Predictive risk factors of cardiorespiratory abnormality for upper gastrointestinal endoscopy in Tibet

Abstract

Background: To explore the predictive factors of cardiorespiratory abnormality in nonsedated patients at high altitude (HA) during upper gastrointestinal endoscopy (UGIE).

Methods: The pulse and saturated oxygen (SaO2) levels of 993 patients undergoing nonsedated UGIE in Tibet were monitored. Bivariate correlation and logistic regression were used to identify predictive risk factors for hypoxemia.

Results: The basal and minimum SaO2 levels during UGIE of the Tibetan group were significantly higher than those of the non-Tibetan group. The minimum SaO2 and maximum pulse in the HA transient residents groups were significantly higher than those in the HA usual residents groups. The incidences of hypoxemia and severe hypoxemia in the Tibetan groups were significantly lower than those of the non-Tibetan groups. Bivariate correlation and logistic regression showed that race, age (≥ 40 years), residence time in HA (<10 years), and basal SaO2 (<89 %) were sufficiently effective to predict hypoxemia. High-risk hypoxemic patients whose residence time in HA was <2 years were more prone to severe hypoxemia. The combination of the four variables showed superior performance in hypoxemia prognosis (AUC-ROC, 0.941; sensitivity, 83.7 %; specificity, 92.5 %) and severe hypoxemia prognosis (AUC-ROC, 0.968; sensitivity, 90.3 %; specificity, 98.0 %).

Conclusions: Race, age, residence time in HA, and basal SaO2 of patients in HA were predictive variables for hypoxemia during UGIE. Non-Tibetan patients with age ≥ 40 years, residence time in HA <10 years, and basal SaO2 <89 % were prone to hypoxemia. Among those groups, patients whose residence time was <2 years were at higher risk for severe hypoxemia.