All data for scoring were collected prospectively and daily for all time points, and were stored in the database. APACHE (acute physiology and chronic health evaluation) II score was calculated prospectively, using the data from the first 24 h after the onset of ARDS. We used the lowest recorded Glasgow coma scale prior to intubation and/or after the patient was extubated to determine CNS failure.
We divided patients into the following two groups by year of ARDS onset: 1987 to 1990; and 1994 to 1999. We did not have complete data sets for 1990 to 1994 due to a temporary change in our research support. We therefore did not include the 1990 to 1994 patients. The oxygenation criterion for selecting ARDS patients in 1987 to 1990 used P(A-a)O2 ratio. The subsequent publication of the North American-European Consensus Criteria led to the widespread adoption of a different and less severe set of criteria using a P/F criterion. In order to compare 1994-to-1999 patients with 1987-to-1990 patients, we established a correspondence between P(A-a)O2 ratio and P/F (Fig 1). A P/F of < 105 corresponded to a P(A-a)O2 of < 0.2. In addition, for the 1994 to 1999 patients only, we compared patients identified with a P/F of 106 to 173 to those identified with P/F < 105. in detail
We used death at the time of hospital discharge as the end point. We analyzed data by independent t test and Pearson \2 analysis, and expressed the results as the mean ± SEM.
We identified a total of 516 ARDS patients with a P/F of < 105 at ARDS onset (1987 to 1990, 256 patients; 1994 to 1999, 260 patients). We identified 288 patients (1994 to 1999) with a P/F range of 106 to 173.
A total of 548 ARDS patients (260 + 288) were identified between 1994 and 1999. Mortality was lower in 1994-to-1999 patients when compared to 1987-to-1990 patients (p < 0.011) [Fig 2]. Fewer deaths occurred in the 1994-to-1999 group for P/F values of both < 105 and 106 to 173. Mortality was lower for those with higher arterial oxygenation efficiency (P/F, 106 to 173) than for those with lower arterial oxygenation efficiency (P/F, < 105) [Table 3]. Between-group differences for age, gender, or risk factor for ARDS were not significant (Table 3).
Figure 1. The Lines originating at the origin encompass the calculations for an Fio2 of 0.4 to 1.0. For all calculations Pao2 = (barometric pressure —47) Fio2 — Paco2 (Fio2 + [1 — Fio2]/R), where barometric pressure = 647 mm Hg (the Salt Lake City altitude), Paco2 = 45 mm Hg, and respiratory quotient = 0.8. The thick, black horizontal arrow indicates a P(A-a)O2 ratio of 0.20, the value used for ARDS screening from 1987 to 1990. The thick, black vertical arrow indicates aP/F of 173, the value used for ARDS screening from 1994 to 1999. When the P(A-a)O2 ratio is < 0.20 at an Fio2 of approximately 0.8, the P/F is 105 (thin vertical arrow). Therefore, we chose this value as the common oxygenation inefficiency selection criterion for the two time periods.
Figure 2. ARDS mortality by year of onset.
Table 3—Patient Characteristics
|P/F||1987-1990 P/F < 105 (n = 256)||1994-1999|
|P/F < 105 (n = 260)||P/F 106-173 (n = 288)||P/F < 173 (n = 548)|
|Dead||54||44 ft||27ft||35 f|
|Age, yr||54 ± 10||50 ± 11||52 ± 10||51 ± 10|
|Male gender, %||48||55||57||56|
|APACHE II score||20 ± 7||20 ± 8||19 ± 8||20 ± 8|