|Year : 2017 | Volume
| Issue : 4 | Page : 210-216
Comparison of nitroglycerine and sodium nitroprusside on serum lactate, mixed venous oxygen saturation and mixed venous and arterial PCO2difference during cardiopulmonary bypass
Indira Gurajala, Padmaja Durga, R Gopinath
Department of Anaesthesiology and Critical Care, Nizams Institute of Medical Sciences, Hyderabad, Telangana, India
|Date of Web Publication||26-Dec-2017|
Dr. Indira Gurajala
Department of Anaesthesiology and Critical Care, Nizams Institute of Medical Sciences, Hyderabad, Telangana
Source of Support: None, Conflict of Interest: None
Background: The primary objective of the study was to evaluate the effect of nitoglycerine (NTG) and sodium nitroprusside (SNP) on serum lactate (S. lactate), mixed venous oxygen saturation (SvO2), and mixed venous arterial carbon dioxide difference (V-ACO2) during cardiopulmonary bypass (CPB). The secondary objectives included the effect on mortality, end organ dysfunction, requirement of vasopressors, duration of mechanical ventilation (MV), intensive care unit (ICU) stay and hospital stay.
Materials and Methods: A prospective randomized single blinded study was conducted in 40 patients aged between 20 years and 70 years who underwent cardiac surgery on CPB. The patients were randomly divided into Group N (n = 20) receiving NTG (0.5–2 mic/kg/min) and group S (n = 20) receiving SNP (0.5–2 mic/kg/min) from the commencement of total CPB up to complete rewarming (nasopharyngeal temperature >36.5°C). Arterial blood gases and S. lactate were measured at baseline, after institution of total bypass, after completion of cooling and rewarming, at weaning off CPB and admission to ICU. Venous blood gas (VBG) was sampled from the venous reservoir immediately after institution of total bypass and completion of rewarming. Urine output, dose of rescue vasodilator, use of inotropes and vasopressor after CPB, end organ dysfunction, duration of MV, ICU, and hospital stay were noted.
Results: Though the SvO2at the end of CPB decreased significantly from the baseline (P < 0.05), it was similar between the groups. There was no difference in V-ACO2too. The S. lactate markedly increased (P = 0.00) from the baseline; however, only the S. lactate at admission to ICU was significantly lower in Group S (P = 0.034). There was no difference in mortality, end organ dysfunction, requirement of vasopressors, duration of MV, ICU, and hospital stay.
Conclusion: The authors showed that S. lactate increased with CPB and this increase did not correlate with SvO2and V-ACO2. NTG and SNP were comparable in their effect on indices of tissue perfusion.
Keywords: Vasodilators, CPB, tissue oxygenation
|How to cite this article:|
Gurajala I, Durga P, Gopinath R. Comparison of nitroglycerine and sodium nitroprusside on serum lactate, mixed venous oxygen saturation and mixed venous and arterial PCO2difference during cardiopulmonary bypass. J NTR Univ Health Sci 2017;6:210-6
|How to cite this URL:|
Gurajala I, Durga P, Gopinath R. Comparison of nitroglycerine and sodium nitroprusside on serum lactate, mixed venous oxygen saturation and mixed venous and arterial PCO2difference during cardiopulmonary bypass. J NTR Univ Health Sci [serial online] 2017 [cited 2018 Jul 18];6:210-6. Available from: http://www.jdrntruhs.org/text.asp?2017/6/4/210/221529
| Introduction|| |
Moderate hypothermia is deliberately induced during cardiopulmonary bypass (CPB) as it provides organ protection against ischemia. The increase in systemic vascular resistance (SVR) secondary to CPB and hypothermia may result in uneven cooling and rewarming, and lead to mismatch in oxygen delivery and consumption. Aerobic metabolism is maintained by the tissues despite inadequate oxygen delivery by increasing the oxygen extraction and this is reflected in decreased mixed venous oxygen saturation (SvO2). When the oxygen delivery falls below a critical level, anaerobic metabolism ensues with lactate production. In the past, venous acidemia and carbon dioxide levels have been identified as markers of global perfusion. Increased difference in mixed venous and arterial carbon dioxide (V-ACO2) is reported in hypovolemic, cardiogenic, and septic shock.,, In fact, V-ACO2 is considered a better index of tissue perfusion and prognosticates morbidity more accurately even when SvO2 is within normal limits. Inadequate tissue perfusion on CPB is common and is one of the factors contributing to postpump multiorgan dysfunction along with systemic inflammatory response. Vasodilators such as nitroglycerine (NTG), sodium nitroprusside (SNP), isoflurane, and propofol are routinely used to improve microcirculation and facilitate uniform cooling and rewarming. Several studies evaluated the effect of these drugs on temperature drop after CPB but their effect on markers of tissue perfusion is unclear.,, The present study is a prospective randomized trial undertaken to compare the effect of NTG and SNP on indices of tissue perfusion during CPB. The primary objective was to evaluate the effect of NTG and SNP on serum lactate (S. lactate), SvO2, and V-ACO2 during CPB. The secondary objectives included the effect on mortality, end organ dysfunction, requirement of vasopressors, duration of mechanical ventilation (MV), intensive care unit (ICU) stay, and hospital stay.
| Material and Methods|| |
After approval from the institutional ethics committee and informed written consent, 40 adult patients scheduled for elective cardiac surgery under CPB were enrolled into two protocols. The patients included were adults (18–70 years) with good left ventricular function belonging to either gender. The patients presenting for emergency cardiac surgery, redo cardiac surgeries requiring femoro-femoral bypass, with preoperative renal or hepatic dysfunction or preoperative acute hemodynamic instability were excluded. The patients were divided into two groups, NTG group (Group N, n = 20) and SNP group (Group S, n = 20) based on computer generated randomization. Only the participants were blinded to the group to which they were allotted. Post-randomization exclusion of cases was done if there was hemodynamic instability requiring inotropic support or if significant bleeding requiring volume resuscitation occurred before institution of CPB or if hematocrit on CPB decreased below 21% necessitating the addition of blood to the CPB circuit.
Patients were premedicated with oral alprazolam 0.5 mg 2 h before surgery. In the operation room, after establishing intravenous (IV) access, fentanyl 0.5 mic/kg and midazolam 1 mg were given. Apart from American Society of Anaesthesiologists standard monitoring, invasive blood pressure through radial artery cannulation and central venous pressures (CVP) via triple lumen in the right internal jugular vein were monitored with the Datex-Ohmeda S/5® Anesthesia Monitor. Anesthesia was induced with IV midazolam (0.1 mg/kg), fentanyl (5 mic/kg), and vecuronium (0.1 mg/kg). After tracheal intubation, MV was adjusted to keep the end-tidal partial pressure of carbon dioxide between 30 mmHg and 35 mmHg. Anesthesia was maintained with intermittent doses of fentanyl (1 mic/kg), midazolam (30 mic/kg) and vecuronium (2 mg), and intermittent positive pressure ventilation with oxygen and air mixture (50%:50%) and end tidal sevoflurane of 1–2%. The ambient temperature was maintained between 22°C and 24°C. The core temperature of the patient was monitored using a thermistor (Datex Ohmeda) in the nasopharynx. Midsternotomy was used to open the chest in all patients. Systemic heparinization with 300 units/kg of heparin was followed by arterial cannulation in the ascending aorta. Depending on the surgery, venous cannulation was done either in the right atrium or separately in inferior vena cava (IVC) and superior vena cava (SVC). On initiation of complete bypass, ventilation was stopped and patients were cooled to a temperature of 30°C. The procedure was accomplished with aortic cross-clamp (ACC) and cardioplegic arrest. Following the removal of ACC, patients were rewarmed to a nasopharyngeal temperature of 36.5°C and weaned off CPB. The hemodynamic adjustment was done with inotropes, vasopressors, and IV fluids to achieve a post CPB heart rate of 70–100 beats/min, mean arterial pressure (MAP) of >70 mmHg and CVP of 8–12 mmHg.
Group N received NTG (0.5–2 mic/kg/min) from the start of complete bypass to rewarming up to a nasopharyngeal temperature of 36.5°C and group S received SNP (0.5–2 mic/kg/min) similarly. The precautions were taken to prevent biodegradation of SNP by covering the syringe and the infusion line with an opaque cover. A hematocrit of 21–27%, pump flows of 2.5 L/m2/min and MAP between 50 mmHG and 90 mmHg were maintained. MAP > than 90 mm Hg was treated with a bolus of 30 mg propofol and MAP <50 mmHg was treated with 50 mic bolus of phenylephrine (PE). The total dose of propofol and PE was recorded. Flows were increased to 3 L/m2/min or more during rewarming period to maintain the MAP.
Arterial blood gases and S. lactate were measured at baseline, immediately after institution of total bypass, completion of cooling, completion of rewarming, at weaning off CPB and admission to ICU. Venous blood gas (VBG) samples were collected immediately after institution of total bypass and completion of rewarming directly from the venous reservoir and SvO2 noted from the VBG. The difference in the partial pressure of carbon dioxide (pCO2) was calculated from the ABG and VBG which were sampled simultaneously at two time points - after institution of total bypass and completion of rewarming. Urine output, dose of propofol, use of phenylephrine on CPB, requirement of inotropes, duration of MV, ICU, and hospital stay were noted. Use of intra-aortic balloon pump (IABP), end organ dysfunction, and in-hospital mortality were also recorded. Inotropic requirement was categorized as low if dopamine ≤5 mic/kg/min and adrenaline ≤0.05 mic/kg/min was used. Use of dopamine and adrenaline in doses higher than those mentioned above or combination of three or more inotropes was classified as high inotropic requirement. Pulmonary dysfunction was diagnosed if hypoxemia [defined as partial pressure of arterial oxygen (PaO2)/Fraction of inspired oxygen (FiO2) <200] or new infiltrates on chest X-ray occurred postoperatively. Increase in serum creatinine by 1.5 times the base line value or an absolute increase of more than 0.3 mg% or a urine output of less than 0.5 mL/kg/h for more than 6 h postoperatively was considered as renal dysfunction. The presence of gross postoperative neurological deficits such as hemiparesis and postoperative sternal wound infections were actively looked for and recorded.
Twenty patients in each group were enrolled. Sample size calculation was performed using power and sample size software by the NCSS-LLC Inc., (Number Cruncher Statistical System) from the mean values of lactate measured on arrival in the ICU (Mean ± SD Group N – 8.93 ± 2.58 and group S – 6.95 ± 1.71). Seventeen patients were needed in each group to achieve 80% power with significance of 0.05 using two sided t-test. Twenty patients in each group were enrolled (to account for the loss of accrual from inability to complete the protocol). The statistical analysis was performed with SPSS 17.0 for Windows. Numerical data were expressed as mean, median, standard deviation and interquartile range, and categorical data were expressed as number and percentages. The intergroup comparison of numerical data was done with Mann–Whitney test and intragroup comparisons was done with Freidman test. The categorical data was compared using Chi-square test. P < 0.05 value was considered to be statistically significant.
| Results|| |
A total of 40 patients were enrolled in the study. Post inclusion, one patient in Group S and two patients in Group N were excluded. Of these, one patient required inotropes before CPB, one patient had bleeding necessitating volume resuscitation and one patient required addition of packed cells on pump due to hemodilution. Thirty seven patients (Group S-19 and Group N-18) were included in final analysis. Group S comprised of older patients and more females compared to Group N. The two groups were comparable in other demographic data such as type of cardiac surgery, durations of CPB, ACC, MV, ICU, and hospital stay [Table 1]. The MAP at the end of rewarming and after weaning off CPB were not different. There was no difference in the requirement of propofol between the groups. No patient in either group required PE to maintain MAP >50 mmHg. There was no difference in the requirement of inotropes or vasopressors during weaning from CPB between the groups. There was no in-hospital mortality in either groups. Two patients had surgical wound infection (one in each group) and one patient in Group S developed postoperative renal dysfunction.
|Table 1: The Demographic and Details of Cardiopulmonary Bypass Management of the Groups S and N|
Click here to view
The details of tissue perfusion indices are given in [Table 2] and [Figure 1]. Though the SvO2 at the end of CPB decreased significantly from the baseline (P< 0.05), it was similar between the groups. V-ACO2 was also not different between the groups at the two measurement points. The S. lactate markedly increased (P = 0.00) from the baseline with the institution of CPB in both groups and continued to increase till the discontinuation of bypass. This increase was more in Group N than Group S, but did not reach statistical significance. The S. lactate at admission to ICU was significantly lower in Group S than in Group N (P = 0.034).
|Figure 1: Mean S. Lactate levels in both groups at different measurement points. Lac: lactate, BL: Baseline, ON: at the start of CPB, 30: at 300 at the end of cooling, 37: at 370 at the end of rewarming, OF: after weaning off bypass, ICU: at admission into ICU.# Significantly increased from the baseline within the groups, * significantly different between the groups.|
Click here to view
| Discussion|| |
Hypothermia is a common practice in cardiac surgery as it is thought to provide organ protection against ischemia. Several surgeons prefer cooling the patient to core temperatures in the range of 28–30°C. Despite rewarming to 37°C before weaning off CPB, there is often a precipitous fall in core temperature which is termed “afterdrop.” The resultant hypothermia may have several deleterious effects such as shivering, increased SVR, and cardiac stress, bleeding, and may be associated with increased risk of postoperative wound infections.,,, Several strategies have been used to mitigate postCPB hypothermia. Use of vasodilators such as NTG, SNP, and phenoxybenzamine on CPB for this purpose is ubiquitous. Noback CR et al. were the earliest to report that SNP reduced the magnitude of afterdrop by as much as 1°C and subsequent authors confirmed these findings.,, Anesthetics with vasodilating properties like isoflurane are also used for this purpose and may have an added advantage of decreasing metabolism. Rajek et al. in their study compared use of forced warm air blankets with SNP and found that cutaneous rewarming was as effective as SNP in amelioration of afterdrop. In recent times, normothermic CPB has been suggested as an alternative temperature management strategy to avoid afterdrop and its side effects.
The increase in SVR due to CPB very likely aggravates tissue ischemia. Ischemia along with systemic inflammatory response is an important factor contributing to multiorgan dysfunction after cardiac surgery. Markers such as SvO2 and S. lactate are frequently used to monitor adequacy of perfusion. We hypothesized that systemic vasodilation improved microcirculation during CPB. Both SNP and NTG are extensively used for vasodilation on CPB and reduction of afterdrop. They are preferred over phenoxybenzamine due to their rapid onset and short duration of action. In addition, studies have found that SNP reduces the incidence of atrial fibrillation, improves cardiopulmonary and renal function by its anti-inflammatory action through release of nitric oxide NO.,,,, However, their effect on markers of perfusion has not been reported. This study was undertaken to compare the effect of NTG (which is the routine vasodilator used on CPB at our institution) with SNP on SvO2, S. lactate, and V-ACO2.
Most perfusionists target SvO2 of more than 70%. However, SvO2 on CPB does not accurately predict regional venous saturation (RvO2). McDaniel et al. in their experimental study on swine found that profound decrease in RvO2 and regional acidemia can occur even when SvO2 is within normal range. This was confirmed by Lindholm et al. in a prospective randomized trial of 30 patients undergoing cardiac surgery. They reported that hepatic and cerebral venous oxygen saturation on CPB are lower than SvO2 and the changes in SvO2 correlate with hepatic but not jugular venous saturation. In our study, SvO2 measured at initiation of total CPB and at completion of rewarming were within normal limits. SvO2 in both groups decreased from baseline but was well above the recommended level of 70%.
In a retrospective study of 1,376 patients undergoing cardiac surgery on CPB, Demers et al. identified a peak blood level of lactate >4 mmol/L during CPB with increased risk of short and long-term morbidity and mortality after cardiac surgery. Munoz et al. found a large increment in S. lactate with institution of CPB. The lactate levels decreased by the time the patient reached ICU. However, in non-survivors and patients with circulatory failure, high levels of S. lactate persisted. The S. lactate in our study increased substantially with initiation of CPB and continued till the termination of CPB. These findings were similar to the findings of Munoz et al., but unlike them, the increase in lactate persisted in the ICU too. The study protocol was planned such that the last measurement of lactate was at admission to ICU, hence the time taken for normalization of lactate was not documented. There was also no correlation between SvO2 and S. lactate levels. These findings are in agreement with that of a prospective observational study by Shabazi et al. in patients undergoing coronary artery bypass grafting which reported that SvO2 did not correlate with S. lactate levels and S. lactate can increase even in the presence of normal SvO2.
V-ACO2 gradient has been studied as a marker of global perfusion and increases in V-ACO2 are reported in cardiogenic, hypovolemic, and septic shock. In cardiac surgery, the utility of V-ACO2 gradient has not been fully elucidated. Ariza et al. measured plasma lactate levels and V-ACO2 gradients in 10 patients following CPB and compared them with conventional indices of tissue perfusion. They reported that blood lactate and CI increased progressively following CPB and an increase in lactate was associated with a decrease in V-ACO2. In contrast, Habicher et al., in their retrospective data analysis of 60 patients, found that in patients with normal SvO2 ≥70%, an increased V-ACO2(≥8 mmHg) was associated with increased postoperative lactate levels and decreased splanchnic function. These findings were associated with a longer need for MV and ICU stay. We found an insignificant increase in V-ACO2 at the end of rewarming when compared to the baseline in both the groups. Though minimal, the increase was more in Group N. However, we did not observe increased requirement of postoperative ventilation or ICU stay in patients with V-AO2 ≥8 mmHg. One procedural difference was that the above quoted study measured V-ACO2 on admission to ICU whereas we recorded values on CPB.
Limitations of Study
As it is the standard protocol to use a vasodilator on CPB in our institution, we did not include a control group. Inclusion of a control would have allowed us to study the effect of CPB on tissue markers without modification by vasodilation.
| Conclusion|| |
CPB by virtue of being highly invasive and unphysiological causes a transient state of global hypoperfusion. Markers such as S. lactate, SvO2, and V-ACO2 have been studied to monitor the adequacy of perfusion. The present study showed that S. lactate increases with institution of CPB and this increase does not correlate with SvO2 and V-ACO2. Both NTG and SNP which are used routinely during CPB to prevent afterdrop are comparable in their effect on these indices of tissue perfusion.
We are thankful to all the Surgeons and Perfusionists, Department of Cardiothoracic Surgery for their cooperation during the conduct of the study.
Financial support and sponsorship
This work was undertaken by the Department of Anesthesiology and Critical Care, Nizams Institute of Medical Sciences, Hyderabad, India and was funded by the Nizams Institute of Medical Sciences as a part of ongoing academic research.
Conflicts of interest
There are no conflicts of interest.
| References|| |
Pacini D, Pantaleo A, Di Marco L, Leone A, Barberio G, Murana G, et al
. Visceral organ protection in aortic arch surgery: Safety of moderate hypothermia. Eur J Cardiothorac Surg. 2014;46:438-43.
Vallée F, Vallet B, Mathe O, Parraguette J, Mari A, Silva S, et al
. Central venous-to-arterial carbon dioxide difference: An additional target for goal-directed therapy in septic shock? Intensive Care Med 2008;34:2218-25.
Adrogué HJ, Rashad MN, Gorin AB, Yacoub J, Madias NE. Assessing acid-base status in circulatory failure. Differences between arterial and central venous blood. N
Engl J Med 1989;320:1312-6.
Kazarian KK, Del Guercio L. The use of mixed venous blood gas determinations in traumatic shock. Ann Emerg Med 1980;9:179-82.
Ospina-Tascón GA, Bautista-Rincón DF, Umaña M, Tafur JD, Gutiérrez A, García AF, et al
. Persistently high venous-to-arterial carbon dioxide differences during early resuscitation are associated with poor outcomes in septic shock. Critical Care 2013;17:R294.
Noback CR, Tinker JH. Hypothermia after cardiopulmonary bypass in man: Amelioration by nitroprusside-induced vasodilation during rewarming. Anesthesiology 1980;53:277-80.
Tuǧrul M, Pembeci K, Camci E, Ozkan T, Telci L. Comparison of the effects of sodium nitroprusside and isoflurane during rewarming on cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1997;11:712-7.
Deakin CD, Petley GW, Smith D. Pharmacological vasodilatation improves efficiency of rewarming from hypothermic cardiopulmonary bypass. Br J Anaesth 1998;81:147-51.
Rajagopalan S, Mascha E, Na J, Sessler DI. The effects of mild perioperative hypothermia on blood loss and transfusion requirement. Anesthesiology 2008;108:71-7.
Seamon MJ, Wobb J, Gaughan JP, Kulp H, Kamel I, Dempsey DT. The effects of intraoperative hypothermia on surgical site infection: An analysis of 524 trauma laparotomies. Ann Surg 2012;255:789-95.
Eberhart LH, Doderlein F, Eisenhardt G, Kranke P, Sessler DI, Torossian A, et al
. Independent risk factors for postoperative shivering. Anesth Analg 2005;101:1849-57.
Frank SM, Higgins MS, Breslow MJ, Fleisher LA, Gorman RB, Sitzmann JV, et al
. The catecholamine, cortisol, and hemodynamic responses to mild perioperative hypothermia. A randomized clinical trial. Anesthesiology 1995;82:83-93.
Ambesh SP, Chattopadhyaya M, Saxena PV, Mahant TS, Ganjoo AK. Combined use of isoflurane and sodium nitroprusside during active rewarming on cardiopulmonary bypass: A prospective, comparative study. J Postgrad Med 2000;46:253-7.
] [Full text]
Rajek A, Lenhardt R, Sessler DI, Brunner G, Haisjackl M, Kastner J, et al
. Efficacy of two methods for reducing postbypass afterdrop. Anesthesiology 2000;92:447-56.
Cavolli R, Kaya K, Aslan A, Emiroglu O, Erturk S, Korkmaz O, et al
. Does sodium nitroprusside decrease the incidence of atrial fibrillation after myocardial revascularisation?: A pilot study. Circulation 2008;118:476-81.
Freyholdt T, Massoudy P, Zahler S, Henze R, Barankay A, Becker BF, et al
. Beneficial effect of sodium nitroprusside after coronary artery bypass surgery: Pump function correlates inversely with cardiac release of proinflammatory cytokines. J Cardiovasc Pharmacol 2003;42:372-8.
Cakira O, Oruca A, Erena S, Buyukbayramb H, Erdincc L, Erena N. Does sodium nitroprusside reduce lung injury under cardiopulmonary bypass? Eur J Cardiothorac Surg 2003;23:1040-5.
Kaya K, Oguz M, Akar AR, Durdu S, Aslan A, Erturk S, Taşöz R¸ et al
. The effect of sodium nitroprusside infusion on renal function during reperfusion period in patients undergoing coronary artery bypass grafting: A prospective randomized clinical trial. Eur J Cardiothorac Surg. 2007;31:290-7.
Mora-Mangano C, Chow JL, Kanevsky M. Cardiopulmonary bypass and the anaesthesiologist. In: Kaplan JA, editor. Essentials of Cardiac Anesthesia. 1st
ed. Philadelphia: Elsevier Saunders; 2008. p. 513-45
McDaniel LB, Zwischenberger JB, Vertrees RA, Nutt L, Uchida T, Nguyen T, et al
. Mixed venous oxygen saturation during cardiopulmonary bypass poorly predicts regional venous saturation. Anesth Analg 1995;80:466-72.
Lindholm L, Hansdottir V, Lundqvist M, Jeppsson A. The relationship between mixed venous and regional venous oxygen saturation during cardiopulmonary bypass. Perfusion 2002;17:133-9.
Demers P, Elkouri S, Martineau R, Couturier A, Cartier R. Outcome with high blood lactate levels during cardiopulmonary bypass in adult cardiac operation. J Thorac Cardiovasc Surg 2000;119:155-62.
Munoz R, Laussen PC, Palacio G, Zienko L, Piercey G, Wessel DL. Changes in whole blood lactate levels during cardiopulmonary bypass for surgery for congenital cardiac disease: An early indicator of morbidity and mortality. J Thorac Cardiovasc Surg 2000;119:155-62.
Shahbazi S, Khademi S, Shafa M, Joybar R, Hadibarhaghtalab M, Sahmeddini MA. Serum lactate is not correlated with mixed or central venous oxygen saturation for detecting tissue hypo perfusion during coronary artery bypass graft surgery: A prospective observational study. Int Cardiovasc Res J 2013;7:130-4.
Ariza M, Gothard JW, Macnaughton P, Hooper J, Morgan CJ, Evans TW. Blood lactate and mixed venous-arterial PCO2 gradient as indices of poor peripheral perfusion following cardiopulmonary bypass surgery. Intensive Care Med 1991;17:320-4.
Habicher M, von Heymann C, Spies CD, Wernecke KD, Sander M. Central venous-arterial pCO2 difference identifies microcirculatory hypoperfusion in cardiac surgical patients with normal central venous oxygen saturation: A retrospective analysis. J Cardiothorac Vasc Anesth 2015;29:646-55.
[Table 1], [Table 2]