Journal of Critical Care (2011) 26, 180–185

Morbidity and mortality characteristics of morbidly obesepatients admitted to hospital and intensive care units☆

Blair D. Westerly MDa,1, Ousama Dabbagh MD, MSPHb,⁎

aDepartment of Medicine, Vanderbilt University School of Medicine, Nashville, TN 37232, USAbDivision of Pulmonary Critical Care Medicine, University of Missouri-Columbia School of Medicine, Columbia,MO 65212, USA

inreS

rein

0d

Keywords:Intensive care unit;Obesity;Morbid obesity;Body mass index;Mechanical ventilation

AbstractBackground: The purpose of this study is to evaluate the outcomes of hospitalized morbidlyobese inpatients.Methods: In this retrospective cohort study, we reviewed the records of all adult morbidly obese patients(defined as body mass index [BM]) N40 kg/m2 upon admission) admitted to tertiary university hospitalfrom 2000 to 2008. Primary outcome was hospital mortality. Secondary outcomes were hospital andintensive care unit (ICU) length of stay (LOS), need for and duration of mechanical ventilation (MV),and tracheostomy rates. We divided patients into quartiles based on their admission BMI. Baselinecharacteristics and outcomes were reported for each quartile.Results: Over the 8-year period, we reviewed 897 admissions for 545 patients. The median number ofadmissions was 1 per patient (mean, 2.44 ± 2.9), with a range of 1 to 20. A total of 40.9% had more thanone admission. Morbidly obese patients were more likely to be admitted to a medical service. HigherBMI quartiles had higher rates of ICU admission, MV, and rate of tracheostomy. Although the higherBMI quartiles had longer hospital LOS, hospital mortality did not significantly differ.Conclusions: As BMI increases, utilization of medical resources also increases such as ICU admission,MV, longer hospital LOS, and tracheostomy. Although overall BMI interquartile mortality rates do notdiffer significantly in our study, utilization of valuable and costly hospital resources is a major challengefacing health care delivery. Our findings indicate the need for increased efforts and novel strategies fortreatment, prevention, and resource allocation to deal with this emerging challenge.© 2011 Elsevier Inc. All rights reserved.

☆ Conflict of interest disclosures: Blair Westerly has no conflict ofterest. Ousama Dabbagh: received on behalf of the University of Missourisearch grants from Pfizer and BMS for other projects. Speaker Bureau:anofi-Aventis. No funding source for this project.⁎ Corresponding author. Tel.: +1 573 882 0808.E-mail address: dabbagho@health.missouri.edu (O. Dabbagh).1 Author's roles: Blair Westerly: design, data collection, and draft and

view of the manuscript. Ousama Dabbagh: concept, design, analysis,terpretation, manuscript review, and overall supervision.

883-9441/$ – see front matter © 2011 Elsevier Inc. All rights reserved.oi:10.1016/j.jcrc.2010.09.005

1. Introduction

The increasing incidence of obesity, defined as a bodymassindex (BMI) greater than 30 kg/m2, has been a national trend intheUnitedStates since themid-1970s [1].By2000,greater than30%ofAmericanswereobese; and an estimated13million hadBMIs greater than 40 kg/m2, meeting criteria for morbidobesity [2]. The direct medical expenditures attributed toobesity, includingpreventative services, diagnostic testing, and

181Morbidity and mortality characteristics of morbidly obese

treatment, accounted for 9.1% of US medical expendituresin 1998 and up to $92.6 billion in 2002 [3]. The burdenof obesity falls not only on the obese themselves, but alsoupon all American taxpayers, as nearly half of theseexpenses were paid by Medicare and Medicaid [3].

Despite existing literature regarding the outcomes of theobese population admitted to hospitals and intensive careunits (ICUs), the current knowledge base insufficientlydescribes the outcomes of morbidly obese patients when theyare admitted to the hospital or ICU. Historically, patientswith BMIs greater than 40 kg/m2 have been consideredcategorically one group and compared with those with lowerBMIs; but the outcomes as BMI increases above thisarbitrary cutoff have not been reported. We examined theoutcomes within the group of morbidly obese patients toevaluate differences that may exist within this group.

2. Methods

This study was conducted by a retrospective chart reviewof all patients admitted to the University of Missouri Hospitalfrom December 11, 2000, to April 1, 2008. The studyprotocol was approved by the Institutional Review Board.Analysis was restricted to adult patients (age≥18 years) whohad height and weight recorded in the electronic medicalrecord to calculate each patient's BMI. Exclusion criteriaincluded pregnancy, calculated BMI less than 40 kg/m2, andunavailability of baseline BMI data.

Baseline data collection obtained from admission informa-tion included age, sex, height, weight, race, admitting service(medical or surgical), and comorbid conditions including thefollowing: a positive smoking history, chronic obstructivepulmonary disease (COPD), diabetes mellitus (DM), conges-tive heart failure (CHF), coronary artery disease (CAD), andrenal insufficiency (defined as creatinine N2.5 mg/dL onadmission). Primary outcome was hospital mortality. Second-

Table 1 Baseline characteristics of the study group by BMI quartiles

1st BMI 40-47.5n = 221

2nd BMI 47.6-54.6n = 224

Age mean (SD) 48.86 (14.22) 50.79 (11.71)Admitting serviceMedical n (%) 155 (70.1) 155 (69.2)Surgical n (%) 66 (29.9) 69 (30.8)Female n (%) 72 (32.6) 92 (41.1)LMWH n (%) 86 (38.9) 74 (33)Smoking n (%) 108 (48.9) 106 (47.5)COPD n (%) 30 (13.6) 27 (12.9)DM n (%) 119 (53.8) 145 (64.7)CHF n (%) 81 (36.7) 115 (51.3)CAD n (%) 77 (34.8) 54 (24.1)Renal insufficiency n (%) 29 (13.3) 33 (14.7)

Renal insufficiency was defined as creatinine N2.5 mg/dL.

ary outcomes included hospital length of stay (LOS), ICULOS, need for and duration of mechanical ventilation (MV),tracheostomy placement, and diagnosis of venous thrombo-embolism (VTE). Venous thromboembolism is defined as thedevelopment of deep vein thrombosis as diagnosed by venousDoppler ultrasound or venogram or pulmonary embolism asdocumented by positive spiral computed tomography, highlyprobable ventilation/perfusion scan, or pulmonary angiogra-phy. Hospital and ICU LOSs were calculated as number ofcalendar days of hospital and ICU stay, respectively.Mechanical ventilation duration was calculated as number ofcalendar days onMV.Hospital mortalitywas defined as deathby any cause during the same admission.

2.1. Statistical analysis

Patients were divided into 4 quartiles according to theBMI. The quartiles were defined according to the distributionof BMI; and therefore, the 4 groups represented 25, 50, 75,and higher percentiles. Baseline characteristics were com-pared among the 4 groups. Continuous variables wereexpressed as means or medians with standard deviations orinterquartile ranges according to normality testing usingKolmogorov-Smirnov tests. Categorical values wereexpressed as proportions or percentages. Comparisons wereperformed using analysis of variance; (and) Kruskal-Wallistest for continuous variables. Categorical variables werecompared using χ2 or Fisher exact tests. Based uponadmission BMI, the cohort was then divided into quartilesof increasing BMI: 40 to 47.5, 47.6 to 54.6, 54.7 to 65.0, andgreater than 65.0 kg/m2. The subgroup of patients who wereadmitted to ICU at any time during their hospitalization wasanalyzed for baseline characteristics, and the primary andsecondary outcomes were compared for each quartile. Theinterquartile groups were compared using the aboveparametric and nonparametric tests, which were thenrepeated for the lowest vs highest interquartile groups.

3rd BMI 54.7-65n = 227

4th BMI N65n = 225

All N = 897 P

49.78 (11.34) 46.38 (10.36) 48.95 (12.09) b.001b.001

119 (52.4) 148 (65.8) 577 (64.3)108 (47.6) 77 (34.2) 320 (35.7)142 (62.6) 140 (62.1) 446 (49.7) b.00187 (38.3) 92 (40.9) 339 (37.8) .35993 (41.7) 90 (40.5) 397 (44.7) .243 (19.1) 41 (18.2) 141 (15.8) .113138 (61.3) 150 (66.7) 552 (61.7) .02994 (42) 97 (43.1) 387 (43.3) .01841 (18.2) 37 (16.4) 209 (23.4) b.00122 (9.7) 36 (16) 120 (13.4) .235

Table 2 Primary and secondary outcome measures for the study group

1st BMI 2nd BMI 3rd BMI 4th BMI All P40-47.5 47.6-54.6 54.7-65 N65*n = 221 n = 224 n = 227 n = 225 n = 897**n = 127 n = 151 n = 147 n = 120 n = 545

Hospital mortality n (%) 9 (4.1) 11 (4.9) 14 (6.2) 17 (7.6) 51 (5.7) .409VTE n (%) 7 (3.2) 10 (4.5) 13 (5.7) 6 (2.7) 36 (4.0) .345Need for ICU n admissions (%) 43 (19.5) 51 (22.8) 61 (26.9) 69 (30.7) 224 (25.0) .037Admitted to ICU 36 (28.3) 49 (32.5) 48 (32.7) 46 (38.3) 179 (32.8)More than 1 ICU admission 4 2 9 9 24MV n (%) 21 (9.5) 29 (12.9) 40 (17.6) 44 (19.6) 134 (14.9) .012Hospital LOS median (range) 3 (2-6) 4 (2-9) 4 (2-9) 5 (3-13) 4 (2-8) b.001Tracheostomy n (%) 10 (4.5) 11 (4.9) 24 (10.6) 30 (13.3) 75 (8.4) .001

* represents number of admissions.** represents number of patients.

182 B.D. Westerly, O. Dabbagh

Comparisons were made between the lowest and highestBMI quartiles using the Mantel-Haenszel tests for categor-ical variables and the Mann-Whitney tests for continuousvariables. Significance was defined as P value b .05, and alltests were 2-sided. Data were analyzed using the SPSSstatistical package (version 17; Chicago, IL, 2008).

3. Results

We included 897 admissions for 545 patients who met theinclusion criteria. The median number of admissions was1 per patient (mean, 2.44 ± 2.9), with a range of 1 to 20. Atotal of 40.9% of these patients had more than one admissionduring this period.

Table 1 shows the clinical characteristics of each quartile.As a group, morbidly obese patients were more likely to beadmitted to a medical service than a surgical service (64.3%vs 35.7%, P b .001). In total 49.7% of patients were females,with distribution weighted toward the highest quartiles.Females represented, respectively, 32.6% vs 41.1% vs 62.

Table 3 Baseline characteristics and outcomes of patients admitted t

1st BMI 40-47.5n = 43

2nd BMI 47.6-54.6n = 51

Age mean (SD) 54.2 (15.14) 52.1 (11.97)Female n (%) 14 (32.6) 54.6 (29.2)LMWH n (%) 22 (51.2) 21 (41.2)Smoking n (%) 23 (53.5) 28 (54.9)COPD n (%) 12 (27.9) 8 (15.7)DM n (%) 25 (58.1) 27 (52.9)CHF n (%) 16 (37.2) 20 (39.2)CAD n (%) 11 (25.6) 12 (23.5)Renal insufficiency n (%) 14 (32.6) 9 (17.6)Admitting serviceMedical n (%) 28 (65.1) 32 (62.7)Surgical n (%) 15 (34.9) 19 (37.3)

Renal insufficiency was defined as creatinine N2.5 mg/dL.

6% vs 62.1% (P b .001). Smoking rates, as well as the use oflow–molecular weight heparins (LMWH), were not signif-icantly different between quartiles (38.9% vs 33% vs 38.3%vs 40.9%, P = .359). Quartiles did not differ in prevalence ofCOPD (13.6% vs 12.9% vs 19.1% vs 18.2%, P = .113) orrenal insufficiency (13.3% vs 14.7% vs 9.7% vs 16%, P =.235 for all the quartiles and P = .422 between the highestand lowest). Each quartile differed with respect to DM(53.8% vs 64.7% vs 61.3% vs 66.7%, P = .029 for all thequartiles and P = .006 between the highest and lowest) butwas not linearly correlated with increasing BMI. Coronaryartery disease was less frequently diagnosed with increasingBMIs (34.8% vs 24.1% vs 18.2% vs 16.4%, respectively,P b .001 for all quartiles and P b .0001 between highest andlowest). This trend was not seen in CHF; patients in thesecond quartile had more CHF compared with those in theother quartiles (36.7% vs 51.3% vs 42% vs 43.1%, P = .018).

3.1. Primary and secondary outcomes

Outcome data for each quartile's admissions are summa-rized in Table 2. Hospital mortality did not significantly

o the ICU

3rd BMI 54.7-65n = 61

4th BMI N65n = 69

All n = 224 P

52.5 (11.47) 47.7 (11.03) 51.3 (12.43) .03340 (65.6) 32 (46.4) 106 (47.3) .00428 (45.9) 32 (46.4) 103 (46.0) .81528 (46.7) 32 (47.1) 111 (50.0) .75319 (31.1) 21 (30.4) 60 (26.8) .23241 (67.2) 47 (68.1) 140 (62.5) .27933 (55.0) 30 (43.5) 99 (44.4) .24016 (26.2) 7 (10.1) 46 (20.5) .0815 (8.2) 11 (15.9) 39 (17.4) .014

.97537 (60.7) 43 (62.3) 140 (62.5)24 (39.3) 26 (37.7) 84 (37.5)

Table 4 Primary and secondary outcomes of patients admitted to the ICU

1st BMI 40-47.5n = 43

2nd BMI 47.6-54.6n = 51

3rd BMI 54.7-65n = 61

4th BMI N65n = 69

All n = 224 P

H-los median (range) 7 (4-17) 11 (5-26) 10 (4-39) 14 (6-27) 10 (5-25) .222Mortality n (%) 7 (16.3) 8 (15.7) 8 (13.1) 15 (21.7) 38 (17) .608VTE n (%) 3 (7.0) 5 (9.8) 5 (8.2) 5 (7.2) 18 (8.0) .952ICU LOS median (range) 4 (2-10) 5 (2-16) 4 (2-19) 6 (3-17) 5 (2-15) .490MV n (%) 21 (48.8) 28 (54.9) 39 (63.9) 44 (63.8) 132 (58.9) .327Tracheostomy n (%) 9 (20.9) 11 (21.6) 22 (36.1) 25 (36.2) 67 (29.9) .124MVD median (range) 0 (0-3) 1 (0-10) 2 (0-13) 2 (0-13) 1 (0-10) .293MV-free days median (range) 28 (0-26) 27 (0-18) 26 (0-15) 26 (0-15) 27 (0-18) .457

H-los indicates hospital LOS; MVD, mechanical ventilation duration (days).

183Morbidity and mortality characteristics of morbidly obese

differ with increasing BMI quartiles (4.1% vs 4.9% vs 6.2%vs 7.6%, P = .409 for all the quartiles and P = .116 betweenthe lowest and highest). Incidence of VTE did not differsignificantly with increasing quartiles (3.2% vs 4.5% vs5.7% vs 2.7%, P = .345 for all quartiles and P = .753between the highest and lowest quartiles). The need for ICUadmission did significantly increase with increasing BMIquartiles (19.5% vs 22.8% vs 26.9% vs 30.7%, P = .037).Using the Mantel-Haenszel statistics, we found that thehighest BMI group had significantly higher odds of ICUadmission compared with the lowest (odds ratio, 1.83; 95%confidence intervals, 1.18-2.84; P = .007). The need for MVwas significantly increased with increasing BMI quartiles(9.5% vs 12.9% vs 17.6% vs 19.6%, P = .012), as was thehospital LOS (3 vs 4 vs 4 vs 5 days, P b .001) andtracheostomy placement (4.5% vs 4.9% vs 10.6% vs 13.3%,P = .001). Compared with the lowest, the highest BMIquartile had the highest odds for need for MV (odds ratio,2.32 [1.33-4.04]; P = .004) and tracheostomy placement(odds ratio, 3.25 [1.55-6.82]; P = .002).

3.2. Patients requiring ICU admission

Table 3 shows the clinical characteristics of those patientsadmitted to the ICU. One hundred seventy-nine (32.8%)patients had a total of 224 ICU admissions. Patients wereyoungest in the highest BMI quartile (ages 54.2 vs 52.1 vs52.5 vs 47.7, P = .033). Again, we found that higher BMIquartiles had higher percentage of females than did lowerquartiles (32.6% vs 29.2% vs 65.6% vs 46.4%, P = .004).There was no statistically significant difference in rates ofsmoking, COPD, CHF, CAD, DM, or LMWH utilizationwith increasing BMI quartiles. The prevalence of renalinsufficiency did significantly differ among the quartiles,with rates of 32.6% vs 17.6% vs 8.2% vs 15.9% (P = .014).Although 62.5% of patients admitted to the ICU wereadmitted to a medical service vs a surgical service, there wereno differences among the quartiles (65.1% vs 62.7% vs60.7% vs 62.3%, P = .975).

Table 4 shows primary and secondary outcomes for ICUadmissions. Hospital mortality was not significantly differentwith increasing BMI quartiles (16.3% vs 15.7% vs 13.1% vs

21.7%, P = .608 for all quartiles and P = .479 between thehighest and lowest). No secondary end points werestatistically different with increasing BMI quartiles. TheICU LOS from lower to higher BMIs was not significantlydifferent (4 vs 5 vs 4 vs 6 days, P = .490). The VTE rates didnot significantly differ among the groups (7% vs 9.8% vs8.2% vs 7.2%, P = .952). The need for MV increased withincreasing BMI, but this was not statistically significant(48.8% vs 54.9% vs 63.9% vs 63.8%, P = .327 for allquartiles and P = .119 between the highest and the lowest).The median duration of MV also did not significantly differbetween the lowest and highest quartiles (P = .109). Theneed for tracheostomy increased with increasing BMI, butdid not reach statistical significance (20.9% vs 21.6% vs36.1% vs 36.2%, P = .124). The highest BMI quartile grouphad a trend toward higher utilization of tracheostomyplacement in comparison to the lowest group (odds ratio,2.15 [0.89-5.2]; P = .087).

4. Discussion

This is the first study to examine a cohort of onlymorbidly obese patients (BMI N40 kg/m2) to examineoutcomes among those with increasing levels of morbidobesity. As obesity and morbid obesity rates increaseacross the United States and around the world, health careworkers and systems face a major challenge imposed bythe utilization of resources by those with BMIs greater than40 kg/m2. It has been demonstrated that hospital mortalityis not significantly higher in morbidly obese patients vstheir normal-weight counterparts [4]. We demonstrate inthis study that mortality does not increase as the degree ofmorbid obesity increases (Table 2), although a clear trendwas present. We have demonstrated that these patients do,however, require expensive resources at alarming rates(Table 2). As BMI increases, the utilization of medicalresources including ICU admission, MV, tracheostomyplacement, and hospital LOS all significantly increase(Table 2).

In the ICUs and wards of America's hospitals, theoutcomes for obese and nonobese patients are significantly

184 B.D. Westerly, O. Dabbagh

different [5-13]. Several studies have indicated that obesepatients have increased inpatient mortality when comparedwith their nonobese counterparts [5-13]. In addition, obesepatients' ICU LOS was longer than normal-weight patients[7,8,13,14]. Furthermore, several studies revealed thatobese patients have an increased duration of MV whencompared with nonobese patients [7,15-18]. Other studiesdemonstrated increased morbidity with invasive ventilationtechniques in the obese [15,16,19]. However, not all studieshave supported the trends of worse outcomes for obesepatients. In fact, at least one study demonstrated thatincreasing BMI is actually associated with decreasedmortality rate after patients had been admitted to the ICU[4]. Other studies have supported this claim and demon-strated decreased mortality in obese groups of theirrespective studies, causing some to hypothesize that theremay be a protective effect of obesity in the critically ill[20-22]. Still, other studies have shown no effect of BMI onICU mortality [15,23,24]. The disparate results of thesestudies leave debate over the true effects of obesity andmorbid obesity on important outcomes as to whether theyare indeed harmful or beneficial.

Utilization of resources has been examined by only a fewinvestigators. Villavicencio et al [25] documented thatcardiac surgery patients with BMIs greater than 50 kg/m2

required high utilization of resources, including prolongedICU and hospital stay, prolonged MV, and increased woundcomplications. Body mass index greater than 54 kg/m2

actually predicted renal failure and prolonged MV [25] inthat trial. In this study, the utilization of resources increasedas BMI increased above 40 kg/m2 in a group of patientsadmitted to the hospital for a variety of etiologies andunderlying disease processes. In addition, the resourceutilization in this study appears to be independent of renalinsufficiency, which was not significantly increasedin increasing BMI quartiles (Table 1), and CAD, whichactually significantly decreased with increasing BMI quar-tiles (Table 1).

The coupling of increased numbers of admissions withhigh rates of resource utilization presents a major economicburden to health care expenditure in an already overburdenedsystem. Whether there are interventions that can beperformed early in these patients' courses to reduce theutilization of such resources remains unexplored. Of course,weight loss is likely the best intervention, given the results ofprevious studies [5-13] demonstrating better outcomes fornonobese patients when compared with obese counterparts.However, nonsurgical approaches to weight loss have beendisappointing; and whether surgical approaches will lead toimproved long-term outcomes has yet to seen. Prognosticand cost implications will be crucial factors in the treatmentof the morbidly obese. Hospital stay, mortality, prognosis,and cost may be significantly different among hospitalstreating this patient population, making true quality compar-isons invalid. In addition, the utilization of transitionalcare before discharge to home or need for home health

services as compared with nonobese controls is alsoundocumented to date.

Because this is a retrospective, observational study, thereare limitations common to such a study design. As is oftenthe case, complete data collection was impaired by theabsence of any information not in the electronic medicalrecord. Most notable in this particular study is the potentialfor excluded admission comorbidities because of lack ofidentification of diagnoses in the admission documents. Weattempted to maximize the accuracy of the data collected byindividually reviewing the patients' hospital records, includ-ing admission notes, progress notes, procedure notes, anddischarge summaries, as opposed to using billing informa-tion. In addition, all available previous physician notes werereviewed for confirmation of diagnoses and laboratoryfindings. Furthermore, we did not collect data aboutobstructive sleep apnea, which could contribute to thecomorbidity of such patients.

As aforementioned, a strength of this study is that it is thefirst to examine morbidly obese patients and categorize themin such a way that does not group them as one homogenousgroup. It is likely that the adverse consequences of morbidobesity continue to worsen as BMI increases, even past thecurrently recognized categories for distinguishing severity ofobesity (obesity, morbid obesity, super-morbid obesity). Wehave shown that the assumption of homogeneity may beinaccurate, as there are differences that exist as patient BMIsexceed 40 kg/m2.

5. Conclusion and clinical implications

As BMI increased as grouped in this study, patientsrequired increased need for ICU admission and MV, longerhospital LOS, and more tracheostomy placements; butmortality rates did not differ significantly. As morbid obesityrates increase, health care workers will face a majorchallenge: morbidly obese patients use expensive resourceswith increasing rates as their BMIs increase. Such utilizationof valuable and costly hospital resources is a major challengefacing health care delivery as morbid obesity increasesthroughout the United States and the rest of the world, andour findings call for national efforts to explore options to dealwith this emerging challenge.

Acknowledgments

Both authors had full access to all of the data in the studyand take responsibility for the integrity of the data and theaccuracy of the data analysis.

The authors would like to acknowledge Brian C Fuller,MD, and Brian J Stout, MD, for their assistance in datacollection. We thank Dr Stevan Whitt for his valuable help inreviewing this manuscript.

185Morbidity and mortality characteristics of morbidly obese

References[1] Centers for Disease Control and Prevention. Overweight and obesity.

Med Net 2007;13:16-7.[2] Flegal KM, Carroll MD, Ogden CL, et al. Prevalence and trends in

obesity among US adults, 1999-2000. JAMA 2002;288:1723-7.[3] Finkelstein EA, Fiebelkorn IC, Wang G. National medical spending

attributable to overweight and obesity: how much, and who's paying?Health Aff (Millwood) 2003 Suppl Web Exclusives;W3:219-26.

[4] Peake SL, Moran JL, Ghelani DR, et al. The effect of obesity on 12-month survival following admission to intensive care: a prospectivestudy. Crit Care Med 2006;34:2929-39.

[5] Bercault N, Boulain T, Kuteifan K, et al. Obesity-related excessmortality rate in an adult intensive care unit: a risk-adjusted matchedcohort study. Crit Care Med 2004;32:998-1003.

[6] Bochicchio GV, Joshi M, Bochicchio K, et al. Impact of obesity in thecritically ill trauma patient: a prospective study. J Am Coll Surg2006;203:533-8.

[7] Brown CV, Neville AL, Rhee P, et al. The impact of obesity on theoutcomes of 1,153 critically injured blunt trauma patients. J Trauma2005;59:1048-51 discussion 1051.

[8] El-Solh A, Sikka P, Bozkanat E, et al. Morbid obesity in the medicalICU. Chest 2001;120:1989-97.

[9] Levi D, Goodman ER, Patel M, et al. Critical care of the obese andbariatric surgical patient. Crit Care Clin 2003;19:11-32.

[10] Muir M, Heese GA, McLean D, et al. Handling of the bariatric patientin critical care: a case study of lessons learned. Crit Care Nurs ClinNorth Am 2007;19:223-40.

[11] Nasraway Jr SA, Albert M, Donnelly AM, et al. Morbid obesity is anindependent determinant of death among surgical critically ill patients.Crit Care Med 2006;34:964-70 quiz 971.

[12] Neville AL, Brown CV, Weng J, et al. Obesity is an independent riskfactor of mortality in severely injured blunt trauma patients. Arch Surg2004;139:983-7.

[13] Yaegashi M, Jean R, Zuriqat M, et al. Outcome of morbid obesity inthe intensive care unit. J Intensive Care Med 2005;20:147-54.

[14] Ciesla DJ, Moore EE, Johnson JL, et al. Obesity increases risk of organfailure after severe trauma. J Am Coll Surg 2006;203:539-45.

[15] Morris AE, Stapleton RD, Rubenfeld GD, et al. The associationbetween body mass index and clinical outcomes in acute lung injury.Chest 2007;131:342-8.

[16] El Solh AA, Jaafar W. A comparative study of the complications ofsurgical tracheostomy in morbidly obese critically ill patients. CritCare 2007;11:R3.

[17] Honarmand A, Safavi M. Do C-reactive protein and body mass indexpredict duration of mechanical ventilation in critically ill traumapatients? Ulus Travma Acil Cerrahi Derg 2008;14:284-91.

[18] Yap CH, Mohajeri M, Yii M. Obesity and early complications aftercardiac surgery. Med J Aust 2007;186:350-4.

[19] Duarte AG, Justino E, Bigler T, et al. Outcomes of morbidly obesepatients requiring mechanical ventilation for acute respiratory failure.Crit Care Med 2007;35:732-7.

[20] Aldawood A, Arabi Y, Dabbagh O. Association of obesity withincreased mortality in the critically ill patient. Anaesth Intensive Care2006;34:629-33.

[21] O'Brien Jr JM, Phillips GS, Ali NA, et al. Body mass index isindependently associated with hospital mortality in mechanicallyventilated adults with acute lung injury. Crit Care Med 2006;34:738-44.

[22] Tremblay A, Bandi V. Impact of body mass index on outcomesfollowing critical care. Chest 2003;123:1202-7.

[23] Landi F, Onder G, Gambassi G, et al. Body mass index and mortalityamong hospitalized patients. Arch Intern Med 2000;160:2641-4.

[24] Ray DE, Matchett SC, Baker K, et al. The effect of body mass index onpatient outcomes in a medical ICU. Chest 2005;127:2125-31.

[25] Villavicencio MA, Sundt III TM, Daly RC, et al. Cardiac surgery inpatients with body mass index of 50 or greater. Ann Thorac Surg2007;83:1403-11.