Advertisement

Improvement of outcomes after coronary artery bypass: A randomized trial comparing intraoperative high versus low mean arterial pressure

      Abstract

      Background The objective of this randomized clinical trial of elective coronary artery bypass grafting was to investigate whether intraoperative mean arterial pressure below autoregulatory limits of the coronary and cerebral circulations was a principal determinant of postoperative complications. The trial compared the impact of two strategies of hemodynamic management during cardiopulmonary bypass on outcome. Patients were randomized to a low mean arterial pressure of 50 to 60 mm Hg or a high mean arterial pressure of 80 to 100 mm Hg during cardiopulmonary bypass. Methods A total of 248 patients undergoing primary, nonemergency coronary bypass were randomized to either low (n = 124) or high (n = 124) mean arterial pressure during cardiopulmonary bypass. The impact of the mean arterial pressure strategies on the following outcomes was assessed: mortality, cardiac morbidity, neurologic morbidity, cognitive deterioration, and changes in quality of life. All patients were observed prospectively to 6 months after the operation. Results The overall incidence of combined cardiac and neurologic complications was significantly lower in the high pressure group at 4.8% than in the low pressure group at 12.9% (p = 0.026). For each of the individual outcomes, the trend favored the high pressure group. At 6 months after coronary bypass for the high and low pressure groups, respectively, total mortality rate was 1.6% versus 4.0%, stroke rate 2.4% versus 7.2%, and cardiac complication rate 2.4% versus 4.8%. Cognitive and functional status outcomes did not differ between the groups. Conclusion Higher mean arterial pressures during cardiopulmonary bypass can be achieved in a technically safe manner and effectively improve outcomes after coronary bypass. (J THORAC CARDIOVASC SURG 1995;110:1302-14)
      Coronary artery bypass grafting (CABG) prolongs life among patients with severe triple vessel or left main coronary artery disease.
      • CASS principal investigators
      • their associates
      Myocardial infarction and mortality in the Coronary Artery Surgery Study (CASS) randomized trial.
      • European Coronary Surgery Study Group
      Long term results of prospective randomised study of coronary artery bypass surgery in stable angina pectoris.
      Refinements in surgical and anesthetic techniques over the past decade have significantly reduced the risk of post-operative cardiac morbidity.
      • Loop FD
      • Lytle BW
      • Cosgrove DM
      • et al.
      Influence of the internal-mammary-artery graft on 10 year survival and other cardiac events.
      • Mangano DT
      • Siliciano D
      • Hollenberg M
      • et al.
      Postoperative myocardial ischemia: therapeutic trials using intensive analgesia following surgery.
      The incidence of neurologic complications,however, has not decreased significantly.
      • Slogoff S
      • Girgis KA
      • Keats AS.
      Etiologic factors in neuropsychiatric complications associated with cardiopulmonary bypass.
      • Shaw PJ
      • Bates D
      • Cartlidge NEF
      • et al.
      Neurologic complications of coronary artery bypass graft surgery: six month follow-up study.
      Efforts to understand the pathophysiology of neurologic and cognitive injury
      • Murkin JM.
      Neurologic dysfunction after CAB or valvular surgery. Is the medium the miscreant?.
      • Shaw PJ
      • Bates D
      • Cartlidge NEF
      • et al.
      Long term intellectual dysfunction following coronary bypass surgery.
      have focused on intraoperative issues unique to cardiac surgery, such as type of oxygenators,
      • Hessel EA
      • Johnson DD
      • Ivey TD
      • Miller DW.
      Membrane versus bubble oxygenator for cardiac operations: a prospective randomized study.
      size and types of arterial filters,
      • Blauth C
      • Arnold J
      • Schulenberg W
      • et al.
      Cerebral microembolism during cardiopulmonary bypass.
      and anesthetic techniques during cardiopulmonary bypass (CPB). Hemodynamic patterns during CABG, especially mean arterial pressures (MAPs) and fixed flow rates, have received less attention.
      • Stockard JJ
      • Bickford RG
      • Schauble JF.
      Pressure-dependent cerebral ischemia during cardiopulmonary bypass.
      Our hypothesis was that intraoperative MAP below the autoregulatory limits of the coronary and cerebral circulations may be a principal determinant of postoperative complications. Therefore, patients maintained at higher MAPs would have better outcomes.
      The objective of this trial was to compare the impact of two strategies of hemodynamic management during CPB on outcome. Patients were randomized to an MAP of either 50 to 60 mm Hg (usual range) or 80 to 100 mm Hg (intervention group). The impact of the MAP strategies on outcomes of mortality, cardiac morbidity, neurologic morbidity, cognitive deterioration, and changes in quality of life were assessed.

      METHODS

      Patient eligibility

      Between October 1991 and February 1994, patients undergoing primary elective multivessel CABG for left main or multivessel coronary artery disease at The New York Hospital–Cornell Medical Center were eligible for enrollment. Exclusion criteria included inability to complete the neuropsychologic tests (blindness, deafness, language difficulties), participation in other studies, and inability to return for follow-up. A total of 423 eligible patients were identified. Two-hundred fifty-one patients agreed to participate. All patients gave informed written consent per institutional review board guidelines.

      Baseline evaluation

      Cardiac and neurologic status, co-morbidity,
      • Charlson ME
      • Ales KL
      • Pompei P
      • MacKenzie CR
      A new method of classification of prognostic comorbidity for longitudinal studies: development and validation.
      depression (assessed by the CES-D
      • Weissman MM
      • Sholomskas D
      • Pottenger M
      • Prusoff BA
      • Locke BZ.
      Assessing depressive symptoms in five psychiatric populations: a validation study.
      ), and physical examination were documented. Cognitive function
      • Williams-Russo P
      • Sharrock NE
      • Mattis S
      • Szatronski TP
      • Charlson ME.
      Cognitive effects after epidural vs general anesthesia: a randomized trial.
      was assessed with an eleven-test neuropsychologic battery (WAIS-R Digit Span, Trail Making A and B, Boston Naming, Benton Visual Retention and Recognition Test, Controlled Oral Word Association, WAIS-R Digit Symbol, Mattis-Kovner Verbal Recall and Recognition, and Finger Tapping Test) evaluating memory (recall and recognition), linguistic, attention, and psychomotor functions. The Ammons Quick Test
      • Ammons RB
      • Ammons CH.
      The Quick Test (QT): provisional manual. Psychoperioperative cognitive assessment in elderly patients. International Neuropsychology Society.
      was used as a proxy for verbal intelligence quotient. Quality of life was measured by the SF-36 Health Survey, which evaluates seven functional domains: physical, social, role, energy, mental, pain, and general health.
      • Stewart AL
      • Greenfield S
      • Hays RD
      • et al.
      Functional status and well-being of patients with chronic conditions: results from the Medical Outcomes Study.

      Randomization procedure and treatment protocol

      Randomization was based on a table of random numbers. On entrance into the operating room, patients were randomized to either low or high MAP during CPB. In the low MAP (control) group, MAP during CPB was maintained between 50 and 60 mm Hg, and in the high MAP (experimental) group, between 80 and 100 mm Hg. In both groups, CPB flow by body surface area and temperature were held constant, and vasoactive drugs were used to maintain MAP in the desired range. Three patients were dropped from the trial. One of these patients did not undergo CABG because of severe aortic calcification and two patients required valve replacement, as detected by transesophageal echocardiography at the time of the operation. Initially, CABG reoperation was not an exclusion and one patient undergoing reoperation was enrolled. Patients undergoing reoperation were subsequently excluded.

      Intraoperative monitoring

      All hemodynamic parameters including blood pressure, heart rate, temperature, and pulmonary artery pressures were downloaded every 10 seconds from a Marquette 7000 monitor (Marquette, Milwaukee, Wis.). All operative events, pharmacologic agents, and bypass parameters were recorded on-line by a research assistant using a locally developed program synchronized with Marquette downloading. Anesthesia was induced with thiopental (1 to 2 mg/kg) and fentanyl (25 μg/kg). Pancuronium provided muscle relaxation. Anesthesia was maintained with a fentanyl bolus (1 to 5 μg/kg, to a total of 50 to 70 μg/kg), midazolam, or isoflurane (pre-CPB and post-CPB periods only). After sternotomy and pericardial incision, heparin was administered to maintain an activated clotting time of more than 480 seconds. After cannulation of the aorta and right atrium, nonpulsatile CPB with a membrane oxygenator and a 40 μm blood filter (Pall Biomedical, East Hills, N.Y.) was instituted. A Bio-Medicus centrifugal pump (Medtronic Bio-Medicus, Eden Prairie, Minn.) or a roller pump (Cobe Cardiovascular Inc., Arvada, Colo.) was used as available. Flow rates were set at 1.6 and 2.4 L/min per square meter during cooling and warming, respectively. The alpha-stat protocol
      • Bashein G
      • Townes BD
      • Nessly ML
      • et al.
      A randomized study of carbon dioxide management during hypothermic cardiopulmonary bypass.
      for blood gasmanagement was used, and body temperature was cooled to 28° to 30°C. Combinations of antegrade and retrograde cold blood cardioplegia were used with a potassium concentration of 28 mEq/L. If the MAP increased above the target level and was unresponsive to fentanyl or midazolam, sodium nitroprusside infusion was administered. If the MAP fell below the target level, phenylephrine was used. If necessary, norepinephrine or metaraminol was added. Intraoperative ischemia was managed by an identical algorithm in both groups (Appendix 1).

      Follow-up

      The study cardiologist and neurologist, blinded to the intraoperative management, performed standardized examinations at 1, 2, and 7 days after the operation and at 6 months. The neuropsychologic battery was administered on postoperative day 7 and at 6 months. At 6 months, an interval history including the patient's symptoms, medications, and hospitalizations was obtained, and the SF-36 and CES-D tests were readministered.

      Definitions of trial outcomes

      The outcomes assessed at 6 months were mortality, cardiac morbidity, neurologic morbidity, deterioration in cognitive status, and deterioration in quality of life. All deaths before 6 months were counted. Cardiac complications were myocardial infarction, pulmonary edema, adult respiratory distress syndrome, low flow state/cardiogenic shock, and cardiopulmonary arrest (Appendix 2). Final designation of cardiac complications was determined by agreement of two cardiologists blinded to the protocol.
      Definite stroke was the principal neurologic complication determined by the neurologist. Stroke included the new onset of a localized and persistent neurologic deficit (e.g., paresis, plegia, aphasia, hemianopsia, cortical blindness).
      Deterioration on three or more cognitive tests was defined as a cognitive complication. For each test, the assessment was based on within-patient change in test performance from preoperative baseline. Changes from preoperative to postoperative function that would be considered clinically important were determined a priori by a panel of experts (Appendix 3). Deterioration in quality of life was defined as a decline of more than five points on the physical component summary score of the SF-36.

      Ware J, Kosinski M, Keller SD. SF-36 Physical and Mental Component Summary Measures—A User's Manual. Boston: New England Medical Center, The Health Institute. [In press].

      Statistical analysis

      Baseline characteristics and trial outcomes in the two groups were compared by means of an intention-to-treat analysis. Subsequently, a pragmatic analysis (based on achieved MAP) was performed in which the frequency of outcomes was examined by the actual MAP achieved, regardless of randomization group. Either the χ2 or Fisher's exact test was used to compare the proportions in the two groups. Confidence intervals were also used to assess the overall difference between the two groups. Logistic regression was used to ensure that differences in baseline clinical characteristics or intraoperative management did not confound the results.

      RESULTS

      Comparability of treatment groups

      A total of 124 patients were randomized to each treatment arm. Baseline characteristics for the two groups are shown in Table I. The mean age was 65.8 ±9.4 years, and 80% were male. Approximately 50% were college graduates and 50% were working before the operation. Half had a history of hypertension, 43% a previous myocardial infarction, and 8% previous angioplasty. The preoperative MAP, an average of the three most recent blood pressures recorded before entrance into the operating room, was 81 mm Hg in both groups. The angiographically determined mean ejection fraction was 48%, and 13% had left main disease. Although 20% of patients had no chest pain, 37% had Canadian Cardiovascular Society class III or IV angina.
      • Campeau L.
      Six percent of patients reported a previous stroke. The two MAP groups were similar with regard to demographic and clinical characteristics. However, patients in the high MAP group were more likely to be working (p= 0.042) and angina-free (p = 0.008) before the operation, whereas those in the low MAP group were more often taking prescription nitrates before the operation (p= 0.007).
      Table IBaseline comparability of the two MAP groups*
      Low MAP (n =124)High MAP (n = 124)
      Demographic
      Age (yr, mean± SD)66.2 ± 10.165.4 ± 8.6
      Male7882
      Caucasian9195
      Widowed1511
      High school education or higher9387
      Working4456
      Cardiac and neurologic history
      Symptom duration (yr, mean ±SD)5.4 ± 7.75.0 ± 7.3
      Previous angioplasty610
      No angina1427
      Canadian Cardiovascular Society class
       I1916
       II2523
       III1511
       IV2723
      Ejection fraction (%, mean ±SD)48.8 ± 12.747.6 ± 12.3
      Left main disease1610
      Previous myocardial infarction3848
      Congestive failure96
      Valvular disease (aortic or mitral)53
      Previous stroke56
      Comorbidity
      Comorbidity score
       0-15659
       2-33228
      > 41213
      Body surface area (m2, mean ±SD)1.9 ± 0.21.9 ± 0.2
      Hypertension4456
      Diabetes2318
      Renal dysfunction35
      Cancer1112
      COPD orasthma613
      Cigarette smoking
       Never2725
       Former5961
       Current93
      Medications
       Diuretics2018
       Calcium channel blockers6066
       Beta blockers5459
       Nitrates6346
       Aspirin5152
       Antihypertensives1915
      *Each entry is given in percent unless otherwise specified.
      COPD, Chronic obstructive pulmonary disease; SD, standard deviation.
      Table II shows the 11 preoperative neuropsychologic test scores for the two treatment groups. There were no significant differences in baseline test scores for any test, including verbal intelligence quotient. The preoperative scores for the domains of the SF-36 are shown in Fig. 1.
      Figure thumbnail gr1
      Fig. 1Preoperative scores for the seven domains of the SF-36 in the high and low MAP groups.
      There were no significant differences in domain or summary scores between the treatment groups. The physical component summary score was 40.5 ± 9.9 in the low MAP and 42.2 ± 11.2 in the high MAP group.
      Table IIPreoperative neuropsychologic test scores
      Low MAPHigh MAP
      MeanSDMeanSD
      Ammons IQ105.714.8105.413.2
      Linguistic function
       Boston Naming Test24.64.424.64.6
       Controlled Oral Word Association37.513.137.613.4
      Memory
       Benton Recall correct5.32.14.92.4
       Benton Recognition7.81.77.61.8
       Mattis-Kovner Verbal Recall10.63.510.53.4
       Mattis-Kovner Recognition2.70.82.70.7
      Psychomotor function
       Trails A43.722.345.224.7
       Trails B104.660.8102.649.3
       Digit Symbol40.913.741.410.7
       Digit Span14.83.914.24.0
       Finger Tapping Test (dominant)46.610.346.210.1
      SD, Standard deviation; IQ, intelligence quotient.
      The details of the management of the two groups during the operative period are shown in Table IIIa. The durations of CPB and of crossclamp and sidebiter clamp application were comparable. Patients in the low MAP group received an average of 3.1 bypass grafts versus 2.9 in the high MAP group, and 80% of patients received an internal mammary graft. A centrifugal pump during CPB was used in more of the patients in the high MAP group (84.7% versus 72.6%, p = 0.02). Flows during CPB normalized for body surface area were identical in the two groups. The mean blood temperature achieved after systemic cooling was 28.4°C in the low MAP group and 28.5°C in the high MAP group.
      Tabled 1Table IIIa.Intraoperative management for the two treatment groups
      Low MAPHigh MAP
      MeanSDMeanSD
      Premedications
       Lorazepam (mg)2.41.02.51.0
       Morphine (mg)7.21.97.31.7
      Induction medications
       Fentanyl (μg/kg)40.120.040.114.0
       Thiopental (mg/kg)1.21.51.41.4
       Heparin (units/kg)315.492.0306.692.6
       Midazolam (mg)5.83.36.34.4
       Pancuronium(mg)14.322.112.66.0
      Pre-CPB cardiac output (L/min)3.51.53.61.5
      CPB time
       CPB duration (min)89.431.584.928.3
       Aortic crossclamp duration (min)46.720.043.116.7
      No. of grafts3.10.82.90.8
      Internal mammary graft (%)77.43.880.73.5
      Pump flows (L/min/m2)
       CPB on—warming1.90.31.90.3
       Warming—crossclamp off2.00.31.90.3
       Crossclamp off—CPB off2.30.22.20.3
      MAPs (mm Hg)
       CPB (full flow)59.25.481.87.8
       CPB (for all flows)51.85.269.57.1
       Aortic crossclamp on (full flow)56.57.181.27.8
       Aortic sidebiter clamp on (for all flows)49.65.867.18.8
      No. of low flow intervals7.73.29.33.8
      Hematocrit during CPB (%)20.52.620.93.1
      Returned to CPB (%)1.61.12.51.4
      SD, Standard deviation.
      The target blood pressures of 50 to 60 mm Hg for the low MAP and 80 to 100 mm Hg for the high MAP group corresponded to the MAP achieved at full flow. However, during CPB, there are episodes of low flow (typically lasting 30 to 60 seconds) necessitated by surgical technique (e.g., when the heart was retracted for examination of the distal vessels or to visualize anastomotic sites obscured by collateral bleeding). During CPB at full flow, an MAP of 59 ± 5 mm Hg was achieved in the low MAP group and an MAP of 82 ± 8 mm Hg in the high MAP group ( p = 0.0001). During CPB, including all brief periods of low flow, the mean MAP was 52 ± 5 mm Hg for the low MAP group and 69 ± 7 mm Hg for the high MAP group.
      Management during CPB required that patients in both groups receive similar drugs, but in different dosages (Table IIIb). The high MAP group received significantly higher mean doses of phenylephrine (the first-line vasopressor) than did the low MAP group (p = 0.0001). Few patients in either group received the second-line vasopressors—epinephrine, norepinephrine, or metaraminol. Thirty-six percent of the low MAP group and 4% of the high MAP group received sodium nitroprusside (the first-line vasodilator), with patients in the low MAP group receiving significantly higher mean doses. The low MAP group also received significantly higher mean doses of nitroglycerin and fentanyl (Bonferroni, p<0.0001 for all).
      Tabled 1Table IIIb.Medications during CPB
      Low MAPHigh MAP
      Medications during CPBMean dose per treated patientSDNo.Mean dose per treated patientSDNo.
      Fentanyl (μg/kg)23.114.211715.19.5104
      Nitroglycerin (μg)1312.42801.378558.5509.459
      Midazolam (mg)6.73.71065.44.797
      Phenylephrine (mg)2.42.11115.24.8112
      Epinephrine (μg)31.831.92852.268.018
      Norepinephrine (μg)25.07.121525.01025.32
      Metaraminol (mg)4.0015.12.36
      Pancuronium(mg)8.73.1818.33.777
      SD, Standard deviation.
      Overall, 93% of the low and 91% of the high MAP group received vasopressors, and 79% of the low and 48% of the high MAP group received vasodilators. There was no relationship between the doses of vasoactive drugs and major cardiac or neurologic outcomes.

      Trial outcomes

      At 6 months, the attrition rate was 4% (11/248 patients). In the low MAP group, three patients refused follow-up, one was lost to follow-up, and one moved. In the high MAP group, four patients refused follow-up and two patients moved.
      The major outcomes are mortality, cardiac morbidity, permanent neurologic deficit, cognitive complications, and deterioration in quality of life. For the 248 patients, the total mortality rate at 6 months was 4% in the low MAP and 1.6% in the high MAP group (p = 0.25). In the low MAP group, three patients (2.4%) died of cardiac or neurologic complications during hospitalization, and by 6 months two additional deaths (multisystem failure, lung cancer) were reported. In the high MAP group, two patients (1.6%) died of cardiac complications after the operation, and there were no additional deaths by 6 months (Table IV).
      Table IVCardiac and neurologic outcomes in the two treatment groups (intention to treat)
      Low MAP (n = 124)High MAP (n = 124)Low - high MAP
      No.%No.%No.%95% CI for % difference
      Fatal stroke21.600.0
      Hemiparesis*21.610.8
      Aphasia32.410.8
      Cortical blindness10.800.0
      Monocular blindness10.800.0
      Other focaldeficit00.010.8
      Total permanent neurologic complications97.232.464.8–0.5, 11.0
      Fatal cardiogenic shock10.821.6
      Shock10.800.0
      Myocard ialinfarction43.210.8
      Total cardiac complications64.832.432.4–2.2, 7.1
      Other death, total (not attributable to cardiac orneurologic causes)21.60021.6–0.6, 3.8
      Total mortality and major cardiac and neurologic morbidity16†12.964.8108.11.0, 15.1
      *Hemiparesis and aphasia; hemiparesis, apraxia, and hemianopsia; hemiparesis, aphasia, and hemisensory deficit.
      †One patient had both a cardiac and a neurologic complication.
      The last column shows the 95% confidence interval (CI) for the low MAP group percent minus high MAP group percent.
      The overall incidence of major cardiac and neurologic outcomes was 12.9% in the low MAP and 4.8% in the high MAP group (p = 0.026); the difference of 8.1% had a 95% confidence interval from 1.0% to 15.1%. Neither the differences in baseline characteristics (work status, freedom from angina, need for nitrates) nor the difference in type of CPB pump used could explain these cardiac and neurologic outcomes. Stroke was more common in the low MAP group than in the high MAP group (7.2% versus 2.4%, p = 0.076). Cardiac complications were also more common in the low MAP group (4.8% versus 2.4%, p = 0.3). Thus all differences in mortality, cardiac morbidity, and neurologic morbidity favored the high MAP group.
      Congestive heart failure (i.e., S3 gallop, typical radiographic changes) occurred after the operation among 1.2.% of the low MAP and 1.6% of the high MAP group without major cardiac events. Minor neurologic complications were defined as focal central nervous system deficits lasting less than 24 hours, including transient ischemic attacks, amaurosis fugax, or transient monocular blindness in patients without a new major focal deficit. In total, 3.2% of the low and 0.8% of the high MAP group sustained a minor neurologic complication. Differences in minor complications also favored the high MAP group.
      Cognitive outcome was assessed by within-patient differences on the eleven-test battery. In total, 23 patients (9.3%) did not complete testing at 6 months: 11 were in the low MAP and 12 in the high MAP group (7 patients had died, 2 patients had a neurologic complication and could not complete testing, 3 patients refused, and 11 were lost to follow-up). Of those completing testing, 12% of the patients in the low and 11% of patients in the high MAP group demonstrated a clinically important decline on three or more tests (95% confidence interval for difference in deterioration in cognitive function between the two groups: -7.6% to 9.1%). Table V shows the mean within-patient change in scores from before to after operation for each neuropsychologic test for the two groups. There was no difference in change in score on any test except the Finger Tapper test ( p= 0.05) between the two MAP groups. Thus the low and high MAP groups did not differ significantly with regard to preservation of cognitive function.
      Table VCognitive function 6 months after the operation: Mean within-patient change in score*
      Low MAPHigh MAP
      MeanSDMeanSD
      Linguistic function
       Boston Naming Test0.42.00.82.3
       Controlled Word Association0.08.30.66.1
      Memory
       Benton Visual Recall0.51.90.81.9
       Benton Recognition0.12.20.12.0
       Mattis-Kovner Recall1.13.41.33.6
       Mattis-Kovner Recognition0.080.790.110.72
      Psychomotor function
       Trails A–2.715.4–4.020.3
       Trails B–3.646.9–13.939.3
       Digit Symbol2.55.93.76.7
       Digit Span0.03.10.42.9
       Finger Tapping Test–1.49.30.97.0
      *Score 6 months after the operation minus the preoperative score.
      SD, Standard deviation.
      Fig. 2 shows the change in score for each of the seven domains of the SF-36.
      Figure thumbnail gr2
      Fig. 2Within patient change in SF-36 for the sevendomains of the SF-36 for the high and low randomization groups.
      Change in functional status could not be assessed for 15 patients in each group (7 patients died, 11 patients were lost to follow-up, 2 patients had a neurologic event, and 10 patients did not complete the SF-36 before or after the operation). The majority of patients demonstrated significant improvement in all seven domains. The degree of improvement did not differ between the groups: 8.3% of patients in the low MAP and 6.5% in the high MAP group reported a decline of more than five points in the physical component summary score. The 95% confidence interval for the difference in deterioration in functional status between the two groups was from -5.0% to 8.6%.
      A pragmatic analysis was conducted to ascertain whether the extent of compliance with the target blood pressure range affected the results. As shown in Table VI, patients in whom higher pressures were achieved had lower complication rates (p= 0.02). Table VII shows that there were no technical complications related to the MAP strategy; there was no increase in bleeding, transfusion requirements, or length of stay in the intensive care unit for patients in the high versus low MAP group.
      Table VIEvents up to 6 months after the operation stratified by actual MAP achieved during CPB (pragmatic analysis)
      Actual MAP achieved (mm Hg)
      <4040-4950-5960-6970-79>80Total
      No.of patients5369054576248
      All cardiac complications20%3%7%0%2%0%9%
      All neurologic complications0%3%10%2%2%0%12%
      All other deaths, (not attributable to cardiac or neurologic causes)0%3%1%0%0%0%2%
      Combined mortality and major cardiac and neurologic morbidity20%8%17%2%4%0%22%
      Each entry is percent of patients in each MAP group who had a complication.
      Table VIIPostoperative factors in the low and high MAP
      Low MAPHigh MAP
      Length of intensive care unit stay (mean hours)77 ± 25060 ± 171
      Length of hospital stay (mean days)17 ± 2513 ± 14
      Packed red blood cells (mean units)3.8 ± 5.44.0 ± 4.7
      Length of intubation (mean hours)24 ± 2132 ± 118
      Chest tube drainage at 24 hours (mean milliliters)1011 ± 701885 ± 500

      DISCUSSION

      Previous trials in adults undergoing cardiac operations for acquired disease have assessed the impact of different anesthetic agents (halothane, enflurane, isoflurane, and sufentanil),
      • Slogoff S
      • Keats AS.
      Randomized trial of primary anesthetic agents on outcome of coronary artery bypass operations.
      • Leung JM
      • Goehner P
      • O'Kelly B
      • et al.
      Isoflurane anesthesia and myocardial ischemia: comparative risk versus sufentanil anesthesia in patients undergoing coronary artery bypass surgery.
      • Helman JD
      • Leung JM
      • Bellows WH
      • et al.
      The risk of myocardial ischemia in patients receiving desflurane versus sufentanil anesthesia for coronary artery bypass graft surgery.
      pH management,
      • Bashein G
      • Townes BD
      • Nessly ML
      • et al.
      A randomized study of carbon dioxide management during hypothermic cardiopulmonary bypass.
      barbiturate infusions during CPB,
      • Zaidan JR
      • Klochany A
      • Martin WM
      • Ziegler JS
      • Harless DM
      • Andrews RB.
      Effect of thiopental on neurologic outcome following coronary artery bypass grafting.
      • Nussmeier NA
      • Arlund C
      • Slogoff S.
      Neuropsychiatric complications after cardiopulmonary bypass: cerebral protection by a barbiturate.
      filters,
      • Aris A
      • Solanes H
      • Camara ML
      • Junque C
      • Escartin A
      • Caralps JM.
      Arterial line filtration during cardiopulmonary bypass: neurologic, neuropsychologic, and hematologic studies.
      • Pugsley W
      • Linger L
      • Paschalis C
      • Treasure T
      • Harrison M
      • Newman S.
      The impact of microemboli during cardiopulmonary bypass on neuropsychological functioning.
      and oxygenators
      • Hessel EA
      • Johnson DD
      • Ivey TD
      • Miller DW.
      Membrane versus bubble oxygenator for cardiac operations: a prospective randomized study.
      on outcome after CABG and have not demonstrated significant differences in neurologic or cardiac complications.
      Historically, MAP during CPB has been maintained in the range of 50 mm Hg,
      • Stockard JJ
      • Bickford RG
      • Schauble JF.
      Pressure-dependent cerebral ischemia during cardiopulmonary bypass.
      and several studies have suggested that there was noadverse impact at this MAP.
      • Ellis RJ
      • Wisniewski A
      • Potts R
      • Calhoun C
      • Loucks P
      • Wells MR.
      • Slogoff S
      • Reul GJ
      • Keats AS.
      Role of perfusion pressure and flow in major organ dysfunction after cardiopulmonary bypass.
      Several studies in noncardiac operations have pointed to the importance of maintaining pressure within the autoregulatory range of the patient.
      • Charlson ME
      • MacKenzie CR
      • Gold JP
      • et al.
      Pre-operative and intra-operative hemodynamic predictors of post-operative myocardial infarction or ischemia in patients undergoing non-cardiac surgery.
      This trial is the first to address the issue of blood pressure management during CPB in patients having CABG.
      The physiology of autoregulation in the coronary and cerebral circulations is fundamental to the hypothesis that intraoperative blood pressure is related to the risk of complications. In the coronary and cerebral circulations, autoregulation preserves perfusion within a relatively narrow range.
      • Strangaard S.
      Autoregulation of cerebral blood flow in hypertensive patients.
      • Lassen NA
      • Christensen MS.
      Physiology of cerebral blood flow.
      Systemic pressures outside the limits of autoregulation may produce changes in perfusion with subsequent tissue damage. Our hypothesis is that patients whose MAP is maintained at higher pressures closer to their autoregulatory range will have decreased postoperative morbidity and deterioration in quality of life after CABG.
      There was prior evidence that MAP can be increased to achieve higher perfusion pressures in a manner that is technically safe while maintaining desirable patient outcomes.
      • Akins CW.
      Noncardioplegic myocardial preservation for coronary revascularization.
      This has been confirmed in the present trial. In this cohort, maintaining higher MAPs did not prolong total time on CPB, decrease distal coronary visualization, or increase duration of crossclamp or sidebiter clamp application. Further, there was no increase in bleeding, transfusion requirements, or length of stay for patients in the high MAP group.
      Blood pressure targets were achieved during periods of full flow. Flow rates were maintained within the prescribed ranges, and all cointerventions were equivalent between the two MAP groups, except for expected differences in the use of vasoactive drugs.
      The total mortality rate at 6 months was 2.8%, with most of the mortality occurring in the perioperative period. At 6 months after the operation, 3.5% had cardiac complications and 4.8% had neurologic complications. These rates compare favorably to other outcome studies of elective CABG.
      • Mangano DT.
      Perioperative cardiac morbidity.
      The rates of cognitive complications reported in other trials vary, primarily because the criteria used for defining the cognitive deterioration differ among these trials.
      • Williams Russo PG
      • Mattis S
      • Szatrowski TP
      • Sharrock N
      • Charlson MD.
      Incidence of post operative cognitive deterioration in cardiac versus non cardiac surgical patients.
      • Newman SP.
      Neuropsychologic and psychological changes.
      The individual occurrence of outcomes (mortality, cardiac, neurologic, cognitive, and functional) all favored the high MAP group. In some instances, very low complication rates contributed to the lack of statistical significance; the trial sample size was calculated in advance assuming higher rates of mortality and morbidity in the control group. When the results were analyzed by comparing the outcomes, the high MAP group had a significantly lower combined incidence of cardiac and neurologic morbidity and mortality.
      A pragmatic analysis was performed to evaluate the relationship of achieved MAP to outcomes. This analysis confirmed that patients with lower achieved MAPs tended to have more cardiac and neurologic morbidity and mortality than those maintained at high MAPs.
      In conclusion, this study demonstrates that higher MAPs can be achieved in a technically safe and reliable manner. This elevation of MAP during CPB effectively improves outcomes after elective CABG.

      DISCUSSION

      Dr. Randall B. Griepp (New York, N.Y.).

      The most important result of this clinical trial is that a high perfusion pressure during CABG reduces the prevalence of perioperative neurologic complications from about 7% to 2%. This conclusion differs from that of other studies in the past and, to my knowledge, is the first demonstration of a cerebral protective effect of high physiologic perfusion pressures during CPB. If these data are applicable to all patients undergoing CABG, many surgeons may wish to change their perfusion protocols. Therefore, a number of questions are appropriate.
      First, was there no downside whatsoever to the high perfusion pressure, no problems with cannulation site bleeding, no increased incidence of clamp injury to the aorta?
      Second, were there any factors other than perfusion pressure and development of aortic plaque as determined by transesophageal echocardiography that were predictive of stroke? In most contemporary studies age has been a powerful predictor of stroke. It would be of interest to know if this study is representative of and corresponds to findings in other studies as well.
      Third, it appears that most of the neurologic complications occurred in the subgroups with severe aortic atherosclerosis as assessed by transesophageal echocardiography. If these subgroups are removed, is there any hint of a difference in neurologic outcome in the remaining majority of patients? If not, is it necessary to use high perfusion pressures in the majority of patients who do not have a high degree of aortic atherosclerosis?
      Fourth, what would you propose as a mechanism of protection against neurologic injury by high perfusion pressures in the presence of severe aortic plaque? It is logical that the brain with intracranial stenotic cerebrovascular disease might fare better with a higher perfusion pressure. However, aortic plaque is thought to predispose to cerebral injury caused by atheroembolism, and how a high perfusion pressure would prevent or ameliorate such events is not obvious. Do you have an explanation?
      A further clinical point. Although perioperative neurologic events were much more frequent in the high pressure group, there was no long-term difference in functional or cognitive status. How important then were the perioperative events?
      Finally, what is your current clinical practice? Do you and your colleagues routinely use a perfusion pressure of 80 to 100 mm Hg in all patients undergoing CABG? If not, in which groups do you not do so?
      My questions are not meant as criticism of this fine article but merely to emphasize the importance of work in this area. It may be too late for some members of our Association to benefit from it, but by the time I need my CABG, I would like to know that the problem of perioperative neurologic dysfunction has been eliminated.

      Dr. Gold

      Thank you, Dr. Griepp, for your insightful questions. The first question dealt with the complications associated with the high mean arterial perfusion pressure. There actually were none. As a matter of fact, there were slightly shorter crossclamp times and less perioperative bleeding. The only difficulty we had was convincing the seven-member surgical faculty that high pressure CPB was indeed safe. There were no complications.
      Factors other than the pressure during the perfusion and the echocardiographic grade of the aorta for predicting stroke included some standard predictors such as age older than 75 years and preexisting neurologic abnormality.
      Indeed, in our study the groups at very low risk, meaning patients with minimal or no atherosclerosis on transesophageal echocardiography, did not have any of the strokes that were detected after the operation. However, with our increasing use of echocardiography, we have noted that the extent of disease in the transverse arch and in the ascending arch with and without epicardial and epiaortic echocardiography and the extent of disease in the descending aorta sometimes is very difficult to measure. Sometimes one will just catch the corner of a plaque that may be a significantly larger structure when fully evaluated. We therefore think that the technique is beneficial to all categories of patients undergoing coronary surgery.
      Regarding the mechanisms of injury, we and others have gathered a wealth of data with the use of transcranial Doppler echocardiography. The embolic components of atherosclerosis of the aorta are significant. However, we believe that atherosclerosis of the aorta, and indeed the presence of coronary disease, are markers for fixed intercerebral vascular stenoses that are better managed with higher perfusion pressures. If you allow the perfusion pressure to remain within the normal autoregulatory range, you overcome some of these stenoses and therefore get better end-organ perfusion, not only of the brain but of the heart, the gut, and other organs.
      With regard to clinical status, this has been a bit of a hard sell. Even within our own institution, in patients with absolutely no visible atherosclerosis of the aorta on transesophageal echocardiography, we keep the pressure between 65 and 75 mm Hg. We do study all of these patients during the operation (if not before). In patients with moderate or advanced disease of the ascending aorta, we would keep the mean perfusion pressure above 80 mm Hg. If I were the patient, I would want the perfusion pressure to be high.

      Dr. Thomas A. Pfeffer (Los Angeles, Calif.).

      I have two questions. The first concerns the technique of performing the proximal anastomoses. Was a partial aortic occluding clamp used for all proximal anastomoses or were some performed with a single total occlusion clamp with all distal and proximal anastomoses performed during the same period of aortic crossclamping? Did the detection of aortic atherosclerotic disease affect this technique? Were any patients considered candidates for an aortic endarterectomy procedure, as advocated by Dr. Kouchoukos?
      Second, although maintaining a desired perfusion pressure is generally easy during CPB, that frequently is not the case in the period immediately after discontinuation of CPB. The patients are usually vasodilated and hemodiluted and we are dependent on our colleagues in anesthesiology for assistance in maintaining adequate perfusion pressure. Were there any differences in these two groups of patients, concerning their blood pressure and hemodynamics, in the period immediately after discontinuation of CPB? Could you comment on the requirement for vasopressors during this period?

      Dr. Gold

      All of these patients underwent creation of a proximal anastomosis with the application of a partial occlusion clamp. Although we are now studying a single-clamp versus a two-clamp protocol using the same type of neurologic and cognitive monitoring that has been proposed in other institutions, for purposes of this study a partial occlusion clamp was always used.
      None of the patients had severe enough disease in the ascending aorta to warrant an aortic endarterectomy. Perhaps in the patients who did have a stroke, we should have considered endarterectomy. Maybe we should reconsider this type of procedure in the future in patients with advanced disease.
      In answer to your last question, immediately after CPB and in the perioperative period in the intensive care unit, there were some minor differences between the two groups of patients. Patients treated with more vasopressor in the intraoperative period required less vasopressor after the operation, required less fluid, had less weight gain, had a shorter period ventilatory support, and actually were released from the intensive care unit slightly (but not statistically) faster. Certainly, there was no downside from using high perfusion pressure.

      Dr. Watts R. Webb (New Orleans, La.).

      Did you measure cerebral blood flow during these studies? The conventional studies have usually shown that autoregulation of blood flow to the brain ranges from 50 up to 150 mm Hg, and, with hypothermia, down to 28°C or less; autoregulation may continue down to as low as 20 mm Hg. I would like to know both the temperatures that you used and the blood flows that you measured during this period of time.

      Dr. Gold

      We did not measure cerebral blood flow. The only flows that we measured were total systemic blood flow during CPB. We used 28°C (moderate hypothermia) for the major part of the aortic crossclamp interval.
      Although your comments about autoregulation of the brain are correct, perhaps in patients with advanced atherosclerosis of the cerebrovascular system some alteration in those numbers needs to be made. If you evaluate integrated electroencephalographic cerebral function and other measurements of cerebral ischemia, you will see that as pressures fall in patients with fixed stenoses, regional cerebral perfusion suffers.

      APPENDIX 1

      Management of intraoperative ischemia

      Before or after CPB, ischemia (ST elevation or depression >1 mm) was managed by a nitroglycerin infusion (0.13 to 1 μg/kg per minute) if there were no hemodynamic changes. If ischemia occurred with an increase in MAP of more than 20%, fentanyl (5 μg/kg), nitroglycerin (0.3 to 1 μg per minute), or sodium nitroprusside was used. For pulse rates greater than 80 beats/min, esmolol (0.1 mg/kg) was used. If the MAP was less than 20% of baseline, phenylephrine (1.5 μg/kg) was given to minimize ischemic electrocardiographic changes (Appendix Fig. 1).
      Figure thumbnail gr3
      Appendix Fig. 1. Management of intraoperative ischemia.

      Management of blood pressure on CPB

      The method for blood pressure management during CPB was as follows for both randomization groups (Appendix Fig. 2): for MAP above target, sodium nitroprusside (50 mg/200 ml) infusions were used as necessary.
      Figure thumbnail gr4
      Appendix Fig. 2. Management of intraoperative MAP. PRN, As required.
      For MAP below target, a phenylephrine bolus (0.5 to 1.0 mg) was given; phenylephrine infusions (10 mg/250 ml) were given if boluses were insufficient. If the patient was unresponsive to phenylephrine, norepinephrine infusion (8 mg/250 ml) was used as a second-line vasopressor. Metaraminol (200 mg/250 ml) was used as a third-line vasopressor if the first- and second-line vasopressors failed.

      APPENDIX 2. DEFINITIONS OF CARDICAC COMPLICAITONS

      Postoperative myocardial infarction:

      New persistent Q waves of more than 0.03 msec and greater than 1 mV in depth required in two contiguous leads on a standard twelve-lead electrocardiogram, in the absence of a new conduction abnormality or a marked change in the QRS axis.
      • The Coronary Drug Project Research Group
      The Coronary Drug project.

      Pulmonary edema:

      (1) Rales occupying two thirds of the lung fields and a typical x-ray picture for pulmonary edema or (2) pulmonary capillary wedge pressure persistently greater than 25 mm Hg. These findings were associated with refractory hypoxemia (arterial oxygen tension/inspired oxygen fraction <200).
      • Murray JF
      • Matthay MA
      • Luce JM
      • Flick MR.
      An expanded definition of adult respiratory distress syndrome.

      Adult respiratory distress syndrome:

      Diffuse interstitial pattern on the roentgenogram, a pulmonary capillary wedge pressure less than 18 mm Hg, and refractory hypoxemia.

      Cardiogenic shock:

      A syndrome of end-organ hypoperfusion (i.e., urine output <10 ml/hr for >2 hours) with a systolic blood pressure less than 90 mm Hg or a mean blood pressure less than 65 mm Hg and a pulmonary capillary wedge pressure greater than 18 mm Hg in combination with a decline in the cardiac index to less than 2.2 L/min per square meter. The elevated central venous pressure or right atrial pressure in the absence of tamponade constituted evidence of right heart failure, if the cardiac output was less than 2.2 L/min per square meter. This includes the finding of systemic hypotension unresponsive to fluid administration, with an elevated pulmonary artery or wedge pressure in the setting of a cardiac index less than 1.5 L/min per square meter and evidence of end-organ hypoperfusion.

      Appendix 3

      Tabled 1The clinically important differences for the neuropsychologic test battery
      Boston NamingDecrease > 4
      Controlled Oral WordAssociationDecrease > 12
      Benton Visual Retention TestDecrease > 2
      Benton Recognition TestDecrease > 2
      Mattis Kovner Verbal RecallDecrease > 3
      Mattis KovnerRecognitionDecrease > 0.7
      Trail Making AIncrease > 22
      Trail Making BIncrease > 84
      WAIS-R DigitSpanDecrease > 3
      Finger Tapping TestDecrease > 6
      WAIS-R Digit SymbolDecrease > 11

      References

        • CASS principal investigators
        • their associates
        Myocardial infarction and mortality in the Coronary Artery Surgery Study (CASS) randomized trial.
        N Engl J Med. 1984; 310: 750-758
        • European Coronary Surgery Study Group
        Long term results of prospective randomised study of coronary artery bypass surgery in stable angina pectoris.
        Lancet. 1982; 2: 1173-1180
        • Loop FD
        • Lytle BW
        • Cosgrove DM
        • et al.
        Influence of the internal-mammary-artery graft on 10 year survival and other cardiac events.
        N Engl J Med. 1986; 314: 1-6
        • Mangano DT
        • Siliciano D
        • Hollenberg M
        • et al.
        Postoperative myocardial ischemia: therapeutic trials using intensive analgesia following surgery.
        Anesthesiology. 1992; 76: 342-353
        • Slogoff S
        • Girgis KA
        • Keats AS.
        Etiologic factors in neuropsychiatric complications associated with cardiopulmonary bypass.
        Anesth Analg. 1982; 61: 903-911
        • Shaw PJ
        • Bates D
        • Cartlidge NEF
        • et al.
        Neurologic complications of coronary artery bypass graft surgery: six month follow-up study.
        Br Med J. 1986; 293: 165-167
        • Murkin JM.
        Neurologic dysfunction after CAB or valvular surgery. Is the medium the miscreant?.
        Anesth Analg. 1993; 76: 213-214
        • Shaw PJ
        • Bates D
        • Cartlidge NEF
        • et al.
        Long term intellectual dysfunction following coronary bypass surgery.
        Q J Med. 1987; 239: 259-268
        • Hessel EA
        • Johnson DD
        • Ivey TD
        • Miller DW.
        Membrane versus bubble oxygenator for cardiac operations: a prospective randomized study.
        J THORAC CARDIOVASC SURG. 1980; 80: 111-122
        • Blauth C
        • Arnold J
        • Schulenberg W
        • et al.
        Cerebral microembolism during cardiopulmonary bypass.
        J THORAC CARDIOVASC SURG. 1988; 95: 668-676
        • Stockard JJ
        • Bickford RG
        • Schauble JF.
        Pressure-dependent cerebral ischemia during cardiopulmonary bypass.
        Neurology. 1973; 23: 521-529
        • Charlson ME
        • Ales KL
        • Pompei P
        • MacKenzie CR
        A new method of classification of prognostic comorbidity for longitudinal studies: development and validation.
        J Chronic Dis. 1987; 40: 373-383
        • Weissman MM
        • Sholomskas D
        • Pottenger M
        • Prusoff BA
        • Locke BZ.
        Assessing depressive symptoms in five psychiatric populations: a validation study.
        Am J Epidemiol. 1977; 106: 203-214
        • Williams-Russo P
        • Sharrock NE
        • Mattis S
        • Szatronski TP
        • Charlson ME.
        Cognitive effects after epidural vs general anesthesia: a randomized trial.
        JAMA. 1995; 274: 44-50
        • Ammons RB
        • Ammons CH.
        The Quick Test (QT): provisional manual. Psychoperioperative cognitive assessment in elderly patients. International Neuropsychology Society.
        Psychol Rep. 1962; 11: 111-161
        • Stewart AL
        • Greenfield S
        • Hays RD
        • et al.
        Functional status and well-being of patients with chronic conditions: results from the Medical Outcomes Study.
        JAMA. 1989; 262: 907-913
        • Bashein G
        • Townes BD
        • Nessly ML
        • et al.
        A randomized study of carbon dioxide management during hypothermic cardiopulmonary bypass.
        Anesthesiology. 1990; 72: 7-15
      1. Ware J, Kosinski M, Keller SD. SF-36 Physical and Mental Component Summary Measures—A User's Manual. Boston: New England Medical Center, The Health Institute. [In press].

        • Campeau L.
        Circulation. 1976; 54 (Letter to the Editor): 522
        • Slogoff S
        • Keats AS.
        Randomized trial of primary anesthetic agents on outcome of coronary artery bypass operations.
        Anesthesiology. 1989; 70: 179-188
        • Leung JM
        • Goehner P
        • O'Kelly B
        • et al.
        Isoflurane anesthesia and myocardial ischemia: comparative risk versus sufentanil anesthesia in patients undergoing coronary artery bypass surgery.
        Anesthesiology. 1991; 74: 838-847
        • Helman JD
        • Leung JM
        • Bellows WH
        • et al.
        The risk of myocardial ischemia in patients receiving desflurane versus sufentanil anesthesia for coronary artery bypass graft surgery.
        Anesthesiology. 1992; 77: 47-62
        • Zaidan JR
        • Klochany A
        • Martin WM
        • Ziegler JS
        • Harless DM
        • Andrews RB.
        Effect of thiopental on neurologic outcome following coronary artery bypass grafting.
        Anesthesiology. 1991; 74: 406-411
        • Nussmeier NA
        • Arlund C
        • Slogoff S.
        Neuropsychiatric complications after cardiopulmonary bypass: cerebral protection by a barbiturate.
        Anesthesiology. 1986; 64: 165-170
        • Aris A
        • Solanes H
        • Camara ML
        • Junque C
        • Escartin A
        • Caralps JM.
        Arterial line filtration during cardiopulmonary bypass: neurologic, neuropsychologic, and hematologic studies.
        J THORAC CARDIOVASC SURG. 1986; 91: 526-533
        • Pugsley W
        • Linger L
        • Paschalis C
        • Treasure T
        • Harrison M
        • Newman S.
        The impact of microemboli during cardiopulmonary bypass on neuropsychological functioning.
        Stroke. 1994; 25: 1393-1399
        • Ellis RJ
        • Wisniewski A
        • Potts R
        • Calhoun C
        • Loucks P
        • Wells MR.
        Reduction of flow rate and arterial pressure at moderate hypothermia does not result in cerebral dysfunction. 79. 1980: 173-180
        • Slogoff S
        • Reul GJ
        • Keats AS.
        Role of perfusion pressure and flow in major organ dysfunction after cardiopulmonary bypass.
        Ann Thorac Surg. 1990; 50: 911-918
        • Charlson ME
        • MacKenzie CR
        • Gold JP
        • et al.
        Pre-operative and intra-operative hemodynamic predictors of post-operative myocardial infarction or ischemia in patients undergoing non-cardiac surgery.
        Ann Surg. 1989; 210: 637-648
        • Strangaard S.
        Autoregulation of cerebral blood flow in hypertensive patients.
        Circulation. 1976; 53: 720-727
        • Lassen NA
        • Christensen MS.
        Physiology of cerebral blood flow.
        Br J Anesth. 1976; 48: 719-734
        • Akins CW.
        Noncardioplegic myocardial preservation for coronary revascularization.
        J THORAC CARDIOVASC SURG. 1986; 88: 174-181
        • Mangano DT.
        Perioperative cardiac morbidity.
        Anesthesiology. 1990; 72: 153-184
        • Williams Russo PG
        • Mattis S
        • Szatrowski TP
        • Sharrock N
        • Charlson MD.
        Incidence of post operative cognitive deterioration in cardiac versus non cardiac surgical patients.
        Clin Res. 1995; 43: 304A
        • Newman SP.
        Neuropsychologic and psychological changes.
        in: Cardiac surgery and the brain. : Edward Arnold, London1993: 34-54
        • The Coronary Drug Project Research Group
        The Coronary Drug project.
        Circulation. 1973; 47: I1-I50
        • Murray JF
        • Matthay MA
        • Luce JM
        • Flick MR.
        An expanded definition of adult respiratory distress syndrome.
        Respir Dis. 1988; 138: 720-723