587166 research-article2015
TAK0010.1177/1753944715587166Therapeutic Advances in Cardiovascular DiseaseD. J. Perkowski et al.
Therapeutic Advances in Cardiovascular Disease
Ten year experience of using a novel metabolic protocol in ‘off pump’ coronary artery bypass revascularization
Original Research
Ther Adv Cardiovasc Dis 2015, Vol. 9(6) 336–341 DOI: 10.1177/ 1753944715587166 © The Author(s), 2015. Reprints and permissions: http://www.sagepub.co.uk/ journalsPermissions.nav
David J. Perkowski, Susan Wagner, Kathryn Muller, Joseph R. Schneider and J. A. St.Cyr
Abstract Purpose: Assessment of both short- and long-term outcomes in patients undergoing offpump coronary artery bypass using a perioperative metabolic protocol. Methods: A total of 975 of 995 adult patients underwent coronary artery bypass ‘off-pump’ from 1997 through 2006. Patients presenting in cardiogenic shock were excluded from this assessment. A perioperative metabolic protocol, which included the implementation of allopurinol, insulin supplementation, magnesium sulfate, supplemental corticosteroids, milrinone, norepinephrine (prn), aspirin, clopidogrel, statins and β-blockers, was used in these patients. Results: The mean age at the time of surgery was 70.5 years and the average number of bypass grafts was 4 per procedure; 18% (n = 176) of the cases had a preoperative intra-aortic balloon pump inserted for hemodynamic instability, tight left main coronary artery stenosis or angina. The 30-day mortality was 1.8% versus a Society of Thoracic Surgeons (STS) predicted mortality of 4.8%. Left main coronary artery disease was present in 38% (n = 371) of the patients. No strokes occurred intra-operatively and the postoperative incidence of stroke was 0.9% (n = 9). Incidence of renal failure requiring dialysis was 0.8% (n = 8). There was a single sternal infection. Mean follow up was 65 months with a survival rate of 90% (n = 955). Re-intervention, which commonly involved PTCA ± stent placement or re-do coronary artery bypass grafting (CABG), was 4% at 1 year and 11.6% (n = 113) during the 65-month follow-up period. Conclusions: Off-pump coronary artery bypass coupled with this novel metabolic protocol was associated with a low operative mortality and acceptable perioperative morbidities, including patients with left main coronary artery disease. These benefits are apparent at both short- and medium-term follow up.
Keywords: coronary artery bypass, metabolic, mortality, off-pump, revascularization
Introduction Surgical coronary artery revascularization is a treatment option for patients with coronary artery disease. Coronary artery bypass (CAB) has been the procedure of choice for patients presenting with left main coronary artery stenosis or individuals with multivessel coronary artery disease with or without ventricular dysfunction. Cardiopulmonary bypass, commonly used during conventional CAB, is associated with known complications such as cerebral vascular accidents, neurocognitive disorders, renal failure and a generalized systemic inflammatory response with potential negative
consequences [Creswell and Karis, 2009; Rastan et al. 2006]. During the past several years, ‘offpump’ CAB (OPCAB) has gained acceptance with the hope that the incidence of these known complications observed with conventional cardiopulmonary bypass might be reduced [Chaudhry et al. 2014; Selke et al. 2005]. The characteristics of patients undergoing CAB have changed over time with the current patient population presenting with more advanced atherosclerotic disease and having an increased number of comorbidities. In addition, many patients
Correspondence to: J. A. St.Cyr, MD, PhD Jacqmar, Inc., 10965 53rd Ave. No., Minneapolis, MN 55442, USA
[email protected] David J. Perkowski, MD Susan Wagner, RN Kathryn Muller, PhD Saddleback Memorial Medical Center, Laguna Hills, CA, USA Joseph R. Schneider, MD, PhD Northwestern Medicine West Region and Northwestern University Feinberg School of Medicine, Wheaton and Chicago, IL, USA
336 http://tac.sagepub.com
DJ Perkowski, S Wagner et al. with coronary artery disease have underlying moderate to severe insulin resistance [Bressler et al. 1996], which can present with such abnormalities as hypokalemia, hypomagnesemia, hyperlidpidemia, hypertension and generalized endothelial dysfunction. Because of these features, a perioperative metabolic management protocol was designed and implemented in this OPCAB patient group with the intention to optimize perioperative management and clinical outcomes. This paper reports on the clinical results using this metabolic protocol in patients undergoing OPCAB, in which all procedures were performed by one surgeon at one medical center.
an additional 24 hours. During the infusion, blood glucose levels were measured hourly with blood pH measured every 4 hours. If the pH dropped below 7.33, the insulin infusion was doubled. Magnesium sulfate (5 g/dose, Hospira, Inc., Lake Forest, IL) was given intravenously over 3 hours prior to surgery with additional doses, if needed, to achieve a serum magnesium level >2mEq/l. Magnesium blood levels were measured after each infusion and repeated as needed for the first 24 hours postoperatively. Supplemental potassium chloride intravenously (IV) (Hospira, Inc., Lake Forest, IL) was administered, if needed, every 2 hours to maintain a potassium level ⩾4mEq/l.
Material and methods A total of 975 from 995 adult candidates scheduled for coronary revascularization underwent OPCAB, performed by one surgeon at Saddleback Memorial Care Center in Laguna Hills, CA, USA from 1996 to 2006. In this 995 cohort, 20 patients were excluded from OPCAB due to presentation with shock or identification of a very high grade left main coronary artery lesion preoperatively that would have compromised manipulation of the heart during the surgical operation. This group included both first time and re-do operations, and all patients were risk stratified according to the Society of Thoracic Surgeons (STS) database. Exclusion criteria included overt cardiogenic shock (1%), as well as an inability to complete the OPCAB operation (1%).
The operating room was kept warm (20–22ºC) to prevent hypothermia. Anesthesia was induced with Propofol (Hospira, Inc., Lake Forest, IL) and pancuronium bromide (Hospira, Inc., Lake Forest, IL). After anesthetic induction, either 500 mg of methylprednisolone sodium succinate (SoluMedrol, Pzifer, Inc., NY) or 1 mg/kg of dexamethasone (Decadron, Wyeth Pharmaceuticals, Madison, NJ) was infused. Intravenously, milrinone, 0.5 µg/kg/min (Primacor, SanofiSynthelabo, NY) was initiated in the operating room. In addition, 2 million units of aprotinin were administered at the start of surgery for its antifibrinolytic effect and its potential to blunt an acute inflammatory response. Norepinephrine 1µg/min (Levophed Bitartrate, Hospira, Inc., Lake Forest, IL) was infused after harvesting the internal mammary artery. Full anticoagulation with unfractionated heparinization (300 IU/kg) was used in all of these surgical patients. For stabilization of the heart for each distal anastomosis, a Medtronic Stabilizer (Medtronic, Minneapolis, MN) was used for the initial 300 patients and a Guidant Expose (Guidant Corporation, San Jose, CA) was used for the last 675 patients. All proximal anastomoses were performed using a partial occluding vascular clamp. Protamine sulfate (1 mg/100 IU heparin IV; APP Pharmaceuticals, LLC, Schaumburg, IL) was given to reverse the heparin anticoagulation at the conclusion of the surgical procedure.
Because of the assumption that an increase in insulin resistance may be present in patients with coronary artery disease, all patients were treated as insulin-resistant. A perioperative metabolic protocol, including therapies directed at insulin resistance, was implemented in every patient. A summary of this protocol follows. Single or multiple doses of oral allopurinol (300 mg/dose, DSM Pharmaceuticals, Inc., Parsippany, NJ) were given, depending upon the timing of the operation, post heart catheterization to reduce the risk of a renal insult. More than a single dose was given depending upon the scheduled operative procedure. An intravenous insulin drip infusion (Eli Lilly and Company, Indianapolis, IN) was initiated in the preoperative period with the intent to achieve a targeted blood glucose level between 100 and 150 mg%. This infusion was continued throughout the surgical procedure and postoperatively was switched to subcutaneous injections for
Postoperatively, the majority (78%) of the patients were extubated within 6 hours following the operative procedure. The implemented metabolic protocol remained in place for 24 hours following surgery. Inotropes were weaned as soon as possible if the patient’s hemodynamic parameters
http://tac.sagepub.com 337
Therapeutic Advances in Cardiovascular Disease 9(6) were acceptable, i.e. keeping the cardiac index > 2 l/minm2. An oral statin was initiated postoperatively to aid in lipid metabolism and potentially subsequent graft patency. Furthermore, oral clopidogrel (75 mg/day) and aspirin (81 mg/ day) were instituted postoperatively to enhance bypass graft patency and to minimize the incidence of a postoperative stroke. Results There was an average of 4 bypass grafts placed/ procedure in all the 995 adult patients. The median age of these patients in this study was 70.5 years of age. Of the patients, 38% presented with left main coronary artery disease and 76% were diagnosed with triple vessel coronary artery disease. A total of 18% had an intra-aortic balloon pump (IABP) inserted preoperatively for hemodynamic instability, tight left main coronary artery stenosis, or persistent angina. Left main coronary artery disease was present in 38% (n = 371) of the patients. The overall surgical mortality was 1.8% (n = 21); 1.9% for men and 3.8% for women. In addition to six pulmonary related and five cardiac deaths, two patients died of a saddle embolus, prior to switching over to full dose heparin for deep vein thrombosis (DVT) prophylaxis. One patient, an 82 year old female, died from a severed spinal cord related to a cervical spicule. One elderly male developed heparin-induced thrombocytopenia. Two patients died from a stroke related to atrial fibrillation and two patients died from liver failure. Comorbidities included one sternal wound infection (0.1%), 8 cases of renal failure requiring temporary hemodialysis (0.8%) and 9 strokes (0.9%, 6 of the 9 strokes in patients with atrial fibrillation). We found that >80% of all our patients required an additional 5 g dose of magnesium sulfate immediately postoperatively, which was infused again over 3 hours. Mean follow up was 65 months with a survival rate of 90% (n = 955). Re-intervention, which commonly involved percutaneous transluminal coronary angioplasty (PTCA) ± stent placement or re-do CABG was 4% at 1 year and 11.6% (n = 133) during the 65 month follow-up period. There were 55 re-do operations with 4 mortalities but only one of these was cardiac related. Discussion Although previous publications have reported clinical results with OPCAB, this is the first known report describing the results of the implementation
of an aggressive metabolic protocol in patients undergoing OPCAB. In addition to the reported benefits of OPCAB, which include a shorter hospitalization and potentially a less myocardial insult with comparable 30 day outcomes compared with conventional cardiopulmonary bypass [Puskas et al. 2003], physiological attention to protecting the vascular endothelium during cardiopulmonary bypass is also important. Endothelial dysfunction results in the production of adhesion molecules such as inter-cellular adhesion molecules (ICAMs) and vascular cell adhesion molecules (VCAMs), which potentially can attract platelets and leucocytes. Factors that influence endothelial dysfunction can include insulin resistance, hypertension, hypokalemia, hypomagnesemia and hyperuricemia [Baron, 1996] and a state of hyperglycemia can promote the production of superoxide radicals, as well as suppressing immune function [Blondet and Beilman, 2007]. Many patients with coronary artery disease are insulin resistant, affecting the production and storage of glycogen. As a result, protein catabolism serves as a source of fuel, which has the potential to increase the production of uric acid [Objalo and Hagerty,1992]. The combination of iodinated contrast during angiography and the production of uric acid during surgery can potentially damage the proximal tubules of the kidneys. Therefore, patients who were deemed surgical candidates were given allopurinol 300 mg orally 1 hour postcatheterization. Allopurinol prevents the conversion of xanthine to uric acid, as well as acting as an antioxidant and free radical scavenger during times of ischemia. The use of corticosteroids perioperatively is important for their interaction with glucocorticoid genes. This stimulates an increased transcription of anti-inflammatory proteins, specifically, interleukin 10 and interleukin 1 receptor antagonists, which inhibit the interaction between glucocorticoid receptors and activated transcription factors such as NF-κB and activator protein-1, which regulate inflammatory gene expression [Boumpas, 1996]. Glucocorticoids prevent the uncoiling of DNA, the availability to access transcription factors, and gene expression. Corticosteroids can increase the production of endothelial cellular enzymes, which can degrade inflammatory peptides and affect adhesion molecule expression, for example, by directly inhibiting ICAM and E-selectin production [Aljada et al. 1999].
338 http://tac.sagepub.com
DJ Perkowski, S Wagner et al. During operative stress, the conversion of glucose to pyruvate remains intact. Without adequate insulin levels, these patients convert pyruvate to lactic acid, which produces a free hydrogen ion, resulting in a state of acidosis, decreasing oxygen delivery to the tissues, both of which can adversely affect cellular function. Supplemental insulin used in this protocol enhances the uptake of glucose by the tissues. Besides the uptake of glucose, insulin also provides additional physiological benefits, such as an increased production of nitric oxide by the endothelium, which results in vasodilation [Cannon, 1998]. A potential state of ketosis does not develop due to adequate levels of methyl malonyl-CoA in the liver; however, a hyperosmolar state can develop [McGarry et al. 1977]. Macrophage function is depressed at glucose levels ⩾250 mg%, affecting wound healing and vulnerability to infection. Other adverse effects of hyperglycemia include glycosylation of hemoglobin with a leftward shift of the oxygen dissociation curve [Brownlee et al. 1998] and sludging of red blood cells in the microcirculation, thereby affecting blood viscosity and decreasing oxygen delivery by red blood cells. With sustained hyperglycemia, glucose will auto-oxidize in the presence of transitional metals such as copper and iron producing a superoxide radical (O2-), reducing to H2O2 and subsequently to H2O through the glutathione pathway. Secondly, the conversion of l-arginine to nitric oxide is blunted. Glucose can be converted to sorbitol, a glucitol, which is metabolized slowly in humans. All three of these reactions require nicotinamide adenine dinucleotide phosphate (NADPH); however, the sorbitol pathway preferentially uses NADPH more so than the other two reactions, affecting normal cellular metabolism [Diaz-Flores et al. 2004]. Magnesium and potassium levels need close attention perioperatively. Magnesium stores are depleted due to excretion along with glucose in these patients with long-standing insulin resistance. Magnesium has many roles, including as an essential cofactor in many intracellular enzymatic reactions, stimulation of the vascular endothelium to produce prostacyclin (a profound vasodilator), stabilization of the cell membrane during ischemia as well as the prevention of swelling of the mitochondria, essential for the function of the sodium–potassium ATPase pump, and magnesium can act as a calcium channel blocker and aids in reducing ventricular dysrhythmias [Booth et al. 2003].
The decision to add both norepinephrine and milrinone in this protocol centered on their complementary and synergistic effects. Milrinone, a phosphodiesterase III inhibitor, is particularly effective for OPCAB because it increases both systolic and diastolic compliance of the heart, increases cardiac output and reduces pulmonary artery pressures [Young and Ward, 1988]. These hemodynamic benefits allow manipulation of the heart for OPCAB, while maintaining an adequate cardiac output. Mechanistically, milrinone induces the sarcoplasmic reticulum of the myocardial cell to sequester calcium thus reducing myocardial wall tension and peripherally creates a state of vasodilation, beneficial in reducing afterload [Benotti et al. 1985]. However, norepinephrine stimulates both alpha and beta-receptors in the myocardial cell, increases nitric oxide production, and aids in stabilizing the sodium–potassium ATPase pump during ischemia [Cohen et al. 1997]. During myocardial ischemia, an excess of hydrogen ions is produced, as well as an alteration in the sodium–potassium pump, allowing sodium to enter the cell. Norepinephrine protects the sodium–potassium pump by reducing the entry of sodium into the cell during ischemia and ultimately reducing the influx of calcium, another cellular benefit. With reperfusion, hydrogen ions are removed from the cell via the sodium–hydrogen pump and ultimately sodium is exchanged for calcium, lessening reperfusion injury [Murphy and Eisner, 2009]. In addition, animal studies have shown that norepinephrine reduces the ischemic zone by 40% in occluded coronary arteries [Cohen et al. 1997]. Although there were no intraoperative strokes with OPCAB, there were 4 postoperative strokes in our first 300 patients. Because of these initial complications, we elected to aggressively treat the inflammatory response associated with surgery. This inflammatory response includes the secretion of cytokines, activation of complement, and activation of leucocytes as well as platelets. In addition, the coagulation cascade is activated, which stimulates kallikrein [Piro et al. 2005]. This surgical setting produces an adherence of leucocytes to the endothelium with the potential for migration into the media layer of the vessel. Platelets are stimulated with enhanced neutrophil arrest and the potential for edema by a release of superoxide granules. Aprotinin, a serine protease, was selected in this study for its antifibrinolytic effect and its potential to blunt an inflammatory response [Royston, 1996]. Theoretically in the
http://tac.sagepub.com 339
Therapeutic Advances in Cardiovascular Disease 9(6) postoperative period, the coagulation cascade functions normally, hemodilution is not dominant and, therefore, patients tend to become hypercoagulable [Mariani et al. 1999]. Based on the earliest experiences, the protocol was revised to include the combination of aprotinin, clopidogrel and aspirin, which were combined to potentially eliminate the incidence of postoperative strokes [Apostolakis et al. 2012]. However, aprotinin is not currently used in our metabolic protocol because it was withdrawn from the US market in 2007 because of potential renal complications. We observed incidences of renal failure, stroke and a single case of a saddle embolism in this study. There is no substantial proof that aprotinin might have played the causative or related role in these complications. Many of our patients exhibited various risk factors, all of which can play a role in potentially sustaining a complication. Dyslipidemia is an independent short- and longterm risk factor in patients after CAB. Endothelial dysfunction due to circulating lipid containing particles can occur as early as 2 hours after a high fat meal [Bae et al. 2003] and, because of this, saphenous vein grafts can have the potential for graft occlusion. Statins have been shown to lower serum cholesterol and improve lipid profiles, including reducing the concentration of harmful lipids, inhibition of progression of atherosclerotic lesions, as well as suppression of inflammatory, immunoaggressive cells in grafted vessels. We have adopted a policy to prescribe statins postoperatively to potentially aid in graft patency and improve patients’ lipid profile. Many proponents of OPCAB have promoted the advantages over conventional coronary bypass procedures. However, most of these protocols have not fully addressed both the metabolic and physiological abnormalities that occur perioperatively in coronary revascularization. We have designed and utilize this novel metabolic protocol targeting these concerns with the intent to optimize perioperative management and clinical outcomes. Using this protocol, our overall surgical mortality was 1.8% versus a risk-adjusted predicted mortality of 4.8% (STS, www.stsdatabase). Similarly, the incidence of perioperative morbidities using this protocol was acceptable with a 90% crude survival rate at 65 months of follow up. We feel that the use of this novel metabolic protocol in OPCAB offers acute and extended benefits for this population, including in higher risk patients.
Conclusion Proponents of OPCAB have promoted the advantages over conventional coronary bypass procedures. However, there is still a need to fully address both the metabolic and physiological abnormalities that can be encountered perioperatively in coronary revascularization. We propose a novel metabolic protocol for OPCAB that has been associated with a low mortality and acceptable perioperative morbidities. This novel metabolic protocol in OPCAB offers both acute and extended benefits for this population, including in higher risk patients. Limitations This presented study involving patients undergoing ‘off pump’ coronary artery revascularization reported favorable findings according to parameters within the STS database benchmarks. The use of a metabolic directed protocol in the perioperative period appeared to offer added benefits for these patients, including a low mortality and acceptable morbidities. However, a major criticism in this report centers on the lack of a concurrent control group during the study period. A comparative, concurrent control group would have provided a more robust comparison to OPCAB without the metabolic intervention. Funding This research received no specific grant from any funding agency in the public, commercial, or notfor-profit sectors. Conflict of interest statement The authors declare no conflicts of interest in preparing this article.
References Aljada, A., Ghanim, H., Assian, E., Mohanty., P, Hamouda, W., Garg, R. et al. (1999) Increased I-kappa B expression and diminished nuclear NF-kappa B in human mononuclear cells following hydrocortisone injection. J Clin Endocrinol Metab 84: 3386–3389. Apostolakis, S., Marin, F. and Lip, G. (2012) Antiplatelet therapy in stroke prevention. Adv Cardiol 47: 141–154. Bae, J., Schwemmer, M., Lee, I., Lee, H., Park, K., Kim, K. et al. (2003) Postprandial hypertriglyceridemia-induced endothelial dysfunction
340 http://tac.sagepub.com
DJ Perkowski, S Wagner et al. in healthy subjects independent of lipid oxidation. Int J Cardiol 87: 259–267. Baron, A. (1996) Insulin and the vasculature–old actors, new roles. J Invest Med 44: 406–412. Benotti, J., Lesko, L., McCue, J. and Alpert, J. (1985) Pharmacokinetics and pharmodynamics of milrinone in chronic congestive heart failure. Am J Cardiol 56: 685–689. Blondet, J. and Beilman, G. (2007) Gylcemic control and prevention of perioperative infection. Curr Opin Crit Care 13: 421–427. Booth, J., Phillips-Bute, B., McCants, C., Podgoreanu, M., Smith, P., Mathews, J. et al. (2003) Low serum magnesium level predicts major adverse cardiac events after coronary artery bypass graft surgery. Am Heart J 145: 1108–1113. Boumpas, D. (1996) A novel action of glucocorticoidsNF-kB inhibition. Br J Rheumanol 35: 709–710. Bressler, P., Bailey, S., Matsuda, M. and DeFronzo, R. (1996) Insulin resistance and coronary artery disease. Diabetolagia 39: 1345–1350. Brownlee, M., Cerami, A. and Vlassara, H. (1998) Advanced glycosylation end products in tissue and the biochemical basis on diabetic complications. New Engl J Med 319: 1315–1321. Cannon, R. (1998) Role of nitric oxide in cardiovascular disease: focus on the endothelium. Clin Chem 44: 1809–1819. Chaudhry, U., Harling, L., Rao, C., Ashrafian, H., Ibrahim, M., Kokotsakis, J, et al. (2014) Off-pump versus on-pump coronary revascularization: metaanalysis of mid- and long-term outcomes. Ann Thorac Surg 98: 563–572. Cohen, M., Thornton, J., Thornton, C., Sato, H., Miki, T. and Downey, J. (1997) Intravenous co-infusion of adenosine and norepinepherine preconditions the heart without adverse hemodynamic effects. J Thorac Cardiovasc Surg 114: 236–242. Creswell, L. and Karis, N. (2009). Complications of cardiopulmonary bypass and cardioplegia. In: Little, A. and Merrill, W. (eds.), Complications in Cardiothoracic Surgery: Avoidence and Treatment. Oxford: Wiley-Blackwell. Diaz-Flores, M., Baiza-Gutman, L., IbanezHernandez, M., Pascoe-Lira, D., Guzman-Greenfel,
A. and Kumate-Rodriguez, J. (2004) Molecular aspects of chronic hyperglycemia-induced tissue damage. Gac Med Mex 140: 437–447. Mariani, M., Gu, Y., Boonstra, P., Grandjean, J., van Oeveren, W. and Ebels, T. (1999) Procoagulant activity after off-pump coronary operation: is the current anticoagulation adequate? Ann Thoracic Surg 67: 1370–1375. McGarry, J., Mannaerts, G. and Foster, K. (1997) A possible role for malonyl-CoA in the regulation of hepatic fatty acid oxidation and ketogenesis. J Clin Invest 60: 265–270. Murphy, E. and Eisner, D. (2009) Regulation of intracellular and mitochondrial sodium in health and disease. Circ Res 104: 292–303. Objalo, C. and Hagerty, D. (1992) Acute uric acid nephropathy following angiography and coronary bypass surgery. Nephron 62: 123–124. Piro, M., Giubilato, G., Pinelli, M., Giordano Sciacca, P. and Biasucci, L. (2005) Endothelium and inflammation. Pamminerva Med 47: 75–80. Puskas, J., Williams, W., Duke, P., Staples, J., Glas, K., Marshall, J. et al. (2003) Off-pump coronary artery bypass grafting provides complete revascularization with reduced myocardial injury, transfusion requirements, and length of stay: a prospective randomized comparison of two hundred unselected patients undergoing off-pump versus conventional coronary artery bypass grafting. J Thorac Cardiovasc Surg 125: 797–808. Rastan, A., Walther, T., Falk, V., Gummert, J., Eckenstein, J. and Mohr, F. (2006) Off-pump coronary artery bypass grafting. State of the art 2006 and results in comparison with conventional coronary artery bypass strategies. Herz 31: 384–395. Royston, D. (2006) Preventing the inflammatory response to open-heart surgery: the role of aprotinin and other protease inhibitors. Int J Cardiol 53(Suppl.): S11–S37. Selke, F., DiMaio, M., Caplan, L., Ferguson, T., Gardner, T., Hiratzka, L. et al. (2005) Comparing on-pump and off-pump coronary artery bypass grafting. Circulation 111: 2858–2864. Young, R. and Ward, A. (1988) Milrinone. A preliminary review of its pharmacological properties and therapeutic use. Drugs 36: 158–192.
Visit SAGE journals online http://tac.sagepub.com
SAGE journals
http://tac.sagepub.com 341
