Effects of Single Dose, Postinduction Dexamethasone on Recovery After Cardiac Surgery

Transcription

1 Effects of Single Dose, Postinduction on Recovery After Cardiac Surgery Jean-Pierre Yared, MD, Norman J. Starr, MD, Frederick K. Torres, MD, C. Allen Bashour, MD, Gregory Bourdakos, MD, Marion Piedmonte, MA, Judith A. Michener, RRT, Jeffrey A. Davis, RRT, and Thomas E. Rosenberger, RRT Deartment of Cardiothoracic Anesthesiology, The Cleveland Clinic Foundation, Cleveland, Ohio Background. Corticosteroids have been recommended to facilitate raid recovery after cardiac surgery. We reviously reorted that dexamethasone given after induction of anesthesia decreases the incidence of ostoerative shivering. We erformed a ost hoc analysis of the data obtained during that study, focusing on secondary outcomes. Methods. A total of 235 adult atients undergoing elective coronary or valvular heart surgery were randomized to receive dexamethasone 0.6 mg/kg or lacebo after induction of anesthesia. Patients who had harmacologically treated diabetes mellitus, had hyersensitivity to dexamethasone, or were receiving treatment with corticosteroids were excluded. Results. We found that, comared with lacebo, atients receiving dexamethasone were more likely to remain tracheally intubated for 6 hours or less (26.4% vs 10.0%, 0.020) and had a lower incidence of early ostoerative fever (20.2% vs 36.8%, 0.009) and new-onset atrial fibrillation during the first 3 days ostoeratively (18.9% vs 32.3%, 0.027). However, we could not demonstrate a statistical difference in the intensive care unit or hosital length of stay, or in overall morbidity and mortality. The dexamethasone-treated atients were also more likely to have a higher blood glucose on admission to the intensive care unit (186 mg/dl vs 143 mg/dl, 0.012). Conclusions. facilitates early tracheal extubation and is associated with a lower incidence of early ostoerative fever and new-onset atrial fibrillation. Aart from a treatable decreased glucose tolerance, dexamethasone treatment was not shown to affect morbidity or mortality significantly. (Ann Thorac Surg 2000;69:1420 4) 2000 by The Society of Thoracic Surgeons Methods for obtaining early tracheal extubation, as well as for early discharge from the intensive care unit (ICU) and hosital after cardiac surgery under cardioulmonary byass (CPB), are receiving a great deal of attention because of the imact of ICU and hosital length of stay on the cost of surgery. Multile strategies have been recommended to facilitate this rocess of fast-track recovery, including the use of low-dose oiate anesthesia [1] and erioerative administration of corticosteroids. Inhibition of the inflammatory resonse to CPB by corticosteroids is thought to accelerate recovery [2, 3], decrease myocardial edema [4], and decrease ostoerative fever as well as fluid and inotroic drug requirements [5]. We recently reorted that a single dose of dexamethasone (0.6 mg/kg) after induction of anesthesia decreases the incidence of ostoerative shivering, and that this effect is indeendent from the duration of CPB and surgery as well as temerature during CPB and uon ICU admission [6]. However, methylrednisolone has been reorted to increase intraulmonary shunt and to delay extubation [7]. We erformed a ost hoc analysis of data obtained during the study of the effects of Acceted for ublication Nov 9, Address rerint requests to Dr Yared, Deartment of Cardiothoracic Anesthesiology, The Cleveland Clinic Foundation, 9500 Euclid Ave, G5, Cleveland, OH 44195; yaredj@cesmt.ccf.org. dexamethasone on shivering, focusing on secondary outcomes such as tracheal intubation time, ICU and hosital length of stay, new-onset atrial fibrillation, and major morbidities. In this reort we resent the imact of dexamethasone treatment on the outcome of cardiacrelated surgery. Material and Methods After Institutional Review Board aroval, and informed consent, a rosective, randomized, double-blind, lacebo-controlled study was undertaken. A total of 236 atients aged 20 years or older and having elective coronary or valvular heart surgery with cardioulmonary byass were randomly assigned to two grous. One grou (DEX) received dexamethasone 0.6 mg/kg, and the other grou (PL) received an equal volume of lacebo (ie, saline) after induction of anesthesia but before skin incision. Patients with a history of hyersensitivity to dexamethasone or of diabetes mellitus treated with insulin or oral medications, or who were receiving theray with corticosteroids were excluded. Patients who were enrolled in the study but returned to the oerating room because of surgical bleeding, or who received additional corticosteroids or other drugs such as arotinin that interfere with the inflammatory resonse, were excluded from the analysis by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 Ann Thorac Surg YARED ET AL 2000;69: DEXAMETHASONE AND OUTCOME OF HEART SURGERY 1421 All atients were remedicated with oral lorazeam 0.02 to 0.04 mg/kg and received standard monitoring. Anesthetic and ostoerative management including benzodiazeines, low-dose oiates, isoflurane, and neuromuscular blockers was comatible with early extubation. Cardioulmonary byass was erformed under either normothermia or moderate hyothermia, according to surgeon reference. All atients received blood cardiolegia and were rewarmed to bladder temerature (T BL) of 37 C before searation from CPB. Uon admission to the ICU, atients were initially ventilated. Central blood temerature (T PA) was measured continuously in the ulmonary artery. According to ICU routine, a hot air blanket (Bair Hugger, Augustine Medical, Inc, Eden Prairie, MN) was alied for T PA less than 35.5 C and discontinued for T PA greater than 36.5 C. Shivering was treated with ancuronium 0.02 to 0.03 mg/kg after unconsciousness had been obtained with midazolam in 1-mg increments. Patients were weaned to continuous ositive airway ressure (CPAP) and extubated according to ICU routine (ie, if atients were resonsive to command, were hemodynamically stable, had aroriate gas exchange and resiratory mechanics, and chest tube drainage did not exceed 50 ml/h for 2 consecutive hours). Residual neuromuscular blockade was reversed when necessary with neostigmine 0.05 mg/kg and glycoyrrolate 0.01 mg/kg. Serum glucose was monitored in the erioerative eriod and hyerglycemia was controlled by insulin according to the reference of the anesthesiologist. The time elased from ICU admission to tracheal extubation, ICU and hosital length of stay, and mortality, as well as the incidence of major neurologic, renal, cardiac, infectious, and ulmonary morbidities were obtained from the Cardiothoracic Anesthesia Database. Additional data were collected about newonset atrial fibrillation in the first 3 ostoerative days, and administration of continuous infusions of insulin, vasoressors (noreinehrine), and inotroic drugs (einehrine, dobutamine, or milrinone) in the ICU on the day of surgery. Early ostoerative fever was defined as eak T PA of 38 C or more in the first 6 hours after ICU admission. Statistical Analysis The rimary endoint of the randomized study was the comarison of the effect of dexamethasone versus lacebo on the incidence of shivering. For the ost hoc analyses, comarison of dexamethasone versus lacebo on outcome, as well as comlications that are categorical (eg, gender), were analyzed using 2 tests unless the cell counts were small, in which case Fisher exact tests were used. Continuous measures were analyzed using t tests if the data seemed to be normally distributed, or the Wilcoxon rank sum tests for variables such as intubation time and length of stay were not normally distributed. All statistical tests were two-tailed; values of 0.05 or less were used to define statistical significance. Table 1. Accruals and Reasons for Exclusion From Analysis For Secondary Outcomes Results Total Randomized (n) Excluded from analysis (n) Reasons Did not receive 4 6 study drug Surgical bleeding 2 1 Received 0 1 arotinin in OR Received additional steroids in OR 2 4 OR oerating room. Between January 15, 1997, and Setember 25, 1997, a total of 236 atients were enrolled in the study. They were divided into two grous (DEX or PL), with 118 atients randomly assigned to each grou. A total of 20 atients were excluded from analysis (Table 1); therefore, outcome data were obtained on 216 evaluable atients and rovide the basis of this reort. The study grou was redominantly male (178 of 216; 82.4%). This was consistent across both treatment grous, with 88 of 110 (80.0%) male atients in the PL grou and 90 of 106 (84.9%) in the DEX grou. Two atients with diet-controlled diabetes mellitus were included in the study, 1 atient in each grou. The two grous were similar in age distribution and in the tye of surgical rocedures that they underwent. In all, 7 atients had unlanned aortic root surgery, 2 of whom required hyothermic circulatory arrest (Table 2). Duration of anes- Table 2. Descritive Data by Treatment Grou (n 110) (n 106) Age (mean SD) (y) Male/Female (n) 88/22 90/16 CABG only (n) Valve(s) only (n) CABG valve(s) (n) Aortic root 3 4 surgery CABG or valve (n) Atrial setal 2 3 defect CABG or valve (n) Circulatory arrest 1 1 during CPB (unlanned) (n) Minimally invasive surgery with CPB (n) 9 11 CABG coronary artery byass grafting; byass. CPB cardioulmonary

3 1422 YARED ET AL Ann Thorac Surg DEXAMETHASONE AND OUTCOME OF HEART SURGERY 2000;69: Table 3. Intraoerative and Postoerative Measurements by Treatment Grou Variable Mean (SD) Anesthesia time (min) 324 (77) (72) CPB time (min) 112 (40) (38) Cross clam time (min) 83 (32) (32) Lowest temerature on CPB ( C) 33.3 (2.6) (2.4) CPB cardioulmonary byass; SD standard deviation. thesia, CPB, and aortic cross-claming, as well as the lowest temerature reached during CPB were not significantly different between grous (Table 3). None of the atients required antagonistic medication for neuromuscular block. There was no significant difference in ICU length of stay between DEX and PL grous ( vs hours, resectively, 0.587), as well as in ostoerative hosital length of stay ( vs days, resectively, 0.259). Duration of tracheal intubation was versus hours for the DEX and PL grous, resectively ( 0.074). Comared with the PL atients, a larger ercentage of DEX atients had a short ( 6 hours) intubation time (10.0% vs 26.4%, resectively, 0.020). The incidence of fever during the first 6 hours ostoeratively was 20.2% versus 36.8% for the DEX and PL grous, resectively ( 0.009). Uon ICU admission the Pao 2 /FiO 2 ratios were not significantly different between grous; however, serum HCO 3 was significantly lower in the DEX grou. Moreover, there were no significant differences in cardiac indexes at base line, as well as uon ICU admission (Table 4). The use of one or more inotroic drug was 26% versus 23% for the PL and DEX grous, resectively ( 0.734), and 1 atient in the PL grou required intraaortic balloon um counterulsation during searation from CPB. Noreinehrine was required in 15.46% versus 18.87% of PL and DEX atients, resectively ( 0.506). Analysis of blood glucose levels at baseline showed no significant difference between the DEX and PL grous ( vs , resectively; 0.753). However, blood glucose was significantly higher in the DEX grou after CPB ( vs mg/dl, resectively; 0.003) and uon ICU admission ( vs mg/dl, resectively; 0.012). The ercentage of atients requiring insulin was significantly higher in the DEX grou only after CPB (5.7% vs 3.0%, resectively; 0.023). The incidence of new-onset atrial fibrillation in the first 3 days ostoeratively was lower in the DEX grou Table 4. Resiratory and Hemodynamic Outcomes Grou Mean (SD) PaO 2 /FiO 2 at ICU admission (92.7) (99.6) HCO 3 baseline (meq/l) (1.84) (2.12) HCO 3 at ICU admission (meq/l) 24.5 (1.79) (2.64) Cardiac index baseline (L min 1 m 2 ) 2.5 (0.7) (0.7) Cardiac index at ICU admission (L min 1 m 2 ) 2.9 (0.7) (0.8) ICU intensive care unit; SD standard deviation. comared with the PL grou (16.0% vs 32.4%, resectively; 0.006). Analysis by surgical subset is resented in Table 5. There were no significant differences in mortality as well as in ulmonary, renal, neurologic, cardiac-related, and infectious morbidity between grous (Table 6). Merging the DEX and PL grous together showed that, comared with atients who exerienced ostoerative shivering, atients without shivering had a lower incidence of fever (19.5% vs 45.8%, resectively; 0.001) and a higher incidence of short ( 6 hours) intubation time (23.6% vs 6.3%, resectively; 0.009). Comment The rimary analysis for this rosective, randomized, controlled trial demonstrated that a single dose of dexamethasone ostinduction reduces the incidence of ostoerative shivering in cardiac surgical atients [6]. The resent ost hoc analysis shows that dexamethasone facilitates early tracheal extubation and reduces the incidence of early ostoerative fever as well as new-onset atrial fibrillation. However, it had no effect on intraulmonary shunting as well as on ICU and hosital length of stay. Moreover, shivering (irresective of the treatment Table 5. Incidence of New Onset Atrial Fibrillation in the First Three Postoerative Days CABG only Valve only Combined CABG and valve CABG cardioulmonary byass graft.

4 Ann Thorac Surg YARED ET AL 2000;69: DEXAMETHASONE AND OUTCOME OF HEART SURGERY 1423 Table 6. Incidence of Postoerative Morbidity Morbidity Death (n) 3 (2.7) 2 (1.9) Pulmonary: intubation time 3 (2.7) 5 (4.7) h (n) (includes reintubation time) Renal failure: oliguria or 3 (2.7) 0 (0) hemodialysis (n) Cardiac: CI 1.8 desite 1 (0.9) 0 (0) inotroic drugs, or MI requiring ventricular assist device (n) Neurologic: focal or global 3 (2.7) 3 (2.7) deficit (n) Infection: blood stream 2 (1.8) 1 (0.9) infection/mediastinitis (n) 0 (0) 0 (0) CI cardiac index; MI myocardial infarction. grou) was associated with delayed extubation and with fever. Conflicting data exist regarding the effect of corticosteroids on oxygen exchange and tracheal intubation time. Although some investigators have found a reduced intubation time in methylrednisolone-treated atients comared with historical controls [2] and imroved oxygenation [8], others have found no effect on intraulmonary shunt fraction [9], and still others have shown an increased intubation time as well as shunting [7]. In the latter study, the methylrednisolone-treated grou received a significantly larger dose of midazolam intraoeratively, and a larger (although not statistically significant) dose of midazolam ostoeratively. This may have contributed to the longer intubation time in the steroidtreated grou. increased the ercentage of atients extubated early ( 6 hours); however, the trend toward a lower mean intubation time aroached, but did not reach, statistical significance. This may reresent a tye II statistical error resulting from a small samle size, but it also may be the result of the resence (in both grous) of a few outliers with very long intubation time, as indicated by the large standard deviation. The reduced intubation time that we observed in the DEX grou may reflect a direct effect of the drug, but is more likely due to the fact that shivering was less common in this grou. Shivering was associated with delayed tracheal extubation, ossibly because of the sedation and neuromuscular blockade required to treat it. We found that dexamethasone did not affect the ratio of arterial to insired oxygen at ICU admission. Imaired oxygenation after cardiac surgery has been attributed to activation of the inflammatory resonse, ulmonary sequestration of neutrohils, and oxidative stress [5, 10, 11]. The inflammatory resonse to CPB can be modulated not only by corticosteroids but also by other factors such as the temerature used during CPB [12, 13]. The temerature maintained during CPB in our study was higher than in other reorts [7], exlaining in art the difference in results. The reduced incidence of fever that we observed in the immediate ostoerative eriod in dexamethasonetreated atients has been reviously described. It has been attributed to inhibition of the release of yrogenic cytokines by corticosteroids [5]. This corticosteroidmediated effect is advantageous in the cardiac surgery atient, as oxygen consumtion is directly related to temerature [14]. The association of dexamethasone treatment with a decreased incidence of new- onset atrial fibrillation in the first 3 days ostsurgery is intriguing. Atrial fibrillation affects 20% to 30% of cardiac surgery atients, with the eak incidence on day 2 or 3 after surgery [15]. It may be associated with serious comlications including strokes, decreased cardiac outut, and imaired oxygen exchange, and it usually necessitates an increase in hosital length of stay [16]. Our results suggest a ossible link between new-onset atrial fibrillation and the ost-cpb inflammatory resonse; however, the mechanism of such an association is unclear. The excessive systemic inflammatory resonse syndrome that is seen in some cardiac surgical atients may result in ostoerative failure of major organs. Corticosteroids are thought to imrove outcome through inhibition of the systemic inflammatory resonse syndrome [17, 18]. Our results show that dexamethasone has no imact on oerative morbidity and mortality. Such findings could imly that a single dose of dexamethasone had no imact on such outcome measurements. However, it is ossible that because these comlications are rare, the samle size was not adequate to allow detection of such a difference. Our data on serious ostoerative comlications are otentially incomlete, as we did not follow-u atients after their hosital discharge. It is therefore ossible that some atients received treatment for such comlications in another facility. Cardioulmonary byass suresses delayed immunity, articularly in older atients [19], and methylrednisolone suresses T-cell mediated function synergistically with CPB [20]. The hyerglycemia and increased need for insulin that we observed in atients receiving dexamethasone have reviously been described [5]. The combination of dexamethasone-induced suression of the immune resonse and hyerglycemia raises concerns about a greater suscetibility to infection [21] and strokes [22] in this grou, and emhasizes the need to monitor and control blood glucose, articularly in atients receiving dexamethasone. Contrary to other reorts, we did not find a significant increase in cardiac index or a decrease in the use of inotroic agents in atients treated with dexamethasone [5, 23]. The metabolic acidemia observed in the DEX grou uon ICU admission is of concern, although it was mild. When comared with historical controls, atients receiving corticosteroids as art of an accelerated recovery rotocol have been shown to have a shorter hosital length of stay [2]. In other cases a nonsignificant decrease in ICU length of stay in atients receiving dexamethasone has been reorted [5]. The lack of difference in ICU

5 1424 YARED ET AL Ann Thorac Surg DEXAMETHASONE AND OUTCOME OF HEART SURGERY 2000;69: and ostoerative hosital length of stay between grous in our study might be due to the fact that we used smaller doses of steroids than in other studies, or it might indicate that dexamethasone had no effect on length of stay. However, it is more likely a reflection of the fact that, for all study atients, we used accelerated recovery rotocols and hosital-secific management decisions, indeendent of study design. In conclusion, this ost hoc study shows that, in addition to the reduced incidence of ostoerative shivering that we demonstrated reviously, dexamethasone facilitates early tracheal extubation, reduces the incidence of early ostoerative fever, and is associated with a lower incidence of new-onset atrial fibrillation. Excet for decreased glucose tolerance and a mild metabolic acidemia on ICU admission, no adverse effect could be attributed to dexamethasone treatment with resect to mortality and morbidity. However, because such comlications are rare, a larger samle size is needed to confirm these findings. An imortant limitation of this study is that we did not measure inflammatory mediators released in resonse to CPB and the study drug. Further research is needed to correlate clinical observations with biochemical markers of inflammation, and to determine whether inhibition of the inflammatory resonse by a different dose of dexamethasone is advantageous. The increased severity of hyerglycemia in dexamethasone-treated atients raises concerns about otential associated adverse effects, and suggests that atients should be monitored and treated effectively to revent hyerglycemia. References 1. Cheng D, Karski J, Peniston C, et al. Early tracheal extubation after coronary artery byass graft surgery reduces costs and imroves resource use. Anesthesiology 1996;85: Engelman RM, Rousou JA, Flack JE, et al. Fast-track recovery of the coronary byass atient. Ann Thorac Surg 1994;58: Dietzman RH, Lunseth JB, Goott B, Berger EC. The use of methylrednisolone during cardioulmonary byass. A review of 427 cases. J Thorac Cardiovasc Surg 1975;69: Vejlsted H, Anderson K, Fischer Hanson B, et al. Myocardial reservation during anoxic arrest. Scand J Thorac Cardiovasc Surg 1983;17: Jansen NJG, van Oeveren W, Broek L, et al. Inhibition by dexamethasone of the reerfusion henomena in cardioulmonary byass. J Thorac Cardiovasc Surg 1991;102: Yared JP, Starr NJ, Hoffman-Hogg L, et al. decreases the incidence of shivering following cardiac surgery. A randomized, double-blind, lacebo controlled study. Anesth Analg 1998;87: Chaney MA, Nikolov MP, Slogoff S. Methylrednisolone augments ulmonary dysfunction following cardioulmonary byass. Anesth Analg 1998;87: Fecht DC, Magovern GJ, Park SB, et al. Beneficial effects of methylrednisolone in atients on cardioulmonary byass. Circ Shock 1978;5: Fillinger MP, Watson RB, Sanders JH, Yeager MP. Pulse steroids: effects on cytokine resonse and recovery after cardiac surgery. Anesth Analg 1997;84: Howard RJ, Crain C, Franzini DA. Effects of cardioulmonary byass on ulmonary leukostasis and comlement activation. Arch Surg 1988;123: Messent M, Sinclair DG, Quinlan GJ, et al. Pulmonary vascular ermeability after cardioulmonary byass and its relationshi to oxidative stress. Crit Care Med 1997;25: Haeffner-Cavaillon N, Rousselier N, Ponzio O. Induction of interleukin-1 roduction in atients undergoing cardioulmonary byass. J Thorac Cardiovasc Surg 1989;98: Tonz M, Mihaljevic T, von Segesser LK, et al. Normothermia versus hyothermia during cardioulmonary byass: a randomized, controlled trial. Ann Thorac Surg 1995;59: Frank SM, Fleisher LA, Olson KF, et al. Multivariate determinants of early ostoerative oxygen consumtion in elderly atients. Anesthesiology 1995;83: Frost L, Molgaard H, Christiansen E, et al. Atrial fibrillation and flutter after coronary artery byass surgery: eidemiology, risk factors and reventive trials. Int J Cardiol 1992;36: Mathew JP, Parks R, Savino JS, et al. Atrial fibrillation following coronary artery byass graft surgery: redictors, outcomes, and resource utilization. JAMA 1996;276: Kawamura T. Cytokines during the erisurgical eriod. Masui-Jaanese J Anesth 1996;45: Cremer J, Martin M, Redl H, et al. Systemic inflammatory resonse syndrome after cardiac oerations. Ann Thorac Surg 1996;61: Rinder C, Mathew J, Davis E, et al. Immunocomromise in the ost-cpb atients: effect of age and sex. Anesthesiology 1995;83: Mayumi H, Zhang Q, Nakashima A, et al. Synergistic immunosuression caused by high-dose methylrednisolone and cardioulmonary byass. Ann Thorac Surg 1997;63: Zerr KJ, Furnary AP, Grunkemeir GL, et al. Glucose control lowers the risk of wound infection in diabetics after oen heart oerations. Ann Thorac Surg 1997;63: Wass CT, Scheithauer BW, Bronk JT, et al. The effect of corticosteroid-associated hyerglycemia and its treatment with insulin, on outcome following near-comlete forebrain ischemia in rats. Anesthesiology 1995;83: Bhandari A, Nikolav MP, Chaney MA. Effects of methylrednisolone on ostoerative hemodynamics in atients undergoing CABG and early extubation. Anesth Analg 1998; 86:114.