Survival predictors in elderly patients with acute respiratory distress syndrome: a prospective observational cohort study

Acute respiratory distress syndrome (ARDS) has a high mortality rate in intensive care units (ICU). The elderly patients remain to be increased of ICU patients. The aim is to investigate the survival predictors of elderly patients with ARDS. We reported a prospective observational cohort research, including the patients with ARDS between October 2012 and May 2015. Demographic, comorbidities, severity, lung mechanics, laboratory data and survival outcomes were analyzed. A total of 463 patients with ARDS were ≥65 years old were enrolled and analyzed. Multivariate logistic regression analysis identified Charlson comorbidity index (CCI) [odds ratio (OR) 1.111, 95% CI 1.010–1.222, p = 0.031], Sequential Organ Failure Assessment (SOFA) score (OR 1.127, 95% CI 1.054–1.206, p < 0.001) and peak inspiratory pressure (PIP) (OR 1.061, 95% CI 1.024–1.099, p = 0.001) which were independently associated with hospital mortality. Regarding the subgroups patients as 65–74 years old, 75–84 years old and ≥85 years old, the baseline characteristics were not significant difference and the hospital mortality rates were also not significant difference. In conclusion, CCI, SOFA score and PIP were identified as survival predictors in elderly patient with ARDS. Assessing comorbidities with CCI is essential in predicting the survival for elderly patients with ARDS.


Data collection.
We enrolled patients into this study if they met the criteria of the Berlin definition of ARDS 1 . Patients were excluded if they were younger than 18 years old, were referred from other hospitals, died within 48 hours, and had incomplete data. Demographics, baseline clinical characteristics and laboratory data were collected on enrollment. The following data were recorded on ICU admission: date of hospital and ICU admission, age, gender, predicted body weight, past underlying diseases history, risk factors and severity of ARDS on the day of diagnosis. The mechanical ventilator settings such as tidal volume, lowest PaO 2 /FiO 2 ratio with the highest PEEP and peak inspiratory pressure (PIP) were recorded during mechanical ventilation when ARDS was recognized within the first 24 hours of ARDS diagnosis. The severity index were recorded within the first 24 hours of ARDS diagnosis including Charlson comorbidity index (CCI) 12 , Acute Physiology and Chronic Health Evaluation (APACHE) II score 13 , Sequential Organ Failure Assessment (SOFA) score 14 , and lung injury score 15 .

Managements of ARDS.
The general mechanical ventilation settings of the patients included a lung protective ventilation strategy using a low tidal volume of 4-8 mL/kg of the predicted body weight, and the PEEP setting guided by low PEEP -FiO 2 table for volume-controlled or pressure-controlled ventilation. Oxygenation was monitored by SpO 2 through pulse oximetry, and the FiO 2 level was adjusted to maintain SpO 2 at more than 90%. Hemodynamics and lung water were monitored if the clinical condition of the patient indicated the need using a PiCCO plus monitor (version 5.2.2; Pulsion Medical System AG, Muenchen, Germany).

Statistical analysis. Data analysis was carried out by SPSS software version 22 (SPSS for Windows, SPSS
Inc., Chicago, IL, USA). Student's t test and ANOVA were used to compare the continuous variables. Categorical data were compared using the chi square test. The risk factor for hospital mortality was analyzed using univariate analysis, and the variables statistically significant (p < 0.05) were included for multivariate analysis by applying multiple logistic regressions based on backward elimination of data. Cumulative survival curves as a function of time were generated using the Kaplan-Meier approach and compared using the log-rank test. P value < 0.05 is considered to be statistically significant.

Results
During the research period, 22,470 admitted adult patients with invasive mechanical ventilation were screened, of whom 1,034 (4.6%) met the criteria of ARDS (Fig. 1). The sources of patients included 9 medical ICUs, 5 post-surgical ICUs, 2 trauma ICUs, 1 burn ICU and emergency department. Eighty-nine patients were excluded, and the remaining 945 patients with ARDS were included for analysis.
The demographic and clinical characteristics of the included population are shown in Table 1. There were no significantly different in gender, CCI, lung injury score and mechanical ventilator settings between the younger (<65 years old) and older (≥65 years old) patients. The older patients had a lower body mass index and higher APACHE II and SOFA scores than the younger population. For the initial oxygenation, the older patients had a higher PaO 2 /FiO 2 ratio (147.9 ± 77.3 vs. 134.3 ± 70.5 mmHg, p = 0.005) and less severe ARDS (34.1% vs. 40.2%, p = 0.041) than the younger patients. The hospital mortality rate was significantly higher in older patients than in younger patients (63.9% vs. 50.2%, p < 0.001). For the ARDS patients without co-morbidities (n = 186), the younger patients (n = 108) had lower hospital mortality rate than older patients (n = 78) (34.3% vs. 57.5%, p = 0.001).
Kaplan-Meier survival curves for hospital survival in the different age groups are shown in Fig. 2. The younger patients (<65 years old) had a significantly higher survival rate than the older patients (≥65 years old) (p = 0.0049). However, the survival rate was not significantly different among the 65-74, 75-84 and ≥85 years old groups (p = 0.774).

Discussion
The main results of this prospective observational cohort study revealed that the older (≥65 years old) patients with ARDS had a lower survival rate than the younger (<65 years old) patients with ARDS. For the elder ARDS patients, the CCI, SOFA score and PIP were significantly and independently associated with hospital mortality. However there was no significant difference in ICU or hospital survival rates among the 65-74 years old, 75-84 years old and ≥85 years old groups.
The Berlin definition classifies the severity of ARDS by the PaO 2 /FiO 2 ratio, and mild, moderate, and severe ARDS are associated with increased mortality (27%, 32% and 45%, respectively) 1 . Recent reports have shown that stratification of severity of ARDS based on baseline value of PaO 2 /FiO 2 did not completely correlate with mortality [16][17][18] . Often as much as 50% of patients classified as having moderate or severe ARDS respond quickly to routine ventilator and oxygenation measures that they do not meet the criteria for moderate/severe ARDS at 24 hours after diagnosis 17 . A 9 -point score based on age, PaO 2 /FiO 2 , and plateau pressure was proposed to predict mortality in patients with ARDS 19 . Compared to patients with ARDS who were younger than 47 years old, those 47-66 years and >66 years old had significantly higher hospital mortality rates (27.5% vs. 44.4% vs. 66.0%, respectively; p < 0.001) 19 . Increasing age is a known risk factor for death in patients with ARDS, and older patients have a higher risk of mortality than younger patients 20 with ARDS who are older than 65 years. In this study, we found that the major determinants of mortality were underlying disease (e.g. CCI), organ function (e.g. SOFA score) and pulmonary condition (e.g. PIP), but not age. Therefore, the impact of age on mortality in patients with ARDS seems to be limited, especially in elderly patients.
A study on patients with ARDS found that patients with serious comorbidities had a mortality rate three times higher than patients without serious comorbidities 22 . The CCI is an index of multiple comorbidities including 22 items which was initially developed in a cohort of 559 internal medicine patients to predict 1 year mortality 16 . In lung cancer patients, the CCI has been shown to be a prognostic predictor 23 , and several studies have reported that the CCI can predict survival and physiological outcomes in patients with ARDS [24][25][26] . In this study on elderly patients with ARDS, we found that the CCI was significantly positively correlated with survival outcomes (OR 1.11, p = 0.031). Therefore, we suggest assessing comorbidities using the CCI to predict survival in elderly patients with ARDS in addition to age.
The SOFA score involves organ dysfunction across six vital organs and it has been shown to be associated with more severe disease and a higher risk of death 14 . Only about 20% of patients with ARDS die from refractory hypoxemia, and approximately 80% of all deaths are caused by multiple organ dysfunction syndromes 27,28 . The SOFA score has been used for patients with ARDS to evaluate organ dysfunction as a surrogate marker of mortality 29 . In terms of liver failure, patients with ARDS and cirrhosis have been reported to have a significantly higher mortality rate (62%) than patients without cirrhosis (43%) (p = 0.02) 30 . An observational study of patients with indirect ARDS found that age, lung injury score, and number of non-pulmonary organ failures (OR 1.67, p = 0.01) were independent risk factors for hospital mortality 31 . The LUNG SAFE study of patients with ARDS found that a higher non-pulmonary SOFA score was associated with poorer outcomes (OR 1.12, p < 0.001) 32 . In our study on elderly patients with ARDS, SOFA score was significantly correlated with hospital mortality (OR 1.18, p < 0.001). The prognosis for elderly patients with ARDS therefore appears to be related to extra-pulmonary organ dysfunction rather than pulmonary dysfunction alone.
A study including 3562 patients with ARDS in nine randomized controlled trials concluded that driving pressure as an index of pulmonary mechanics of the respiratory system was the strongest predictor of mortality 33 . Another study on 56 patients with ARDS reported that treatment strategies leading to decreased transpulmonary driving pressure at 24 hours may be associated with an improved 28 -day mortality rate 34 . In addition to driving pressure, plateau pressure has also been reported to be a predictor of mortality in patients with ARDS 19,32 . A prospective, descriptive, and validation study reported that the hospital mortality rates of patients with ARDS with a   19 . Peak inspiratory pressure, which is an easily measurable parameter of lung mechanics was also associated with hospital mortality in the LUNG SAFE study (OR 1.02, p = 0.002) 32 . For our elderly patients with ARDS, we found that PIP was significantly correlated with hospital mortality (OR 1.068, p = 0.001). In theory, the PIP is different from plateau pressure. However, the peak airway has a good collinearity with plateau pressure especially when patients were to be deep sedated or paralyzed in early stage of ARDS. Due to its convenience and feasibility, PIP may be useful as a prognostic index in real world patient care.
There are several limitations to this study. First, this study was conducted at one referral medical center, and our results may not be generalizable to patients in community hospitals or other models of intensive care. Nevertheless, the number of enrolled elderly patients with ARDS was reasonably high, and thus we believe our findings are of value. Second, few studies have investigated elderly patients with ARDS, and these studies have mostly focused on critically ill patients as a whole. We chose 65 years of age as a cutoff value mainly because previous studies on critically ill patients have used this cutoff to define "elderly" patients. We further classified the elderly patients into three arbitrary age groups of 65-74, 75-84 and ≥85 years old without considering morbidities, functional status and other disabilities, and this may have affected the outcomes. Third, different health care systems in different countries will have different policies for intensive care for critically ill elderly patients. Finally, it is possible to have a selection bias from the patients' collection. Some patients were possibly rejected to be admitted in ICU because of underlying comorbidity by the physician in charge.
In conclusion, CCI, SOFA score and PIP were predictors of hospital survival in elderly patients with ARDS. The risk of mortality in the elderly patients ARDS was associated with the degree of lung injury and also with the underlying disease and presence of other organ dysfunction. When making decisions regarding life-sustaining therapy for elderly patients with ARDS, both comorbidities and advanced age should be taken into consideration.   Table 4. Demographics and baseline characteristics of the patients with ARDS aged ≥65 years old (n = 463). ARDS: acute respiratory distress syndrome; BMI: body mass index; PBW: predicted body weight; APACHE: acute physical and chronic health evaluation; SOFA: sequential organ function assessment; PIP: peak inspiratory pressure; PEEP: positive end expiratory pressure; PaO 2 /FiO 2 : alveolar oxygen pressure/fraction of inspiratory oxygen. All values are expressed as number of patients (%) or mean ± SD. * Statistically significant difference between those aged 65-74 years old and those aged ≥85 years old.