Background
Severe complicated influenza including pneumonia, myocarditis and neurologic complications are still a burden on intensive care units (ICU) nowadays, especially viral or secondary bacteria pneumonia-induced acute respiratory distress syndrome (ARDS) [
1,
2]. During the winter season in 2016, there was an outbreak of influenza in Taiwan. Totally, 1735 subjects were admitted to ICUs due to severe complicated influenza pneumonia according to the data from the Centers for Disease Control of Taiwan [
3]. Patients with influenza pneumonia needing mechanical ventilation were at high risk of rapid progression to ARDS. For the 2009 pandemic H1N1 virus infection, 49–72% of patients admitted to ICUs had complications with ARDS [
4,
5].
There are several therapeutic options for refractory hypoxemia in patients with severe ARDS [
6,
7], but only a few options have been confirmed with clinical validity by previous studies, including higher positive end-expiratory pressure (PEEP) [
8,
9], lower tidal volume [
10], neuromuscular blocking agents [
11] and prone positioning [
12]. Prone positioning was first suggested in 1974 [
13]; however, the clinical benefit of prone positioning in patients with ARDS was not confirmed until 2013 when the PROSEVA study showed decreased 28-day and 90-day mortality and increased ventilator-free days only when it was started early and there were sufficiently long sessions [
12]. Further, meta-analysis by Cochrane database also revealed that prone positioning would reduce the mortality rate when used with lung-protective strategies and longer duration in patients with severe ARDS [
14,
15].
Few studies have explored the effect of prone positioning focused on influenza pneumonia-related ARDS patients. Xu et al. [
16] studied H7N9 influenza patients with prone positioning, and decrease in carbon dioxide retention was noted, but no clinical outcome was mentioned. Moreover, what factors that can predict the efficacy of prone positioning in severe ARDS are not entirely clear [
17].
The aim of this study is to investigate the survival predictors of prone positioning in patients with severe ARDS caused by influenza pneumonia.
Discussion
The aim of this multicenter retrospective study was to evaluate the effect of prone positioning focusing on patients with influenza pneumonia-related ARDS. After multivariate Cox regression analysis, PSI, renal replacement therapy and increased dynamic driving pressure were associated with 60-day mortality in patients with influenza pneumonia-related ARDS receiving prone positioning.
Most of the studies evaluating the effect of prone positioning were in ARDS patients with heterogeneous risk factors [
14,
15]. For specific conditions such as burns, prone positioning has been demonstrated to safely implement and improve oxygenation (in burn patients with severe ARDS) in a burn intensive care unit [
24]. The present study was more homogenous and specific to patients with ARDS caused by influenza pneumonia. Systematic review and meta-analysis studies in prone positioning have revealed decreased mortality in patients with severe acute hypoxemic respiratory failure, but not in less severe hypoxemia. Survival benefits were noted using a range of PaO
2/FiO
2 ratio thresholds up to approximately 140 mm Hg [
25] or less than 200 mm Hg [
26]. In the present study, the PaO
2/FiO
2 ratio was 95.9 ± 54.5 mm Hg before prone positioning. However, the PaO
2/FiO
2 ratio was not significantly different between 60-day survivors and 60-day non-survivors (102.3 ± 59.8 mm Hg vs. 81.3 ± 37.6 mm Hg,
p = 0.153). In terms of the response of prone positioning to ARDS, the different entities of the risk factor possibly produce different outcomes. In addition to severity of hypoxemia, further clinical trials would assist in clarifying the survival benefits of prone positioning in the specific risk factors.
Some studies have shown that acute kidney injury (AKI) was common and an independent risk factor for mortality in patients with influenza A [
27‐
30]. In patients with severe ARDS caused by H1N1 influenza pneumonia, a recent study also revealed AKI was common and demonstrated significantly increased mortality [
31]. The 53% mortality rate among the 38 patients requiring renal replacement therapy was significantly higher than the 0% mortality rate among the 19 patients not requiring renal replacement therapy. The present study in patients receiving prone positioning caused by influenza pneumonia-related ARDS demonstrated that the requirement for renal replacement therapy had nearly 6 times the mortality rate (hazard ratio 6.248) than patients not requiring renal replacement therapy. In order to reduce the mortality in patients with severe ARDS caused by H1N1 influenza pneumonia, it is important to prevent development of AKI and need for renal replacement therapy by avoiding nephrotoxic agents and supplying sufficient renal perfusion and oxygenation.
Amato and colleagues analyzed 9 randomized controlled trials in ARDS patients and demonstrated that driving pressure was the strongest predictor of mortality [
32]. A secondary analysis of data from 787 ARDS patients enrolled in two independent randomized controlled trials revealed that when ventilating patients with low tidal volume, driving pressure was a risk factor for death in ARDS patients, as was plateau pressure or compliance of respiratory system [
33]. Airway driving pressure was significantly related to lung stress and could detect lung over-stress with acceptable accuracy (
r2 = 0.581
p < 0.0001 and
r2 = 0.353
p < 0.0001 at 5 and 15 cm H
2O of PEEP) in ARDS patients [
23]. Furthermore, the APRONET study on prone positioning of ARDS patients found that prone positioning was associated with low complication rates, significant increase in oxygenation, and a significant decrease in driving pressure [14 (11–17 cm H
2O) to 13 [
10‐
16] cm H
2O,
p = 0.04] [
34]. Our previous study for severe ARDS patients with ECMO revealed that higher dynamic driving pressure [hazard ratio 1.070 (1.026–1.116),
p = 0.002] during the first 3 days of ECMO was one of the factors independently associated with ICU mortality [
35]. The present study in influenza pneumonia-related ARDS patients receiving prone positioning also found that increased dynamic driving pressure (hazard ratio 1.372, 95% confidence interval 1.095–1.718;
p = 0.006) was identified as one of the independent predictors associated with 60-day mortality. It was suggested that ventilatory support with lung-protective strategy with low tidal volume and optimal PEEP level be applied, and these be then adjusted according to the driving pressure, ideally less than 15 cm H
2O, although this limit should be addressed in future studies [
36]. Despite some studies associating driving pressure with physiological and clinical outcomes, it is necessary to evaluate the driving pressure as a primary end point during ventilatory setting in ARDS patients in the near future.
The LUNG SAFE study showed that the use of prone positioning actually depended on the severity of hypoxemia, from 1% in mild to 5.5% in moderate and to 16.3% in severe ARDS [
37]. A prospective international prevalence study (the APRONET study, ARDS Prone Position Network) found that the rates of prone positioning were up to 5.9%, 10.3% and 32.9% in mild, moderate and severe ARDS [
30]. In our study, the rates of prone positioning were 18%, 15% and 31% in mild, moderate and severe ARDS, respectively. The substantially different rates in the use of the prone positioning may reflect the management bias of prone positioning in patients with ARDS between the different studies. Furthermore, among our eight involved hospitals, the rate of prone positioning varied from 0% (0/37) to 67% (2/3) and the bias even existed between different hospitals in the same study. It is important to be homogenous on the indication and management in the selected prone position as one of the standard interventions in severe ARDS.
This study has some limitations. Firstly, since this study is retrospective, some patients or data might be missing. Secondly, the primary end point of this study was 60-day mortality, and the value of computed power was 0.585. This was a retrospective study, and 65 patients with severe ARDS receiving prone positioning were analyzed. Although more patients were needed to increase the power of this study, the limitation was from the nature of retrospective study within a 3-month period. Thirdly, prone positioning is not a routine intervention in the management of ARDS and has no standard procedure such as how many hours a day, how to perform it or how to protect the patients. In this study, even though every patient had prone positioning for more than 16 h a day, the exact duration showed little difference between each hospital. Fourthly, the change in physiological measurements pertains to a difference between supine and prone position, and hence, the impact of chest wall is not taken into account. Finally, in this study, we focused on influenza-related ARDS patients, and whether the result can be extrapolated to all patients with ARDS is unknown, requiring further investigation. To confirm the benefit of prone positioning in ARDS especially in influenza pneumonia, further prospective randomized control studies are needed with strict standard procedures and patient selection.
Authors’ contributions
KCK, KWC, MCC and KYY had the idea of the study and made substantial contributions to conception and design, and drafted the manuscript. SJL, YCC and KYY made substantial contributions to conception and design, and analysis and interpretation of data, and reviewed the manuscript. HCW, LCC, WCP and CKP were involved in drafting the manuscript and revising it critically for important intellectual content. HCH, WCC, HCW and WFF made substantial contributions to acquisition of data, and analysis and interpretation of data. YMC, HCW and CCS made substantial contributions to conception and design, acquisition of data and analysis and interpretation of data. CLW, MJT, KCK and KYY made substantial contributions to conception and design, and analysis and interpretation of data and reviewed the manuscript. All authors read and approved the final manuscript.