Incidence and outcomes of acute lung injury in the surgical intensive care unit of a tertiary care hospital in Bangkok, Thailand

Our surgical intensive care unit (SICU) is a 14-bed facility in a hospital that has 2110 beds. The majority of the patients admitted to the SICU are from general surgical wards or operating theatres. Neurosurgical, cardiovascular–thoracic, and trauma patients are admitted to other, independent ICUs belonging to the respective subspecialty departments. Our SICU is a closed unit covered by intensivist and multidisciplinary consultation teams. The nurse-to-patient ratio is 1:1 during day and 1:2 during night. The teams of physicians consist of 1 or 2 critical care fellows and 4 to 5 anesthesiology residents who are supervised by an attending physician. The outline of this observational cohort study is shown in Figure 1.

Figure 1

This study was prospective and observational. After receiving approval from the institutional review board of Siriraj Hospital (approval No. Si 271/2010, 31 May 2010), we screened all surgical patients aged ≥18 years, who were admitted to the general SICU during the study period to May 31, 2013. Patients who required ventilatory support for ≥24 h were included. The investigators excluded those patients who were terminally ill and had received permanent ventilatory assistance before their admission to the SICU. The included patients or their family members provided written informed consent to participate in the study. All data collection was continued throughout the ICU stay.

The AECC definition of ALI and ARDS was used for incidence reporting in this study. The diagnostic criteria were the presence of acute hypoxemia with a ratio of partial pressure of arterial oxygen to the fraction of inspired oxygen (pulmonary arterial oxygen tension [PaO₂]/fraction of inspired oxygen [FiO₂]) of 300 mmHg or less (for acute lung injury) or of 200 mmHg or less (for ARDS): bilateral infiltrates seen on a frontal chest radiograph and no clinical evidence of left atrial hypertension or (if measured) pulmonary artery occlusion pressure of 18 mmHg or less. A diagnosis of pneumonia was made if the patients met all the following criteria within 3 days. (1) New infiltrate or cavitation with air-fluid level persisting for at least 24 h. (2) Fever (>38.3°C) or hypothermia (<36.0°C) and white blood cell count >10⁵ or <4 × 10³ or >25% increase above last value or bands >10%. (3) Bacteriologic confirmation (demonstrated by at least one) of positive blood culture with same organism identified in sputum or other respiratory culture, or protected brush specimen with >10³ colony forming unit (cfu)/mL pathogen, or bronchoalveolar lavage with >10⁴ cfu/mL pathogen, or nonbronchial bronchoalveolar lavage with >10³ cfu/mL pathogen, or positive gram stain with >3+ of one type of bacteria, or positive semiquantitative sputum culture with >3+ of growth of one type of pathogenic organism (if not quantitative, then it must have moderate or heavy growth).

Of the 2,523 patients admitted to the SICU, 495 (20%) required mechanical ventilatory support for >24 h during the study period (June 1, 2010 to May 31, 2013). Among the age group <18 years, terminally ill patients and those who required mechanical ventilator support for <24 h, 2028 patients were excluded (Figure 1). The reasons for exclusion are shown in Figure 1. Sixteen patients (3%) had ALI/ ARDS on the day of admission. Of the 495 patients without ALI/ARDS on the day of admission, 15 (3%) developed ALI/ARDS. The onset of ALI/ARDS was at day 2 approximately (IQR 1, 6). The causes of ALI and ARDS were sepsis (54%) and pneumonia (46%), with an average PaO₂/FiO₂ ratio on the day of ALI and ARDS of 183 ± 63.4.

Demographic data, including age, gender, body mass index, race, smoking, and American Society of Anesthesiologists Physical Status were not significantly different between the ALI/ARDS group and the non-ALI/ARDS group. There were no significant differences for the majority of patients in both groups in terms of past medical history, except for the higher percentage of end stage renal disease in the ALI/ARDS group (27% vs 6%, P = 0.01). Additionally, the patients who developed ALI/ARDS had significantly higher APACHE II scores (19.4 ± 8.8 vs 13.6 ± 6.6, P = 0.001) and developed sepsis more frequently at the SICU admission (60% vs 34%, P = 0.05). Only half of the patients in ALI/ARDS group were postoperative, whereas nearly 80% of the patients in the non-ALI/ARDS group had undergone operations (P = 0.05) (Table 1).

There was no significant difference in the hemoglobin level at SICU admission between the groups. Patients who developed ALI/ARDS received more red blood cell transfusions (1609 mL vs 644 mL, P = 0.03). Positive fluid balance during the first 24 h (P = 0.08) and fresh frozen plasma administration (P = 0.06) were not significantly different between the groups (Table 1). Regarding ventilatory support data, patients who developed ALI/ARDS received lower tidal volume on day 1 (7.7 ± 1.6 vs 9.3 ± 2.0 mL per kg, P = 0.003) and higher positive end expiratory pressure (PEEP) on day 2 (6.8 ± 2.0 vs 5.3 ± 1.6 cmH₂O, P = 0.008). Controlled mechanical ventilation (CMV) and synchronized intermittent mechanical ventilation were the only two modes of ventilatory support implemented for all patients with ALI and ARDS (Table 2). Pneumonia and acute kidney injury were more severe among the patients with ALI and ARDS (40% vs 9%, P = 0.002 and 33% vs 10%, P = 0.02 respectively). There were no differences in the incidence of sepsis, septic shock, gastrointestinal bleeding and myocardial ischemia between the two groups (Table 3).

The patients who developed ALI/ARDS demonstrated higher numbers of ventilator days (median, 17 days vs 5 days, P = 0.002) and longer SICU length of stay (20 days vs 7 days, P = 0.01). In addition, patients in ALI/ARDS group showed higher in-hospital mortality and mortality at 30 day (60% vs 23.5%, P = 0.003 and 47% vs 21%, P = 0.03, respectively) (Table 3).

Diagnosis of ARDS involves a combination of various criteria including clinical findings, oxygenation, hemodynamics, and radiographic criteria. These allow the inclusion of a highly heterogeneous group of patients who showed various types of lung injuries, which resulted in the large discrepancy in incidence of ALI/ ARDS. Although AECC criteria have been used to identify ALI/ARDS for decades, the use of these criteria has been controversial for many years [13]. The reported incidence of ALI/ARDS using the AECC definition showed a wide disparity among countries and centers.

The present study demonstrated a lower incidence of ALI/ARDS (3% or 5 new cases per year) comparing with the data reported in previous studies [6]. Two recently published studies from the USA reported the incidence of ALI/ARDS, using the AECC definition. Rubenfeld et al. [12] conducted a prospective cohort study in 21 hospitals in Washington State from 1999–2000 using a validated screening protocol. The crude incidence of ALI from this study was 78.9 per 100,000 person-years. A study by Li et al. [11] was also performed in the USA. They conducted a retrospective analysis of patients admitted over an 8-year period (2001–2008) in two hospitals by using an electronic ARDS screening tool and reported the incidence of ALI/ARDS as 33.8 per 100,000 person-years. However, there were some differences between these 2 studies including the duration of the study period (2 vs 8 years). The timing of studies also differed because Li et al. performed their study in the era of lung protective ventilation strategy, whereas Rubenfeld et al. conducted their study before the publication of the ARDSnet trial. Additionally, the study designs were different; one study was prospective and the other was retrospective. Notably, the reported incidence of ALI/ARDS was remarkably higher in the USA than in Europe. Perhaps the differences with respect to demographics and healthcare systems might account for the magnitude of the difference in the incidence of ALI/ARDS between the continents. The marked variations in ICU management systems, including the amount of ICU beds, ICU utilization, ICU staffing, and types of specific diseases admitted to the ICUs might also be influential factors. ICUs in the USA admit a higher number of patients than those in Europe and Canada [14].

As mentioned, our SICU consists of 14 beds, and there are approximately 700–800 patients admitted per year. This is a relatively low number compared with those reported in both the USA and Europe. Our study population only included general surgery patients, excluding both thoracic and trauma, for which a higher incidence of ALI was reported in previous studies [15]. Additionally, the majority of the patients were postoperative, which seems to be associated with lower severity when compared with nonsurgical patients [16]. Half of patients who developed ALI/ ARDS were nonsurgical as compared with only 20% of patients who did not develop ALI/ARDS. Thus, the inclusion of internal medicine, thoracic surgery, and trauma patients might have resulted in a higher incidence of ALI/ARDS, as reported in previous studies [15].

The 15 patients who developed ALI/ARDS had an early onset (day 2) and higher severity. Nevertheless, the APACHE II score was only 19.4 ± 8.8 and the average PaO₂/FiO₂ ratio on the day of ALI/ARDS onset was 183 ± 63.4, which was classified as moderate degree of ALI/ARDS according to the Berlin criteria [7]. Moreover, the tidal volume per predicted body weight ranged from 7.7–8.5 mL/kg, and the PEEP was ≤7 cmH₂O at the onset of ALI/ARDS. These data indicate less severity of the disease than that indicated in previous reports [9]. We did not find any other causes of ALI/ARDS, except for sepsis and pneumonia, which were also the most common etiologies of ALI/ARDS in the ALIEN study [9].

In spite of the lower incidence of ALI/ARDS, the hospital mortality rate in patients who developed ALI/ARDS was remarkably high (47%) in the present study, but comparable to all major series that ranged from 40% to 50% approximately [8-10]. Fortunately, the mortality rate has been decreasing over time since 1967 (by nearly 70%) [6]. The majority of ARDS patients die from multiple system organ dysfunction, resulting from the development of systemic inflammatory reactions predisposed by lung injury [17]. In our study, severe hypoxemia contributed much less to patient death. Papazian et al. [17] demonstrated that early administration of neuromuscular blocking agents in patients with ARDS can improve survival. The possible mechanisms of benefit involved either facilitating patient-ventilator synchrony or the reduction of systemic inflammatory responses. To date, lung protective ventilation strategy, including using low tidal volume, a plateau pressure of ≤30 cmH₂O and moderate to high PEEP, has been the only proven evidence to decrease mortality in patients with ARDS under ventilatory support [1].

Several limitations should be mentioned. First, as a result of the intermittently suspended services in the SICU because of an uncontrolled multidrug resistant infection, renovation of the SICU, and severe flooding in Bangkok, the amount of patient admissions was lower than we expected from the pilot study. Second, there was no high-dependency unit at the time of data collection; most of patients who were admitted to our SICU were not complicated and did not require mechanical ventilator support longer than 24 h. Lastly, our design was unable to demonstrate the role of lung protective ventilation strategy or the administration of neuromuscular blocking agents in changes in the incidence or outcomes in ALI/ARDS patients. Nevertheless, to our knowledge, this is the first study in Thailand showing the incidence and outcomes of ALI/ARDS patients in the SICU of a tertiary care hospital.

The incidence of ALI/ARDS according to the AECC definition in a general SICU in our hospital was low, but the mortality rate was high. ALI/ARDS developed early in the SICU, and the most common causes were sepsis and pneumonia. A study with a larger sample size should be conducted to identify independent risk factors for the development of ALI/ ARDS.