Critical Care Management 10:00 AM – 11:20 AM

Wednesday, November 2, 1994 Critical Care Management 1 0:00 AM - 11:20 AM DECREASED PREVALENCE OF HIV

Title: SARCOIDOSIS

Author(s):*Murray

SEROPOSITI1VITY IN

Gilman,M.D. FCCP, Tina

Rizacdf,

MPH,

John Boring,III,Ph.D.,PFulmonary Division and School of Public Health, Emory University School of Medicine, Atlanta, Georgia Sarcoidosis and the human itmaunodeficiency virus (MIV) are two distinct diseases, which in the United States most commonly affect young adult, inner-city, African Americans, Despite demographic similarities, the coexistence of these two disease entities has rarely been reported in the medical literature. A retrospective cohort study was conducted at Grady -emorial Hospital to determine the prevalence of HIV infection in a well-defined sarcoidpopulation and to evaluate whether a diagnosis of sarcoidosis is protective against HIV infection. Seventy-five cases of sarcoid, documented either by clinical or pathologic criteria, were identified as cases. All had HIV testing performed based upon the clinical suspicion of their health care providers. A control group Cn=225)r obtained from a list of patients with prior HIV testing at Grady Memorial Hospital was matched on a 3:1 basis to the sarcoid cases according to age, sex, and race. Uni-variate, stratified and logistic regression were utilized to determine the significance of the association for individual risk factors for HIV and to control for confounding. The prevalence of HIV was 2.7% in the sarcoid group versus 18.2% in the non-sarcoid control group. Our results demonstrate that 1) patients with a history of sarcoidosis were significantly less likely to have a positive HI'J test, and 2) having sarcoidosis showed a protective effect against HIV (risk ratio 0.101;95% confidence interval 0.019-0.533). The protective association remained after controlling for a series of known risk factors for HIV. These data suggest that a history of sarcoidosis is independently associated with a decreased likelihood of acquiring an HIV infection.

Title: GELATINASES ARE ABUNDANT IN BAL OF PATIENTS WITH VARIOUS LUNG DISEASES. RA Fiero*, H Mirza, Author(s) Dimassimo S Zucker, AN Hurewitz, FCCP. Dept of Medicine SUNY-Stony Brook /Northport VA Medical Center, NY. Metalloproteinases (MMPs) participate in matrix remodelling throughout the body. Gelatinases are MMPs which degrade denatured collagen and other matrix proteins. To investigate the role of gelatinases in lung disease, we obtained alveolar lavage (BAL) fluid from 16 patients with cancer, COPD or interstitial disease. Gelatin zymography permitted a characterization of gelatinases from BAL samples (supematants, cell lysates, and media of alveolar macrophages grown in serum-free RPMI over 48 hours). Immunoassay with monoclonal antibodies demonstrated the amount of gelatinase B and the MMP inhibitors, TIMP 1 and 2. BAL fluid showed greatest activity at 92 kDa (gelatinase B) with less gelatinolysis at 68 kDa (gelatinase A) and 125 kDa (?gelatinase B - ,2 microglobulin complex from neutrophils). Activities of these enzymes varied among individuals. T$1 small sample size did not permit comparison between groups of diseases. Lysed BAL cells contained gelatinase B and 125 kDa gelatinase but no gelatinase A. Alveolar macrophages produced gelatinase B in amounts similar to the corresponding BAL supernate but no gelatinase A or gelatinase B-microglobulin complex. Immunoassay of BAL supemates for gelatinase B, TIMP 1 and TIMP 2 showed concentrations of 50±32, 889±288 and 16±4 ng/ ml, respectively (mean ± SEM). We conclude that alveolar lavage fluid from patients with lung disease contains abundant gelatinases. Whether the disribution of these gelatinases correlates with dinical diagnosis or other laboratory features of the BAL is presently under study. [Supported, in part by Veterans Administration].

PROSPECTIVE, RANDOMIZED, COMPARATIVE TRIAL OF CONTINUOUS T Ptl VSe5 BOLUS FUROSEMIDE INFUSION IN THE INTENSIVE CARE UNIT. Author(s):D. Schuller ', J.P. Lynch. John Cochran VAMC. Wash Univ Sch of Med, St. Louis, MO Positive fluid balance of critically-ill patients isa common problem in the ICU associated with a poor outcome (CHEST 1991;100:1068-75). Failed attempts to accomplish netdiuresis are in partdue to the lack of a standardized regimen of diuretic dosage adjustment to assure that a 'diuresis goal" is maintained until the "therapeutic endpoint" is achieved. In this study, we prospectively compared the relative benefit of two different strategies of diuretic use in the ICU. 27 critically-ill patients with an indication for "aggressive diuresis" were randomized into either a bolus (B) or continuous (C)furosemide infusion group. All patients received an initial bolus infusion of furosemide (40 mg IV) and fluid restriction was implemented. In the "B" group (n=15) patients received frequentbolus infusions with increments ofthedose following a flow diagram to achieve a net diuresis goal of -=1 ml/Kg/hr. In the "C" group (n=12), patients were started on a furosemide drip (0.1 mg/Kg/hr) and the dose adjusted hourly to achieve the same net diuresis goal. In both groups, patients were kept on the study until a therapeutic endpoint was met at the discretion of the primary physician. The two groups were similar with respect to baseline characteristics and severity of illness scores. Although on average, patients in the 'B" group trended to stay a shorter time in the study (22 + 10 vs 31 + 14 hrs, p=0.06), and received a smaller cumulative furosemide dosage (336 + 375 vs 672 + 496 mg, p<0.04), no significant difference in net hourly diuresis (185 + 143 vs 169 + 93 mlJhr, p=0.71) or outcome was observed. Both groups had a significant decrease in weight from baseline (4.1 + 3 Kg, p<0.0002 and 4.9 + 5 Kg, p<0.005; respectively), and with the exception of 4 patients (2 in each group), all maintained a negative fluid balance throughout the study with an houriy net diuresis >=1 ml/Kg/hr. We conclude, that both protocols of diuretic use in the ICU are feasible, well tolerated and equally effective in accomplishing a negative fluid balance.

THE IMPACT OF HIV TESTING ON BLOOD UTILIZATION IN THE INTENSIVE CARE UNIT IN PATIENTS WITH AUthOr(S)GASTROINTESTINAL BLEEDING Mario Ravry, Harold Paz*, FCCP, Hahnemann University, Phila., PA

Title:

In order to determine whether the AIDS epidemic had influenced physician use of blood products in the management of gastrointestinal hemorrhage and to develop new criteria for transfusions in such patients, we studied 148 patients with gastrointestinal hemorrhage admitted to the intensive care unit. Forty eight of these patients were admitted before the onset of HIV testing of the blood supply (Group 1) and 100 cases were admitted after the HIV testing was begun (Group were not transfused (8 in Group 1, 10 in Group II). IM. Of the 148 patients,not18transfused had higher median hemoglobin levels con The patients that were admission and higher median hemoglobin nadirs during hospitalization than patients who were transfused. The remaining 130 patients were transfused (40 in Group I, 90 in Group II). Transfused patients in Group II did not have significantly lower median hemoglobin levels on admission (7.9 [4.2-12.5] gm/dl) than transfused patients in Group I (9.3 [4.1-13.5] gm/dl) (p = 0.058). A significantly lower median hemoglobin concentration prior to the first transfusion event (7.4 [4.2-10.3] gm/dl) was found in Group II than was found in Group I (8.5 [4.2-12.1] gm/dl) (p = 0.016). There were no significant difference in the total number of units of packed red blood cells, fresh frozen plasma or platelets transfused in each group. There was no significant difference in mortality between the two groups with 11 patients (22.9%) dying in Group I and 23 patients (23.0%) dying in Group H. The cause of death in 13 of the 34 patients was related to cardiovascular and hemodynamic complications of gastrointestinal bleeding. There was no significant difference in mean age (Group 1: 60.5; Group II: 59.4) or mean hemoglobin nadir (Group I: 7.0 gm/dl; Group II: 7.1 gm/dl) for those that died in the two groups. These data would suggest that physicians are transfusing patients at hemoglobin levels which are lower than before HIV testing began. However, there has been no decrease in the total median amount of blood products transfused since HIV testing was begun. This would suggest the need for rational guidelines on the use of blood patients.

CHEST / 106 / 2 / AUGUST, 1994 / Supplement

157S

Wednesday, November 2, 1994 Critical Care Management, continued DOPAMINE AND PROLACTIN CONCENTRATIONS IN MICU PATIENTS Title: RECEIVING HIGH DOSE DOPAMINE AND CONTROLS JS Marinac, Pharm.D.*, SK Willsie, D.O., FCCP, SC Hamburger, M.D., Author(S): BL Herndon, Ph.D., L O'Toole, M.D., G Reisz, M.D., FCCP. University of Missouri Kansas City, School of Medicine, Department of Medicine, Kansas City, Missouri.

lbjectives: To determine the basal dopamine concentrations in MICU patients who receive high dose dopamine and those who do not, & evaluate these findings with respect to APACHE II scores & changes in prolactin levels. Characterize the pharmacokinetics of dopamine in treated patients.

Methods: All patients admitted to the MICU were evaluated for entry. Baseline APACHE II scores were calculated. Patients with underlying disorders or those receiving medications known to interfere with dopamine or prolactin were excluded. Before initiation of dopamine, blood samples were taken for basal dopamine and prolactin levels. Early morning blood was collected daily for presumed steady state dopamine (ie: > 4 hours of dopamine therapy at a single infusion rate > 5 uglkglmin) and prolactin levels, & again on the first day following discontinuation of dopamine. In control patients who did not receive dopamine, early morning blood samples for were obtained on at least three consecutive days. Ten patients met entry criteria; four dopamine, six controls. Results: Fourteen samples were obtained in the dopamine patients and 18 samples in the control group. The APACHE II scores of the two groups were similar (26.3 ± 5.6 vs 22.0 ± 6.8) (p-ns). Large intersubject variability in basal dopamine levels was found, yet basal dopamine levels were significantly higher in the subsequently dopamine-treated group than in the control group (812 ± 992 pgtmL vs 260 ± 441 pglmLl, (p < 0.05), which are higher than that reported in critically ill children and in healthy adults (488 pglmL and 103 pglmL respectively). 1,2 The mean predopamine prolactin level in the dopamine patients was similar to an age, gender and APACHE II matched control group patient (40.6 pg/mL vs. 45.8 pglmL) (p-ns). Changes in dopamine levels were inversely associated with fluctuations in serum prolactin in both the dopamine & control patients, although no linear relationship was found. Plots of dopamine dose versus dopamine level revealed a linear correlation (r- 1.00, p < 0.001, & r - 0.99, p < 0.05) in 2 dopamine subjects.

Conclusions: Admission dopamine levels in patients who required dopamine pressor support were significantly higher than MICU patients with similar APACHE II scores. Changes in exogenous and endogenous dopamine levels were inversely associated with changes in serum prolactin. The findings and implications of this pilot study warrant further investigation. 2. J Pediatr 1986;110:293-8. 1. Crit Care Med 1989;17:1203.

BURNING BRIDGES: NEUROMUSCULAR BLOCKER (NMB) USE AND TRACHEAL INTUBATION (TI) IN TRAUMA PATIENTS H Bosscher, S Heard, FCCP, Dept of Anesthesiology, UMMC, Worcester, MA 01655

Introduction: NMB use during TI in trauma patients (PTS) with potential airway injuries (Al) may be a proximate cause of airway complications. Airway management techniques and NMB use in TI of PTS were reviewed. Methods: Records of 161 TI collected prospectively over a 3.5 year period were reviewed. Al: trauma to facial soft tissues, facial fractures or burns and/or cervical spine injury. Groups were classified as: 1) uncomplicated TI (UTI); 2) difficult TI (DTI); multiple attempts, Sa02 < 90%, esophageal intubation, need for anesthesiologist; 3) cricothyrotomy (CR). Data were analyzed by chisquare and Spearman rank correlation. Results: Agitation (28%), respiratory distress (27%) and mental status changes (22%) were the major indications for TI. Most TI were oral (94/u). Table shows complication rates (%): CR DTI 21 (13) 5 (3) 161 135 (84) TI w/NMB 121 4 (3) 101 (83) 16 (14) TI w/o NMB 1 (2) 34 (85) 5 (13) 40 7 (16) 1 (2) TI w/Al & NMB 43 35 (82) TI w/AI w/o NMB 19 15 (79) 1 (5) 3 (16) No statistical relationship between NMB and airway complications was ALL

UTI

n

Title: HEMOGLOBIN;

A RANDOMIZED CLINICAL TRIAL TO DETERMINE HOW MUCH IS ENOUGH IN ICU?

Author(s): PC Hebert, GA Wells, G Pagliarello, J Marshall, C Martin, M Tweeddale for the Canadian Critical Care Trials Group.

Historically, a transfusion threshold of 100 g/L has been advocated. However, in critically ill patients, there is conflicting evidence and a lack of clinical trials examining different approaches to transfusions. In a randomized clinical trial, we allocated 69 patients with hemoglobin (Hb) values < 90g/L admitted to one of 5 tertiary care ICUs to either a conservative (n=34) transfusion strategy with Hb values of 70-90 g/L or a liberal (n=35) transfusion strategy with Hb values of 100-120 g/L. Baseline characteristics including age (60 vs 58 yrs), gender (59% vs 47% males), APACHE El (19.5 vs 20), and primary diagnosis were similar between groups (p>0.05). 70% of all patients were admitted to the ICU from surgical services. The most frequent diagnoses overall were respiratory failure (29%), multiple trauma (23%), gastrointestinal diseases (14%) and high risk postoperative admissions (1 1%). 28% of patients required a Swan Ganz catheter and 30% required mechanical ventilation. The overall 30 day all cause mortality was 27%. Average Hb values over 20 days were 88 glL in the conservative group and 103 g/L in the liberal group (p<0.0001). The total number of units transfused were 24 vs 32 respectively (p=0.61). 30 day all cause mortality rates were 27% in both groups (p=0.934). There were also no differences in length of hospital (p=0.23) or ICU (p=0.32) stay between groups. We observed a trend towards an increased mean survival times (14 vs 10 days, p=0.06) favouring the conservative group. In conclusion, mortality rates and length of ICU and hospital stays were not significantly affected by successfully implemented transfusions strategies. Therefore, a more conservative approach to the administration of blood may be safe in critically ill patients. The trend to longer mean survival times noted using a conservative approach warrants further investigation. Supported by the Canadian Red Cross Society, Blood Services and the Physicians' Services Incorporated.

Ttle: FACTORS INFLUENCING

INTUBATION TIME AND A-a 02 DIFFERENCE IN POST-OPEN HEART PATIENTS. Author(s): M. Al-Dairi, MD,* and B. Lahiri, MD, FCCP. Univ. of Connecticut School of Medicine, Farmington, CT, and St. Francis Hospital and Medical Center, Hartford, CT. We evaluated factors influencing post-operative respiratory insufficiency in 60 open heart surgery patients. Age, sex, smolking history, number of grafted vessels, fluid balance, and cardiopulmonary bypass punp time were related to duration of intubation and A-a 02 difference at approximately 24 hours following surgery. Intubation time was related to: 1) fluid balance (r=0.59, p.<0.001), 2) pump time (r--0.53, p.<0.001), and 3) age (r=0.34, p.=0.008). Likewise, 24 hour A-a 02 difference was related to these three variables: r=0.40, p.=0.002, r=0.44, p.=0.001, and r=0.42, p.=0.001, respectively. The cconbination of these three variables explained 53% of the variation in intubation time and 38% of the variation in 24 hour A-a 02 difference. Gender, smoking history, and nuTher of vessels grafted were not related to intubation time or A-a 02 difference. Esitive fluid balance, prolonged pump time, and age were strong predictors of pulmonary insufficiency following open heart surgery.

found. Conclusion: DTI (16%) are common occurrences but do not appear to be exacerbated by NMB use. Individual occurrences of difficult TI, however, may have been caused the by use of NMB.

158S

Abstracts, 60th Annual International Scientific Assembly

Wednesday, November 2, 1994 Title: SAFE TRANSPORT OF PATIENTS WITH ARDS: USE OF A PORTABLE POWER SUPPLY TO MAINTAIN PRESSURE CONTROL VENTILATION Author(s): Joel R. Peerless, MD, FCCP., Dan Pavlik, MEd, RRT, Ken Zielinski, Dan Gladyzewski, Ruth Neale, RN Cese Western Reserve University at MetroHealth Medical Center Cleveland, Ohio Introduction: Intrahospital transport of patients with severe ARDS is often necessary for diagnostic studies or surgical procedures. These patients often require high levels of inspired oxygen IFl 021, ventilating pressures IVP) and minute ventilation (Ve). Use of portable ventilators or hand ventilation may result in inadequate gas exchange during transport. Methods: We configured a UL approved, compact, rechargable Uninterruptible Power System (UPS) to supply the Puritan Bennett 7200 ventilator with heated humidifier and compressor (to avoid the need for an air cylinder during transport). We conducted 1 5 transports (CT-10, OR-4, Angiography-1i in nine patients with severe ARDS who required pressure control ventilation.

Results: Severity of Iniury PO/FI02 Lung Injury Score Ventilator Settings

FI02 1mm/Hg)

VP (cm) PEEP (cm) Vs IL/min)

Range 64-172 2.6-3.6

Mean(SD) 117 (28.2) 3.2 (0.29)

50-100

67 (17.2) 38 (8.01 11.3 12.2) 11.3 (2.1)

30-65

7.5-16 8.4-16.7

All transports were completed in less than 15 minutes. There was no loss of power supply during any transport. 13/15 (87%) transports were without incident. In two instances patients required increased ventilator settings to correct

hypoxemia. Conclusion: Patients with severe ARDS are at high risk for respiratory deterioration during transport to other areas of the hospital. Use of a portable power supply can ensure that the ventilatory support provided in the ICU is maintained throughout transport, and that increased levels of support can be provided when needed.

Title:EFFICACY AND C(OMFORr OF MASK POSITIVE PRESSURE VENTILATION VIA A UNIQUE FULL FACE MASK Author(s): William Marino* MD FCCP Our Lady of Mercy Medical Center, Bronx, N.Y. Non-invasive positive pressure ventilation

(NIPPV) has become a standard conponent of therapy for respiratory failure in various conditions, but the masks used to provide this are often subject to airleaks and discomfort or even skin damage at pressure points such as the bridge of the nose. A mask covering the entire face with a gasket seal around the rim of the face and no pressure points has thus been evaluated as a means of providing NIPPV. 15 patients admitted to this institution for respiratory failure (who had failed to achieve an acceptable response to NIPPV via other masks) were treated with NIPPV via the Respironics #MTFMV11 prototype mask. 7 were treated with Pressure Support Ventilation (PSV) and 8 were treated with Assist-Control Ventilation (A/C). Those patients treated with PSV dropped their mean pCO2 on independent breathing from 76mmrHg to 45mmHg within 5 days of intermittent treatment. The 8 treated with A/C went from 70mmOrHg to 50mmHg in the same time period. 12 of these patients, all of whom had chronic respiratory failure, were converted to nasal mask ventilation and have done well subsequently. All, at the time of conversion, expressed strong preference for the MTFMV-11 mask because of its greater comfort. The MTFMV-11 full face mask is an effective and very comfortable means of providing NIPPV.

DuPont Young Investigator Award Winner Title: Author(s):

WORK OF BREATHING ANALYSIS DURING TACHYPNEIC EXTUBATION

TRIAL FAILURES REDUCES FALSE NEGATIVE CLASSIFICATION

O Kirton MD FCCP*, B DeHaven RRT FCCM, J Morgan MBA RRT, A Hart MD, D Shatz MD University of Miami/Jackson Memnorial Medical Center, The Ryder Trauma Center, Miami, FL

OBJECTIVE: There is increased awareness of imposed work of breathing potentially contributing to apparent failure to wean from ventilation. This study evaluates tachypnea as a useful indicator of failure during the room air CPAP extubation trial when associated with elevated imposed work of breathing. METHODS: We prospectively studied mechanically ventilated Trauma ICU patients for nine months. After patients were able to maintain airway and acceptable consciousness level, room air CPAP (R/A CPAP) extubation trals( Fl02=0.21,CPAP=5cmH2O,PSV=O) were initiated. When passed (PaO2 > 55,PaCO2 c 45 w/ prior eucarbia,pH > 7.35, respiratory rate < 30/min), extubation followed. If patients failed due to hypoxia, ventilatory support resumed. If tachypnea was the reason for failure, work of breathing was measured with the Bicore CP100. If WOBp was < 1.1, extubation proceeded despite tachypnea. If WOBp > 1.1, imposed work (WOBimp) was measured, and if WOBp-WOBimp was < 0.8, patients were extubated. RESULTS: Of 418 extubations, 17% (n=73) were extubated despite tachypnea from 3252/min (mean 38/min) in conjunction with either a WOBp < 1.1 J/l or an increased WOBimp. The reintubation rate was 6.9% and 6.8% respectively. DISCUSSION: Extubation failure rates were similar for patients passing R/A CPAP, those who would have failed by tachypnea, and prior reports regarding extubation failures with the R/A CPAP technique. Tachypnea as an indicator of respiratory distress is probably quit sensitive, but may not be specific to outcomes of weaning trials. Increased inspiratory resistance or decreased lung compliance may be compensated for by decreased tidal volume and increased respiratory rate. Some tachypnea was possibly due to increased WOBimp, as elevated respiratory rate usually abated within 18 hours after extubation. CONCLUSION: Reliance on tachypnea as an indicator or predictor of sucessful extubation is likely inappropriate for a iarge number of patients. Patient risks, determined by the extubation failures and reintubation rate, are unchanged from conventional room air CPAP extubation trials. Patient costs for WOB monitoring is less than 1/3 daily mechanical ventilator costs. Should only one ICU day ($4300.00) be saved per 'failed' but sucessfully extubated patient, costs of $292,400.00 were avoided.

TRACHEAL-PRESSURE TRIGGERING A DEMAND-FLOW CPAP SYSTEM DECREASES WORK OF BREATHING Author(s): G Messinger MD*, MJ Banner PhD, PB Blanch RRT, AJ Layon MD FCCP. Depts. of Anesth, Med, and Physiol, Univ of Fla College of Med, Gainesville, FL 32610-0254.

tle: *

Mechanical ventilation of critically ill patients often incurs patient-ventilator dyssynchrony, respiratory muscle loading, and fatigue due to insensitivity of the ventilator triggering mechanism. We compared the work of breathing (WOB) during spontaneous ventilatipn with different combinations of pulmonary mechanics among three methods of triggering the ventilator ON by using either conventional pressure triggering from inside the ventilator, flowby triggering, or tracheal-pressure triggering at the carinal end of the endotracheal tube. In an in-vitro model of the respiratory system consisting of a bellows (lungs) in a plastic canister (chest wall), spontaneous ventilation was simulated with a piston-driven"pump. Continuous positive airway pressure of 5 cm H20 was maintained (Puritan-Bennett 7200ae) at a peak inspiratory flow rate demand of 60 L/min and a resistance of 5 and then 20 cm H20/L/sec. Real-time measurements of pleural pressure, volume, and WOB were recorded by a computerized respiratory monitor (Bicore, CP-100, Irvine, CA) by using the Campbell + 3S diagram. Data were analyzed by 2factor ANOVA. Compared with con3 2 ventional and flow-by triggering, tra+ -:a cheal-pressure triggering decreased * 2 2 WOB (fig.) and may enhance patientTM ventilator synchrony.

Fig. At airway resistance of o5r 20 cm H20/L/sec, WOB is higher (P < 0.05)

with flow-by* (striped bar) and conientionalt (open bar) than tracheal-pressure (solid bar) triggering of mechanical ven-

tilation.

Mii °

X/

0,5

, cm,H20/L

20 vcnLis

CHEST / 106/2/ AUGUST, 1994 / Supplement

159S

Wednesday, November 2,1994 Critical Care Management, continued

Title: WITHRAWAL OF THERAPY IN INTENSIVE CARE

Author(s): JSTurner*,CJMorgan, WL Michell, SR Benatar Adult ICU, Royal Brompton National Heart and Lung Hospital, London, England and Groote Schuur Hospital, Cape Town, South Africa Life support was withdrawn in less than 50% of patients dying in an ICU in the USA (Smedira et al. N Engl J Med 1990;322:309-315); figures may be different elsewhere. This study in a tertiary cardiothoracic referral centre in England examines the frequency of withdrawal of therapy, medical grounds for withdrawal, the people involved in making the decisions, the way the decisions are implemented and the ICU outcome. Data were prospectively collected on all ICU patients who died or had life support withdrawn. These included demographics, APACHE IIscore, and data relevant to the above aims. The study was approved by the institutional Ethics Committee. Results: There were 58 deaths out of 814 ICU discharges; 10 were unexpected cardiac arrests and 48 (83%) involved limitation of therapy. Of these, the mean age was 60.0 years (20-78 years), and mean APACHE II score 19.8 (8-30). The primary diagnoses included cardiac surgery, pneumonia, thoracic surgery, ARDS, and

cardiomyopathy. Reasons for withdrawal were multiple organ failure in 20 patients, an irreversible limiting organ failure in 17, two organ failure with overwhelming sepsis or chronic health in 7, chronic health plus poor prognosis in 3, and brain death in 1. Both medical and nursing staff were involved in all decisions, and the families in all but two where events happened too quickly. In 18 patients therapy was immediately reduced and in 30 it was not increased; in 14 of the latter this was followed by reduction. The actual therapies involved were most often inotropes and ventilation, but also included antibiotics, dialysis, antiarrhythmics, TPN, blood products, IV fluids, cardioversion, and further surgery. The mean time from admission to withdrawal was 11.5 days (141 days). The mean time to outcome (ICU death in all patients) was 12.0 hours (< I hour to 96 hours), though in the 18 patients in whom therapy was reduced at once the mean time was 6.6 hours (maximum 24 hours). Conclusions: Withdrawal of therapy occurred commonly with wide consensus reached before adecision was made. There were a nurnber of reasons for withdrawal with several therapeutic options; the time to death was generally short.

NASOENTERIC FEEDING TUBE PLACEMENT USING ELECTROMYOGRAM AND Author(s): ERYTHROMYCIN INFUSION *Howard Levy, MD, FCCP, James Heyes Ph.D, PE, Todd Tomba, BSEE. University of New Mexico, Albuquerque, NM and Ross Products Division, Abbott Laboratories, Columbus, OH. Post-pyloric feeding is desirable in critically ill patients who often

Title:

have

gastroparesis. Post-pyloric placement is difficult, requiring

fluoroscopy, endoscopy or time consuming blind attempts. This study evaluated the success rate and time required to place post-pyloric tubes using Erythromycin infusion and gastrointestinal electromyogram (EMG) signal. Methods: Unweighted Flexiflo IOF feeding tubes were modified by the placement of an electrode 8cm from the tip to record EMG signals (Ross Products Division, Columbus, OH). Gastric (G) signals are high amplitude with a frequency of 3/min while the duodenum (D) and jejunum (J) are low amplitude and 11-13/min. Erythromycin was infused at a dose of 3mg/Kg to enhance G motor activity and emptying. The nasoenteric tube was placed in the stomach and its position confirmed by auscultation and EMG. The tube was then slowly advanced until D/J EMG was detected. Demographic data and time to enteric placement and tube position on X-ray were recorded. Results: 25 mechanically ventilated patients (13 respiratory failure, 6 pancreatitis, 6 ARDS) had 34 placements. There were 16 males and 9 females with mean age 53.8 (range 22-85). 29 placements were successful (24 patients), 18 J and 1 D, requiring an average 7.5 minutes (range 318). 3 attempts were unsuccessful in the I patient who could not receive erythromycin due to allergy (total duration 27 mins). 2 unsuccessful attempts (1 G and 1 esophageal) were repeated after X-ray and placed in J and D respectively (total 18mins each). Conclusion: Erythromycin infusion and EMG guidance can facilitate rapid nasoenteric feeding tube placement in critically ill patients at the bedside.

160S

PREDICTION OF CALORIC REQUIREMENTS USING A SIMPLE GENDER TITLE: NONSPECIFIC EQUATION WITH EMPHASIS ON AVOIDING OVERFEEDING

Mark A Fusco MD*, Loren D Nelson MD FCCP. Dept of Surg, VanderbiltUniversity, Nashville TN and methods: Estimating energy expenditure (EE) is difficult, often Obiectives requiring multiple equations and "stress factors" to approximate EE measured by indirect calorimetry (MEE). Since indirect calorimetry is not always available, we sought to develop a gender independent equation to predict the EE in a large group of hospitalized patients. Indirect calorimetry studies were performed on 1879 patients between Jan. 1990 and Dec. 1992 using a MGM-11 metabolic cart (Utah Medical) (3000 data sets). Studies in subjects less than 16 or with incomplete information concerning gender, Ht, Wt, or age were eliminated leaving 1618 patients (2644 data sets). Stepwise linear regression was performed using EE as the independent variable and age(yr), Ht("), and Wt(kg) as the subjects dependent variables. The analysis was repeated on the male and female Estimated EE was calculated separately yielding two gender dependent (GD) equations. method and the appropriate GD in all patients using the gender independent equation. Each of the EE determinations were compared to the MEE using an unpaired two tailed Student's t test. The developed equations were compared to other common prediction methods using bias and precision (95% confidence intervals). The percentage of patients who would have had a greater than 20% overestimation of their caloric the likelihood of overfeeding. requirements by each method was calculated to estimate Differences between these values were assessed by a Chi squared test. linear Results: The equations derived from Group regression analysis were as as a whole: EE = -983 -4(age) +32(Ht) +11(Wt) (Eq. A); Males: EE = 406 -5(age) = The mean (Eq. C). +8(Wt) +12(Ht) +13(Wt) (Eq. B); Females: EE 412 -4(age) +12(Ht) EE for the entire group using the GI method was 1811 ± 310 (± std deviation). The GD method resulted in a mean of 1849 ± 330. Neither of these methods yielded results which were significantly different from the MEE (1832 ± 536; a=0.05). Both methods compared v zed for bias and preci sion. Avorably to the widely use estimates of EE hen anal AUTHOR(S):

(GIl)

follows:

Bias

Precision (t x std error) Pts overfed >20%(%)

A 21 13.9 9.0

aBor

C

-17

13.5

an 25xWt 33 19.1 15

pi1.6xHB arris-Benedict(HB) -645

18.3

-62

19.3

21 <1 HB had the lowest incidence of overfeeding. However, HB underestimated MEE on it methods with the A was When remaining compared equation average by 645 calories. was less likely to overestimate MEE ( p <0.01). Conclusions: The EE can be reliably estimated without accounting for gender. The simplified single equation: EE = -983 -4(age) + 32(Ht) + 11(Wt) can be used to estimate

11.3

EE when indirect calorimetry is not available. When compared to MEE, the acceptable bias and only a 9% incidence of significant overfeeding.

an

Titit'ABGALLOPURINOL TO REDUCE ISCHEMIC

equation has

INJURY DURING

Author(s)llan Goldman*, M.D., Harjinder Grewal, M.D., Linda Budzilowicz, R.N., Alvin Greengart, M.D. Edgar Lichstein, M.D., FCCP, Jacob Shani, M.D., Robert Frankel, M.D., FCCP, Israel Jacobowitz, M.D., FCCP, Ketan Shevde, M.D., Gerald Hollander, M.D., FCCP. Maimonides Medical Center, SUNY Health Science Center at

Brooklyn, Brooklyn,

New York

Recent reports suggest that allopurinol pretreatment offers cardioprotection during open heart surgery. We examined the effect of allopurinol in patients undergoing elective bypass. In a randomized double blinded study, fifty four patients were assigned to allopurinol (300mg per/day) or placebo. Treatment was begun the day prior to surgery and continued through the second postoperative day. A vasopressor scoring system was used to define the degree of support required. A score of indicated no vasopressor support, whereas, a score of IV indicated multiple vasopressor drugs

need for IABP. Results: Thirty two patients received allopurinol and 22 pts received placebo. These groups were similar with regard to age, sex, risk factors, prior Ml, NYHA classification, ejection fraction, and number of bypassed vessels. Eight of the 12 patients with the highest vasopressor scores (Gp III & IV) were in the allopurinol group. III IV II Vasopressor Score 8 0 16 8 Allopurinol 0 4 11 7 Placebo Conclusion: Allopurinol patients were not protected and, in fact, required greater hemodynamic support, as measured by the need for vasopressors postoperatively.

or

Abstracts, 60th Annual International Scientific Assembly