Arterial pH and PaCO2 (at presentation or pre NIV) are poor predictors of death or length of stay (LOS) in patients receiving non invasive ventilation (NIV) for acute type 2 respiratory failure (T2RF)

Abstract

Introduction

Patients with decompensated T2RF and arterial pH <7.26 are thought to have worse outcomes and there is debate over treatment in a level 2 setting. This study at a District General Hospital aimed to identify if pH, PaCO2 or level 2 care has significant correlation to LOS or mortality.

Method

Retrospective analysis of patients requiring NIV over 12 months(November 2012-13), and the relationship between admission and pre-NIV pH and PaCO2 to mortality and LOS.

Results

185 patients identified (89 male, 96 female). Overall mortality was 44.3%. There was no statistically significant difference between mean pre-NIV pH (7.25, SD0.10 vs 7.20,SD0.11 ) or admission pH(7.29,SD0.13 vs 7.30,SD0.12) in survival group compared to deceased group. Similarly, no significant difference was seen in pre-NIV PaCO2(10.97,SD3.72 vs 10.24,SD3.60) or admission PaCO2(9.76,SD4.08 vs 9.24,SD3.90) in the two groups. Analysis with spearman's rank correlation co-efficient showed no correlation between LOS with pre-NIV pH, pre-NIV PaCO2, admission pH or admission PaCO2. Median LOS was 11 days.

41.6% (77 patients) were level 2 with a mortality of 42.25%, ward patient mortality was higher 46.15% (p0.61) but not statistically significant. Mean pre-NIV pH for HDU patients was 7.22,SD0.12, compared with 7.26,SD0.70 for ward. Median LOS was 11 days for HDU and 10 days for ward.

Conclusion

Pre-NIV pH and PaCO2 in patients with T2RF are not reliable predictors of mortality or LOS. This review indicates that appropriate patients with T2RF should be treated with NIV regardless of severity of the arterial pH and PaCO2.