Table 1.
Main outcomes after injury
| Outcomes after injury | Means and proportions | Overall (n = 14,637) | ICU (n = 13,763) | Rehab unit (n = 874) | Number of studies | Subgroup differences | |
|---|---|---|---|---|---|---|---|
| p-value# | p-value# (excluding Branco et al. and Anand et al.) | ||||||
| Weaning success, n (%) | Crude | 1299/1942 (66.9%) | 674/1068 (63.1%) | 625/874 (71.5%) | 24 (10 + 14) | ||
| Meta-analytic | 67.8% [54.2–79%] | 63.1% [45.2–77.9%] | 71.5% [51.1–85.8%] | 0.5059 | – | ||
| Partial weaning, n (%) | Crude | 75/195 (38.5%) | 2/44 (4.5%) | 73/151 (48.3%) | 6 (1 + 5) | ||
| Meta-analytic | 31.9% [7.4–73.3%] | 4.5% [1.1–16.4%] | 43.3% [11.1–82.4%] | 0.0582 | – | ||
| Partial or total weaning, n (%) | Crude | 1351/1942 (69.6%) | 676/1068 (63.3%) | 675/874 (77.2%) | 24 (10 + 14) | ||
| Meta-analytic | 75.4% [63.9–84.1%] | 63.7% [45.6–78.6%] | 82.3% [69.6–90.4%] | 0.0610 | – | ||
| Tracheostomy, n (%) | Crude | 8344/13371 (62.4%)* | 8142/13165 (61.8%)* | 202/206 (98.1%) | 27 (19 + 8) | ||
| Meta-analytic | 93.9%* [77.2–98.6%] | 80.9% [51.4–94.4%] | 100% [9.1–100%] | 0.2201 | 0.1829 | ||
| Decannulation, n (%) | Crude | 150/291 (51.5%) | 33/131 (25.2%) | 117/160 (73.1%) | 10 (3 + 7) | ||
| Meta-analytic | 69.4% [39.3–88.8%] | 30% [10.9–60.1%] | 82.9% [53–95.4%] | 0.0135 | – | ||
| Duration of mechanical ventilation, days | Meta-analytic | 30.9 [24.8–37.1] | 26.9 [19.8–34.1] | 46.1 [28.3–63.9] | 22 (14 + 8) | 0.0503 | 0.0850 |
| 26.9 [19.8–34.1] | 96.5 [65.1–127.8]§ | 22 (14 + 8) | < 0.0001 | < 0.0001 | |||
| ICU/rehab days | Meta-analytic | 25.7 [20.2–31.2] | 22.9 [17.3–28.6] | 78.3 [49.1–107.6] | 22 (20 + 2) | 0.0003 | 0.0003 |
| Hospital days | Meta-analytic | 44 [37.3–50.7] | 44 [37.3–50.7] | – | 13 (13 + 0) | - | - |
| Pneumonia, n (%) | Crude | 814/1927 (42.1%) | 738/1630 (45.3%) | 76/297 (25.6%) | 21 (14 + 7) | ||
| Meta-analytic | 38.3% [25.7–52.8%] | 40.0% [27.4–54.0%] | 35.5% [11.3–70.5%] | 0.8132 | - | ||
| Mortality, n (%) | Crude | 1223/13756 (8.9%) | 1210/13331 (9.1%) | 13/425 (3.1%) | 25 (20 + 5) | ||
| Meta-analytic | 6.3% [4.1–9.6%] | 7.7% [5.2–11.2%] | 0.8% [0–18.5%] | 0.1751 | 0.1825 | ||
Meta-analytic means and proportions were computed through the “metamean” and “metaprop” functions (R package “meta”), respectively
*Crude and meta-analytic proportions differed appreciably because of the very large sample size of one of the included studies, Branco et al. [32] (5256 patients, weighting for 73% of the total in the crude calculation), which enrolled patients with very low prevalence of tracheostomy (21%) and whose weighting in the pooled meta-analytic result was more comparable with those of the other studies
#p-values yielded by the subgroup analyses: ICUs vs rehabilitation units with and without Branco et al. [32] and Anand et al. [35]
§Including duration of mechanical ventilation prior to admission and during the stay in rehabilitation