis and possibly in total hemoglobin mass (Hbmass),13,14 while an animal study suggested an expansion of plasma volume (PV), albeit non-significant compared to hypoxic controls.16 The ability of this daily low-dosage of ACZ to sufficiently increase ventilation, improve hypoxemia, and to reduce Hct in CMS patients living at higher altitudes (where partial oxygen pressure is lower) remains uncertain. The changes in Hbmass and blood volumes that may underlie this change in Hct also remain to elucidate. Statins are other molecules that could be beneficial against CMS through their pleiotropic effects on vascular function17 and possibly on erythropoiesis18 but their effects have never been properly tested in this context. Thus, we conducted in La Rinconada (a gold mining city located in southern Peru at 5100–5300 m and considered as the highest city in the world19), a double-blinded, placebo-controlled study, to assess the efficacy of ACZ and statins to decrease Hct, improve CMS condition and to understand the underlying changes in intravascular volumes.

While the final evaluations of this clinical trial (initially planned after 9 months of treatment) were disturbed by the COVID-19 pandemic,20 most of the patients included were reevaluated after 3 weeks of treatment in order to assess 1) the potential short-term efficacy of the treatments on Hct and 2) the potential underlying mechanisms of these treatments on intravascular volumes. We hypothesized that ACZ only would reduce the Hct following 3 weeks of treatment, ant that this early change would be mediated by both a decrease in Hbmass and a PV expansion.

This study assessing the early efficacy and hematological changes induced by drugs potentially improving CMS condition highlights 1) the early efficacy of ACZ to reduce Hct (and [Hb]), even at an extreme altitude > 5000 m and 2) the underlying changes in intravascular volumes. In the ACZ group, the early decrease in Hct observed in the present study was comparable to previous results in highlanders albeit living at an altitude ∼1000 m lower.13–15 As expected, and previously shown,12–15 ACZ administration stimulated ventilation and improved oxygenation, mainly consecutively to a mild metabolic acidosis induced by the inhibition of renal tubular carbonic anhydrase26 and without any identified significant side-effects (e.g., hypokalemia). The relative rise in PaO2 observed in the ACZ group (+13%) was similar to previously observed by others ∼1000 m of altitude lower.15 Despite the absence of any significant change in Hct, decreases in CMS severity scores in the placebo and atorvastatin groups were almost similar than those observed in the ACZ group (Table 2). A similar placebo effect has previously been observed13 and might be, in part, explained by the difficult and subjective evaluation of the symptoms (and their intensity) in CMS population, where the relationship between symptoms’ severity and Hct (or [Hb]) can be inconsistent in cross-sectional studies.22,27,28

In accordance with our hypothesis regarding the effect of ACZ on intravascular volumes in highlanders, the significant increase in PV and the strong correlation between this increase and the decrease in Hct, suggested that the early decrease in Hct was not solely mediated by a decrease in erythrocytosis, but also by fluid regulation leading to an increase in PV, albeit the correlation between Hct and calculated intravascular volumes should be interpreted with caution due to a between-variable dependency. While a decrease in PV could have been expected due to the mild diuretic effect of ACZ,29 this intriguing finding seems in line with previous results observed in a rodent model of chronic hypoxia exposure.16 In this experimental study, ACZ-treated rats during exposure to chronic hypoxia exhibited no significant change in PV compared to normoxic controls, while non-treated hypoxic controls showed the more classically expected hypoxic PV contraction. In lowlanders, reduction in PV is a well-known and rapid physiological response to high-altitude exposure (allowing an increase in arterial oxygen content) with a prompt recovery of PV upon reoxygenation.30 The exact underlying mechanism(s) of the early PV expansion under ACZ therapy observed in our study in CMS highlanders remain(s) to be investigated but may involve, at least in part, an activation of renin-angiotensin-aldosterone system (RAAS) possibly mediated by the oxygenation improvement.16 In Andeans, PV contraction has been shown to contribute, together with an increase in Hbmass, to the CMS polycythemia31,32 and is associated with a decrease in RAAS activity.31 Furthermore, we previously showed that CMS and non-CMS highlanders in La Rinconada had similar Hbmass but that CMS highlanders tended to have a smaller PV, with an inverse relationship between CMS symptoms and PV.22 Taking together, these results support the potential implication of PV in CMS pathophysiology as well as a potential therapeutic target.

The apparent discrepancy between the significant decrease in RBCV (which also correlated with the decrease in Hct) and the non-significant decrease in Hbmass (Fig. 2A and 2B) in the ACZ group could be explained by the small sample size coupled with some inter-individual variability. The more pronounced decrease in RBCV than in Hbmass may also reflect a reduction in reticulocyte and young RBC counts (which have a greater volume than mature and older RBCs33) secondary to the onset of erythropoiesis decline, as suggested by the decreasing trend in plasma EPO level (Table 2). Nevertheless, EPO concentrations must be interpreted cautiously due to a large variability, especially in CMS patients.34 The absence of a complete blood count including reticulocyte count in this study does not allow us to confirm this hypothesis.

At last, the significant increase in Hbmass in the placebo group during this short-term study was unexpected and remains unexplained. Similarly, without any associated measurement of the RASS activity, the absence of concomitant contraction of PV to keep BV constant could not be explained. Participants were well-balanced among the treatment groups and resided at the same altitude (i.e., between 5100 and 5300 m) during the 3-week period. A measurement error seems unlikely for various reasons. First, identical and rigorous CO-rebreathing method was applied at both time-points (see Supplementary Methods) and second, similar pre-measurement fraction of venous carboxyhemoglobin (%HbCO) between the two time-points (eTable 2), allowing us to exclude any drift in %HbCO measurement during the study period. Furthermore, the significant decrease in absolute change in %HbCO during the CO-rebreathing measurement between W3 and BL in the placebo group only (eTable 2) was concordant with the increase in Hbmass, which was also concomitant with the increasing trend in sTfR (Table 2), suggesting a higher erythropoietic activity.24 Although none of subjects included reported any recent prolonged stay at lower altitude and had similar professional activities and life habits, we cannot exclude the fact that some of them may have spent some days at lower altitude (e.g. weekend in Juliaca, 3800 m) or may have been exposed to higher levels of hypoxia during their mining activities. However, the randomized double-blinded study design makes a systematic recruitment bias in one arm unlikely and suggests that differences between groups reflect the effect of the treatment.

Several limitations should be acknowledged. Due to the COVID-19 pandemic and lockdowns, we were not able to proceed with the pre-planned long-term reassessment. As a consequence, we could not determine if a significant Hbmass reduction would occur later in the ACZ or statin groups and we could not evaluate the potential long-term effects of the drugs (especially atorvastatin) on secondary outcomes such as vascular reactivity or oxidative-nitrosative stress.7 The good tolerance and efficacy of ACZ remains to be confirmed over a longer period, even if it has previously been demonstrated up to 6 months albeit at lower altitude.14 Furthermore, mechanisms underlying the early change in PV such as the RAAS activity remain to be elucidated.31 Finally, conducting drug-studies exclusively in male CMS, as in the present study and others,12–15 may restrain the generalizability of results to female CMS, who are also affected by CMS, albeit with a much lower prevalence, at least before menopause.2,3 Studies conducted in female CMS sufferers are unfortunately lacking.

In summary, 3 weeks of ACZ administration significantly improved oxygenation and reduced Hct in CMS highlanders living at extreme altitude > 5000 m. This early effect on Hct was correlated with both an increase in PV and a decrease in RBCV. Atorvastatin uptake had no effect on hematological parameters after 3 weeks of treatment but longer-term beneficial effects on vascular function may be expected.17 Further studies are required to understand the mechanisms of the ACZ-induced PV increase as well as to determine the role of changes in PV in the adaptative or maladaptive hematological responses to chronic hypoxia in high-altitude dwellers.

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

ClinicalTrials.gov identifier: NCT04251364

BC, ES, PR, SD, CAH, AP, AASG, IH, DG, JVB, PC, AP and SV designed the study and/or acquired the data. BC performed the statistical analysis. BC, ES, PR, SD, JVB, PC, AP and SV analyzed and interpreted the data. BC and SV wrote the manuscript; ES, PR, SD, CAH, AP, AASG, IH, DG, JVB, PC and AP revised it critically for important intellectual content. All authors gave final approval and agreed to be accountable for all aspects of work ensuring integrity and accuracy.