RESULTS
A total of 1308 patients were screened for enrollment, of which 1014
(77%) did not meet inclusion criteria (Figure 1). The most common
diagnosis for ARF was pneumonia (Table 1).
TLS for baseline ranged from 0 to 31 points (mean=), with the highest
mean observed in those with ARDS and the lowest in patients with
COPD/Asthma (Table 2). Posterior lung scores contributed a greater
proportion to the TLS than anterior lung scores in all groups except
ARDS. The contribution of B-score to TLS was greater than
atelectasis/consolidation across all diagnostic categories, especially
in CHF, ARDS, and aspiration (Table 2).
Tertiles of baseline TLS (0-9, 9.5-18, 18.5-36) were positively
associated with mortality (p=0.0), median ICU (p=0.00) and hospital LOS
(p=0.0), and ICU ventilator hours (p0.00), and inversely associated with
baseline PaO2/FiO2 ratio (p0.00) (Figure 4).
Patients in the three clinical groupings had significantly different
total lung, B, anterior, and posterior scores (p <0.001,
0.005, <0.001, 0.00, respectively) with the highest scores in
the parenchymal group (Table 3). The patients in the no pulmonary
disease group had a non-pulmonary etiology for intubation on final chart
review not apparent at the time of initial inclusion.
Average time to perform the PU on 10 randomly selected days (n=28 ) was
135 seconds (range=88-205). Among the 64 virtual patients scored by the
4 participating physicians during pre-study training, inter-rater
agreement was very good for TLS (kappa = 0.83), almost perfect for
assignment of the A classification (kappa=0.96), very good for
atelectasis/consolidation (kappa=0.84), and substantial for the B1, B2,
and B3 classifications (kappa=0.77, 0.61, 0.74, respectively).