Abstract
Background:
Prior studies suggest that mortality differs by sex and race in patients who develop acute lung injury (ALI). Whether differences in presentation account for these disparities remains unclear. We sought to determine whether sexual and racial differences exist in the rate of ALI development and ALI-related mortality after accounting for differences in clinical presentations.
Methods:
This was a multicenter, observational cohort study of 5,201 patients at risk for ALI. Multivariable logistic regression with adjustment for center-level effects was used to adjust for potential covariates.
Results:
The incidence of ALI development was 5.9%; in-hospital mortality was 5.0% for the entire cohort, and 24.4% for those patients who developed ALI. Men were more likely to develop ALI compared to women (6.9% vs 4.7%, P < .001) and had a nonsignificant increase in mortality when ALI developed (27.6% vs 18.5%, P = .08). However, after adjustment for baseline imbalances between sexes these differences were no longer significant. Black patients, compared to white patients, presented more frequently with pneumonia, sepsis, or shock and had higher severity of illness. Black patients were less likely to develop ALI than whites (4.5% vs. 6.5%, P = .014), and this association remained statistically significant after adjusting for differences in presentation (OR, 0.66; 95 % CI, 0.45-0.96).
Conclusions:
Sex and race differences exist in the clinical presentation of patients at risk of developing ALI. After accounting for differences in presentation, there was no sex difference in ALI development and outcome. Black patients were less likely to develop ALI despite increased severity of illness on presentation.
Acute lung injury (ALI) is a common and lethal critical care syndrome occurring after an acute illness or injury such as sepsis, shock, pneumonia, or aspiration1‐3 Evidence suggests that sex and racial disparities exist in ALI-related mortality.4 However, it is not clear whether blacks and men have higher rates of developing ALI or rather higher mortality from ALI. A study suggested that the racial disparities were mediated by increased severity of illness at the time of ALI development.5 To date, no study in a population with multiple risk factors has examined whether sex and racial differences exist in ALI development and related outcomes.
In a multicenter, observational cohort study, we evaluated the relationship between race and sex and ALI development in at-risk patients. Our goals were to determine whether sex and racial differences exist at the time of presentation and to determine whether sex and racial differences exist in the rate of ALI development and outcomes in those who develop ALI. Results of the study were previously reported in abstract form.6
Materials and Methods
Study Design
Ours was a planned substudy of a multicenter, observational cohort study designed to predict ALI development in at-risk patients.7 Institutional review boards (IRBs) at all participating institutions approved the study with a waiver of informed consent for medical records review. The Mount Sinai School of Medicine IRB approved this substudy (No. 090-0019 0001 01). The Albert Einstein College of Medicine IRB exempted this substudy from full IRB review.
Study Population and Data Collection
The study enrolled consecutive patients prospectively at 19 hospitals and retrospectively at three hospitals over 6 months, beginning in March 2009. Adults admitted to the hospital were screened for one or more of the following study-defined ALI risk factors: sepsis, shock, pancreatitis, pneumonia, aspiration, high-risk trauma, or high-risk surgery.1,6‐11 We excluded patients who were < 18 years of age, patients fulfilling the criteria for ALI at the time of ED presentation or admission for surgery,12 outside-hospital transfers, and patients who expired in the ED or were admitted for comfort care only and subsequent visits. To focus on sex and race in the United States, we limited our study to the 20 US centers.
Patient-level data were abstracted using predrafted case report forms from the medical record. Patient characteristics were collected using standardized predefined criteria during the first 6 h after initial ED presentation, or at the time of admission in high-risk surgical patients, and included demographics, comorbidities including alcohol and tobacco use, prehospitalization prescription medications, admission vital signs and laboratory measures, and ALI risk factors. Pertinent processes of care measures included the administration of packed RBCs (PRBCs) during the initial 4 hospital days and the use of ventilatory assistance at any time during the hospitalization. The study population and data collection have been described previously in greater detail.7
Exposure Variables
Patient race, ethnicity, and sex were assigned by study personnel based on review of the medical record. Race was categorized into one of four categories: white, black, Asian, and other. Ethnicity was coded as Hispanic or non-Hispanic.
Outcomes
The primary outcome was ALI development during the hospitalization. The American-European Consensus Conference definitions for ALI and ARDS were used: the presence of bilateral pulmonary infiltrates on chest radiograph and the absence of left atrial hypertension as the primary explanation for pulmonary edema and hypoxemia (ratio of PaO2 to FIO2 < 300 for ALI and < 200 for ARDS).12 Secondary outcomes included in-hospital mortality and hospital and ICU length of stay.
Statistical Analysis
We used multivariable logistic regression to adjust for potential confounding in the relationships between race and sex and ALI development and between race and sex and in-hospital mortality. We considered the following variables as potential confounders: age,13 BMI,14,15 cause of lung injury,6‐9,11,16 comorbid conditions (alcohol abuse, tobacco use, diabetes, end-stage renal disease, class 4 congestive heart failure),17‐19 prescription medications (aspirin)20 and tidal volumes prescribed during initiation of mechanical ventilation,21 illness severity as measured by the baseline APACHE (Acute Physiology and Chronic Health Evaluation) II score,22 RBC transfusion,23,24 and lung injury prediction (LIP) score.7 Patients were dichotomized into having or not having received PRBC based on transfusion of any volume of PRBC during the first 4 hospital days. We included each potential confounder in the ALI development model. For the ALI-related mortality model, we added potential confounders associated with the outcome of interest in univariate analyses at a significance level < .20 to the base in-hospital mortality model and maintained the confounder if the OR estimate was altered by > 10%.25 We did not adjust for the use of mechanical ventilation, because we considered such use to be in the causal pathway to ALI development.21
We assessed for center-level effects in the association between race and ALI development using the modified Breslow-Day test for homogeneity of ORs.26 To account for center-level effects, we used mixed-effects multivariable logistic regression models with hospital-specific random effects.27 To determine whether sex acted as an effect modifier in the associations with race, we included an interaction term in the logistic regression equation. In sensitivity analyses, to further examine the association between race and ALI development, we used conditional logistic regression models. Furthermore, we excluded subjects from centers that enrolled patients retrospectively from our hospital-specific random effects models. We present the results as ORs and 95% CIs. We defined significance as a P value < .05 and used Stata, version 10.0 software to perform statistical analyses (StataCorp LP).
Results
Over 6 months, we studied 5,201 patients at high risk of ALI development, including 1,135 black patients (21.8%). Women comprised 45% of the cohort. The incidence of ALI development was 5.9% (95% CI, 5.2%-6.6%). In-hospital mortality was 5.0% for the entire cohort (95% CI, 4.4%-5.6%) and 24.4% for those patients who developed ALI (95% CI, 19.7%-29.6%).
Baseline Characteristics by Sex and Race
As shown in Table 1, men, compared with women, were younger (P < .001), more likely to be white (P < .001), and less likely to be obese (P < .001). Regarding risk modifiers for ALI development, men were more likely to be alcohol abusers (P < .001) and less likely to have diabetes (P = .017), and sex-based differences in the predisposing conditions for ALI existed. Men had higher LIP scores (P < .001) and were more likely to receive mechanical ventilation (P < .001).
Table 1.
—Baseline Characteristics by Sex
| Characteristics | Male (n = 2,886) | Female (n = 2,315) | P Valuea |
| Demographics | |||
| Age, y | 55 (42-69) | 58 (44-72) | < .001 |
| Hispanic ethnicity | 251 (10.3) | 227 (11.5) | .19 |
| Race | |||
| White | 1,974 (70.3) | 1,445 (64.7) | Ref |
| Black | 569 (20.3) | 566 (25.3) | < .001 |
| Asian | 39 (1.4) | 41 (1.8) | .11 |
| Other | 226 (8.1) | 183 (8.2) | .34 |
| Admission source | |||
| Home | 2,288 (80.4) | 1,828 (80.1) | .816 |
| Outside facility | 559 (19.6) | 454 (19.9) | |
| Comorbidities | |||
| CHF class 4 | 83 (2.9) | 78 (3.4) | .31 |
| Diabetes | 627 (21.7) | 568 (24.5) | .02 |
| ESRD | 112 (3.9) | 96 (4.2) | .63 |
| COPD | 306 (10.6) | 249 (10.8) | .86 |
| Asthma | 148 (5.1) | 277 (12.0) | < .001 |
| Metastatic solid cancer | 142 (4.9) | 126 (5.4) | .40 |
| Chest radiation | 26 (0.9) | 43 (1.9) | .003 |
| GERD | 320 (11.1) | 365 (15.8) | < .001 |
| Alcohol abuseb | 166 (8.3) | 36 (2.0) | < .001 |
| Obesityc | 681 (23.6) | 630 (27.2) | .003 |
| Tobacco use | |||
| None | 1,194 (45) | 1,259 (58.2) | < .001 |
| Former or current | 1,458 (55.0) | 904 (41.8) | < .001 |
| ALI predisposing conditions | |||
| Shock | 236 (8.2) | 163 (7.0) | .13 |
| Aspiration | 135 (4.7) | 69 (3.0) | .002 |
| Sepsis | 886 (30.7) | 917 (39.6) | < .001 |
| Pancreatitis | 156 (5.4) | 154 (6.7) | .06 |
| Pneumonia | 622 (21.6) | 574 (24.8) | .006 |
| High-risk surgeryd | 831 (28.8) | 602 (26.0) | .025 |
| Emergency surgery | 184 (6.4) | 118 (5.1) | .05 |
| High-risk trauma | 615 (21.3) | 319 (13.8) | < .001 |
| ALI risk modifiers | |||
| APACHE II | 9 (5-14) | 9 (5-13) | .90 |
| LIPS | 3 (2-4.5) | 2.5 (1.5-4.0) | < .001 |
| LIPS > 4 | 740 (25.6) | 473 (20.4) | < .001 |
| Process of care | |||
| PRBC transfusion | 440 (15.2) | 331 (14.3) | .34 |
| Noninvasive ventilation | 274 (10.1) | 245 (11.0) | .32 |
| Mechanical ventilation | 911 (32.8) | 28.7 (9.6) | < .001 |
Data are presented as No. (interquartile range) or No. (%). ALI = acute lung injury; APACHE = Acute Physiology and Chronic Health Evaluation; CHF = congestive heart failure; ESRD = end stage renal disease; GERD = gastroesophageal reflux disease; LIPS = lung injury prediction score; PRBC = packed RBC. Ref = reference.
For comparisons, whites are baseline.
Alcohol abuse is defined as > 14 drinks/wk.
Obesity is defined as BMI > 30.
High-risk surgery is defined as cardiac, aortic, thoracic, spinal, and acute abdominal procedures.
Black patients, compared with whites, were younger (P < .001), more likely to be women (P < .001), and more likely to be obese (P < .05) and presented with higher APACHE II scores (P < .001). White patients were more likely to consume alcohol (P < .001) and less likely to have diabetes (P < .001) and had higher LIP scores than did black patients (P < .001). During the hospitalization, whites were more likely to receive PRBC transfusion and mechanical ventilation (P < .001) despite lower APACHE II scores. The frequency of risk conditions differed significantly by race (Fig 1, Table 2).
Figure 1.
Development of lung injury by racial category and selected risk conditions. Total column height represents the percentage of patients within each racial group with the ALI risk factor. The light gray portion on top of each column represents the proportion who went on to develop ALI, and the bottom dark gray portion represents the proportion who did not develop ALI. For all seven risk categories, P < .05 by χ2 test comparing the proportion of subjects within each risk category by race. For the risk factor of sepsis alone, P < .05 by χ2 for ALI development by race. ALI = acute lung injury.
Table 2.
—Baseline Characteristics by Race
| Characteristics | White (n = 3,419) | Black (n = 1,135) | P Valuea | Asian (n = 80) | P Valuea | Other Race (n = 409) | P Valuea |
| Demographics | |||||||
| Age, y | 60 (45-73) | 52 (42-65) | < .001 | 58 (38-70) | .08 | 46 (34-62) | < .001 |
| Male sex | 1,974 (57.7) | 569 (50.1) | < .001 | 39 (48.8) | .11 | 226 (55.3) | .34 |
| Hispanic ethnicity | 86 (3.0) | 14 (1.5) | .01 | 0 | .14 | 277 (79.6) | < .001 |
| Admission source | |||||||
| Home | 2,613 (77.3) | 953 (85.3) | < .001 | 69 (86.3) | .06 | 360 (88.5) | < .001 |
| Outside facility | 768 (22.7) | 164 (14.7) | < .001 | 11 (13.8) | .06 | 47 (11.5) | < .001 |
| Comorbidities | |||||||
| CHF class 4 | 86 (2.5) | 61 (5.4) | < .001 | 1 (1.3) | .47 | 8 (2.0) | .49 |
| Diabetes | 686 (20.1) | 338 (29.8) | < .001 | 21 (26.3) | .17 | 105 (25.7) | .008 |
| ESRD | 69 (2.0) | 121 (10.7) | < .001 | 2 (2.5) | .76 | 13 (3.2) | .13 |
| COPD | 412 (12.1) | 101 (8.9) | .004 | 2 (2.5) | .009 | 28 (6.9) | .002 |
| Asthma | 238 (7.0) | 126 (11.1) | < .001 | 4 (5.0) | .49 | 39 (9.5) | .06 |
| Metastatic solid cancer | 184 (5.4) | 61 (5.4) | .99 | 4 (5.0) | .88 | 11 (2.7) | .019 |
| Chest radiation | 50 (1.5) | 15 (1.3) | .73 | 0 | .31 | 2 (0.5) | .11 |
| GERD | 499 (14.5) | 128 (11.3) | .005 | 6 (7.5) | .07 | 41 (10.0) | .012 |
| Alcohol abuseb | 156 (6.1) | 31 (4.0) | .02 | 0 (0) | .046 | 11 (3.8) | .11 |
| Obesityc | 859 (25.1) | 320 (28.2) | .04 | 11 (13.8) | .02 | 91 (22.2) | .20 |
| Tobacco use | |||||||
| None | 1,609 (50.0) | 485 (47.3) | .14 | 61 (82.4) | < .001 | 216 (60.8) | < .001 |
| Former or current | 1,610 (50.0) | 540 (52.7) | .14 | 12 (17.6) | < .001 | 139 (39.2) | < .001 |
| ALI predisposing conditions | |||||||
| Shock | 248 (7.3) | 105 (9.3) | .03 | 11 (13.8) | .03 | 30 (7.3) | .95 |
| Aspiration | 154 (4.5) | 27 (2.4) | < .001 | 6 (7.5) | .21 | 8 (2.0) | .016 |
| Sepsis | 1,005 (29.4) | 556 (49.0) | < .001 | 26 (33.0) | .55 | 160 (39.1) | < .001 |
| Pancreatitis | 160 (4.7) | 72 (6.3) | .03 | 5 (6.3) | .51 | 63 (15.4) | < .001 |
| Pneumonia | 742 (21.7) | 324 (28.6) | < .001 | 16 (20.0) | .71 | 68 (16.6) | .018 |
| High-risk surgeryd | 1,065 (31.2) | 198 (17.4) | < .001 | 22 (27.5) | .49 | 98 (24.0) | .003 |
| Emergency surgery | 179 (5.2) | 74 (6.5) | .1 | 7 (8.8) | .17 | 22 (5.4) | .9 |
| High-risk trauma | 688 (20.1) | 152 (13.4) | < .001 | 17 (21.3) | .8 | 58 (14.2) | .004 |
| ALI risk modifiers | |||||||
| APACHE II | 8 (5-13) | 9 (5-15) | < .001 | 8 (4-14) | .93 | 7 (4-13) | .0011 |
| LIPS | 2.5 (2- 4.5) | 2.5 (1.5-3.5) | < .001 | 3 (1.8-4.3) | .85 | 2 (1-3.5) | < .001 |
| LIPS > 4 | 856 (25.0) | 222 (19.6) | < .001 | 20 (25.0) | .99 | 74 (18.1) | < .001 |
| Process of care | |||||||
| PRBC transfusion | 556 (16.3) | 148 (13.0) | .009 | 14 (17.5) | .77 | 37 (9.1) | < .001 |
| Noninvasive ventilation | 398 (12.3) | 84 (7.7) | < .001 | 5 (7.4) | .21 | 21 (5.3) | < .001 |
| Mechanical ventilation | 1,084 (32.8) | 249 (22.6) | < .001 | 24 (33.8) | .86 | 114 (28.3) | .07 |
Data are presented as No. (interquartile range) or No. (%). See Table 1 legend for expansion of abbreviations.
For comparisons, whites are baseline.
Alcohol abuse is defined as > 14 drinks/wk.
Obesity is defined as BMI > 30.
High-risk surgery is defined as cardiac, aortic, thoracic, spinal, and acute abdominal procedures.
A total of 409 patients were categorized as “other” race. Compared with whites, these patients were younger (P < .001), more likely to be Hispanic and to have diabetes and sepsis as risk modifiers, and less likely to have pneumonia (Table 2). They also had lower LIPS and APACHE II scores (P < .001).
Outcomes by Sex and Race
As shown in Table 3, baseline characteristics, predisposing conditions, and ALI risk modifiers differed depending on whether ALI developed. Men, compared with women, were more likely to develop ALI (6.9% vs 4.7%, P < .001) and experienced a longer ICU (P = .002) but not hospital (P = .10) length of stay (Table 4). Men who developed ALI had a nonsignificant increase in in-hospital mortality (27.6% vs 18.5%; P = .08; 95% CI, 0.94%-2.99%).
Table 3.
—Patient Demographics and Predisposing Conditions and ALI Development
| Characteristics | No ALI (n = 4,894) | ALI (n = 307) | P Value |
| Demographics | |||
| Age, y | 57 (43-71) | 56 (43-68) | .17 |
| Male sex | 2,687 (54.9) | 199 (64.8) | < .001 |
| Hispanic ethnicity | 451 (11.0) | 27 (9.4) | .43 |
| Race | |||
| White | 3,197 (67.4) | 222 (73.8) | Ref |
| Black | 1,084 (22.9) | 51 (16.9) | < .05 |
| Asian | 76 (1.6) | 4 (1.3) | .59 |
| Other | 385 (8.1) | 24 (8.0) | .63 |
| Admission source | |||
| Home | 3,894 (80.7) | 222 (72.6) | < .001 |
| Outside facility | 929 (19.3) | 84 (27.4) | < .001 |
| Predisposing conditions for ALI | |||
| Shock | 330 (6.7) | 69 (22.5) | < .001 |
| Aspiration | 171 (3.5) | 33 (10.8) | < .001 |
| Sepsis | 1,685 (34.4) | 118 (38.4) | .15 |
| Pancreatitis | 301 (6.2) | 9 (2.9) | < .001 |
| Pneumonia | 1,102 (22.5) | 94 (30.6) | < .001 |
| High-risk surgerya | 1,214 (24.8) | 67 (21.8) | .24 |
| Emergency surgery | 265 (5.4) | 37 (12.0) | < .001 |
| High-risk trauma surgery | 862 (17.6) | 72 (23.5) | < .05 |
| ALI risk modifiers | |||
| Alcohol abuseb | 191 (5.3) | 11 (5.5) | .17 |
| Obesityc | 1,211 (24.7) | 100 (32.6) | < .01 |
| Diabetes mellitus | 1,135 (23.2) | 60 (19.5) | NS |
| Smoking | |||
| None | 2,324 (51.2) | 129 (46.4) | .12 |
| Former | 1,070 (23.6) | 63 (22.7) | .72 |
| Active | 1,143 (25.2) | 86 (30.9) | < .05 |
| APACHE II | 8 (5-13) | 13 (9-20) | < .001 |
| LIPS | 2.5 (1.5-4) | 5 (3.5-7) | < .001 |
| Process of care | |||
| PRBC transfusion | 670 (13.7) | 101 (32.9) | < .001 |
| Use of noninvasive ventilation | 428 (9.2) | 91 (33.2) | < .001 |
| Mechanical ventilation | 1,243 (26.3) | 269 (87.9) | < .001 |
Data are presented as No. (interquartile range) or No. (%). NS = not significant. See Table 1 legend for expansion of other abbreviations.
High-risk surgery is defined as cardiac, aortic, thoracic, spinal, and acute abdominal procedures.
Alcohol abuse is defined as > 14 drinks/wk.
Obesity is defined as BMI > 30.
Table 4.
—Association Between Sex and Race and ALI Development and In-Hospital Mortality in Those Who Develop ALI
| Outcomes | Whole Cohort (N = 5,201) | Male (n = 2,886) | Female (n = 2,315) | P Value | White (n = 3,419) | Black (n = 1,135) | P Valuea | Asian (n = 80) | P Valuea | Other Race (n = 409) | P Valuea |
| ALI/ARDS | 307 (5.90) | 199 (6.9) | 108 (4.7) | < .001 | 222 (6.5) | 51 (4.5) | .01 | 4 (5) | .59 | 24 (5.9) | .63 |
| ALI/ARDS mortality | 75 (24.4) | 55 (27.6) | 20 (18.5) | .08 | 55 (24.8) | 11 (21.6) | .63 | 2 (50.0) | .25 | 4 (16.7) | .38 |
| Death, all cause | 259 (4.98) | 144 (4.99) | 115 (4.97) | .97 | 178 (5.2) | 56 (4.9) | .72 | 5 (6.3) | .68 | 13 (3.2) | .08 |
| ICU length of stay | 2 (1-5) | 2 (1-5) | 2 (0-5) | .002 | 2 (1-5) | 3 (1-5) | .10 | 3 (0-4) | .98 | 2 (0-4) | .02 |
| Hospital length of stay | 6 (3-10) | 6 (3-10) | 6 (3-9) | .10 | 6 (3-9) | 6 (4-10) | .04 | 4.5 (3-9) | .36 | 6 (4 -10) | .11 |
Data are presented as No. (interquartile range) or No. (%). See Table 1 legend for expansion of abbreviations.
For comparisons, whites are baseline.
Whites, compared with blacks, were more likely to develop ALI (6.5% vs 4.5%, P = .014); however, there was no difference in ALI-associated mortality (24.8% vs 21.6%, P = .63) (Table 4). Hispanic ethnicity was associated with neither ALI development (5.6% in Hispanics [27 of 478] and 6.6% in non-Hispanics [259 of 3,928], P = .43), nor ALI-associated mortality (P = .84).
There was no evidence of effect modification by sex in the relationship between race and ALI development (P = .35) or mortality (P = .90). When stratified by race and sex, we found that the rate of ALI development was highest among white men at 7.3%, followed by black men at 5.8%, white women at 5.4%, and then black women at 3.2%. Regarding ALI-associated mortality, in-hospital mortality was highest among white men at 27.1%, followed by black men at 24.2%, white women at 20.5%, and then black women at 16.7%.
In multivariable analyses, sex was not associated with ALI development after adjustment. However, black race, compared with white, was independently associated with a decreased likelihood of developing ALI after adjustment (Table 5). Because the modified Breslow-Day test for OR homogeneity across centers was significant (P = .08), suggesting significant heterogeneity across centers, we adjusted for center-level effects. The association between race and ALI development persisted after adjusting for center-level effects with hospital-specific random effects (n = 5,043; adjusted OR, 0.66; 95% CI, 0.45-0.97; P = .033) and when we used conditional logistic regression (n = 5,008; adjusted OR, 0.65; 95% CI, 0.44-0.96; P = .029). Further, the association persisted after excluding subjects from the three retrospective centers in our random effects models (n = 4,744; adjusted OR, 0.67; 95% CI, 0.45-0.99; P = .04).
Table 5.
—Multivariable Logistic Regression Model Demonstrating Association Between Sex and Race and ALI Development (n = 5,043)
| Variablea | Adjusted OR | 95% CI | P Value |
| Male sex | 1.24 | (0.96-1.63) | .10 |
| Race | |||
| White | Ref | Ref | |
| Blackb | 0.69 | (0.49-0.98) | .04 |
| Asian | 0.79 | (0.28-2.28) | .67 |
| Other | 1.08 | (0.67-1.74) | .75 |
See Table 1 legend for expansion of abbreviations.
Adjusted for age, BMI, comorbid conditions (diabetes mellitus, end-stage renal disease, class 4 CHF), aspirin use, APACHE II score, RBC transfusion, and LIPS. Alcohol abuse and tobacco use, when available, were included in the LIPS.
The association between black race and ALI development remained significant after adjusting for initial tidal volume per kilogram of predicted body weight in ventilated patients (n = 1,148; OR, 0.57; 95% CI, 0.36-0.90; P = .02).
In multivariable analyses, neither sex nor race was independently associated with ALI-related mortality after adjusting for potential confounders (Table 6). The relationships remained nonsignificant after adjusting for center-level effects (data not shown).
Table 6.
—Multivariable Logistic Regression Model Demonstrating Association Between Sex and Race and In-Hospital Mortality in Patients With ALI (n = 301)
| Variablea | Adjusted OR | 95% CI | P Value |
| Male sexb | 1.54 | (0.81-2.92) | .19 |
| Race | |||
| White | Ref | Ref | |
| Black | 0.87 | (0.39-1.93) | .73 |
| Asian | 3.48 | (0.35-34.98) | .29 |
| Other | 0.70 | (0.21-2.33) | .56 |
See Table 1 legend for expansion of abbreviation.
Adjusted for age, sepsis, shock, aspiration, and APACHE II.
The association between sex and in-hospital mortality remained nonsignificant after adjusting for initial tidal volume per kilogram of predicted body weight in ventilated patients (n = 224; OR, 2.08; 95% CI, 0.88-4.90; P = .10).
Discussion
In this multicenter cohort study of patients at high risk of ALI, we found that sex-based differences existed at clinical presentation prior to ALI development. After accounting for these differences, there existed no sex-based disparities in the rate of ALI development or ALI-related mortality. We found that race-based differences existed at presentation and during the hospitalization that modified the risk of ALI development. After accounting for these differences, we found that blacks remained significantly less likely to develop ALI compared with similarly at-risk whites.
This study has a number of strengths. First, this is a large multicenter study of patients at risk of ALI that used a valid definition to identify ALI cases. Prior studies relied on International Classification of Diseases, Ninth Revision codes, which have a low sensitivity for detecting ALI.4,28,29 Second, this cohort contained patient-level data, allowing for the assessment of baseline clinical differences and process-of-care differences. Last, this cohort examined the development of and mortality from ALI, allowing for a better understanding of where sex and racial disparities may lie.
Our study is the first, to our knowledge, to focus on the relationship between sex and race and ALI development in a population with one or more ALI risk factors. Prior studies focused on patients in whom ALI had developed or were limited to patients with trauma.30,31 Our finding that race-based differences exist in the rate of ALI development support findings in the trauma population.30 In the population-based epidemiologic study by Moss et al,4 men and blacks had increased ALI mortality. However, because this study focused on decedents with ALI, and details prior to ALI development were unavailable, it is unclear whether the observed disparities were due to a greater incidence of conditions that predispose to ALI, a more severe presentation, higher ALI development in patients at similar risk of ALI, or a similar incidence of ALI but a higher mortality after ALI development. We found that men and blacks are more likely to present to the hospital with conditions (eg, sepsis, shock) associated with the highest risk of ALI development.9,10,23,32‐35 In addition, we found significant baseline differences in clinical presentation by sex and race. However, after adjustment and center-level effects, men were no longer at a higher risk of developing ALI, whereas blacks were less likely to develop it.
Contrary to the findings of Moss et al,4 we did not find a relationship between ALI-related mortality and sex or race after adjusting for differences in clinical presentation. Our findings support the work of Erickson et al5 and Brown et al,36 who found that differences at hospital presentation accounted, in part, for disparities in ALI mortality. We acknowledge that our sample size may have limited our ability to detect whether disparities existed in ALI-related mortality, and our results may have differed had we included patients who presented with ALI or broadened the study population to include any patient who developed ALI rather than only at-risk patients, or had we examined long-term outcomes rather than hospital mortality. Given the disparities in discharge practices and posthospitalization care in other populations,33‐35,37‐39 and the greater use of long-term acute-care hospitals among men and blacks, long-term mortality may have differed from hospital mortality.34,37,38
It is not clear why blacks may have a higher incidence of sepsis but lower ALI development than do at-risk whites. One possibility is that chronic conditions that alter immune function and predispose to sepsis, such as diabetes, are more common in black patients,20,33 and yet diabetes appears to protect against ALI development.19,23,40 In this study, blacks remained significantly less likely to develop ALI when diabetes was included in the model. Another explanation is that sex and racial-based care differences during the hospitalization might influence development of and outcome in ALI. We found disparities in the rate of RBC transfusions and the use of mechanical ventilation, both of which have been associated with increased development of ALI.23,41 Our study suggests the need to focus on both prehospital factors and care provided during the hospitalization to identify and eliminate sex and racial disparities.42
Our study has several limitations. First, self-reported race is a poor surrogate for individual ancestry and genetic background; however, it has been important in defining different groups in society that experience health disparities. We acknowledge that such health disparities are multifactorial and can stem from genetic, socioeconomic, religious, cultural, and educational differences between groups. Unfortunately, such data were not available for analysis in this study. Second, our parent study excluded 166 patients presenting with ALI, and data regarding their sex and race are not available. Third, racial and ethnic determination depended on medical records, rather than an interview with the patient. Further, ethnic determination was missing in 15% of the patients and “other” race constituted 409 patients (8% of the study population), which may have biased our results in this subgroup in unforeseen ways. Fourth, given the observational study design, a potential for uncontrolled confounding exists. Although we adjusted for transfusion, other process-of-care measures, such as early goal-directed therapy and timely antibiotics for severe sepsis, known to differ by sex and race,43‐45 were not available. Fifth, we did not ascertain longer-term outcomes and transitions of care to long-term facilities, and further studies are required to assess this relationship.
Conclusions
We found that sex- and race-based differences exist at presentation in patients at high risk of ALI. These differences accounted for the sex-based difference in the rate of development of ALI and ALI-related mortality. These differences did not account fully for the association between race and ALI development, because black patients remained less likely to develop ALI. Finally, although we did not find disparities in short-term outcomes, future studies are required to study whether long-term outcomes differ as a result of differences in posthospitalization care.
Supplementary Material
Online Supplement
Acknowledgments
Author contributions: Dr Gong is the guarantor and assumes overall responsibility for the study.
Dr Lemos-Filho: contributed to the planning and execution of the statistical analysis and the manuscript preparation, including first draft and manuscript review.
Dr Mikkelsen: contributed to the manuscript preparation and review and the review and execution of the statistical analysis.
Dr Martin: contributed to the planning and execution of the study, the planning of the statistical analysis, and the manuscript review.
Dr Dabbagh: contributed to the planning and execution of the study, the planning of the statistical analysis, and the manuscript review.
Dr Adesanya: contributed to the planning and execution of the study, the planning of the statistical analysis, and the manuscript review.
Dr Gentile: contributed to the planning and execution of the study, the planning of the statistical analysis, and the manuscript review.
Dr Esper: contributed to the planning and execution of the study, the planning of the statistical analysis, and the manuscript review.
Dr Gajic: contributed to the planning and execution of the study, the planning of the statistical analysis, and the manuscript review.
Dr Gong: contributed to the development of the research plan, the conduct of the study, the supervision and guidance of the statistical analysis, the review of the results, and the manuscript writing and review.
Financial/nonfinancial disclosures: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.
Additional information: The e-Appendix can be found in the “Supplemental Materials” area of the online article.
Abbreviations
- ALI
acute lung injury
- APACHE
Acute Physiology and Chronic Health Evaluation
- IRB
institutional review board
- LIP
lung injury prediction
- PRBC
packed RBC
Footnotes
A complete list of study participants is located in e-Appendix 1 (342.9KB, pdf) .
Funding/Support: This study was supported by the United States Critical Illness and Injury Trials Group (USCIITG) [NIGMS U01 GM083407, NHLBI HL60710, HL086667, NHLBI 108712 and HL084060].
For editorial comment see page 881
Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.
References
- 1.Hudson LD, Steinberg KP. Epidemiology of acute lung injury and ARDS. Chest. 1999;116(suppl 1):74S-82S [DOI] [PubMed] [Google Scholar]
- 2.Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med. 2005;353(16):1685-1693 [DOI] [PubMed] [Google Scholar]
- 3.Ware LB, Matthay MA. The acute respiratory distress syndrome. N Engl J Med. 2000;342(18):1334-1349 [DOI] [PubMed] [Google Scholar]
- 4.Moss M, Mannino DM. Race and gender differences in acute respiratory distress syndrome deaths in the United States: an analysis of multiple-cause mortality data (1979- 1996). Crit Care Med. 2002;30(8):1679-1685 [DOI] [PubMed] [Google Scholar]
- 5.Erickson SE, Shlipak MG, Martin GS, et al. ; National Institutes of Health National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome Network Racial and ethnic disparities in mortality from acute lung injury. Crit Care Med. 2009;37(1):1-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lemos-Filho LB, Martin GS, Dabbagh O, et al. Race, gender, and the development of and mortality in ALI/ARDS. Am J Respir Crit Care Med. 2011;183A5595 [Google Scholar]
- 7.Gajic O, Dabbagh O, Park PK, et al. U.S. Critical Illness and Injury Trials Group: Lung Injury Prevention Study Investigators (USCIITG-LIPS) Early identification of patients at risk of acute lung injury: evaluation of lung injury prediction score in a multicenter cohort study. Am J Respir Crit Care Med. 2011;183(4):462-470 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ferguson ND, Frutos-Vivar F, Esteban A, et al. Clinical risk conditions for acute lung injury in the intensive care unit and hospital ward: a prospective observational study. Crit Care. 2007;11(5):R96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fowler AAHR, Hamman RF, Good JT, et al. Adult respiratory distress syndrome: risk with common predispositions. Ann Intern Med. 1983;98(5 pt 1):593-597 [DOI] [PubMed] [Google Scholar]
- 10.Hudson LD, Milberg JA, Anardi D, Maunder RJ. Clinical risks for development of the acute respiratory distress syndrome. Am J Respir Crit Care Med. 1995;151(2 pt 1):293-301 [DOI] [PubMed] [Google Scholar]
- 11.Stapleton RD, Wang BM, Hudson LD, Rubenfeld GD, Caldwell ES, Steinberg KP. Causes and timing of death in patients with ARDS. Chest. 2005;128(2):525-532 [DOI] [PubMed] [Google Scholar]
- 12.Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;149(3 pt 1):818-824 [DOI] [PubMed] [Google Scholar]
- 13.Johnston CJ, Rubenfeld GD, Hudson LD. Effect of age on the development of ARDS in trauma patients. Chest. 2003;124(2):653-659 [DOI] [PubMed] [Google Scholar]
- 14.Gong MN, Bajwa EK, Thompson BT, Christiani DC. Body mass index is associated with the development of acute respiratory distress syndrome. Thorax. 2010;65(1):44-50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Morris AE, Stapleton RD, Rubenfeld GD, Hudson LD, Caldwell E, Steinberg KP. The association between body mass index and clinical outcomes in acute lung injury. Chest. 2007;131(2):342-348 [DOI] [PubMed] [Google Scholar]
- 16.Sevransky JE, Martin GS, Mendez-Tellez P, et al. Pulmonary vs nonpulmonary sepsis and mortality in acute lung injury. Chest. 2008;134(3):534-538 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Calfee CS, Matthay MA, Eisner MD, et al. Active and passive cigarette smoking and acute lung injury after severe blunt trauma. Am J Respir Crit Care Med. 2011;183(12):1660-1665 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Moss M, Bucher B, Moore FA, Moore EE, Parsons PE. The role of chronic alcohol abuse in the development of acute respiratory distress syndrome in adults. JAMA. 1996;275(1):50-54 [PubMed] [Google Scholar]
- 19.Moss M, Guidot DM, Steinberg KP, et al. Diabetic patients have a decreased incidence of acute respiratory distress syndrome. Crit Care Med. 2000;28(7):2187-2192 [DOI] [PubMed] [Google Scholar]
- 20.Kor DJ, Erlich J, Gong MN, et al. ; U.S. Critical Illness and Injury Trials Group: Lung Injury Prevention Study Investigators Association of prehospitalization aspirin therapy and acute lung injury: results of a multicenter international observational study of at-risk patients. Crit Care Med. 2011;39(11):2393-2400 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.The Acute Respiratory Distress Syndrome Network Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med. 2000;342(18):1301-1308 [DOI] [PubMed] [Google Scholar]
- 22.Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13(10):818-829 [PubMed] [Google Scholar]
- 23.Gong MN, Thompson BT, Williams P, Pothier L, Boyce PD, Christiani DC. Clinical predictors of and mortality in acute respiratory distress syndrome: potential role of red cell transfusion. Crit Care Med. 2005;33(6):1191-1198 [DOI] [PubMed] [Google Scholar]
- 24.Netzer G, Shah CV, Iwashyna TJ, et al. Association of RBC transfusion with mortality in patients with acute lung injury. Chest. 2007;132(4):1116-1123 [DOI] [PubMed] [Google Scholar]
- 25.Maldonado G, Greenland S. Simulation study of confounder-selection strategies. Am J Epidemiol. 1993;138(11):923-936 [DOI] [PubMed] [Google Scholar]
- 26.Tarone RE. On heterogeneity tests based on efficiency scores. Biometrika. 1985;70(2):519-522 [Google Scholar]
- 27.Localio AR, Berlin JA, Ten Have TR, Kimmel SE. Adjustments for center in multicenter studies: an overview. Ann Intern Med. 2001;135(2):112-123 [DOI] [PubMed] [Google Scholar]
- 28.Howard AE, Courtney-Shapiro C, Kelso LA, Goltz M, Morris PE. Comparison of 3 methods of detecting acute respiratory distress syndrome: clinical screening, chart review, and diagnostic coding. Am J Crit Care. 2004;13(1):59-64 [PubMed] [Google Scholar]
- 29.Thomsen GE, Morris AH. Incidence of the adult respiratory distress syndrome in the state of Utah. Am J Respir Crit Care Med. 1995;152(3):965-971 [DOI] [PubMed] [Google Scholar]
- 30.Ryb GE, Cooper C. Race/ethnicity and acute respiratory distress syndrome: a National Trauma Data Bank study. J Natl Med Assoc. 2010;102(10):865-869 [DOI] [PubMed] [Google Scholar]
- 31.Heffernan DS, Dossett LA, Lightfoot MA, et al. Gender and acute respiratory distress syndrome in critically injured adults: a prospective study. J Trauma. 2011;71(4):878-883 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Burton DC, Flannery B, Bennett NM, et al. ; Active Bacterial Core Surveillance/Emerging Infections Program Network Socioeconomic and racial/ethnic disparities in the incidence of bacteremic pneumonia among US adults. Am J Public Health. 2010;100(10):1904-1911 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Esper AM, Moss M, Lewis CA, Nisbet R, Mannino DM, Martin GS. The role of infection and comorbidity: Factors that influence disparities in sepsis. Crit Care Med. 2006;34(10):2576-2582 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348(16):1546-1554 [DOI] [PubMed] [Google Scholar]
- 35.Mayr FB, Yende S, Linde-Zwirble WT, et al. Infection rate and acute organ dysfunction risk as explanations for racial differences in severe sepsis. JAMA. 2010;303(24):2495-2503 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Brown LM, Kallet RH, Matthay MA, Dicker RA. The influence of race on the development of acute lung injury in trauma patients. Am J Surg. 2011;201(4):486-491 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Joynt KE, Orav EJ, Jha AK. Thirty-day readmission rates for Medicare beneficiaries by race and site of care. JAMA. 2011;305(7):675-681 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kahn JM, Benson NM, Appleby D, Carson SS, Iwashyna TJ. Long-term acute care hospital utilization after critical illness. JAMA. 2010;303(22):2253-2259 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Medicine Io Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Acadamies Press;2001 [PubMed] [Google Scholar]
- 40.Esper AM, Moss M, Martin GS. The effect of diabetes mellitus on organ dysfunction with sepsis: an epidemiological study. Crit Care. 2009;13(1):R18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Zilberberg MD, Carter C, Lefebvre P, et al. Red blood cell transfusions and the risk of acute respiratory distress syndrome among the critically ill: a cohort study. Crit Care. 2007;11(3):R63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Smedley B, Stith A, Nelson A. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Institute of Medicine, Committee on Understanding and Eliminating Racial and Ethnic Disparities in Health Care. Washington, DC: National Acadamies Press;2003 [PubMed] [Google Scholar]
- 43.Mayr FB, Yende S, D’Angelo G, et al. Do hospitals provide lower quality of care to black patients for pneumonia?. Crit Care Med. 2010;38(3):759-765 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Mikkelsen ME, Gaieski DF, Goyal M, et al. Factors associated with nonadherence to early goal-directed therapy in the ED. Chest. 2010;138(3):551-558 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Polsky D, Jha AK, Lave J, et al. Short- and long-term mortality after an acute illness for elderly whites and blacks. Health Serv Res. 2008;43(4):1388-1402 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Online Supplement