Introduction: The use of high-flow nasal cannulas (HFNC) in patients with hypoxemic ventilatory failure reduces the need for mechanical ventilation and does not increase mortality when intubation is promptly applied. The aim of the study is to describe the behavior of HFNC in patients who live at high altitudes, and the performance of predictors of success/failure of this strategy. Methods: Prospective multicenter cohort study, with patients aged over 18 years recruited for 12 months in 2020 to 21. All had a diagnosis of hypoxemic respiratory failure secondary to pneumonia, were admitted to intensive care units, and were receiving initial management with a high-flow nasal cannula. The variables assessed included need for intubation, mortality in ICU, and the validation of SaO2, respiratory rate (RR) and ROX index (IROX) as predictors of HFNC success / failure. Results: One hundred and six patients were recruited, with a mean age of 59 years and a success rate of 74.5%. Patients with treatment failure were more likely to be obese (BMI 27.2 vs 25.5; OR: 1.03; 95% CI: .95-1.1) and had higher severity scores at admission (APACHE II 12 vs 20; OR 1.15; 95% CI: 1.06-1.24). Respiratory rates after 12 (AUC .81 CI: .70-.92) and 18 h (AUC .85 CI: .72-0.90) of HFNC use were the best predictors of failure, performing better than those that included oxygenation. ICU mortality was higher in the failure group (6% vs 29%; OR 8.8; 95% CI:1.75-44.7). Conclusions: High-flow oxygen cannula therapy in patients with hypoxemic respiratory failure living at altitudes above 2600 m is associated with low rates of therapy failure and a reduced need for mechanical ventilation in the ICU. The geographical conditions and secondary physiological changes influence the performance of the traditionally validated predictors of therapy success. Respiratory rate <30 proved to be the best indicator of early success of the device at 12 h of use.
Keywords: chronic hypoxemia; high altitude; high-flow nasal cannula; hypoxemic ventilatory failure; pneumonia.