Table 3.
Meta-analyses of randomized trials that evaluate mortality in ARDS
| Author, year, reference |
Intervention | n | Participants | Deaths | Time | Risk ratio (95% CI)# |
Fixed (F) or random (R) Effect and heterogeneity |
Interpretation by the authors |
|---|---|---|---|---|---|---|---|---|
| Singh, 2013 [66] | Inhaled β2- agonists vs. pl |
2 | 313/293 | 97/76 | Hospital | 1.22 (0.95- 1.56) |
R (I2=0%) | No survival benefit. |
| 2 | 182/182 | 66/52 | 28 d | 1.04 (0.50- 2.16) |
R (I2=83%) | |||
|
Santa Cruz,
2013 [60] |
High vs. low PEEP without other interventions |
3 | 1136/1163 | 378/429 | Hospital | 0.90 (0.81- 1.01) |
F (I2=0%) | Trend toward mortality benefit. |
| Zhang, 2013 [65] | Exogenous surfactant vs. pl |
8 | 1101/1043 | 368/349 | 28-30 d | 1.00 (0.89- 1.12) |
F (I2=0%) | Intervention was not associated with reduced mortality. No difference among the different types of surfactant. |
|
Alhazzani, 2013
[64] |
Cisatracurium vs. pl |
3 | 223/208 | 70/93 | ICU | 0.70 (0.55- 0.89) |
R (I2=0%) | Cisatracurium reduced 28 days, ICU and hospital mortality |
| 3 | 223/208 | 76/98 | Hospital | 0.72 (0.58- 0.91) |
R (I2=0%) | |||
| 3 | 223/208 | 57/81 | 28 d | 0.66 (0.50- 0.87) |
R (I2=0%) | |||
| Meng, 2012[54] | Exogenous surfactant vs. pl |
9 | 1285/1289 | 396/392 | 28-30 d | OR: 1.02 (0.86-1.20) |
F (I2=0%) | Intervention did not improve survival |
|
Afshari, 2011
[42] |
Inhaled nitric oxide vs. pl |
14 | 660/590 | 265/228 | Variable (1-365 d) |
1.06 (0.93- 1.22) |
F (I2=0%) | No benefit on survival |
| 9 | 578/504 | 208/578 | 28 d | 1.12 (0.95- 1.31) |
F (I2=0%) | |||
| Burns, 2011[47] | Pressure and volume-limited ventilation vs. traditional MV |
10 | 888/861 | 312/366 | Hospital | 0.84 (0.70- 1.00) |
R (I2=43%) | Borderline (p=0.05) statistically significant reduction in mortality. |
|
Dasenbrook,
2011 [48] |
Higher vs. lower PEEP |
4 | 1166/1194 | 311/356 | 28 d | 0.90 (0.79- 1.02) |
F (I2=11%) | No significant difference in 28 d survival. |
|
Abroug, 2011
[40] |
Prone vs. supine positioning |
7 | 862/813 | NA | ICU | 0.91 (0.75- 1.12) |
R (I2=0%) | No significant effect on ICU mortality. Sub-analysis showed a survival benefit in those with more severe forms of ARDS |
| Dee, 2011 [67] | Inflammation- modulating diet vs. control diet |
3 Sa me stu die s |
171/173 | 42/72 | Hospital | 0.58 (0.42- 0.79) |
R (I2=0%) | Intervention improved survival |
| Briel, 2010 [46] | Higher vs. lower PEEP |
3 | 1136/1163 | 324/381 | ICU | 0.87 (0.78- 0.97) |
Log-binomial regression |
No improvement in hospital survival. Survival improved in more severe forms of ARDS. |
| 374/409 | Hospital | 0.94 (0.86- 1.04) |
||||||
| Iwata, 2010 [51] | Sivelestat vs. pl | 4 | 379/379 | NA | 28-30 d | 0.95 (0.72- 1.26) |
R (I2=0%) | No significant survival benefit at 28-30 d but worse survival at |
| 2 | 253/258 | NA | 6 m | 1.27 (1.00- 1.62) |
R (I2=0%) | 6 m | ||
| Sud, 2010 [61] | Prone vs. supine positioning (severe hypoxemia) |
7 | 295/260 | 157/163 | Hospital | 0.84 (0.74- 0.96) |
R (I2=0%) | Prone positioning reduced mortality in patients with severe hypoxemia. Overall, no significant effect. |
| Prone vs. supine positioning (less severe hypoxemia) |
7 | 590/578 | 248/230 | Hospital | 1.07 (0.93- 1.22) |
R (I2=0%) | ||
|
Lamontagne,
2010 [53] |
Corticosteroid therapy vs. pl |
12 | 471/495 | 147/176 | Hospital | 0.84 (0.66- 1.06) |
R (I2=29%) | Low-dose corticosteroid therapy may reduce all-cause mortality |
| Lower corticosteroid dose vs. pl |
9 | 374/396 | 95/128 | Hospital | 0.68 (0.49- 0.96) |
R (I2=30%) | ||
| Sud, 2010 [3] | HFOV vs. conventional MV |
6 | 189/176 | 73/87 | Variable (Hospital or 30 d) |
0.77 (0.61- 0.98) |
R (I2=0%) | Intervention might improve survival |
|
Putensen, 2009
[59] |
Lower vs. higher TV at similar PEEP |
3 | 518/515 | 177/211 | Hospital | OR: 0.75 (0.58-0.96) |
F (I2=18%) | Low TV reduced hospital mortality. Higher PEEP did not improve mortality |
| Higher vs. lower PEEP at low TV |
3 | 1136/1163 | 378/429 | Hospital | OR: 0.86 (0.72-1.02) |
F (I2=0%) | ||
| Lower TV + higher PEEP vs. higher TV and lower PEEP |
2 | 79/69 | 30/42 | Hospital | OR: 0.38 (0.20-0.75) |
F (I2=0%) | ||
| Tang, 2009 [62] | Corticosteroids vs. pl |
4 | 191/150 | 45/53 | Hospital | 0.51 (0.24- 1.09) |
R (I2=51%) | Low-dose steroids was not associated with improved survival |
|
Phoenix, 2009
[57] |
Higher vs. lower PEEP |
6 | 1233/1251 | 415/482 | Early mortality (Hospital and 28 d) |
0.87 (0.79- 0.97) |
R (I2=0%) | PEEP may provide a mortality benefit. |
| Only studies with groups with similar tidal volumes |
3 | 1136/1163 | 378/429 | Hospital | 0.90 (0.81- 1.01) |
R (I2=0%) | ||
|
Kopterides, 2009
[52] |
Prone vs. supine positioning |
4 | 662/609 | 245/230 | ICU | 0.97 (0.77- 1.22) |
R (I2=32%) | No survival differences, however ICU mortality was lower in severely ill patients. |
| Oba, 2009 [55] | High PEEP vs. low PEEP |
5 | 1215/1232 | 408/464 | Hospital | 0.89 (0.80- 0.99) |
F (I2=0%) | Survival benefit in hospital mortality, but statistical and clinical heterogeneity. Effect greater in patients with higher ICU severity scores |
| 3 | 889/914 | 253/296 | 28 d | 0.88 (0.76- 1.01) |
F (I2=0%) | |||
|
Pontes-Arruda
2008 [58] |
Inflammation- modulating diet vs. control diet |
3 | 152/144 | 37/62 | 28 d | OR: 0.40 (0.24-0.68) |
F (I2=0%) | Mortality reduction in those treated. |
| Peter, 2008 [56] | Corticosteroid vs. pl |
5 | 303/268 | 127/141 | Variable (Hospital- 60d)) |
OR: 0.62 (0.23-1.26) |
R (SD=0.53) | No significant survival benefit |
|
Tiruvoipati,
2008 [63] |
Prone vs. supine positioning |
4 | 662/609 | 263/246 | Variable (ICU-6 m) |
OR: 0.98 (0.70-1.30) |
R (I2=18%) | No significant survival benefit |
|
Alsaghir, 2008
[45] |
Prone vs. supine positioning |
3 | 241/225 | 113/113 | ICU | OR: 0.79 (0.45-1.39) |
R (I2=40%) | No difference in mortality. Subgroup analysis suggested a beneficial effect in patients with higher illness severity |
| 3 | 641/590 | 238/223 | 28-30 d | OR: 0.95 (0.71-1.28) |
R (I2=28%) | |||
| 4 | 662/609 | 301/279 | 90 d | OR: 0.99 (0.77-1.27) |
R (I2=10%) | |||
|
Agarwal, 2007
[43] |
Corticosteroids vs. pl (early ARDS) |
3 | 147/153 | 85/105 | Variable (Hospital / 30 d) |
OR: 0.57 (0.25-1.32) |
R (I2=53%) | No benefit in survival |
| Corticosteroids vs. pl (late ARDS) |
3 | 118/117 | 33/41 | Variable (Hospital / 30 d) |
OR: 0.58 (0.22-1.53) |
R (I2=42%) | ||
|
Adhikari, 2007
[41] |
Nitric oxide vs. pl | 9 | 577/509 | 199/162 | Hospital | 1.10 (0.94- 1.30) |
R (I2=0%) | No mortality benefit |
|
Agarwal, 2006
[44] |
Noninvasive ventilation with conventional treatment |
3 | 55/56 | 17/20 | ICU | 0.96 (0.80- 1.12) |
R (I2=0%) | No survival benefit. No difference between intratracheal instillation and aerosolized methods |
|
Davidson,
2006[49] |
Exogenous pulmonary surfactant vs. pl |
6 | 631/639 | 235/255 | 28-30 d | OR: 0.97 (0.73-1.3) |
F (NA) | Intervention did not improve survival |
|
Eichacker, 2002
[50] |
Low vs. control (higher TV and plateau pressure) tidal volumes |
2 | 461/453 | 145/189 | Variable (Hospital- 28 d) |
0.75 (0.63- 0.89)* |
NA | Significant heterogeneity in outcomes that precluded a single summary effect. |
| Low vs. control (lower TV and plateau pressure) tidal volumes |
3 | 144/144 | 70/62 | Variable (Hospital- 60 d) |
1.13 (0.88- 1.45)* |
Abbreviations: d: day, I2: heterogeneity, ICU: intensive care unit, m: month, MV: mechanical ventilation, NA: not available, OR: odds ratio, PEEP: positive end-expiratory pressure.SD: standard deviation among studies, TV: tidal volume.
*
approximate values obtained from their figure 1.
#
unless specified the value provided is risk ratio, otherwise odds ratio or risk reduction is reported.