Abstract

Introduction: Pneumonia is still the leading cause of death in individuals with spinal cord injury (SCI). Knowledge of modifiable predictors of pneumonia would be important to improve prevention.

Aims: To analyze which respiratory function parameter is the best discriminator between individuals with and without pneumonia and to estimate a risk threshold for this factor.

Methods: Retrospective cohort study. FVC, FEV1, PEF, Pimax, Pemax of individuals with or without pneumonia were measured. ROC analysis were used to find the best discriminator between individuals with and without pneumonia. For the best discriminator odds ratio and a pneumonia risk threshold and its sensitivity/specificity were calculated.

Results: In total, 307 patients were analyzed (70 with/237 without pneumonia). In the motor complete sub-group Pimax was the best discriminator between individuals with and without pneumonia (AUC 0.86; 95%CI: 0.78-0.93). 115% of the SCI specific reference value was identified as pneumonia risk threshold (sensitivity 74.4%/specificity 83.4%). The odds of having a pneumonia for individuals with Pimax below 115% of the SCI reference value was 16.1 (95%CI: 7.2-36.0) times the odds of having a pneumonia with Pimax above 115% of reference values. For individuals with incomplete SCI, a PEF of 7.04l/s and a Pimax of 93.5cmH2O were identified as best discriminators (AUC 1.0) between individuals with and without pneumonia (sensitivity/specificity 100%).

Conclusions: Pimax is the best discriminator between those with and without pneumonia for all SCI individuals. Inspiratory resistance training should be promoted to prevent patients with inspiratory muscle strength below threshold values from pneumonia.