Incidence and outcomes of acute respiratory distress syndrome in intensive care units of mainland China: a multicentre prospective longitudinal study

In 1967, Ashbaugh and colleagues [1] proposed a new syndrome in adults called acute respiratory distress syndrome (ARDS). Since then, many studies have investigated ARDS. In 2012, the Berlin ARDS definition [2] was published and found that the greater the severity of ARDS was, the higher the mortality rate would be. To decrease the mortality rate of ARDS, researchers have attempted to improve and implement respiratory support strategies, including incorporating a small tidal volume [3], high positive end-expiratory pressure (PEEP) [4], prone position ventilation [5], the lung recruitment manoeuvre [6], the use of neuromuscular blockers [7], high-frequency oscillatory ventilation (HFOV) [8, 9], and extracorporeal membrane oxygenation (ECMO) [10‐12]. Some of the techniques have resulted in excellent progress while others were still with uncertain effect. However, there is limited information on the use of these strategies in the treatment of ARDS patients, and prospective studies from mainland China are especially lacking [13, 14].

We aimed to address some clinically important questions regarding ARDS epidemiology and management in mainland China. To date, very few studies [14] have mentioned the incidence of ARDS in some regions of China, let alone the use of lung protective interventions and adjuncts. Providing insight into the use of these interventions could enable the development of more effective interventions in clinical practice.

Therefore, we undertook the CHARDS (China Acute Respiratory Distress Syndrome epidemiology) study to assess the medical/respiratory ICU epidemiology and respiratory support of ARDS and to understand how clinicians use mechanical ventilation and adjunctive interventions in routine clinical practice.

In this prospective registry study carried out in 18 ICUs in mainland China, ARDS was still an important public health problem, with a low ICU incidence, geographic variation, and high hospital mortality. The low use of recommended adjuncts and the high ratio of withdrawal of life-sustaining care, among moderate to severe ARDS patients, were also found.

In this study, the average ARDS incidence was low in the MICUs/RICUs of mainland China, and there was nearly 17-fold variation in the incidences of different ICUs. Prior epidemiological studies reported an ARDS incidence ranging from 2 to 25% of ICU patients [15‐18]. In the LUNGSAFE study [15], which included ICUs from 50 countries, the geographic variation was also high, and the incidence in Asia was lower. In contrast to other studies, our study showed a lower ratio of mild ARDS and a higher ratio of moderate ARDS relative to severe ARDS, while in the LUNGSAFE study [15], the proportions of mild, moderate, and severe ARDS were 30%, 46.6%, and 23.4%, respectively. The data included in the Berlin definition [19] also showed proportions of 22%, 64%, and 14% for mild, moderate, and severe ARDS, respectively. The reason for the relatively low incidence of ARDS and high variation incidence between different ICUs in this study is threefold. First, the patients admitted to the different ICUs had different kinds of diseases. The leading ICU received severe cases of pneumonia from all over the country and had the highest incidence of ARDS, while patients in the ICUs with lower incidences of ARDS were limited to patients with chronic respiratory diseases such as COPD or interstitial lung diseases. Second, in the LUNGSAFE study [15], Bellani et al. found that ARDS was underdiagnosed, and clinician recognition of ARDS was the lowest in patients with mild ARDS (51.3%; 95% CI, 47.5–55.0%). As the severity of ARDS increased, the proportion of clinician recognition increased accordingly. In our study, some of the patients with mild ARDS may not be recognized by clinicians in general wards or emergency room and had no chance to be transferred to an ICU, which may be both the reason of low incidence of ARDS and low ratio of mild ARDS cases. Third, there were also other ICUs in most of the included hospitals, such as surgical, general, or emergency ICUs, and some of the ARDS patients (especially for extrapulmonary ARDS patients) have been admitted there. There was also the possibility that because the screening protocol was implemented by the investigators but not the computer algorithm, some ARDS cases may have been missed by investigators. Other explanations of the low incidence may be the use of different definitions of ARDS in other studies (AECC or Berlin definition) and the prospective or retrospective designs of the studies.

Since 2000, variable ventilation strategies have been proven to be effective according to ARDS mechanical ventilation guidelines, including small tidal volume, appropriate PEEP, limit plateau pressure, and prone position ventilation. Some other strategies are inconclusive and still need more investigation, including the use of neuromuscular blockers, RMs, and ECMO. In our study, these proven or recommended approaches to mechanical ventilation and adjunctive measures appeared to be underused. The implementation of low tidal volume ventilation was in the highest agreement with the guidelines. 78.25% of the patients in this study had a tidal volume of ≤ 8 ml/kg PBW recommended as the guideline, which indicates the wide adoption of implementing a small tidal volume strategy by Chinese doctors. The other two important parameters consistent with the low tidal volume are the limitation of plateau pressure and driving pressure, which guarantee the avoidance of barotrauma. We found from the multivariate regression analysis that patients with a driving pressure > 15 cmH₂O had worse outcomes, while this was not the case for plateau pressure. This was in accordance with Amato’s study [20], which motivates us to focus more on the limitation of driving pressure. Higher PEEP was used in patients with severe ARDS compared with those with mild or moderate ARDS, as reported in prior studies [15], but concerns have been raised by the relatively low PEEP found in our study compared to the recommended PEEP in ARDSnet [3] and other studies [6]. Adjunctive measures were used infrequently for RM, prone positioning, and neuromuscular blockade. Among 343 ventilated patients with P/F < 150 mmHg within 1 week, only 85 (24.8%) patients used PPV. However, recent studies [21‐23] about RM have shown only improvement for oxygenation but no benefit for mortality. It is possible that the relatively low use of adjunctive measures reflects the uncertain effect of the treatment or the low compliance to the guidelines among clinicians. Additionally, expensive measurements such as ECMO were not widely applied in the included centres (only in 3 ICUs).

Our study showed high ICU and hospital mortality rates of patients with ARDS. In fact, the mortality rates of ARDS have remained at approximately 36–50% since the syndrome was first described [6, 15, 17, 24‐28]. The reason for the high mortality rate may be fourfold. First, there was a high incidence of withdrawal of life-sustaining care, which has seldom been reported before. When we excluded withdrawal patients, the hospital mortality rate was 34.9% (152/435). When we included withdrawal patients who died soon after withdrawal, the mortality rate was 46.3% (244/527). Suchyta et al. [29] found that withdrawal was more likely for patients older than 55 years (21/51) than for those 55 years or younger (3/32; p < 0.001). However, most of the families made such decisions due to economic reasons or based on Chinese traditions in our study. Second, there was relatively low compliance with some of the recommended guidelines. Third, the therapeutic levels varied in different regions of China. Finally, limitations in the facilities resulted in the loss of opportunities to receive further treatment, such as ECMO. We also found the hospital mortality was not different between mild and moderate group; the main reason may be that 42 in 51 mild ARDS patients progressed to moderate or severe group, and also the number of the included ARDS cases was too small to be statistically significant.

This study had a number of limitations. Although the included hospitals are general hospitals across Mainland China, most of the included ICUs are respiratory ICUs, and the risk factors may be constrained to intrapulmonary elements; therefore, the extra pulmonary element-induced ARDS patients may have been underrepresented. Additionally, the number of included ICUs was relatively small, which may have led to selection bias and could not represent the reality in mainland China, limiting the generalizability of our findings. As an epidemiological study, we could not obtain patient data from the enrolled ICUs, so it is possible that we missed some patients who met the inclusion criteria for ARDS in participating centres, especially for patients with mild ARDS. To ensure our data quality, we assigned two doctors to check the quality of the case report forms from different centres every 3 months. To standardize the inclusion procedure, we offered all the investigators web-based training and midterm conferences. Two important respiratory mechanical parameters, namely, plateau pressure and driving pressure, were reported in only 52.8% (211/400) of all IPPV patients with ARDS. The reason for so few patients undergoing this measure may be that most of them were ventilated with the spontaneous ventilation mode. Additionally, doctors did not recognize that plateau pressure and driving pressure were important parameters for ARDS patients and that refractory tachypnoea may also affect the measurement of plateau pressure and driving pressure.

Among included ICUs in Mainland China, the ARDS incidence was lower than those found in other studies. Most concerning is the high mortality and withdrawal rates of life sustaining care in patients with moderate to severe ARDS. The ARDS patients were undertreated and had a relatively low level of compliance with the guidelines. The findings indicate the potential for improvement in early recognition of mild ARDS outside the ICU and standardization of ventilation management of ARDS patients.