Impact of Body Mass Index on Outcomes Following Critical Care

Impact of Body Mass Index on Outcomes Following Critical Care* Alain Tremblay, MDCM; and Venkata Bandi, MBBS

Study objectives: To determine the impact of body mass index (BMI) on outcomes in critically ill patients. Design: Retrospective analysis of a large multi-institutional ICU database. Measurements: The influence of BMI classification (underweight, < 20 kg/m2; normal [control subjects], 20 to 25 kg/m2; overweight, 25 to 30 kg/m2; obese, 30 to 40 kg/m2; severe obesity, > 40 kg/m2) on hospital survival, functional status at hospital discharge, and ICU/hospital length of stay (LOS) was analyzed via multivariate analysis, adjusting for age, gender, type of hospital admission, and severity score (ie, simplified acute physiologic score [SAPS] II and mortality prediction model [MPM] at time zero). Univariate analysis also was performed according to the quartile of the severity score. All comparisons were to the normal BMI group. Results: Of 63,646 patient datasets, 41,011 were complete for height, weight, and at least one of the two severity scores. We found increased mortality in underweight patients (odds ratio [OR] of death: SAPS group, 1.19; MPM group, 1.26) but not in overweight, obese, or severely obese patients. ICU and hospital LOS were increased in both the severely obese (OR of discharge: ICU, 0.81 and 0.84, respectively; hospital, 0.83 and 0.87, respectively) and underweight groups (OR of discharge: ICU, 0.96 and 0.94, respectively; hospital, 0.91 and 0.90, respectively). Only in the SAPS group did the obese group have increased ICU LOS (OR, 0.96) and hospital LOS (OR, 0.96). Functional status at discharge was impaired in underweight patients (OR of disability: ICU, 1.11; hospital, 1.19). Overweight patients had decreased discharge disability (OR of disability: SAPS, 0.93; MPM, 0.94), while the results in the obese group were discordant between the two severity score groups (SAPS, not significant; MPM, 0.91; p < 0.05 for all ORs). Conclusions: Low BMI, but not high BMI, is associated with increased mortality and worsened hospital discharge functional status. LOS is increased in severely obese patients and, to a lesser extent, in underweight patients. Patients in the overweight and obese BMI groups may have improved mortality and discharge functional status. (CHEST 2003; 123:1202–1207) Key words: body mass index; ICU; length of stay; mortality; nutrition; obesity; severity of illness; underweight; weight Abbreviations: APACHE ⫽ acute physiology and chronic health evaluation; AUC ⫽ area under the receiver operating curve; BMI ⫽ body mass index; CI ⫽ confidence interval; H-L ⫽ Hosmer-Lemeshow goodness-of-fit; LOS ⫽ length of stay; MPM-0 ⫽ mortality prediction model at time 0; OR ⫽ odds ratio; SAPS ⫽ simplified acute physiologic score

mass index (BMI), calculated as weight (in B ody kilograms)/height (in meters, squared), is an anthropometric index that is well-correlated to total body fat content. Despite some contradictory results,1 population studies have demonstrated Ushaped curves relating BMI to mortality,2,3 suggesting increased mortality at both extremes of body weight. *From the Department of Pulmonary and Critical Care Medicine, Baylor College of Medicine, Houston, TX. Manuscript received November 20, 2001; revision accepted June 27, 2002. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: [email protected]). Correspondence to: Alain Tremblay, MDCM, Division of Respiratory Medicine, Health Sciences Center, 3330 Hospital Dr NW, Calgary, AB, T2N 4N1 Canada; e-mail: [email protected] 1202

Multiple investigators4 –10 have examined the impact of both low and high BMI on outcomes for patients with specific diseases or patients who have undergone specific procedures. Other studies11,12 have examined the relationship of BMI with outcomes in hospitalized patients, but no published study has specifically examined this relationship in a population of patients who have been admitted to an ICU. Despite a wealth of clinically and scientifically applicable severity-of-illness scoring systems, none of the scores commonly used in critical care research have incorporated anthropometric measurements in their predictive algorithms, and, to our knowledge, no anthropometric measurements were evaluated during the initial development of these scoring systems.13–16 Clinical Investigations in Critical Care

We sought to examine the impact of BMI on in-hospital mortality, functional status at time of ICU discharge, and length of stay (LOS) following ICU admission in a large population of critically ill patients. Our hypothesis was that adverse outcomes would be observed in patients who were below and above the normal range of BMI. Materials and Methods All data were obtained from a large, multiinstitutional, critical care patient data set (Project Impact Critical Care Data System; Society of Critical Care Medicine; Des Plaines, IL [see http:// www.projectimpacticu.cc for project details]). The database was queried on March 28, 2000, following the approval of the study protocol by the Project Impact Research Committee. Specifically, data on age, height, ICU admission weight, gender, type of ICU admission (ie, medical, elective surgical, or emergent surgical), ICU and hospital LOS, mortality prediction model at time 0 (MPM-0), and simplified acute physiologic score (SAPS) II scores and status at hospital discharge (ie, independent, dependant, or dead) were requested. No data allowing identification of individual patients, hospital, or physician were supplied. Data for the database were collected prospectively by dedicated data coordinators at each of the participating sites from patient charts. Hospital admission weight and height were recorded from actual measurements or, when not available, from the best clinical estimate of patient care providers. BMI index was calculated from the available height and weight data. The cohort was divided into five BMI subgroups, as follows: underweight, ⬍ 20 kg/m2; normal, (control subjects), 20 to 25 kg/m2; overweight, 25 to 30 kg/m2; obese, 30 to 40 kg/m2; and severe obesity, ⬎ 40 kg/m2. Patients were entered into two separate cohorts for analysis, according to the availability of severity-of-illness scores (SAPS-II group and MPM-0 group). Patients were excluded from the analysis if the data on weight and height were absent or if neither severity score was available. For all analyses, BMI groups were compared to the normal

BMI (control) group. Comparisons between other BMI groups were not performed. Hospital survival and functional status for each BMI group were calculated according to quartiles of SAPS-II or MPM-0 scores (with the first quartile having the lowest predicted survival) and were compared with the ␹2 test and Bonferroni correction for multiple comparison between groups. ICU and hospital LOS data for survivors in each BMI group according to quartiles of SAPS-II or MPM-0 scores were compared with the Kruskal-Wallis H test followed by the Dunnett t test for multiple comparisons against a single control group. Logistic regression was performed to assess the combined impact of age, gender, type of ICU admission, BMI group, and severity score for survival and functional status. Area under the receiver operating curve (AUC), model ␹2 test. and HosmerLemeshow goodness-of-fit (H-L) statistics were used to select the best model. Cox regression models was used for LOS data censoring at the time of death to avoid a positive effect on LOS from early mortality. A p value of ⬍ 0.05 was accepted as being statistically significant. Data were analyzed with computer software (SPSS, version 9.0.0; SPSS; Chicago, IL).

Results At the time of query, the Project Impact database contained 63,646 patient records. Of these, 41,011 (64.4%) had complete height and weight data, allowing the calculation of the BMI and at least one severity-of-illness score (MPM-0, 37,470 records [58.9%]; and SAPS-II, 36,153 records [56.8%]). The demographic characteristics of these groups are shown in Table 1. Survival Data The survival data as analyzed by quartiles of severity scores for each BMI group are summarized in Figure 1. The following odds ratios (ORs) for

Table 1—Demographic Characteristics*

Characteristics Male gender, % Age, yr (SD) BMI subgroups, % Underweight Normal Overweight Obese Severely obese BMI Admission type, % Medical Emergent surgical Elective surgical Hospital mortality, % ICU LOS, d Hospital LOS, d

Initial Cohort (n ⫽ 63,646)

MPM-0 Group (n ⫽ 37,470)

SAPS-II Group (n ⫽ 36,153)

Excluded Group (n ⫽ 22,635)

55.8 60.1 (18.6)

54.4 59.6 (18.6)

54.7 59.7 (18.5)

61.2 60.6 (18.9)

12.4 33.1 29.8 19.0 5.7 27.1 (8.3)

12.7 33.0 29.6 18.9 5.9 27.1 (8.4)

60.7 13.2 26.1 17.0 1.8 (0–200) 7.0 (1–383)

61.0 13.9 25.1 14.6 2.0 (0–200) 8.0 (1–383)

61.5 12.2 26.2 14.8 1.7 (0–200) 7.0 (1–383)

63.2 9.9 26.9 12.9 1.4 (0–119) 6.0 (1–309)

*Values given as mean (SD) or median (range), unless otherwise indicated. www.chestjournal.org

CHEST / 123 / 4 / APRIL, 2003

1203

spectively, and the model ␹2 tests were p ⬍ 0.0001 and p ⬍ 0.0001, respectively, for the SAPS-II and MPM-0 equations. The impact of membership in the various BMI groups is listed in Table 2. The underweight groups in both the MPM-0 and SAPS-II cohorts had increased mortality, while the overweight group had decreased mortality in the MPM-0 group only. LOS

Figure 1. Survival to hospital discharge according to quartiles of severity score is shown. Top: SAPS-II group. Bottom: MPM-0 group.

mortality were statistically significant. In the second and third quartiles, the ORs of mortality of the underweight group were 1.25 (95% confidence interval [CI], 1.04 to 1.49) and 2.07 (95% CI, 1.53 to 2.80), respectively, for the SAPS-II group and 1.38 (95% CI, 1.15 to 1.65) and 1.68 (95% CI, 1.27 to 2.22), respectively, for the MPM-0 group. The obese group in the second MPM-0 quartile had an OR of 0.78 (95% CI, 0.65 to 0.95), and the overweight group in the third MPM-0 quartile had an OR of 0.76 (95% CI, 0.60 to 0.97), although this was not seen in the corresponding SAPS-II subgroups. Logistic regression revealed that age (SAPS-II group only), gender, SAPS-II/MPM-0 scores, and type of admission all predicted mortality. The AUCs were 0.865 (95% CI, 0.860 to 0.870) and 0.831 (95% CI, 0.825 to 0.836), respectively, the H-L statistics were 31.9 (p ⫽ 0.0001) and 69.9 (p ⬍ 0.0001), re1204

LOS data as analyzed by quartiles of severity score in survivors are summarized in Figure 2. The following comparisons reached statistical significance. ICU LOS was increased in the overweight, obese, and severely obese groups in the first quartiles of SAPS-II and MPM-0 scores, in the obese group in the second quartile of MPM-0 scores, and in the severely obese group in the second and third quartiles of SAPS-II scores and in the second and fourth quartiles of MPM-0 scores. Hospital LOS was increased in the overweight, obese, and severely obese groups in the first quartiles of SAPS-II and MPM-0 scores, in the underweight and severely obese groups in the second MPM-0 quartile and the third SAPS-II quartile, in the underweight group of the fourth MPM-0 quartile, and in the severely obese group in the second and fourth SAPS-II quartiles. Cox regression analysis including the entire groups (censoring at the time of death) and controlling for age, gender, type of admission, and severity score revealed ORs for ICU discharge in the SAPS-II and MPM-0 groups of 0.96 (95% CI, 0.93 to 0.999) and 0.94 (95% CI, 0.90 to 0.97), respectively, for the underweight group, 0.96 (95% CI, 0.93 to 0.995) and not significant, respectively, for the obese group, and 0.81 (95% CI, 0.77 to 0.86) and 0.84 (95% CI, 0.80 to 0.90), respectively, for the severely obese group. For hospital discharge, the ORs were 0.91 (95% CI, 0.88 to 0.95) and 0.90 (95% CI, 0.87 to 0.93), respectively, for the underweight group, 0.96 (95% CI, 0.93 to 0.995) and not significant, respectively, for the obese group, and 0.83 (95% CI, 0.79 to 0.88) and 0.87 (95% CI, 0.83 to 0.92), respectively, for severely obese.

Table 2—Hospital Mortality SAPS-II Group

MPM-0 Group

BMI Subgroups

OR

95% CI

OR

95% CI

Underweight Overweight Obese Severely obese

1.19 0.92 0.96 1.09

1.08–1.32 0.85–1.01 0.87–1.06 0.93–1.27

1.26 0.89 0.93 1.03

1.14–1.39 0.82–0.97 0.85–1.03 0.88–1.21

Clinical Investigations in Critical Care

Figure 2. LOS data according to quartiles of severity score for (top left, A) the ICU SAPS-II group, (top right, B) the ICU MPM-0 group, (bottom left, C) the hospital SAPS-II group, and (bottom right, D) the hospital MPM-0 group.

Functional Status in Survivors at Hospital Discharge Functional status data were analyzed by quartiles of severity score. There were no differences in the risk of being at the partially or totally dependant functional level on hospital discharge in the SAPS-II group. In the MPM-0 group, the underweight group had an increased risk of dependence in the second and fourth quartiles, while the obese and severely obese group showed decreased risk in the third quartile only. Logistic regression revealed that age (SAPS-II group only), gender, SAPS-II/MPM-0 scores, and type of admission all predicted disability. The AUCs were 0.684 (95% CI, 0.678 to 0.691) and 0.655 (95% CI, 0.649 to 0.462), respectively, the H-L statistics were 72.5 (p ⬍ 0.0001) and 32.6 (p ⫽ 0.0001), respectively, and the model ␹2 test values were p ⬍ 0.0001 and p ⬍ 0.0001, respectively, for the SAPS-II and MPM-0 equations. The impact of membership in the various BMI groups is listed in Table 3. An increase in the risk of dependency status in the underweight group and a decrease in the risk of dependency status in the overweight group for both the SAPS-II and MPM-0 cohorts and in the obese group (MPM-0 only) were seen. www.chestjournal.org

Discussion We analyzed a large cohort of critically ill patients for the impact of BMI on outcomes following ICU admission. Our only exclusion criterion was missing BMI or severity score data, making the group relatively unselected, although the excluded patients did show some differences in gender, hospital mortality, and hospital LOS (Table 1). This raises the concern that the approximately one third of patients excluded from the analysis because of missing data may be different than the analyzed cohort. The most important effects of BMI on the measured outcomes were seen in the underweight group (BMI, ⬍ 20 kg/m2). An increased OR for mortality was seen in both the SAPS-II and MPM-0 cohorts.

Table 3—Disability at Hospital Discharge SAPS-II Group

MPM-0 Group

BMI Subgroup

OR

95% CI

OR

95% CI

Underweight Overweight Obese Severely obese

1.11 0.93 0.94 1.09

1.02–1.20 0.88–0.99 0.87–1.01 0.97–1.21

1.19 0.94 0.91 1.02

1.10–1.29 0.88–0.998 0.85–0.98 0.91–1.14

CHEST / 123 / 4 / APRIL, 2003

1205

As well, survivors in this group were significantly less likely to return to their hospital preadmission functional status. Interestingly, in a univariate analysis according to type of admission, increased mortality was seen in the underweight patients in the medical and emergent surgical groups, but not in the elective surgical group (data not shown). Care must be taken in interpreting these findings, as we did not control for recent weight loss or underlying chronic illness such as cancer, which may independently influence outcomes. Nevertheless, this is consistent with previous reports concerning hospitalized patients,11,12 as well as concerning patients with pneumonia,6 those needing hemodialysis,9 those with HIV,10 and those who have undergone heart transplantation.8 Population studies1,3 also suggest increased mortality in low-BMI subgroups. Despite our large sample size, we did not identify increased mortality in the overweight, obese, or severely obese patient groups. In fact, the overweight group in the MPM-0 cohort showed decreased mortality, and both the overweight and obese groups showed improved functional status at hospital discharge. While this seems to contradict data from population-based studies1,3 and diseasespecific studies,4,5as well as general assumptions in the critical care literature,17,18 it is consistent with prior data from studies of hospitalized patients.11,12 This does not exclude the possibility of increased complication rates in obese patients, but it may be that other factors such as nutritional reserve play a beneficial compensatory role in these patients. This may explain the higher LOS seen in the severely obese patients. Of course, our study does not analyze the risk of obese patients initially developing an illness requiring ICU admission, but only their outcome once it has developed. Some authors have suggested that obesity may confer a greater risk only in younger hospitalized patients,12 although the negative effects of high BMI have been demonstrated across all age groups in population studies,7 even if the risk does decrease with age in some reports.19 The lack of increased mortality in obese and severely obese patients in our study persisted even when patients who were ⬎ 65 years of age and those who were ⬍ 65 years of age were analyzed separately (data not shown). Increased ICU LOS and hospital LOS were seen in the severely obese group and to a lesser degree in the underweight group. Given the lack of increased mortality in the severely obese group, this increased LOS may be attributed to prolonged recovery from illness (eg, prolonged weaning) or to the increased incidence of nonfatal complications. Unfortunately, our study was not designed to detect these factors. Further study of these issues is warranted. 1206

We know of three studies, which are in abstract form, that have attempted to examine the relationship of BMI to ICU outcomes in a univariate fashion. The first study20 was negative (ie, no outcome differences between BMI groups) but also was underpowered to detect the magnitude of differences seen in the current analysis. The second study21 divided a group of patients somewhat arbitrarily into two groups according to BMIs of ⬎ 28 or ⬍ 28 kg/m2, which cannot give clear results as to outcomes at the extremes of BMI. Higher hospital mortality (but not ICU mortality) and LOS were seen in the group with BMI ⬎ 28 kg/m2. Using the same cutoff value in our patient population, hospital mortality was higher in the group with BMI ⬎ 28 kg/m2 than in the group with BMI ⬍ 28 kg/m2 (17.8% vs 15.4%, respectively; p ⬍ 0.001 [␹2 test]). The third study22 compared 117 patients with BMIs of ⬎ 40 kg/m2 with 132 agedmatched, nonobese control subjects and found increased mortality (30% vs 17.5%, respectively) in the obese group. Although APACHE (acute physiology and chronic health evaluation) II scores were similar in this univariate analysis, the obese group was much more likely to be receiving mechanical ventilation, suggesting that other factors were involved in their poorer outcomes. Clearly, these contradictory results are due to the different patient populations studied and the methodological approaches of the analyses. We think that the large sample size drawn from multiple institutions combined with a more rigorous multivariate analysis make the data from the present study more reliable. The accuracy of the data obtained for critically ill patients is a significant concern. Most ICU practitioners will acknowledge that weight and height data in the ICU are often estimated rather than measured. Body weight at the time of ICU admission also may be significantly different than a patient’s normal body weight because of volume depletion or overload. This intrinsic imprecision of our data can be acknowledged only as a limitation of this and other studies on this topic. Recent changes in BMI also may play an important role in outcomes, but these data were not available for patients in our population. Our original database query did not include data on specific diagnoses. This raises the concern that patients in different BMI groups had different types of diseases, possibly confounding the analysis. On the other hand, the severity-of-illness scores utilized in this study have been validated across a wide population of critically ill patients and represent an estimated mortality risk that is independent of the primary diagnosis. Data analysis on such a large number of patients leads to the danger that type II errors may occur (ie, the finding of a significant difference by chance Clinical Investigations in Critical Care

when there is in fact none). The use of statistical corrections for multiple analyses and the fact that all of our results are consistent with our initial hypothesis make this less likely. Other results with CIs including 1 were not described as being statistically significant despite similar absolute ORs to other comparisons to avoid type I errors. The raw data are nevertheless presented in table form for the reader to review. None of the results with borderline significance were in the opposite direction to other statistically significant results. A final criticism of our data centers on the use of severity scores in multivariate analyses. The poor calibration of mortality prediction models when applied to a population different than the one on which the models were developed is well-known.23,24 This is represented in our data by the poor H-L statistics in the logistic regression analyses. Nevertheless, as we are not attempting to predict individual patient outcomes, but rather attempting to adjust for severity as best we can, we think that this is a minor issue. Conclusion In a retrospective analysis of a large multiinstitutional database of patients who had been admitted to ICUs, we found increased mortality in underweight patients but not in overweight, obese, or severely obese patients. LOSs in the ICU and hospital were increased in both the severely obese and underweight patient groups. Functional status at hospital discharge was impaired in underweight patients. Overweight and obese patients may have more favorable outcomes in terms of mortality and hospital discharge functional status. Further study is warranted to identify the specific explanations for the above differences. As well, the inclusion of BMIs in the development of severity scoring systems should be considered. ACKNOWLEDGMENT: We thank Dr. D. Chalfin and all participants in Project Impact for making these data available, Dr. K. Dunn and Dr. C. Contant for their advice on statistical issues, and Dr. L. Natho for her careful review of the manuscript.

5 6

7 8

9 10

11 12 13 14 15 16 17 18 19 20 21

References 1 Manson JE, Stampfer MJ, Hennekens CH, et al. Body weight and longevity: a reassessment. JAMA 1987; 257:353–358 2 Pi-Sunyer FX, Dietz WH, Becker DM, et al. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Bethesda, MD: National Heart, Lung, and Blood Institute, 1998 3 Troiano RP, Frongillo EA Jr, Sobal J, et al. The relationship between body weight and mortality: a quantitative analysis of combined information from existing studies. Int J Obes Relat Metab Disord 1996; 20:63–75 4 Hoit BD, Gilpin EA, Maisel AA, et al. Influence of obesity on www.chestjournal.org

22

23 24

morbidity and mortality after acute myocardial infarction. Am Heart J 1987; 114:1334 –1341 Smith-Choban P, Weireter LJ, Maynes C. Obesity and increased mortality in blunt trauma. J Trauma 1991; 31:1253– 1257 Hedlund J, Hansson LO, Ortqvist A. Short and long term prognosis for middle-aged and elderly patients hospitalized with community-acquired pneumonia: impact of nutritional and inflammatory factors. Scand J Infect Dis 1995; 27:32–37 Calle EE, Thun MJ, Petrelli JM, et al. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999; 341:1097–1105 Grady K, White-Williams C, Naftel D, et al. Are preoperative obesity, and cachexia risk factors for post heart transplant morbidity, and mortality: a multi-institutional study of preoperative weight-height indices. J Heart Lung Transplant 1999; 18:750 –763 Fleischmann E, Teal N, Dudley J, et al. Influence of excess weight on mortality, and hospital stay in 1346 hemodialysis patients. Kidney Int 1999; 55:1560 –1567 Shor-Posner G, Campa A, Zhang G, et al. When obesity is desirable: a longitudinal study of the Miami HIV-1-Infected Drug Abusers (MIDAS) Cohort. J Acquir Immune Defic Syndr 2000; 23:81– 88 Galanos AN, Pieper CF, Kussin PS, et al. Relationship of body mass index to subsequent mortality among seriously ill hospitalized patients. Crit Care Med 1997; 25:1962–1968 Landi F, Onder G, Gambassi G, et al. Body mass index and mortality among hospitalized patients. Arch Intern Med 2000; 160:2641–2644 Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13:818 – 829 Knaus WA, Wagner DP, Draper EA, et al. The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991; 100:1619 –1636 Lemeshow S, Teres D, Klar J, et al. Mortality probability models (MPM II) based on an international cohort of intensive care unit patients. JAMA 1993; 270:2478 –2489 Le Gall JR, Lemeshow S, Saulinier F. A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270:2957–2963 Boulanger BR, Milzman DP, Rodriguez A. Trauma and pre-existing disease: Part II. Specific management concerns. Crit Care Clin 1994; 10:613– 622 Marik P, Varon J. The obese patient in the ICU. Chest 1998; 113:492– 498 Stevens J, Cai J, Pamuk E, et al. The effect of age on the association between body-mass index and mortality. N Engl J Med 1998; 338:1–7 Ray DE, Matchett SC, Kaufman JH, et al. Effect of body mass index on patient outcomes in a medical intensive care unit [abstract]. Am J Respir Crit Care Med 2000; 161 (suppl):A398 Abbasi AA, Mak R, Sullivan J, et al. Body mass index as predictor of outcome in critically ill patients in intensive care units [abstract]. Crit Care Med 2000; 28 (suppl):A115 El-Solh AA, Sikka P, Muhammad A, et al. Outcome of critically ill obese patients admitted to the medical intensive care unit: the CIMO (The Critical Care Initiative for the Study of Hospitalized Patients with Morbid Obesity) study [abstract]. Crit Care Med 2000; 28 (suppl):A113 Patel PA, Grant BJB. Application of mortality prediction systems to individual intensive care units. Intensive Care Med 1990; 25:977–982 Moreno R, Apolone G, Reis Miranda D. Evaluation of the uniformity of fit of general outcome prediction models. Intensive Care Med 1998; 24:40 – 47 CHEST / 123 / 4 / APRIL, 2003

1207