Acute respiratory distress syndrome (ARDS) is a heterogeneous disorder affecting 10% of ICU patients (1). The Berlin definition of ARDS classified severity based on Pao2 to Fio2 ratios into three categories: mild, moderate, and severe. These correspond to mortality rates of 27%, 32%, and 45%, respectively (2).
Low tidal volume ventilation, prone positioning, and neuromuscular blockade have been reported to improve mortality in clinical trials of ARDS (3–5), although the benefits of neuromuscular blockade were not seen in recent multicenter trial (6). All of these interventions are recommended in clinical practice guidelines (7). However, implementation of these interventions outside trial settings presents challenges which may impact their effectiveness in the community. Failure to implement low tidal volume ventilation in clinical practice was documented repeatedly in multiple studies for many years following publication of evidence of its effectiveness (8,9).
Ventilation in the prone position was shown in a randomized controlled trial (Proning Severe Acute Respiratory Distress Patients [PROSEVA] trial) of moderate to severe ARDS in French ICUs with greater than 5 years’ experience using this technique to reduce mortality from 32% in the supine position group to 16% in the prone position group (5). After publication of this trial, we began to implement mechanical ventilation in the prone position for ARDS in our ICU in patients similar to those shown to benefit in this trial.
We sought to describe our early outcomes from this quality improvement project and determine how our early experience compared with the result of the PROSEVA trial patients in the prone group.
MATERIALS AND METHODS
Our institutional interprofessional clinical practice guideline for placing patients in prone position has been described previously (10). This guideline addressed indications for, and logistics of placing patient in prone position. Nurses and respiratory therapists practiced placing patients in prone position in a simulation environment. An institutional guideline on lung-protective ventilation for ARDS was also in place. However, actual treatment decisions were at the discretion of treating clinicians.
We then conducted a retrospective cohort study of patients who underwent mechanical ventilation in the prone position in our medical ICU between June of 2013 and October of 2016. Patients were identified by respiratory therapy charting. Patients in other ICUs, not meeting Berlin definition for ARDS, or who on chart review were found not to have undergone prone positioning were excluded.
Demographic and clinical data were abstracted from the electronic health record and entered into a Research Electronic Data Capture (REDCap) database (11). Collected clinical data included ICU admission diagnosis; cause of ARDS; comorbidities; use of vasopressors, inhaled nitric oxide, and dialysis; do-not-resuscitate (DNR) status; hospital mortality; physical therapy assessments; and complications such as skin wounds. DNR status means that there was an order in the electronic record instructing the nurses not to perform cardiopulmonary resuscitation in the event of cardiac arrest. In our institution, the decision to write this order is often accompanied by additional decisions to limit or withdraw aggressive ICU care, which would be separately documented in a note. However, for this project, only the DNR order was captured for analysis.
Descriptive statistics were used to characterize the cohort. Standardized differences were used to compare characteristics of PROSEVA patients in the prone group and Christiana Care patients. Continuous outcomes between Christiana Care patients and PROSEVA patients in the prone group were compared using unpaired t tests, while categorical outcomes were compared using z tests. Analysis was performed with SAS software (version 9.4, SAS Institute Inc., Cary, NC).
The study was approved by the Christiana Care Health System Institutional Review Board.
We identified 81 patients who underwent mechanical ventilation in the prone position in our medical ICU during the study period. The most common mode of ventilation was assist control volume control (56 patients) followed by assist control pressure control (22 patients). One patient underwent high frequency oscillatory ventilation, five patients received airway pressure release ventilation, and another five patients had more than one ventilator mode used. The median time between intubation and the first prone position session was 23.8 hours with a range from 2.1 hours to 411 hours (17 d). The duration of the first proning session in our cohort ranged from 1.5 to 40.5 hours with an average of 17.5 ± 7.9. The mean tidal volume among patient in assist control volume control was 5.8 mL/kg predicted body weight.
As shown in Table 1, on average the Christiana Care patients received lung-protective tidal volumes and a mean positive expiratory pressure of 14 centimeters of water. The mean Pao2/Fio2 was in the severe ARDS range at 99. More than half of the patients received steroids and neuromuscular blockade and the frequency of comorbid psychiatric disease was very high in these patients. The standardized difference calculations indicate that while some of the patients’ characteristics between Christiana Care and PROSEVA are small (|Difference| < 0.2) such as Pao2/Fio2 ratio others are much larger such as sepsis (|Difference| > 1.3) (12).
There were no accidental extubations during proning. Only three patients were identified as having skin wounds during their hospitalizations. All three patients had wounds on the sacrum/coccyx and one also had a wound at his tracheostomy site. Based on the locations of injury, these did not appear to be related to prone positioning. There were no endotracheal tube dislodgements. Eight patients had dislodgement of IV access including peripheral, central, and arterial lines. Thirteen patients were noted to have dependent edema of the eyes, lips, and tongue. We found that 38 out of 46 survivors (83%) of our patients were able to ambulate prior to discharge from the hospital.
Although guidelines and meta-analysis support the effectiveness of mechanical ventilation in the prone position for patients with moderate to severe ARDS, reports of outcomes from implementation of this practice outside of trials are rare (7,13). To better understand our outcomes, we systematically compared our cohort of proned patients to the subset of PROSEVA trial patients randomized to prone positioning. As shown in Table 1, characteristics of our patients were similar to those in the PROSEVA trial. However, PROSEVA patients had higher prevalence of sepsis diagnosis (82.2% vs 12.3%) and lower prevalence of pneumonia (62.4% vs 74.1%). PROSEVA patients also had lower mean positive end-expiratory pressure (10 cm of water vs 14 cm of water) upon proning initiation. It seemed that neuromuscular blockade was used more frequently in PROSEVA study patients in the prone group (91.0% vs 85.2%) and steroids less frequently (39.6% vs 64.2%). Interestingly, as many as 37% of our cohort had comorbid psychiatric disease or substance abuse. The frequency of these comorbidities was not reported in the PROSEVA trial.
Overall mortality was lower in the PROSEVA study prone group patients than in our cohort (16.0% vs 43.2%), however, after excluding patients with a DNR order, the mortality of our cohort (13.7%) was very similar to that of the PROSEVA patients. This raises the question about which patient factors need to be accounted for when looking at ARDS mortality rates in the medical literature.
The average duration of the first proning session in our cohort was 17.5 ± 7.9, which is similar to the average duration of 17 hours per day in the PROSEVA trial. As shown in Table 2, among survivors, time to extubation and length of stay were shorter in our cohort.
Although many characteristics of both patient populations were similar, there are several important differences. First, the PROSEVA trial had a 12–24 hour stabilization period prior to randomization and only included patients within 36 hours of ARDS diagnosis (5). In contrast, we included all patients who underwent mechanical ventilation in the prone position during the study period regardless of the ARDS duration (including one patient who was placed in the prone position after 17 d of mechanical ventilation), and without any stabilization period. Additionally, we did not exclude patients who had treatment limitations, whereas these patients were excluded from the PROSEVA trial if such limitations were in place at the time of randomization (5).
The PROSEVA trial also excluded patients on home oxygen or chronic noninvasive ventilation, noninvasive ventilation for more than 24 hours prior to enrollment, and chronic disease with life expectancy less than 1 year; while our observational trial of patients actually treated with prone positioning included such patients (5). There also may be cultural differences in end-of-life care discussions between our institution in the United States and the European ICUs involved in the trial. The absence of ventilator-associated pneumonia diagnoses in our cohort compared with the 22.8% incidence in the PROSEVA study is almost certainly due to cultural and economic factors influencing definition of this syndrome in the United States versus Europe.
Although the above differences may explain why we could not achieve results similar to PROSEVA, our data suggest hypotheses for future work. For example, the large number of patients in our group whose first proning session was less than 16 hours represents a potential area for quality improvement efforts. The same is potentially true for patients who were not placed in the prone position for many days after intubation. Unfortunately, we did not record the reason for early termination of the first proning session or the time of ARDS diagnosis. The seemingly increased steroid use in our cohort combined with the higher mortality are interesting in the context of some research suggesting that steroids are beneficial in ARDS. Additionally, the higher frequency of pneumonia in our population raises the possibility that our cohort of ARDS patients may have had more pulmonary rather than nonpulmonary causes of ARDS, perhaps representing distinct phenotypes within this syndrome (14,15). In addition to pulmonary and nonpulmonary etiologies of ARDS, multiple other approaches to phenotyping ARDS have been proposed (16). In addition to their often-discussed repercussions for clinical trials, further exploration of ARDS phenotypes have implications for future improvement science and quality improvement work.
It is encouraging that the majority of survivors in our population were able to ambulate prior to hospital discharge. This outcome was achieved despite the use of neuromuscular blockade and steroids in the majority of patients. It may reflect the young age of our patient population. The low incidence of complications from mechanical ventilation in the prone position also suggest that this intervention can be implemented safely.
Our study has several limitations. First, as a retrospective cohort study of only proned patients, we do not know how many patients who were candidates for prone positioning were not proned or why those decisions were made. It is also possible that the respiratory therapy documentation failed to capture some patients who were proned. Additionally, we do not know the details of goals of care discussions between the treating team and patients’ family which likely had a significant impact on outcomes, particularly on mortality. Finally, our results may not be generalizable to other institutions or other types of ICUs.
In summary, overall mortality in our cohort was higher than that observed in the PROSEVA trial, but when excluding those with treatment limitation, the survival was similar. Although patient characteristics were comparable in both cohorts, differences between the populations related to timing of enrollment, end-of-life decision making, and clinical care may explain the differences. Despite this, we feel that examining prone positioning outcomes for ARDS patients in a community setting has value for planning future studies and quality improvement. Our study reflects the challenges of implementing in community settings recommendations from clinical trials, conducted in very controlled environments with highly selected patients. Future research to determine optimal timing of prone initiation and how end-of-life decisions affect ARDS data would be of interest.
1. Bellani G, Laffey JG, Pham T, et al.; LUNG SAFE Investigators; ESICM Trials GroupEpidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries.JAMA2016315788–800
2. Ranieri VM, Rubenfeld GD, Thompson BT, et al.; ARDS Definition Task ForceAcute respiratory distress syndrome: The Berlin definition.JAMA20123072526–2533
3. Brower RG, Matthay MA, Morris A, et al.; Acute Respiratory Distress Syndrome NetworkVentilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome.N Engl J Med20003421301–1308
4. Papazian L, Forel JM, Gacouin A, et al.; ACURASYS Study InvestigatorsNeuromuscular blockers in early acute respiratory distress syndrome.N Engl J Med20103631107–1116
5. Guérin C, Reignier J, Richard JC, et al.; PROSEVA Study GroupProne positioning in severe acute respiratory distress syndrome.N Engl J Med20133682159–2168
6. Moss M, Huang DT, Brower RG, et al.; National Heart L, Blood Institute PCTNEarly neuromuscular blockade in the acute respiratory distress syndrome.N Engl J Med20193801997–2008
7. Fan E, Del Sorbo L, Goligher EC, et al.; American Thoracic Society, European Society of Intensive Care Medicine, and Society of Critical Care MedicineAn official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine clinical practice guideline: Mechanical ventilation in adult patients with acute respiratory distress syndrome.Am J Respir Crit Care Med20171951253–1263
8. Bourdeaux CP, Thomas MJ, Gould TH, et al. Increasing compliance with low tidal volume ventilation in the ICU with two nudge-based interventions: Evaluation through intervention time-series analyses.BMJ Open20166e010129
9. Needham DM, Colantuoni E, Mendez-Tellez PA, et al. Lung protective mechanical ventilation and two year survival in patients with acute lung injury: Prospective cohort study.BMJ2012344e2124
10. Mitchell DA, Seckel MA. Acute respiratory distress syndrome and prone positioning
.AACN Adv Crit Care201829415–425
11. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support.J Biomed Inform200942377–381
12. Kim HY. Statistical notes for clinical researchers: Effect size.Restor Dent Endod201540328–331
13. Munshi L, Del Sorbo L, Adhikari NKJ, et al. Prone position for acute respiratory distress syndrome. A systematic review and meta-analysis.Ann Am Thorac Soc201714S280–S288
14. Gattinoni L, Pelosi P, Suter PM, et al. Acute respiratory distress syndrome caused by pulmonary and extrapulmonary disease. Different syndromes?Am J Respir Crit Care Med19981583–11
15. Luo L, Shaver CM, Zhao Z, et al. Clinical predictors of hospital mortality differ between direct and indirect ARDS.Chest2017151755–763
16. Sinha P, Calfee CS. Phenotypes in acute respiratory distress syndrome: Moving towards precision medicine.Curr Opin Crit Care20192512–20