To the Editor:
Prone ventilation improves oxygenation in most, but not all, patients with the acute respiratory distress syndrome (ARDS) (1–4). Until recently, however, this improvement has not been associated with improved survival. The ARDS Management Arms (ARMA) study comparing low with standard tidal volume ventilation actually found that survival was better in subjects who had worse oxygenation (5).
Two previous studies concluded that the changes in gas exchange that occurred with prone ventilation did not predict survival in ARDS (6, 7) but both only used prone ventilation 5–8 h/d, and neither found that prone ventilation improved overall survival. Recently, Guérin and colleagues (8) reported that prone ventilation for an average of 16 h/d improved oxygenation and reduced the mortality of patients with ARDS by 50%. We sought to determine if the improvements in gas exchange they observed predicted survival.
This was a retrospective analysis of data collected prospectively by Guérin and colleagues (8). Inclusion criteria are described in the original article. Arterial blood gases (ABG) were analyzed prior to turning prone, 1 hour after turning, and after completing the first session of prone ventilation while the patients were still prone. Changes in ABGs were examined relative to 28-day mortality. The PaO2/FiO2 (P/F) ratio and PaCO2 were analyzed as means ± SD, in quintiles of units or mm Hg, respectively, in quintiles of subjects, by classifying subjects as “responders” or “non-responders” per Gattinoni and colleagues (6) (i.e., an increase in P/F of ≥20 mm Hg or a decrease in PaCO2 of ≥1 mm Hg), and in subgroups based on changes in the P/F ratio coupled with changes in the PaCO2 per Lemasson and colleagues (7) (i.e., an increase in P/F of ≥20% or <20% and a decrease in PaCO2 of ≥1 or <1 mm Hg).
Analyses were performed using SAS Enterprise Guide 4.3 (SAS Institute, Inc., Cary, NC) or SPSS Statistics 17.0 (SPSS, Chicago, IL). Data were compared using paired or unpaired t tests and chi-squared analysis. P < 0.05 was considered significant. Gas exchange variables measured prior to turning prone that were significantly different between survivors and those who died were analyzed by receiver operating characteristic (ROC) curves.
Paired ABGs were available for 232 of 237 subjects (98%) who received prone ventilation for 1 hour, and for 228 of 237 subjects (96%) at the completion of the first prone session; 194 (82%) survived. Gas exchange prior to instituting prone ventilation was similar in those who survived and those who died with the exception that survivors had a higher pH (7.33 ± 0.10 vs. 7.27 ± 0.14, P = 0.003), a lower PaCO2 (49 ± 13 vs. 55 ± 17, P = 0.036), and a slightly higher FiO2 (0.78 ± 15 vs. 0.84 ± 16, P = 0.021). The area under the ROC curve for pH was 0.62; at a pH of 7.29 (the pH of maximum sensitivity and minimum 1 − specificity), the sensitivity of pH for predicting survival was 0.67 and the specificity was 0.53. The area under the ROC curve for PaCO2 was 0.59; at a PaCO2 of 45 mm Hg (the maximum sensitivity and minimum 1 − specificity), the sensitivity of PaCO2 for predicting survival was 0.44 and the specificity was 0.68. Accordingly, these differences in gas exchange seen prior to initiating prone ventilation were neither sensitive nor specific for predicting survival.
After 1 hour of prone ventilation, subjects who survived had a higher pH, but no difference was observed with respect to any change in ABGs on turning prone between subjects who survived or died (Table 1). The P/F ratio decreased on turning prone in 43 subjects at 1 hour, and 39 (91%) of these survived. Of the 38 subjects who died, 4 (11%) had P/F ratios that decreased an average of −25 ± 35 mm Hg, and of the 194 subjects who survived, 39 (20%) had P/F ratios that decreased an average of −14 ± 14 mm Hg after turning prone (P = 0.25). No difference in survival was observed when examining the change in P/F or the change in PaCO2 by quintiles of change (P = 0.41 and P = 0.21, respectively), by quintiles of subjects (P = 0.55 and P = 0.94, respectively, data not shown), or when combining the change in P/F with the change in PaCO2 (P = 0.670) (Table 1).
|Variable||Survived (N = 194)||Died (N = 38)||P Value|
|pH, units||7.34 ± 0.10||7.27 ± 0.12||0.004|
|Δ from pre–prone positioning||0.01 ± 07||0.01 ± 0.05||0.764|
|PaCO2, mm Hg||49 ± 14||52 ± 12||0.149|
|Δ from pre–prone positioning||−0.5 ± 9.4||−2.7 ± 9.1||0.182|
|PaO2, mm Hg||119 ± 65||118 ± 59||0.950|
|Δ from pre–prone positioning||38.9 ± 63||38.8 ± 59||0.987|
|FiO2, %||74 ± 16||78 ± 17||0.205|
|Δ from pre–prone positioning||−3.9 ± 11.4||−6.5 ± 11.6||0.220|
|PEEP, cm H2O||12 ± 3||12 ± 2||0.654|
|Δ from pre–prone positioning||−1.6 ± 3.0||−1.3 ± 3.2||0.602|
|P/F, mm Hg||166 ± 83||152 ± 62||0.321|
|Δ from pre–prone positioning||60 ± 79||55 ± 60||0.618|
|P/F increase, mm Hg||0.78|
|≥ 20, N (%)||123 (85)||25 (17)|
|< 20, N (%)||71 (85)||13 (15)|
|PaCO2 decrease, mm Hg||0.86|
|≥ 1, N (%)||99 (83)||20 (17)|
|< 1, N (%)||95 (84)||18 (16)|
|Survival by quintile of P/F response after 1 h of prone ventilation, mm Hg, survivors/total patients (%)||0.410|
|−86 to −1||39/43 (91)|
|0 to 85||99/121 (82)|
|86 to 171||37/47 (78)|
|172 to 257||13/15 (87)|
|258 to 343||6/6 (100)|
|Survival by quintile of change in PaCO2 after one hour of prone ventilation, mm Hg, survivors/total (%)||0.210|
|−30 to −12||13/20 (65)|
|−11 to + 7||160/189 (85)|
|+8 to +26||16/18 (89)|
|+27 to +45||4/4 (100)|
|+45 to + 64||1/1 (100)|
|Survival by combined P/F and PaCO2 response, survivors/total (%)||0.670|
|+P/F response, no PaCO2 response||15/16 (94)|
|+P/F response, +PaCO2 response||18/23 (78)|
|No P/F response, no PaCO2 response||80/97 (82)|
|No P/F response, +PaCO2 response||81/97 (87)|
ABGs prior to, and at the completion of, the first session of prone ventilation were similar in those who survived and those who died with the exception that those who died had a lower pH (7.37 ± 0.08 vs. 7.30 ± 0.12, P < 0.0001). No differences between the changes in ABGs in survivors versus those who died were found by any method of analysis (data not shown).
This is the first study assessing the relationship between the effect of prone ventilation on gas exchange and survival in a population of patients with ARDS in which prone ventilation improved survival. Although gas exchange improved and mortality was reduced, we found no association between the improvement in gas exchange and survival.
We suggest that prone ventilation improves survival in ARDS by reducing ventilator-induced lung injury (VILI) as first theorized in 1997 (9), three years prior to publication of the ARMA study that reached the same conclusion with respect to low tidal volume ventilation (8). VILI is thought to occur as a result of cyclical airspace opening and closing and/or from lung overdistension. Prone positioning reduces the gravitational pleural pressure gradient (10–13). This results in a more uniform distribution of end-expiratory lung volume and, accordingly, a more uniform distribution of tidal volume, which, in turn, reduces cyclical airspace opening and closing (14) and the surfactant depletion that occurs as a result of larger volume excursions (15) in dependent lung regions, and also reduces the overdistension that occurs in nondependent regions.
We conclude that the increase in survival seen in patients with ARDS who receive prone ventilation does not depend on whether the change in position improves gas exchange and infer that it results from the ability of prone positioning to reduce VILI. Accordingly, we suggest that prone ventilation should not be considered as salvage therapy for patients with severe hypoxemia but that it should be routinely initiated early in ARDS, certainly when the P/F ratio is ≤150, and continued until the P/F ratio exceeds 150 per the results of Guérin and colleagues (8). Numerous laboratory studies suggest that instituting prone ventilation at even higher P/F ratios could be beneficial (15).
|1.||Gattinoni L, Tognoni G, Pesenti A, Taccone P, Mascheroni D, Labarta V, Malacrida R, Di Giulio P, Fumagalli R, Pelosi P, et al.; Prone-Supine Study Group. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001;345:568–573.|
|2.||Guerin C, Gaillard S, Lemasson S, Ayzac L, Girard R, Beuret P, Palmier B, Le QV, Sirodot M, Rosselli S, et al. Effects of systematic prone positioning in hypoxemic acute respiratory failure: a randomized controlled trial. JAMA 2004;292:2379–2387.|
|3.||Mancebo J, Fernández R, Blanch L, Rialp G, Gordo F, Ferrer M, Rodríguez F, Garro P, Ricart P, Vallverdú I, et al. A multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care Med 2006;173:1233–1239.|
|4.||Taccone P, Pesenti A, Latini R, Polli F, Vagginelli F, Mietto C, Caspani L, Raimondi F, Bordone G, Iapichino G, et al.; Prone-Supine II Study Group. Prone positioning in patients with moderate and severe acute respiratory distress syndrome: a randomized controlled trial. JAMA 2009;302:1977–1984.|
|5.||The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301–1308.|
|6.||Gattinoni L, Vagginelli F, Carlesso E, Taccone P, Conte V, Chiumello D, Valenza F, Caironi P, Pesenti A; Prone-Supine Study Group. Decrease in PaCO2 with prone position is predictive of improved outcome in acute respiratory distress syndrome. Crit Care Med 2003;31:2727–2733.|
|7.||Lemasson S, Ayzac L, Girard R, Gaillard S, Pavaday K, Guérin C. Does gas exchange response to prone position predict mortality in hypoxemic acute respiratory failure? Intensive Care Med 2006;32:1987–1993.|
|8.||Guérin C, Reignier J, Richard J-C, Beuret P, Gacouin A, Boulain T, Mercier E, Badet M, Mercat A, Baudin O, et al.; PROSEVA Study Group. Prone positioning in severe acute respiratory distress syndrome. N Engl J Med 2013;368:2159–2168.|
|9.||Albert RK. The prone position in acute respiratory distress syndrome: where we are, and where do we go from here. Crit Care Med 1997;25:1453–1454.|
|10.||Lai-Fook SJ, Rodarte JR. Pleural pressure distribution and its relationship to lung volume and interstitial pressure. J Appl Physiol (1985) 1991;70:967–978.|
|11.||Mutoh T, Guest RJ, Lamm WJE, Albert RK. Prone position alters the effect of volume overload on regional pleural pressures and improves hypoxemia in pigs in vivo. Am Rev Respir Dis 1992;146:300–306.|
|12.||Lamm WJE, Graham MM, Albert RK. Mechanism by which the prone position improves oxygenation in acute lung injury. Am J Respir Crit Care Med 1994;150:184–193.|
|13.||Albert RK, Hubmayr RD. The prone position eliminates compression of the lungs by the heart. Am J Respir Crit Care Med 2000;161:1660–1665.|
|14.||Cornejo RA, Díaz JC, Tobar EA, Bruhn AR, Ramos CA, González RA, Repetto CA, Romero CM, Gálvez LR, Llanos O, et al. Effects of prone positioning on lung protection in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 2013;188:440–448.|
|15.||Albert RK. The role of ventilation-induced surfactant dysfunction and atelectasis in causing acute respiratory distress syndrome. Am J Respir Crit Care Med 2012;185:702–708.|