Acute respiratory failure is one of the most common causes for the emergency admission of patients to intensive care units [
The medical expenditure on treatment of critically ill patients with acute respiratory failure is substantial. The estimated health care costs related to critical care are approximately 0.7 percent of the annual gross domestic product in US [
Traditionally, a number of severity scoring systems such as Acute Physiology and Chronic Health Evaluation (APACHE) score, Simplified Acute Physiologic Score (SAPS), Mortality Prediction Model (MPM), and Sequential Organ Failure Assessment (SOFA) score are frequently adopted for severity evaluation and as outcome predictors in ICU patients [
This retrospective study was conducted at one of the medical intensive care units (ICU) in Kaohsiung Chang Gung Memorial Hospital, a 2,400-bed, tertiary teaching hospital in southern Taiwan. The study was conducted from January 1, 2010, to July 31, 2010, for adult patients (aged 18 years or more) admitted consecutively to the designated medical ICU who fulfilled criteria for severe acute respiratory failure. Severe acute respiratory failure was defined as acute respiratory insufficiency requiring ventilator support for at least 24 hours. The study was approved by the Institutional Review Board of Chang Gung Memorial Hospital, and the requirement for patient consent was waived.
Clinical data retrieved from medical records included age, gender, initial admitting diagnosis, source of admission, hospital duration, previous ICU admission within 2 years, underlying comorbidities, tracheostomy status prior to current admission, Sequential Organ Failure Assessment (SOFA) score on the day of admission, and Acute Physiological Assessment and Chronic Health Evaluation II (APACHE II) score. Respiratory and ventilator parameters retrieved included PaO2, FiO2, PEEP, and mPaw on the 1st and the 3rd day of mechanical ventilation. Both oxygenation index (OI) and PaO2/FiO2 ratio on the 1st and 3rd day of ICU admission were calculated and recorded, respectively, for each patient. OI was defined as 100 × mPaw/(PaO2/FIO2). The change in OI values from the 1st day to the 3rd day was obtained by subtracting day 1 OI from day 3 OI. The primary outcome was overall hospital mortality. Secondary outcome measure was ventilator weaning failure, which was defined as the failure of complete autonomy from the ventilator for more than 5 days.
Categorical variables were analyzed using the chi-square test or Fisher’s exact test where appropriate, and continuous variables were compared using Student’s
A total of 145 consecutive patients were admitted into the designated MICU within a 7-month period (January 1, 2010 to July 31, 2010), of which 34 patients did not fulfill the criteria of severe acute respiratory failure. Of the 111 patients who met the criteria for severe acute respiratory failure and received mechanical ventilation, 11 had critical data that were either incomplete or missing and were not entered into the statistical analysis. The remaining 100 patient’s baseline characteristics and ventilator parameters on the 1st and 3rd day were shown in Table
Baseline characteristics of patients.
Characteristics |
|
---|---|
Age (years) | 72.14 (15.77) |
Gender ( |
|
Male | 58 (58) |
Female | 42 (42) |
Hospital duration (days) | 38.1 (30.5) |
Hospital mortality ( |
38 (38) |
SOFA* score on admission | 4.71 (2.71) |
APACHE* II score | 25.82 (6.74) |
Source ( |
|
Emergency department | 55 (55) |
Hospital ward | 38 (38) |
Other ICU* or other hospitals | 7 (7) |
Previous MICU* admission ( |
30 (30) |
Comorbidities ( |
|
Congestive heart failure | 10 (10) |
Hypertension | 62 (62) |
Chronic obstructive pulmonary disease | 22 (22) |
Pulmonary tuberculosis | 9 (9) |
CRF* with tracheostomy | 11 (11) |
Liver cirrhosis | 10 (10) |
Chronic kidney disease | 19 (19) |
Diabetes mellitus | 35 (35) |
Cerebrovascular accident | 30 (30) |
Hepatitis B | 6 (6) |
Hepatitis C | 4 (4) |
Cancer history | 22 (22) |
Respiratory and ventilator parameters | |
Fi |
0.58 (0.21) |
Pa |
140.46 (88.21) |
PEEP* day 1 (cmH2O) | 5.55 (2.63) |
mPaw* day 1 (cmH2O) | 11.99 (3.23) |
PaO2/FiO2 day 1 (mmHg) | 267.95 (181.11) |
OI* day 1 (cmH2O/mmHg) | 7.06 (6.01) |
FiO2 day 3 | 0.43 (0.16) |
PaO2 day 3 (mmHg) | 109.79 (32.84) |
PEEP day 3 (cmH2O) | 5.8 (2.65) |
mPaw day 3 (cmH2O) | 11.46 (3.02) |
PaO2/FiO2 day 3 (mmHg) | 287.92 (116.3) |
OI day 3 (cmH2O/mmHg) | 5.74 (5.82) |
OI change from day 1 to day 3 | −1.31 (6.34) |
*APACHE: Acute Physiological Assessment and Chronic Health Evaluation, CRF: chronic respiratory failure, FiO2: fraction inspired oxygen, ICU: intensive care unit, MICU: medical intensive care unit, mPaw: mean airway pressure, OI: oxygenation index, PaO2: arterial oxygen tension, PEEP: positive end expiratory pressure, SOFA: Sequential Organ Failure Assessment.
Variables are expressed as mean (standard deviation) and categorical data are expressed as number (percentage).
Etiologies of severe acute respiratory failure.
Etiologies of respiratory failure ( |
|
---|---|
Acute myocardial infarction | 1 (1) |
Aortic aneurysm rupture | 1 (1) |
Community-acquired pneumonia | 30 (30) |
Hospital-acquired pneumonia | 12 (12) |
Healthcare-associated pneumonia | 6 (6) |
Chemical intoxication | 1 (1) |
Carbon monoxide poisoning | 1 (1) |
COPD* acute exacerbation | 8 (8) |
Acute stroke | 2 (2) |
Empyema | 1 (1) |
Organophosphate intoxication | 1 (1) |
Seizure disorder | 1 (1) |
Hyperosmolar hyperglycemic state | 2 (2) |
Infective endocarditis | 1 (1) |
Decompensated liver cirrhosis | 1 (1) |
Acute pulmonary embolism | 1 (1) |
Intra-abdominal infection | 2 (2) |
Sepsis unknown focus | 4 (4) |
Skin/soft tissue infection | 3 (3) |
Upper gastrointestinal bleeding | 5 (5) |
Tracheal stenosis | 1 (1) |
Urosepsis | 13 (13) |
Uremia | 1 (1) |
Ventilator-associated pneumonia | 1 (1) |
Table
Characteristics of survivors and nonsurvivors: bivariate analysis#.
Characteristics | Hospital mortality |
Survivor |
|
---|---|---|---|
Age (years) | 68.37 (17.55) | 74.45 (14.22) | 0.1 |
Gender ( |
|||
Male | 22 (57.9) | 36 (58.1) | 0.99 |
Female | 16 (42.1) | 26 (41.9) | |
SOFA* score on admission | 5.63 (3.07) | 4.15 (2.3) | 0.007 |
APACHE* II score | 27.05 (7.44) | 25.06 (6.21) | 0.15 |
Source ( |
|||
Emergency department | 18 (47) | 37 (60) | |
Hospital ward | 17 (45) | 21 (34) | 0.48 |
Other ICU* or other hospitals | 3 (8) | 4 (6) | |
Previous MICU* admission ( |
13 (42) | 17 (27.4) | 0.47 |
Comorbidities ( |
|||
Congestive heart failure | 5 (13.2) | 5 (8.1) | 0.41 |
Hypertension | 23 (60.5) | 39 (62.9) | 0.81 |
Chronic obstructive | 7 (18.4) | 15 (24.2) | 0.5 |
pulmonary disease | |||
Pulmonary tuberculosis | 2 (5.3) | 7 (11.3) | 0.31 |
CRF* with tracheostomy | 3 (7.9) | 8 (12.9) | 0.44 |
Liver cirrhosis | 6 (15.8) | 4 (6.5) | 0.13 |
Chronic kidney disease | 7 (18.4) | 12 (19.4) | 0.91 |
Diabetes mellitus | 12 (31.6) | 23 (37.1) | 0.57 |
Cerebrovascular accident | 9 (23.7) | 21 (33.9) | 0.28 |
Hepatitis B | 4 (10.5) | 2 (3.2) | 0.14 |
Hepatitis C | 1 (2.6) | 3 (4.8) | 0.59 |
Cancer history | 5 (13.2) | 5 (8.1) | 0.41 |
*APACHE: Acute Physiological Assessment and Chronic Health Evaluation, CRF: chronic respiratory failure, ICU: intensive care unit, MICU: medical intensive care unit, SOFA: Sequential Organ Failure Assessment.
#Continuous variables were analyzed by Student’s
Variables are expressed as mean (standard deviation) and categorical data are expressed as number (percentage).
Respiratory and ventilator parameters of survivors and nonsurvivors on day 1 and day 3 of mechanical ventilation#.
Day 1 and day 3 Respiratory and ventilator parameters | |||
---|---|---|---|
Characteristics | Hospital mortality |
Survivor |
|
Fi |
0.66 (0.21) | 0.53 (0.19) | 0.002 |
Pa |
163.02 (104.78) | 126.63 (73.87) | 0.14 |
PEEP* day 1 (mmHg) | 5.95 (3.94) | 5.31 (1.28) | 0.38 |
mPaw* day 1 (mmHg) | 12.16 (3.35) | 11.89 (3.18) | 0.69 |
PaO2/FiO2 day 1 (mmHg) | 271.84 (184.39) | 265.56 (180.55) | 0.87 |
OI* day 1 | 7.81 (7.62) | 6.59 (4.78) | 0.33 |
| |||
FiO2 day 3 | 0.51 (0.21) | 0.38 (0.1) | 0.001 |
PaO2 day 3 (mmHg) | 102.72 (42.17) | 114.12 (24.94) | 0.07 |
PEEP day 3 (cmH2O) | 6.11 (3.71) | 5.58 (1.66) | 0.42 |
mPaw day 3 (cmH2O) | 12.89 (3.42) | 10.58 (2.38) | <0.001 |
PaO2/FiO2 day 3 (mmHg) | 245.84 (144.83) | 313.7 (86.32) | 0.01 |
OI day 3 (cmH2O/mmHg) | 8.82 (8.16) | 3.86 (2.26) | <0.001 |
| |||
OI change from day 1 to day 3 (cmH2O/mmHg) | 1.0 (8.446) | −2.74 (4.08) | <0.001 |
*FiO2: fraction inspired oxygen, mPaw: mean airway pressure, OI: oxygenation index, PaO2: arterial oxygen tension, PEEP: positive end-expiratory pressure.
#Continuous variables were analyzed by Student’s
Variables are expressed as mean (standard deviation) and categorical data are expressed as number (percentage).
Using univariate analysis of factors capable of predicting overall hospital mortality, nonsurvivors were found to have a higher initial SOFA score (AOR 1.24, 95% confidence interval (CI) 1.05–1.47,
Multivariate analysis of predictors of hospital mortality in patients with severe acute respiratory failure.
Variables | Odds ratio | CI |
|
---|---|---|---|
SOFA* score | 1.27 | 1.04–1.55 | 0.02 |
Pa |
1.01 | 1.00–1.02 | 0.05 |
OI* day 3 | 1.49 | 1.13–1.95 | 0.004 |
OI change from day 1 to day 3 | 1.11 | 0.99–1.24 | 0.07 |
*FiO2: fraction inspired oxygen, OI: oxygenation index, PaO2: arterial oxygen tension, SOFA: Sequential Organ Failure Assessment.
ROC curves were plotted to identify cutoff values that would best determine hospital mortality of ICU patients (Figure
Comparison of cutoff value, sensitivity, specificity, AUC*, and
Factor | Cutoff | Sensitivity | Specificity | AUC |
|
---|---|---|---|---|---|
SOFA* score | 5.5 | 53% | 73% | 0.647 | 0.01 |
Day 3 OI* | 3.79 | 69% | 71% | 0.724 | <0.001 |
*AUC: Area under curve, SOFA: Sequential Organ Failure Assessment, OI: oxygenation index.
ROC curve analysis for predictability of hospital mortality between Day 3 OI and SOFA score.
To illustrate the correlation between oxygenation index and survival time, we stratified day 3 oxygenation index into 4 groups (0 < OI < 5, 5 < OI < 10, 10 < OI < 15, and OI ≥ 15) using the Kaplan-Meier method. There was significant association between value of day 3 OI and the survival time as demonstrated by the log-rank test (
Kaplain Meier curve of stratified day 3 OI groups versus survival time.
Table
Characteristics of patients with different weaning outcomes: bivariate analysis#.
Characteristics | Weaning success |
Weaning failure |
|
---|---|---|---|
Age (years) | 72.59 (15.04) | 79.81 (10.14) | 0.08 |
Gender ( |
|||
Male |
27 (58.7) |
9 (56.3) |
0.86 |
SOFA* score on admission | 4.39 (2.30) | 3.44 (2.22) | 0.16 |
APACHE* II score | 24.7 (6.62) | 26.13 (6.43) | 0.43 |
Source ( |
|||
Emergency department | 27 (58.7) | 10 (62.5) | |
Hospital ward | 17 (37.0) | 4 (25) | 0.42 |
Other ICU* or other hospitals | 2 (4.3) | 2 (12.5) | |
Previous MICU admission ( |
10 (21.7) | 7 (43.8) | 0.09 |
Comorbidities ( |
|||
Congestive heart failure | 5 (10.9) | 0 (0) | 0.17 |
Hypertension | 28 (60.9) | 11 (68.8) | 0.57 |
Chronic obstructive | 9 (19.6) | 6 (37.5) | 0.15 |
pulmonary disease | |||
Pulmonary tuberculosis | 5 (10.9) | 2 (12.5) | 0.86 |
CRF* with tracheostomy | 4 (8.7) | 4 (25) | 0.09 |
Liver cirrhosis | 4 (8.7) | 0 (0) | 0.22 |
Chronic kidney disease | 11 (23.9) | 1 (6.3) | 0.12 |
Diabetes mellitus | 15 (32.6) | 8 (50) | 0.22 |
Cerebrovascular accident | 9 (19.5) | 12 (75) | <0.001 |
Hepatitis B | 2 (4.3) | 0 (0) | 0.4 |
Hepatitis C | 3 (6.5) | 0 (0) | 0.3 |
Cancer history | 10 (21.7) | 0 (0) | 0.04 |
*APACHE: Acute Physiological Assessment and Chronic Health Evaluation, CRF: chronic respiratory failure, ICU: intensive care unit, MICU: medical intensive care unit, SOFA: Sequential Organ Failure Assessment.
#Continuous variables were analyzed by Student’s
Variables are expressed as mean (standard deviation) and categorical data are expressed as number (percentage).
Day 1 and day 3 respiratory and ventilator parameters of patients with different weaning outcomes#.
Day 1 and Day 3 Respiratory and ventilator parameters | |||
---|---|---|---|
Characteristics | Weaning success |
Weaning failure |
|
Fi |
0.55 (0.2) | 0.46 (0.17) | 0.1 |
Pa |
129.11 (79.84) | 119.5 (54.74) | 0.69 |
PEEP* Day 1 (mmHg) | 5.41 (1.33) | 5.0 (1.10) | 0.27 |
mPaw* Day 1 (mmHg) | 12.0 (3.22) | 11.56 (3.14) | 0.64 |
PaO2/FiO2 Day 1 (mmHg) | 259.02 (180.78) | 284.36 (184.45) | 0.63 |
OI* Day 1 | 6.99 (5.15) | 5.46 (3.4) | 0.27 |
| |||
FiO2 day 3 | 0.38 (0.08) | 0.4 (0.14) | 0.57 |
PaO2 Day 3 (mmHg) | 114.63 (24.8) | 112.64 (26.09) | 0.79 |
PEEP Day 3 (cmH2O) | 5.41 (1.33) | 5.0 (1.1) | 0.43 |
mPaw day 3 (cmH2O) | 10.37 (2.44) | 11.19 (2.14) | 0.24 |
PaO2/FiO2 Day 3 (mmHg) | 313.94 (76.81) | 313.02 (112.25) | 0.97 |
OI Day 3 (cmH2O/mmHg) | 3.67 (2.03) | 4.39 (2.82) | 0.27 |
| |||
OI change from day 1 to day 3 (cmH2O/mmHg) | −3.32 (4.46) | −1.06 (1.97) | 0.008 |
*FiO2: fraction inspired oxygen, mPaw: mean airway pressure, OI: oxygenation index, PaO2: arterial oxygen tension, PEEP: positive end expiratory pressure.
#Continuous variables were analyzed by Student’s
Variables are expressed as mean (standard deviation).
Univariate analysis identified history of CVA as the potential risk factor for weaning failure (AOR = 12.33, 95% CI 3.21–47.38,
Multivariate analysis of predictors of weaning failure in patients with severe acute respiratory failure.
Variables | Odds ratio | CI* |
|
---|---|---|---|
Cerebrovascular accident | 10.16 | 2.37–43.58 | 0.002 |
OI change from day 1 to day 3 | 0.74 | 0.51–1.08 | 0.12 |
*CI: confidence interval, OI: oxygenation index.
This study has several findings. First, adults with severe acute respiratory failure receiving mechanical ventilation who did not survive were more likely to have initial high SOFA score, high FiO2, high day 3 mPaw, low day 3 PaO2/FiO2, high day 3 OI, and increasing OI from 1st to 3rd day of mechanical ventilation. When multivariate logistic regression analysis was implemented, SOFA score and 3rd day OI were found to be independent risk factors for hospital mortality, but the PaO2/FiO2 ratio was not. Second, OI had a better combined sensitivity and specificity than SOFA score in predicting mortality as reflected in our ROC analysis. Third, our study demonstrated significant correlation between the value of OI and survival time. When values of day 3 OI were stratified into 4 groups and plotted against survival time on the Kaplan-Meier graph, the group with highest OI value had a shortest survival time, while lower OI groups survived longer. Fourth, CVA was the only independent predictor of weaning failure from ventilator in our study.
OI was originally used in pediatric field as an index for prediction of mortality of infants with hypoxic respiratory failure [
Numerous studies [
Our results demonstrated OI predicts mortality better than does PaO2/FiO2 in ventilated patients. Originally described in 1974 by Horovitz and colleagues, PaO2/FiO2 ratio was introduced in an attempt to overcome the limitations of alveolar-arterial (A-a) O2 pressure gradient and arterial alveolar (a-A) oxygen tension ratio (a/A ratio) and enable the evaluation of PaO2 at varying FiO2 [
In the past decade, low tidal volume ventilation (LTVV) has gained more popularity among clinicians as a measure to reduce mortality in ARDS. In patients with acute respiratory failure who do not meet the criteria but with risks for development of ARDS or ALI, evidence showed that LTVV with high PEEP may prevent development of ALI/ARDS [
Our study showed preexisting CVA was an independent predictor for ventilator weaning failure. Although acute stroke patients who require mechanical ventilation are known to carry poor outcomes [
Change in OI in the 1st 3 days was shown to correlate with weaning outcome in our study, though multivariate analysis failed to establish its role as an independent predictor. The relationship between OI and weaning outcome was discussed in several studies. In the study by Tseng et al., they demonstrated that congestive cardiac failure (
Our study has several limitations. First, a retrospective review of existing data was conducted, inevitably, disadvantage such as missing key data in small amount of patients would occur. This may reduce the representativeness of the sample. Second, the relatively small sample size implies a single data may have a greater influence on final results. Despite this, it did not affect our final conclusion or inference because our main results were highly significant since their
This study suggested that elevated OI measured in the first 3 days of mechanical ventilation and high SOFA score are independent predictors of mortality in patients with acute respiratory failure requiring mechanical ventilation. OI is comparable with, if not superior than, general severity scores such as APACHE II and SOFA score in predicting mortality. Our study also suggested that by conducting serial OI measurements and monitoring trends over time may provide more useful information than any single measurement. In the future, prospective studies measuring serial OIs in a larger scale of study cohort will be required to further consolidate our findings.
There are no conflict of interests to declare.