Kenneth P. Steinberg

The effects of ibuprofen on the physiology and survival of patients with sepsis

BACKGROUND In patients with sepsis the production of arachidonic acid metabolites by cyclooxygenase increases, but the pathophysiologic role of these prostaglandins is unclear. In animal models, inhibition of cyclooxygenase by treatment with ibuprofen before the onset of sepsis reduces physiologic abnormalities and improves survival. In pilot studies of patients with sepsis, treatment with ibuprofen led to improvements in gas exchange and airway mechanics. METHODS From October 1989 to March 1995, we conducted a randomized, double-blind, placebo-controlled trial of intravenous ibuprofen (10 mg per kilogram of body weight [maximal dose, 800 mg], given every six hours for eight doses) in 455 patients who had sepsis, defined as fever, tachycardia, tachypnea, and acute failure of at least one organ system. RESULTS In the ibuprofen group, but not the placebo group, there were significant declines in urinary levels of prostacyclin and thromboxane, temperature, heart rate, oxygen consumption, and lactic acidosis. With ibuprofen therapy there was no increased incidence of renal dysfunction, gastrointestinal bleeding, or other adverse events. However, treatment with ibuprofen did not reduce the incidence or duration of shock or the acute respiratory distress syndrome and did not significantly improve the rate of survival at 30 days (mortality, 37 percent with ibuprofen vs 40 percent with placebo). CONCLUSIONS In patients with sepsis, treatment with ibuprofen reduces levels of prostacyclin and thromboxane and decreases fever, tachycardia, oxygen consumption, and lactic acidosis, but it does not prevent the development of shock or the acute respiratory distress syndrome and does not improve survival.

Chronic alcohol abuse is associated with an increased incidence of acute respiratory distress syndrome and severity of multiple organ dysfunction in patients with septic shock.

ObjectiveAlcohol is one of the most commonly used drugs in the world and causes dysfunction in many vital organs. However, the effects of chronic alcohol abuse on acute lung injury and nonpulmonary organ dysfunction are relatively unexplored. The goal of this study was to determine the effects of chronic alcohol abuse on the incidence and severity of the acute respiratory distress syndrome and multiple organ dysfunction syndrome in patients with septic shock. DesignMulticenter prospective epidemiologic study. SettingIntensive care units of four university urban hospitals. PatientsA total of 220 critically ill patients with septic shock. Measurements and Main FindingsThirty percent of the patients (66 of 220) were identified as having a history of chronic alcohol abuse based on a positive response to an alcohol screening questionnaire. The incidence of acute respiratory distress syndrome in patients with a positive history of chronic alcohol abuse was 70% (46 of 66), compared with 31% (47 of 154) in individuals without a history of chronic alcohol abuse (p < .001). After adjusting for differences in the source of infection, sex, age, chronic hepatic dysfunction, diabetes, severity of illness, nutritional status, and smoking status, the effects of chronic alcohol abuse on the incidence of acute respiratory distress syndrome remained significant (p < .001; odds ratio, 3.70; 95% confidence interval, 1.83–7.71). The effect of the source of infection (pulmonary vs. nonpulmonary) on the development of acute respiratory distress syndrome also remained significant in this multivariable analysis (p < .001; odds ratio, 3.68; 95% confidence interval, 1.95–7.18). Based on the highest daily Sequential Organ Failure Assessment score, patients with a history of chronic alcohol abuse had more severe nonpulmonary organ dysfunction when compared with nonalcoholics (9.42 ± 3.89 vs. 8.05 ± 4.10, p = .01). After adjusting for source of infection, sex, age, nutritional status, history of diabetes, and smoking status, the effects of chronic alcohol abuse on the incidence of nonpulmonary organ dysfunction also remained significant (p = .03; odds ratio, 2.07; 95% confidence interval, 1.09–3.97). ConclusionsWe conclude that chronic alcohol abuse is an independent risk factor for acute respiratory distress syndrome and increases the severity of nonpulmonary organ dysfunction in patients with septic shock.

Type III Procollagen Peptide in the Adult Respiratory Distress Syndrome: Association of Increased Peptide Levels in Bronchoalveolar Lavage Fluid with Increased Risk for Death

The adult respiratory distress syndrome frequently results in a fibroproliferative response and excessive lung extracellular matrix accumulation that may preclude recovery [1, 2]. An analysis of lung tissue from patients who died of the adult respiratory distress syndrome provided biochemical evidence that collagen, the major extracellular matrix component of lung, was greatly increased [3]. Immunohistologic evaluation of lung tissue from patients with the adult respiratory distress syndrome showed an abundance of type I and type III collagen, with type III collagen predominating earlier in the disease course [4]. Collagen accumulation in the adult respiratory distress syndrome results, at least in part, from increased procollagen synthesis, a mechanism that has been shown in numerous animal models of acute lung injury [5, 6]. Further, excessive lung collagen synthesis and accumulation may contribute to the high fatality rates associated with the adult respiratory distress syndrome by promoting progressive respiratory dysfunction. Alternatively, excessive lung collagen synthesis might indirectly influence outcome by impeding resolution of respiratory failure and increasing the risk for subsequent fatal complications, including multiple organ dysfunction. The relation between collagen synthesis and clinical outcomes in the adult respiratory distress syndrome has been difficult to examine directly. However, the N-terminal peptide of type III procollagen (procollagen III), cleaved from the precursor procollagen molecule during synthesis, appears to be a useful marker of collagen synthesis. Increased levels of procollagen III have been detected in serum from patients with sarcoidosis [7], idiopathic pulmonary fibrosis [8-10], and the adult respiratory distress syndrome [11, 12], as well as other conditions involving tissue fibrosis, such as cirrhosis [13], wound healing [14], trauma [12], and myelofibrosis [15]. In patients with idiopathic pulmonary fibrosis, increased procollagen III concentrations in bronchoalveolar lavage fluid were strongly associated with indices of clinical disease severity [10]. Although increased levels of procollagen III have been detected in lavage fluid from a few patients with the adult respiratory distress syndrome [16], their association with clinical outcome has not been examined. We analyzed procollagen III levels from bronchoalveolar lavage fluid in patients with the adult respiratory distress syndrome and studied the relation of increased lavage procollagen III levels to fatality rates. Methods Patients All patients between the ages of 18 and 72 years who were admitted to intensive care units at Harborview Medical Center [Seattle, Washington] between September 1986 and April 1991 were screened prospectively for the onset of the adult respiratory distress syndrome. Patients were screened using the following criteria: 1) for critical hypoxemia with cutoff Pao 2/Fio 2 ratios of 150 mm Hg or less or of 200 mm Hg or less using 5 cm H2O or more of positive end-expiratory pressure; 2) for diffuse parenchymal infiltrates involving at least three quadrants on chest radiographs; 3) for pulmonary artery wedge pressure [when available] of 18 mm Hg or less or no clinical evidence of congestive heart failure; and 4) no other obvious explanation for these findings [17, 18]. Because of the possible risk for complications related to bronchoalveolar lavage, patients with the adult respiratory distress syndrome were excluded if they met any of the following criteria: 1) Pao 2 less than 80 mm Hg with Fio 2 of 1.0; 2) evidence of acute ischemic heart disease; 3) severe hypotension (systolic blood pressure < 90 mm Hg); 4) cardiac dysrhythmias (heart rate > 140 beats/min or complex ventricular ectopy); 5) sustained increased intracranial pressure greater than 20 mm Hg; and 6) endotracheal tube internal diameter less than 7.0 mm. Patients were not excluded because of high minute ventilation, high levels of positive end-expiratory pressure, or presence of barotrauma. Informed consent was obtained from either the patient or the legal surrogate. The study was approved by the University of Washington Human Review Committee. Before bronchoalveolar lavage was done, the following clinical data were obtained: levels of Fio 2 and Pao 2, static compliance, and level of positive end-expiratory pressure. These data were used to calculate a modified lung injury score for the adult respiratory distress syndrome, as described by Murray and colleagues [19], except that a chest radiograph score was not included. However, all patients had alveolar infiltrates in three or four quadrants. Thus, the Murray acute lung injury score would be 0.75 to 1.0 points greater than our modified score. Patients in our study with lung injury scores of 1.75 or more met Murray criteria for severe lung injury. Risk factors associated with development of the adult respiratory distress syndrome were defined as previously described [17]. These included the sepsis syndrome, trauma, aspiration of gastric contents, drug overdose, and multiple transfusions. Trauma risk was defined as the presence of multiple long bone or pelvic fractures, pulmonary contusion, or trauma-associated multiple transfusions ( 15 units in 24 hours of emergency resuscitation) [17]. For this analysis, we combined clinical risks for aspiration of gastric contents, drug overdose, and multiple transfusions without trauma into the category other risks. Risk factors were identified prospectively when the patient was entered into the study. The first 18 patients in the study had a single bronchoalveolar lavage. Subsequently, we attempted to do lavage serially at days 3, 7, and 14 after the onset of the adult respiratory distress syndrome unless the patient died, was extubated, or became too unstable to tolerate a lavage, as indicated by the criteria above. Patients were followed until death or hospital discharge. Survival was defined as discharge from hospital. Organ failure and cause of death, as defined by Montgomery and colleagues [18], were analyzed in a subset of patients enrolled during 1990. Bronchoalveolar Lavage and Analyses All patients were intubated at the time of lavage. They were ventilated with Fio 2 levels of 1.0 for 10 to 15 minutes before and during the procedure. An adaptor was placed on the patients endotracheal tube, and a fiberoptic bronchoscope was passed through the endotracheal tube into the lower airway and was wedged into a subsegment of either the right middle lobe or lingula. Five 30-mL aliquots of sterile pyrogen-free 0.9% NaCl at room temperature were instilled (150 mL total) and recovered by gentle suction. Six normal volunteers had lavage using a similar technique. Serum samples also were obtained from patients at the time of lavage and were stored at 70C. Lavage samples were transported immediately to the laboratory for analysis. The fluid was filtered through gauze moistened with 0.9% NaCl, and the total recovered volume was recorded. Total cell counts were done in a hemacytometer, and differential cell counts were done on cytospin preparations stained with Diff-Quick (American Scientific Products, McGaw Park, Illinois). Cell viability was measured by trypan blue exclusion. After an aliquot was taken for cellular analysis, the lavage fluid was centrifuged at 200 g for 15 minutes to pellet the cells. Aliquots of the supernatant lavage fluid were put into polypropylene tubes and stored at 70C. Total protein was measured on an aliquot of the supernatant using either the modified Lowry method [20] or the bicinchoninic acid method [21]. Type III Procollagen Peptide Analysis The concentration of procollagen III in bronchoalveolar lavage fluid specimens or serum was determined by radioimmunoassay (RIAgnost PAP, Behringwerke, Marburg, Germany) according to the manufacturers instructions using 20 L of lavage fluid or serum. The radioimmunoassay for procollagen III was linear over a range of 0.4 to 9.5 U/mL. Serum control samples provided by the manufacturer contained 1.6 to 1.7 U/mL. Samples in which the procollagen III concentration was greater than the standard detection range were diluted 1:4 in 0.9% NaCl. Samples in which procollagen III concentrations were less than the detection range were assigned a value of 0.4 U/mL for subsequent data analysis. To determine the variability of repeat determinations, five lavage samples were analyzed in triplicate. The variability of triplicate determinations of procollagen III concentration measured in five different lavage samples was 5% or less. In one analysis, an aliquot of the lavage sample was subjected to a total of three freeze-thaw cycles to determine stability of the peptide under these conditions. No increase in the variability of measured values was detected. Cross-sectional and Serial Analyses Lavage fluid procollagen III concentrations in survivors and nonsurvivors initially were compared using the nonparametric Wilcoxon rank-sum test [22]. We then calculated the relative risk (RR) for fatality in patients with a lavage fluid procollagen III level of 1.75 U/mL or more compared with those with values less than 1.75 U/mL; 95% CIs were determined using the method of Rothman [23]. We also did a stratified analysis to determine if the relation between procollagen III concentration and fatality differed by risk factor for the adult respiratory distress syndrome (sepsis, trauma, other) or degree of lung injury, using the lung injury score. The difference in RRs for death among risk groups for the adult respiratory distress syndrome and lung injury severity groups was assessed by the Breslow-Day test of homogeneity [24]. We did logistic regression analyses of data obtained on days 3, 7, and 14 after onset of the adult respiratory distress syndrome to measure the effect of procollagen III levels on the risk for death while controlling for the effect of demographic and physiologic variables that, independ

Recovery rate and prognosis in older persons who develop acute lung injury and the acute respiratory distress syndrome.

What is the problem and what is known about it so far? Elderly people are much more likely to develop respiratory failure and require the use of a respirator than younger people. On the other hand, the results of respirator use in the elderly do not appear to be as good as in younger people, but the exact relationship between age and outcome is uncertain. The National Heart, Lung, and Blood Institute (NHLBI) sponsored a large study on acute lung injury that leads to respiratory failure. The data were made available to other researchers for further analysis. Why did the researchers do this particular study? To find out how advancing age affects survival from acute lung injury and whether recovery takes longer in older patients. Who was studied? 902 patients at 24 U.S. hospitals who participated in the NHLBI study between 1996 and 1999. Patients were included in the study if they had acute lung injury requiring use of a respirator, had low oxygen levels, and had a chest x-ray indicating that the respiratory problem was not due to heart failure. Of the participants, 729 were younger than 70 years of age and 173 were 70 years of age or older. How was the study done? Patients were evaluated each day to see if a respirator was still needed. Depending on the results of these evaluations, patients were allowed to breathe on their own for 5 minutes; if they did well, additional time off the respirator was prescribed and respiratory assistance was eventually discontinued. The time required to reach each landmark of recovery was recorded, as were the length of stay in the intensive care unit and the survival rate. What did the researchers find? Seventy percent of the younger patients and 40% of the older patients were discharged from the hospital alive within the first 180 days. At the time of entry into the study, the severity of lung injury was similar in both age groups. Older patients needed the respirator longer, were more likely to require reinstitution of respirator therapy after initial improvement, and stayed in the intensive care unit longer than younger patients. At 28 days after initiation of respirator use, the survival rate was lower with each decade of advancing age. What were the limitations of the study? Several nonrespiratory health problems (such as brain abnormalities) were not carefully monitored but could have affected the results. In addition, older patients may have been more likely than younger patients to have had life support withdrawn because of poor general condition. Furthermore, preexisting illness was difficult to take into account in interpreting the results. What are the implications of the study? Although elderly patients seem to recover initially from acute lung injury at a rate similar to younger patients, they are more likely to return to mechanical ventilation and are more likely to die of their acute lung illness. Age alone should not be used as a criterion to deny respirator care. Rather, more effort should be made to improve outcomes in elderly people who develop acute lung failure.

Randomized, placebo-controlled trial of lisofylline for early treatment of acute lung injury and acute respiratory distress syndrome

Objective To determine whether the administration of lisofylline (1-[5R-hydroxyhexyl]-3,7-dimethylxanthine) would decrease mortality in patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS). Design A prospective, randomized, double-blind, placebo-controlled, multicenter study. Setting Intensive care units at 21 hospitals at the ten centers constituting the ARDS Clinical Trials Network. Patients A total of 235 patients who met eligibility criteria were enrolled in the study (116 into the lisofylline group, 119 into the placebo group). Interventions Patients were randomized to receive either lisofylline or placebo. The dose of lisofylline was 3 mg/kg with a maximum dose of 300 mg intravenously every 6 hrs. The intravenous solution of study drug was administered over 10 mins every 6 hrs. Dosing was continued for 20 days or until the patient achieved 48 hrs of unassisted breathing. Measurements and Main Results The trial was stopped by the Data Safety Monitoring Board for futility at the first scheduled interim analysis. The patient groups had similar characteristics at enrollment. No significant safety concerns were associated with lisofylline therapy. There was no significant difference between groups in the number of patients who had died at 28 days (31.9% lisofylline vs. 24.7% placebo, p = .215). There was no significant difference between the lisofylline and placebo groups in terms of resolution of organ failures, ventilator-free days, infection-related deaths, or development of serious infection during the 28-day study period. The median number of organ failure–free days for the five nonpulmonary organ failures examined (cardiovascular, central nervous system, coagulation, hepatic, and renal) was not different between the lisofylline and placebo groups. Although lisofylline has been reported to decrease circulating free fatty acid levels, we did not find any such treatment effect compared with placebo. Conclusions In this study, there was no evidence that lisofylline had beneficial effects in the treatment of established ALI/ARDS.

Effects of ibuprofen on the physiology and survival of hypothermic sepsis

OBJECTIVES The objective was to compare the clinical and physiologic characteristics of febrile septic patients with hypothermic septic patients; and to examine plasma levels of cytokines tumor necrosis factor alpha (TNF-alpha and interleukin 6 (IL-6) and the lipid mediators thromboxane B2 (TxB2) and prostacyclin in hypothermic septic patients in comparison with febrile patients. Most importantly, we wanted to report the effect of ibuprofen treatment on vital signs, organ failure, and mortality in hypothermic sepsis. SETTING The study was performed in the intensive care units (ICUs) of seven clinical centers in the United States and Canada. PATIENTS Four hundred fifty-five patients admitted to the ICU who met defined criteria for severe sepsis and were suspected of having a serious infection. INTERVENTION Ibuprofen at a dose of 10 mg/kg (maximum 800 mg) was administered intravenously over 30 to 60 mins every 6 hrs for eight doses vs. placebo (glycine buffer vehicle). MEASUREMENTS AND MAIN RESULTS Forty-four (10%) septic patients met criteria for hypothermia and 409 were febrile. The mortality rate was significantly higher in hypothermic patients, 70% vs. 35% for febrile patients. At study entry, urinary metabolites of TxB2, prostacyclin, and serum levels of TNF-alpha and IL-6 were significantly elevated in hypothermic patients compared with febrile patients. In hypothermic patients treated with ibuprofen, there was a trend toward an increased number of days free of major organ system failures and a significant reduction in the 30-day mortality rate from 90% (18/20 placebo-treated patients) to 54% (13/24 ibuprofen-treated patients). CONCLUSIONS Hypothermic sepsis has an incidence of approximately 10% and an untreated mortality twice that of severe sepsis presenting with fever. When compared with febrile patients, the hypothermic group has an amplified response with respect to cytokines TNF-alpha and IL-6 and lipid mediators TxB2 and prostacyclin. Treatment with ibuprofen may decrease mortality in this select group of septic patients.

Diabetic patients have a decreased incidence of acute respiratory distress syndrome

Objective Our ability to predict which critically ill patients will develop acute respiratory distress syndrome (ARDS) is imprecise. Based on the effects of diabetes mellitus on the inflammatory cascade, we hypothesized that a history of diabetes might alter the incidence of ARDS. Design A prospective multicenter study. Setting Intensive care units at four university medical centers. Patients One hundred thirteen consecutive patients with septic shock. Interventions None. Measurements and Main Results All patients were prospectively followed during their intensive care course for the development of ARDS. A history of diabetes was identified in 28% (32/113) of the patients. In this study, nondiabetics were more likely to develop septic shock from a pulmonary source (48%, 39/81) compared with diabetics (25%, 8/32) (p = .02). Forty-one percent (46/113) of the patients with septic shock developed ARDS. Forty-seven percent of the nondiabetic patients developed ARDS compared with only 25% of those with diabetes (p = .03, relative risk = 0.53, 95% confidence interval = 0.28–0.98). In a multivariate logistic regression analysis, when we adjusted for several variables including source of infection, the effect of diabetes on the incidence of ARDS remained significant (p = .03, odds ratio = 0.33, 95% confidence interval = 0.12–0.90). Conclusions In patients with septic shock, a history of diabetes is associated with a lower risk of developing ARDS compared with nondiabetics.

Integrating Palliative and Critical Care: Evaluation of a Quality Improvement Intervention

RATIONALE Palliative care in the intensive care unit (ICU) is an important focus for quality improvement. OBJECTIVES To evaluate the effectiveness of a multi-faceted quality improvement intervention to improve palliative care in the ICU. METHODS We performed a single-hospital, before-after study of a quality-improvement intervention to improve palliative care in the ICU. The intervention consisted of clinician education, local champions, academic detailing, feedback to clinicians, and system support. Consecutive patients who died in the ICU were identified pre- (n = 253) and postintervention (n = 337). Families completed Family Satisfaction in the Intensive Care Unit (FS-ICU) and Quality of Dying and Death (QODD) surveys. Nurses completed the QODD. The QODD and FS-ICU were scored from 0 to 100. We used Mann-Whitney tests to assess family results and hierarchical linear modeling for nurse results. MEASUREMENTS AND MAIN RESULTS There were 590 patients who died in the ICU or within 24 hours of transfer; 496 had an identified family member. The response rate for family members was 55% (275 of 496) and for nurses, 89% (523/590). The primary outcome, the family QODD, showed a trend toward improvement (pre, 62.3; post, 67.1), but was not statistically significant (P = 0.09). Family satisfaction increased but not significantly. The nurse QODD showed significant improvement (pre, 63.1; post, 67.1; P < 0.01) and there was a significant reduction in ICU days before death (pre, 7.2; post, 5.8; P < 0.01). CONCLUSIONS We found no significant improvement in family-assessed quality of dying or in family satisfaction with care, we found but significant improvement in nurse-assessed quality of dying and reduction in ICU length of stay with an intervention to integrate palliative care in the ICU. Improving family ratings may require interventions that have more direct contact with family members.

Modulation of neutrophil apoptosis by granulocyte colony-stimulating factor and granulocyte/macrophage colony-stimulating factor during the course of acute respiratory distress syndrome

Objective: To determine whether bronchoalveolar lavage fluid (BALF) from patients either at risk for the acute respiratory distress syndrome (ARDS) or with sustained ARDS modulates neutrophil apoptosis; to measure the BALF concentrations of the apoptosis inhibitors granulocyte colony‐stimulating factor (G‐CSF) and granulocyte/macrophage colony‐stimulating factor (GM‐CSF) before and after the onset of ARDS; and to determine whether the BALF concentrations of G‐CSF and/or GM‐CSF are associated with clinical outcome. Design: Prospective cohort study. Setting: Tertiary university hospital. Patients: Twenty patients at risk for ARDS and 45 patients with established ARDS. Interventions: Patients at risk for ARDS underwent bronchoalveolar lavage within 24 hrs of being identified, then again 72 hrs later. Patients with ARDS underwent bronchoalveolar lavage within 24 hrs of meeting ARDS criteria, then again on days 3, 7, and 14 of the disease. Measurements and Main Results: Normal peripheral blood neutrophil were incubated overnight in BALF from normal volunteers, from patients at risk for ARDS, or from patients with ARDS. neutrophil apoptosis was determined by flow cytometric analysis of annexin V binding. G‐CSF and GM‐CSF were measured in BALF by immunoassays. Compared with normal BALF, BALF from patients on days 1 and 3 of ARDS inhibited neutrophil apoptosis, but BALF from patients at later stages of ARDS, or from patients at risk for ARDS, did not. The BALF concentrations of both G‐CSF and GM‐CSF were elevated early in ARDS and decreased toward later stages. Patients who lived had significantly higher concentrations of GM‐CSF in the BALF than those who died. Conclusions: We conclude that the antiapoptotic effect of ARDS BALF on normal neutrophil is highest during early ARDS, and decreases during late ARDS. G‐CSF and GM‐CSF are present in BALF from patients with ARDS, and their concentrations parallel the antiapoptotic effect of ARDS BALF. These data support the concept that the life‐span of neutrophil in the air spaces is modulated during acute inflammation. GM‐CSF in the air spaces is associated with improved survival in patients with ARDS.

Evaluation of a standardized order form for the withdrawal of life support in the intensive care unit

Objective:The intensive care unit remains a setting where death is common, and a large proportion of these deaths are preceded by withdrawal of life support. We describe a quality improvement project implementing and evaluating a “withdrawal of life support order form” to improve quality of end-of-life care in the intensive care unit. Design:Before-after evaluation. Setting:County-owned, university-operated, tertiary, level I trauma center. Subjects:Subjects were 143 nurses and 61 physicians. Interventions: We conducted a before-after evaluation of the order form’s implementation. The order form has sections on preparations, sedation/analgesia, withdrawal of mechanical ventilation, and the principles of life support withdrawal. To evaluate the form, we surveyed intensive care unit clinicians regarding satisfaction with the form, measured nurse-assessed quality of dying and death with a 14-item survey (scored 0 for worst possible death to 100 for best possible), and performed chart review to assess narcotic and benzodiazepine use and time from ventilator withdrawal to death. Measurements and Main Results:We surveyed 143 nurses and 61 physicians about satisfaction with the form. Among nurses reporting that the form was used (n = 73), most (84%) reported that the order form was helpful and they were most satisfied with the sedation and mechanical ventilation sections. Almost all physicians found the form helpful (95%), and >70% of physicians found three of the four sections helpful (sedation, mechanical ventilation, and preparations). We obtained quality of dying and death scores for 41 patient deaths before and 76 deaths after the intervention. These scores did not significantly change (mean preintervention score, 78.3; mean postintervention score, 74.2; p = .54) before and after the intervention. Total doses of narcotics and benzodiazepines increased after implementation of the order form in the hour before ventilator withdrawal, the hour after ventilator withdrawal, and the hour before death (p ≤ .03). There was no change in the median time from ventilator withdrawal to death (preintervention 37 mins, postintervention 39 mins; p = .49). Conclusions:Nurses and physicians found the withdrawal of life support order form helpful. The order form did not improve nurses’ assessment of patients’ dying experience. Medications for sedation increased during the postorder form period without evidence of significantly hastening death. Although the order form was helpful to clinicians and changed medication delivery, demonstrating clear improvements in quality of dying may require larger sample sizes, more sensitive measures, or more effective interventions.