Jacobs syndrome

A 15-Year Cohort Study in a Managed Care Setting

Study objective: To examine the association of cigarette smoking and alcohol consumption with hospital presentation of ARDS in a well-defined, multiethnic population.

Design: Retrospective cohort study.

Setting: Health maintenance organization in Northern California.

Participants: A total of 121,012 health plan subscribers (54.2% women), aged 25 to 89 years.

Outcome measure: Hospital presentation of ARDS (validated by medical chart review) from baseline in 1979 to 1985 through the end of 1993 (median, 9.9 years).

Results: There were 56 cases of ARDS (33 in men, 23 in women). The case fatality rate was 39% in both genders. ARDS was independently related to increasing age (rate ratio of 10 years, 1.38; 95% confidence interval [CI], 1.12 to 1.71), to current smoking of [is less than] 20 cigarettes/d (rate ratio vs never cigarette smokers, 2.85; 95% CI, 1.23 to 6.60), and to current cigarette smoking of [is greater than or equal to] 20 cigarettes/d (rate ratio vs never smokers, 4.59; 95% CI, 2.13 to 9.88). No association was observed between alcohol consumption and ARDS.

Conclusions: The results of this study suggest a relationship (with evidence of dose-response effect) between cigarette smoking and ARDS. Assuming a causal relationship, approximately 50% of ARDS cases were attributable to cigarette smoking. (CHEST 2000; 117:163-168)

Key words: alcohol; ARDS; epidemiology; risk factors; smoking

Abbreviations: BMI = body mass index; ICD-9 = Ninth Revision of the International Classification of Diseases

The ARDS is a severe clinical picture defined by the presence of dyspnea, tachypnea, hypoxemia, radiographic evidence of bilateral chest infiltrates, and decreased respiratory compliance, and is caused by inflammatory-cell-mediated endothelial lung injury.[1-3] It is a rapidly progressive syndrome, with an overall fatality rate of 40 to 60%.[1,4,5]

Since there is no specific treatment modality, the identification of patients at higher risk for the development of ARDS is paramount. Well-established clinical (or proximal) predictors of ARDS include sepsis, aspiration of gastric contents, and major trauma.[5-11] However, less is known about nonproximal factors that may increase the likelihood of eventually developing ARDS.

Cigarette smoking and alcohol consumption are potential nonproximal risk factors for ARDS. First, smoking increases the risk of many lung and systemic disorders predating ARDS,[12-16] and secondly, cigarette smoke contains highly reactive hydroxyl radicals capable of causing membrane peroxidation,[17] damage to DNA,[18] and inflammatory reactions.[12-16] In turn, chronic heavy alcohol consumption has been linked to trauma outcomes,[19,20] and to increased risk of complication during the course of hospitalization.[21]

The aim of this study was to test the hypothesis of whether cigarette smoking and alcohol consumption are independent risk factors for subsequent presentation of ARDS in a well-defined, multiethnic population.

Cohort Description

Participants were members of the Kaiser Permanente Medical Care Program (Northern California Region), a San Francisco Bay area-based health maintenance organization. The subscribers are socioeconomically diverse, but their average education level is higher than the general population in the area served by the health plan.[22] As a result, the lowest (but also the highest) end of the socioeconomic spectrum are underrepresented.

At the multiphasic health checkup (a voluntary medical examination at the time of enrollment), members filled questionnaires about demographics, medical history, smoking and alcohol use, and a blood sample was drawn for chemistry determinations. Participants were classified into never smokers (those who reported having never used any tobacco product); former cigarette smokers (ever smokers for at least 1 year, but no longer smoking cigarettes in the current year); and current cigarette smokers (still smoking cigarettes regularly, at least five cigarettes per week, almost every week). Current cigarette smokers were subclassified into smokers of [is less than] 20 cigarettes/d and smokers of [is greater than or equal to] 20 cigarettes/d. Likewise, subjects were classified as never drinkers (lifelong abstainers); former drinkers (ever drinkers for at least 1 year, but no longer consuming alcohol); current drinkers of less than three drinks per day; and current drinkers of more than or equal to three drinks per day during the past year.

Weight and height were measured following standardized procedures.[23] Body mass index (BMI) was calculated as weight (kg) divided by height ([m.sup.2]). Lung function tests including FVC and [FEV.sub.1] were performed using a Vertek VR5000 Lung Function computer (Electro/Med. Instruments; Houston, TX) according to methods previously described.[23] Lung time(ion tests as a component of the multiphasic health checkup were discontinued in 1981. For participants with repeated health examinations, only data from the first available multiphasic examination were used.

The total sample for this study contained 121,012 health plan subscribers (55,389 men and 65,790 women) between the ages of 2,5 and 89 years at the time of their first multiphasic health checkup at San Francisco or Oakland between 1979 and 1985. No exclusions were performed because of prevalent disease or missing data on study variables. This study was approved by the Institutional Review Board of The Kaiser Foundation Research Institute.

Outcome Measure and Medical Record Review

Since 1971, Kaiser Permanente of Northern California has maintained an automated database of all overnight hospitalizations in the 23 medical centers affiliated with the health plan. The database contains personal information, dates of hospital admission and discharge, discharge vital stares, the principal discharge diagnosis, and up to 12 possible secondary discharge diagnostic codes. For the purposes of this study, we used a computer algorithm that extracted codes 518.5 ("pulmonary insufficiency following trauma and surgery") and 518.82 ("other pulmonary insufficiency, not elsewhere classified") of the Ninth Revision of the International Classification of

Diseases (ICD-9),[24] regardless of whether they were a principal or secondary diagnosis, for the period from 1979 to 1993. Overall, 130 hospitalizations were found that listed codes 518.5 or 518.82. Following this initial screening of inpatient records, the medical charts of these 130 patients were requested to the Kaiser facilities in the region and reviewed by a study physician (C. I.) who was unaware of the risk factor status of the participants. The hospital charts of (bur patients (two men and two women) were not available for review (three were sent to physician's offices, and one to a non-Kaiser Permanente medical facility). Thus, medical record abstraction was performed on 126 charts. ARDS cases were defined as hospitalizations with ICD-9 codes 518.5 or 518.82, which met the European-American ARDS consensus conference definition: acute onset of respiratory distress (ie, within hours); bilateral infiltrates on chest radiograph; hypoxemia (ratio of [O.sub.2] saturation to concentration of [O.sub.2] in the inspired air [Pa[O.sub.2]/fraction of inspired oxygen] [is less than] 200 mm Hg); and no clinical evidence of left atrial hypertension (ascertained from physician's diagnostic impressions as noted in the discharge summary report).[1] Of the 126 reviewed cases, 56 "true-positives" met all four criteria for ARDS. Thus, the positive predictive value of ICD-9 codes 518.5 or 518.82 was 44% (56/126). Because charts without ICD-9 codes 518.5 or 518.82 were not reviewed, the negative predictive value, sensitivity, and specificity of diagnosis relying on ICD-9 codes could not be established. The most common reasons for the 70 false-positive cases were lack of evidence of bilateral infiltrates and/or hypoxemia.

A secondary aim of the medical chart review was to ascertain the proximal or precipitating factors leading to ARDS. This information was obtained directly from the "Hospital Course," section of the discharge summary report.

As an attempt to characterize the number of admissions in the ICU over the follow-up period, we identified procedures typically employed in the ICU setting, including assisted ventilation (ICD-9 codes 93.90, 93.91, 96.70, 96.71 and 96.72); arterial catheterization (38.91); central venous line pressure monitoring (89.62); pulmonary artery pressure monitoring (89.63); and Swan-Ganz catheterization (89.64). In addition, we identified patients who underwent one of more of these procedures and died during the course of his or her hospitalization.

Statistical Methods

Sex- and age-specific incidence rates of hospital presentation of ARDS were calculated both per [10.sup.5] persons and per [10.sup.5] person-years of follow-up. Person-time was defined for each participant as the time elapsed from baseline multiphasic health checkup to the earliest of the following: hospitalization for ARDS (n = 56); end of health plan membership (n = 59,570); or December 31, 1993, the closing date (n = 61,386). End of membership was defined as disenrollment for [is greater than] 2 consecutive years. Compared to individuals who remained in the health plan until the closing date, those who left the plan before the dosing date were, on average, younger (41 [+ or -] 15 vs 45 [+ or -] 13 years) and had lower BMI (24.6 [+ or -] 4.6 vs 25.3 [+ or -] 4.7 kg/[m.sup.2]), but did not differ importantly in ethnic background, education level smoking, and alcohol consumption habits (data not shown). The median length of follow-up time was 9.9 years, with a maximum of 15 years.

We used the Cox proportional hazards approach to model the age-adjusted association of the predictor variables of interest (cigarette smoking and alcohol consumption) with the outcome.[25] A multivariate model was then used that included also gender, race, education level, and BMI as possible confounders. To account for possible nonlinearity, BMI was entered in the regression as three categorical levels representing quartiles. To prevent loss of data, dummy variables were created for missing observations on education level, race, cigarette smoking, alcohol consumption, and BMI, respectively; none were related to the outcome. All analyses were conducted using appropriate software (SAS Version 6.11; SAS Institute; Cary, NC).[26]

We estimated attributable risk of ARDS due to smoking as the excess cases divided by the total observed cases, in percent. We first estimated the expected (background) number of cases if everyone had been a never smoker by applying the never smoker rate to the numbers of people in the various smoker and former smoker categories. The number of excess cases is then total observed cases minus the number of expected cases.

RESULTS

A total of 56 confirmed cases of ARDS (33 in men, 23 in women) were ascertained during the study period (from 1979 to 1993). The in-hospital case fatality rate was 39.4% in men (13/33) and 39.1% in women (9/23). The distribution of proximal or precipitating factors leading to ARDS were as follows: diffuse pulmonary infection (n = 17; 30.3%), sepsis syndrome (n = 16; 28.6%), aspiration (n = 15; 26.8%), hypertransfusion (n = 3; 5.3%), cardiopulmonary bypass (n = 2; 3.6%), lung contusion (n = 1; 1.8%), severe nonthoracic trauma (n = 1; 1.8%), and methotrexate-induced pneumonitis (n = 1; 1.8%).

Although we are very likely underestimating the number of ARDS cases, our data suggests an incidence of 4.6 cases/100,000 persons/yr. This estimate is derived from the finding of 56 cases in 9.9 years in a population of 121,012 persons.

Over the course of the study, we identified 401 hospitalizations with an assisted-ventilation code only; 924 hospitalizations with arterial, central venous line, pulmonary, or Swan-Ganz catheterization only; and 127 hospitalizations with both an assisted-ventilation code and a code for at least one of the catheterization modalities listed above. Based on these numbers, the estimated incidence rate of assisted ventilation was 33.5/100,000 persons/yr; the incidence rate of selected catheterization procedures was 77.1/100,000 persons/yr; and the incidence rate of both assisted ventilation and selected catheterization procedures was 10.6/100,000 persons/yr. The in-hospital mortality among those receiving assisted ventilation only, selected catheterization procedures, or both was 32.4%, 33.8%, and 44.9%, respectively.

Compared to the rest of the cohort, ARDS cases were older, more likely to be men, and white (Table 1). There was a greater prevalence of current cigarette smoking (both [is less than] 20 and [is greater than or equal to] 20 cigarettes/d) and of current alcohol consumption of three or more drinks per day among ARDS cases compared to cohort counterparts. No difference was observed for BMI.

Table 1--Baseline Characteristics of ARDS Cases vs the Rest of the Cohort. Kaiser Permanente Medical Care Program, Northern California Region (1979-1985)

(*) p value for t test comparison of continuous variables and [chi square] comparison of categorical variables.

([dagger]) Presented as means (SD).

([double dagger]) Hispanic or Latino nonwhite, American Indian or Alaska Native, Native Hawaiian or Pacific Islander.

The false-positive cases (those with ICD-9 codes 518.5 or 518.82 but not meeting the study criteria for ARDS) were older (57 [+ or -] 11 years), more likely to be black (29%), and more likely to be current smokers of [is less than] 20 cigarettes/d (10%) and of [is greater than or equal to] 20 cigarettes/d (23%).

The incidence rates of hospital presentation of ARDS were higher in men than in women (Table 2). In women, the rates increased monotonically with increasing age; in men, the rates increased from ages 25 to 44 years to ages 45 to 64 years, but not from ages 45 to 64 years to ages [is greater than] 65 years.

Table 2--Gender- and Age-Specific Incidence Rates of ARDS. Kaiser Permanente Medical Care Program, Northern California Region (1979-1993)(*)

(*) Person-years and number of events are allocated to age categories based on age at baseline.

Results of the age- and multivariate-adjusted analyses are given in Table 3. In the univariate models, the risk of ARDS increased with advancing age, male gender, current smoking, and former alcohol consumption. There was a clear dose-response association with cigarette smoking, with a 5.7-fold increased risk among heavy smokers (ie, [is greater than or equal to] 20 cigarettes/d), and a borderline association with former cigarette smoking. In the multivariate pooled analysis, ARDS continued to be related to increasing age, and to cigarette smoking in a dose-response fashion. It was also observed that, when other risk factors were considered, blacks were at lower risk than their white counterparts, and that former alcohol consumption was no longer related to the outcome. No apparent independent relationship existed with average consumption of three alcoholic drinks or more per day. There was a suggestion of a U-shaped association with BMI.

Table 3--Relative Risks and 95% Confidence Intervals for ARDS Associated With Selected Baseline Variables. Kaiser Permanente Medical Care Program, Northern California Region (1979-1993)(*)

(*) Analysis based on 56 events in 121,012 persons.

([dagger]) Separate models for each categorical or continuous predictor, adjusting for age and gender.

([double dagger]) All predictors in the same model.

([sections]) Hispanic or Latino nonwhite, American Indian or Alaska Native, Native Hawaiian or Pacific Islander.

([parallel]) Number of events, by smoking status: never (14), former (15), current < 20 cigarettes/d (10), current [is greater than or equal to] 20 cigarettes/d (16), unknown (1); number of events, by alcohol consumption status: never (5), former (5), current < 3 drinks/d (32), current [is greater than or equal to] 3 drinks/d (9), unknown (5).

In a sensitivity analysis including all cases with codes 518.5 and 518.82 (n = 130), the multivariate-adjusted relative risks associated with current cigarette smoking of [is less than] 20 and [is greater than or equal to] 20 cigarettes/d (vs never smoking) were 1.57 (95% confidence interval, 0,86 to 2.84) and 3.82 (95% confidence interval, 2.34 to 6.27), respectively.

If all participants had been never smokers, and the never smoker rate had prevailed, only 6 of 15 cases would have occurred among former smokers, only 4 of 10 cases among smokers of [is less than] 20 cigarettes/d, and only 4 of 16 cases among smokers of [is greater than or equal to] 20 cigarettes/d. Thus risk of ARDS attributable to cigarette smoking is estimated to be 50%.

We also examined in the full cohort the association of ARDS with other physiologic variables, such as blood cholesterol, systolic BP, and WBC counts; these were found to be unrelated to the outcome and thus were not considered in the final analysis. In a subset of 46,138 men and women who had valid spirometry data (among whom, 21 cases of ARDS developed), no associations were found between FVC or [FEV.sub.1] and ARDS (data not shown).

DISCUSSION

In this insured, ethnically diverse population of the San Francisco Bay area, the incidence of ARDS was low but comparable to recent estimates with similar strict criteria for ARDS.[27-29] The incidence of 4.6 cases of ARDS per 100,000 persons/yr in our population was about half the estimated incidence of hospitalizations with both assisted ventilation and selected catheterization procedures (10.6/100,000 persons/yr). Furthermore, the case fatality rate of ARDS (39%) was only slightly lower to that of inpatients who were known to have undergone assisted ventilation and selected catheterization procedures (45%).

The main finding of the study was an independent dose-response association between current cigarette smoking and the subsequent hospital presentation of ARDS. This predictive association may be explained by the fact that cigarette smokers, as shown in previous studies but not specifically examined here, are more likely to develop acute precipitating factors of ARDS such as pneumonia, sepsis syndrome, or injury, and/or to undergo digestive or cardiopulmonary surgery.[30-32] These causal pathways are also supported by evidence of impaired immunocompetence in smokers.[33,34]

In addition to the effect mediated through precipitating factors, smoking could cause alveolar damage and thus directly contribute to respiratory insufficiency and ARDS. Cigarette smoke contains high concentrations of reactive oxygen species capable of causing membrane peroxidation,[17] DNA adducts,[18] and inflammation.[12-16,35]

The risk of ARDS associated with smoking is likely to be an underestimation due to study participants who might have quit smoking during follow-up. However, this source of bias could also operate in the other direction if, for example, ex-smokers may have gone back to smoking or because some subjects could have started smoking after entry.

The relationship with increasing age is not surprising, given the reduced host defense mechanisms with advanced age,[36,37] and the age-related increased incidence of underlying illnesses, including pneumonia.

On the other hand, we found no association between alcohol consumption and ARDS incidence. One study has demonstrated that a prior history of chronic alcohol abuse increases the risk of developing ARDS in patients with an identified clinical at-risk diagnosis.[38] Moreover, some in vitro studies have shown that ethanol induces changes in the lung[39] and alters neutrophil function in a dose-dependent fashion.[40,41] A limitation is that the present study did not specifically address chronic alcohol abuse.

A noteworthy yet unexpected finding was a lower risk of ARDS among blacks. This result could be due to chance, and should be replicated in future studies before any interpretation can be made.

The use of ICD-9 codes 518.5 and 518.82 is likely to underestimate the true incidence of ARDS in this population. Clearly, cases coded as pneumonia, pulmonary edema, or sepsis could have also met criteria for ARDS. However, after conducting a medical chart review, we were able to identify 56 true-positive eases and rule out 70 false-positive cases. This approach is a valid one because we were interested in risk relationships, and not in determining the overall incidence of ARDS. It is our believe that, unless smoking was associated with the probability of being a true-positive case (vs being a false-negative case), then confounding should not occur.

Another limitation of the study was that the chart reviewer was not blinded to the study hypothesis. In addition, our case-ascertainment method yielded only two trauma-related ARDS cases. Thus, the current findings cannot be generalized to ARDS associated with trauma. In summary, older age and cigarette smoking were significant risk factors for the development of ARDS in our large multiethnic population. Thus, it appears that the decision to be a smoker may substantially elevate the risk of a near-fatal or fatal outcome in a critical clinical situation. These epidemiologic findings are important because they suggest that many cases of ARDS may be preventable through avoidance of smoking, probably because less smoking would lead to fewer intermediate conditions, such as pneumonia, trauma, and cardiopulmonary surgery.

REFERENCES

[1] Bernard GR, Artigas A, Brigham KL, et al. Report of the American-European Consensus Conference on the acute respiratory distress syndrome: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am Rev Respir Dis 1994; 151:818-824

[2] Knaus WA, Sun X, Hakim RB, et al. Evaluation of definitions for adult respiratory distress syndrome. Am J Respir Crit Care Med 1994; 150:311-317

[3] Fein A, Wiener-Kronish JP, Niederman M, et al. Pathophysiology of the adult respiratory distress syndrome: what have we learned from human studies? Crit Care Clin 1986; 2:429-453

[4] Anzueto A, Baughman RP, Guntupalli KK, et al. Aerosolized surfactant in adults with sepsis-induced acute respiratory distress syndrome. N Engl J Med 1996; 334:1417-1421

[5] Krafft P, Fridrich P, Pernerstorfer T, et al. The acute respiratory distress syndrome: definitions, severity and clinical outcome; an analysis of 101 clinical investigations. Intensive Care Med 1996; 22:519-529

[6] Hudson LD. Epidemiology of the adult respiratory distress syndrome. Semin Respir Crit Care Med 1994; 15:254-259

[7] Parsons PE, Worthen GS, Moore EE, et al. The association of circulating endotoxin with the development of the adult respiratory distress syndrome. Am Rev Respir Dis 1989; 140:294-301

[8] Niederman MS, Fein AM. Sepsis syndrome, the adult respiratory distress syndrome, and nosocomial pneumonia: a common clinical sequence. Clin Chest Med 1990; 11:6:33-6,56

[9] Doyle RL, Szaflarski N, Modin GW, et al. Identification of patients with acute lung injury: predictors of mortality. Am J Respir Crit Care Med 1995; 152:1818-1824

[10] Messent M, Sullivan K, Keogh BF, et al. Adult respiratory distress syndrome following cardiopulmonary bypass: incidence and prediction. Anesthesia 1992; 47:267-268

[11] Pepe PE, Potkin RT, Reus DH, et al. Clinical predictors of the adult respiratory distress syndrome. Am J Surg 1982; 144:124-130

[12] McKarns SC, Smith CJ, Morton MJ, et al. Correlation of hematologic markers of inflammation and lung function: a comparison of asymptomatic smokers and nonsmokers. Hum Exp Toxicol 1996; 15:523-532

[13] Wesselius LJ, Nelson ME, Bailey K, et al. Rapid lung cytokine accumulation and neutrophil recruitment after lipopolysaccharide inhalation by cigarette smokers and nonsmokers. J Lab Clin Med 1997; 129:106-114

[14] Finkelstein R, Fraser RS, Ghezzo H, et al. Alveolar inflammation and its relation to emphysema in smokers. Am J Respir Crit Care Med 1995; 152:1666-1672

[15] McCrea KA, Ensor JE, Nail K, et al. Altered cytokine regulation in the lungs of cigarette smokers. Am J Respir Crit Care Med 1994; 150:696-703

[16] Mills CM, Hill SA, Marks R. Altered inflammatory responses in smokers. BMJ 1993; 307:911

[17] Vruwink KG, Gershwin ME, Sachet P, et al. Modification of human LDL by in vitro incubation with cigarette smoke or copper ions: implications for allergies, asthma and atherosclerosis. J Investig Allergol Clin Immunol 1996; 6:294-300

[18] Degawa M, Stern SJ, Martin MV, et al. Metabolic activation and carcinogen-DNA adduct detection in human larynx. Cancer Res 1994; 54:4915-4919

[19] Iribarren C, Reed DM, Wergowske G, et al. Serum cholesterol and mortality due to suicide and trauma in the Honolulu Heart Program. Arch Intern Med 1995; 155:695-700

[20] Boffeta P, Garfinkel L. Alcohol drinking and mortality among men enrolled in an American Cancer Society prospective study. Epidemiology 1990; 1:342-348

[21] Jurkovich GJ, Rivara FP, Gurney JG, et al. The effect of acute alcohol intoxication and chronic alcohol abuse on outcome from trauma. JAMA 1993; 270:51-56

[22] Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health 1992; 82:703-710

[23] Collen MF, ed. Multiphasic health testing services. New York, NY: John Wiley & Sons, 1978; 153-189

[24] International Classification of Diseases, 9th Revision. 5th ed. Salt Lake City, UT: Medicode Inc., 1998; 1:138

[25] Cox DR. Regression models and life tables. J R Stat Soc 1972; 34:187-202

[26] SAS/STAT Software: changes and enhancements through Release 6.12. Cary, NC: SAS Institute, 1997; 871-948

[27] Villar J, Slutsky AS. The incidence of the adult respiratory distress syndrome. Am Rev Respir Dis 1989; 140:814-816

[28] Thomsen GE, Morris AH. Incidence of the adult respiratory distress syndrome in the state of Utah. Am J Respir Crit Care Med 1995; 152:965-971

[29] Garber BG, Hebert PC, Yelle JD, et al. Adult respiratory distress syndrome: a systemic overview of incidence and risk factors. Crit Care Med 1996; 24:687-695

[30] Nakajima M, Manabe T, Niki Y, et al. Cigarette smoke-induced acute eosinophilic pneumonia. Radiology 1998; 207: 829-831

[31] Salive ME, Wallace RB, Ostfeld AM, et al. Risk factors for septicemia-associated mortality in older adults. Public Health Rep 1993; 108:447-453

[32] Kaul TK, Fields BL, Riggins LS, et al. Adult respiratory distress syndrome following cardiopulmonary bypass: incidence, prophylaxis and management. J Cardiovasc Surg 1998; 39:777-781

[33] Bridges RB, Chow CK, Rehm SR. Micronutrient status and immune function in smokers. Ann N Y Acad Sci 1990; 587:218-231

[34] Meliska CJ, Stunkard ME, Gilbert DG, et al. Immune function in cigarette smokers who quit smoking for 31 days. J Allergy Clin Immunol 1995; 95:901-910

[35] Matsumoto K, Aizawa H, Inoue H, et al. Eosinophilic airway inflammation induced by repeated exposure to cigarette smoke. Eur Respir J 1998; 12:387-394

[36] Roberts-Thompson IC, Yvonschaiyud V, Whittingham S. Aging, immune response and mortality. Lancet 1974; 2:368-370

[37] Wayne SJ, Rhyne RL, Garry Pi, et al. Cell-mediated immunity as a predictor of morbidity and mortality in the aged. J Gerontol Med Sci 1990; 45:M45-M48

[38] Moss M, Bucher B, Moore FA, et al. The role of chronic alcohol abuse in the development of acute respiratory distress syndrome in adults. JAMA 1996; 275:50-54

[39] Liau DF, Hashim SA, Pierson RN III, et al. Alcohol-induced lipid change in the lung. J Lipid Res 1981; 22:680-686

[40] Spagnuolo PJ, MacGregor RR. Acute ethanol effect on chemotaxis and other components of host defense. J Lab Clin Med 1975; 86:24-31

[41] Hallengren B, Forsgren A. Effect of alcohol on chemotaxis, adherence and phagocytosis of human polymorphonuclear leukocytes. Acta Med Scand 1978; 204:43-48

(*) From the Kaiser Permanente Division of Research (Drs. Iribarren and Sidney), Oakland, CA; the Division of Epidemiology, School of Public Health (Drs. Jacobs and Gross), University of Minnesota, Minneapolis, MN; and the Department of Medicine, Division of Pulmonary and Critical Care Medicine and Cardiovascular Research Institute (Dr. Eisner), University of California, San Francisco, CA.

This study was supported by contract RO1-AG-12264-01A1 from the National Institute on Aging, National Institutes of Health, Bethesda, MD.

Manuscript received February 23, 1999; revision accepted June 16, 1999.

Correspondence to: Carlos Iribarren, MD, MPH, PhD, Kaiser Permanente Division of Research, 3505 Broadway, Oakland, CA 94611; e-mail: cgi@dor.kaiser.org

COPYRIGHT 2000 American College of Chest Physicians
COPYRIGHT 2000 Gale Group