Premium
Oxygen‐Tension Based Indices as Predictors of Survival in Critically III Dogs: Clinical Observations and Review
Author(s) -
Pelt Deborah R. Van,
Wingfield Wayne E.,
Wheeler Stevan L.,
Salman Mowafak D.
Publication year - 1991
Publication title -
journal of veterinary emergency and critical care
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.886
H-Index - 47
eISSN - 1476-4431
pISSN - 1479-3261
DOI - 10.1111/j.1476-4431.1991.tb00090.x
Subject(s) - medicine , oxygen tension , arterial oxygen tension , arterial blood , fraction of inspired oxygen , arterial blood gas analysis , cardiology , oxygen , pulmonary shunt , multicollinearity , anesthesia , lung , linear regression , mechanical ventilation , chemistry , organic chemistry , machine learning , computer science
Summary With respiratory therapy in critically ill veterinary patients becoming more commonplace, a consistent Indicator of pulmonary function status is necessary. Although calculation of the pulmonary shunt fraction correlates well with the degree of pulmonary dysfunction, lts detetmination requires placement of a pulmonary arterial catheter, an invasive procedure that may not be practical in many clinical situations. Using Information obtained from the atterial blood gas, many other oxygen‐tension derived Indices have been suggested as noninvasive measurements of the efficiency of pulmonary gas exchange, lncluding the alveoiar‐arterial oxygen tension difference, the ratio of arterial to alveolar oxygen tension, the ratio of arterial to inspired oxygen tension, and the ratio of alveolar‐arterial gradient to arterial oxygen tension. A total of 427 blood gas values from 195 different patients were evaluated. A loglstic regression model using a stepwise algorithm was constructed to assess potential multicollinearity and interaction between factors. The only factors that contributed significantly to the model predictive of survival were age (p < 0.015), base excess (p < 0.029), and the alveolar‐arterial oxygen tension difference (p < 0.014).