|Year : 2016 | Volume
| Issue : 2 | Page : 201-206
Deliberate hypotensive anesthesia during maxillofacial surgery: A comparative study between dexmedetomidine and sodium nitroprusside
Rehab S El-Kalla MD , Mona B El Mourad
Anesthesia and Surgical Intensive Care Department, Faculty of Medicine, Tanta University, Tanta, Egypt
|Date of Submission||16-Jan-2015|
|Date of Acceptance||05-Aug-2015|
|Date of Web Publication||11-May-2016|
Rehab S El-Kalla
Anesthesia and Surgical Intensive Care Department, Faculty of Medicine, Tanta University, El Gheish St., Tanta 31257, Gharbia
Source of Support: None, Conflict of Interest: None
The aim of the present study was to evaluate the anesthetic properties such as the onset of required hypotension, the quality of the surgical field, the percentage of inhaled anesthesia, intraoperative consumption of fentanyl, time to recovery, and adverse events of dexmedetomidine (DEX) as a hypotensive agent in comparison with sodium nitroprusside (SNP) in maxillofacial surgery.
Patients and methods
A total of 70 ASA I or II patients, aged 20-60 years, scheduled for maxillofacial surgery were randomly assigned to receive either DEX 1 mg/kg before induction of anesthesia for over 10 min followed by 0.2-0.5 mg/kg/h infusion during maintenance (DEX group; n = 35 patients) or SNP 0.25 mg/kg/min infusion after induction of anesthesia (SNP group; n = 35 patients) to maintain mean arterial blood pressure (MAP) between 50 and 65 mmHg. Hemodynamic variables and end-tidal isoflurane concentration were recorded, quality of surgical field was assessed by the same surgeon who was blinded of the selected hypotensive agent, and intraoperative fentanyl consumption and recovery time were recorded.
Heart rate was significantly lower (P < 0.05) in DEX group. Both drugs were effective in achieving the desired level of hypotension (MAP: 50-65 mmHg). However, the time needed to reach the target MAP was significantly shorter in the SNP group than in the DEX group. End-tidal isoflurane concentration and intraoperative fentanyl consumption were significantly lower in the DEX group than in the SNP group, and the quality of the surgical field was better in the DEX group compared with the SNP group, but the time for recovery was significantly longer in the DEX group than in the SNP group.
We concluded that DEX is an effective and safe agent with anesthetic benefits for controlled hypotension, and that compared with SNP DEX offers the advantage of better quality of the surgical field and decreased anesthetic and fentanyl requirements intraoperatively. However, DEX was associated with significantly longer time to reach the target MAP and delayed recovery from anesthesia compared with SNP.
Keywords: dexmedetomidine, hypotensive anesthesia, maxillofacial surgery, sodium nitroprusside
|How to cite this article:|
El-Kalla RS, El Mourad MB. Deliberate hypotensive anesthesia during maxillofacial surgery: A comparative study between dexmedetomidine and sodium nitroprusside. Ain-Shams J Anaesthesiol 2016;9:201-6
|How to cite this URL:|
El-Kalla RS, El Mourad MB. Deliberate hypotensive anesthesia during maxillofacial surgery: A comparative study between dexmedetomidine and sodium nitroprusside. Ain-Shams J Anaesthesiol [serial online] 2016 [cited 2021 May 9];9:201-6. Available from: http://www.asja.eg.net/text.asp?2016/9/2/201/182226
| Introduction|| |
Maxillofacial surgeries are wide range of procedures on head and neck, combining different specialties with extensive intraoperative blood loss. These procedures need good cooperation of all members of the team to reduce tissue bleeding and edema. Therefore, controlled hypotension during general anesthesia is used to decrease blood loss, and to clear and dry the surgical field .
Controlled hypotension is defined as a reduction of the systolic blood pressure to 80-90 mmHg, a reduction of mean arterial blood pressure (MAP) to 50-65 mmHg, or a 30% reduction of baseline MAP .
Various agents [e.g. magnesium sulfate , sodium nitroprusside (SNP), nicardipine, nitroglycerin, esmolol, a2 -agonists , labetalol, and high doses of potent inhaled anesthetics] have been used to achieve controlled hypotension . Ideal hypotensive agents are characterized by their rapid onset, ease of administration, dose dependency, rapid elimination, and nontoxicity .
SNP is used for induction and maintenance of controlled hypotension. It is a vasodilator with a rapid onset. After contact with hemoglobin, the SNP molecule is immediately degradated, and blood pressure rapidly returns to baseline values after discontinuation of SNP infusion .
Dexmedetomidine (DEX) is a potent a2 -adrenoceptor agonist with sedative, analgesic, and sympatholytic properties . The reduction in heart rate (HR) and MAP is due to decreased norepinephrine release . Its use is also characterized by a decrease in opioid and anesthetic requirements .
In this prospective randomized controlled trial our aim was to evaluate the anesthetic properties such as the onset of required hypotension, the quality of the surgical field, the percentage of inhaled anesthesia, consumption of fentanyl intraoperatively, time to recovery, and adverse events of DEX as a hypotensive agent in comparison with SNP in maxillofacial surgery.
| Patients and methods|| |
The Institutional Review Board's approval (code number: 187/06/10) was obtained for the present study. After taking their written informed consent, 70 adult ASA I or II patients, 20-60 years of age and of both sexes, scheduled for elective maxillofacial surgery were included in the study. The study was carried out at Tanta University Hospital (Tanta, Egypt) from January 2011 to July 2012. Exclusion criteria for this study included patients with clinically significant neurologic, cardiovascular, renal, or hepatic diseases; with uncontrolled hypertension; patients with pre-existing coagulation defect or on anticoagulant therapy; patients who were pregnant or breast feeding; and patients allergic to a2 adrenergic agonists or SNP.
Patients were allocated into two groups: group I (SNP group), comprising 35 patients; and group II (DEX group), comprising the remaining 35 patients. Randomization was done using sealed opaque envelopes containing the group of the assignment in a ratio of 1 : 1. A blinded nurse, not participating in the study or data collection, read the numbers in the envelopes to allocate groups to the patients.
SNP group received intravenous SNP at infusion rate of 0.25 mg/kg/min, which was adjusted to maintain MAP of 50-65 mmHg.
DEX group received intravenous DEX by a loading dose of 1 mg/kg over 10 min followed by a continuous infusion of 0.2-0.5 mg/kg/h, which was adjusted to maintain MAP of 50-65 mmHg.
Preoperatively, patients were subjected to careful history taking and clinical examination; investigations included complete blood picture, prothrombin time and activity, liver and renal function tests, and ECG. All patients were premedicated with intravenous midazolam (Dormicum; Roche, Hoffmann, La Roche Ltd) 0.02 mg/kg 15 min before induction of anesthesia.
On arrival at the operating room, a 22-G arterial catheter was inserted in the radial artery of the nondominant hand under complete aseptic technique and local lidocaine 1% anesthesia administration at the area of insertion for continuous measurement of blood pressure and HR and for blood gas sampling. Standard monitors (Infinity Kappa; Draeger Medical System, USA) were applied and baseline values (values taken after premedication) of invasive MAP, HR, end-tidal CO 2 , and peripheral oxygen saturation (SPO 2 ) were recorded.
Before the induction of anesthesia, patients assigned to the DEX group received a loading dose of 1 mg/kg DEX diluted in 100 ml 0.9% saline infused over 10 min, followed by continuous intravenous infusion of 0.2-0.5 mg/kg/h.
All patients were preoxygenated for 3-5 min before the induction of anesthesia, using 1 mg/kg fentanyl (Janssen Pharmaceuticals), propofol (Fresenius) 2 mg/kg, and rocuronium (Esmeron, Organon) 0.6 mg/kg to facilitate armored endotracheal intubation.
Patients assigned to SNP group received an intraoperative continuous infusion of SNP immediately after induction of anesthesia at a rate of 0.25 mg/kg/min. The infusion rate of both studied drugs was gradually increased to gain target MAP between 50-65 mmHg before skin incision. The maximum allowed infusion rate was 0.5 mg/kg/h for the DEX group and 2.5 mg/kg/min for the SNP group. Patients who failed to gain the target MAP, even after receiving maximum dose, were excluded from the study and other therapy was given. Anesthesia was maintained with isoflurane (AIT) in 100% oxygen. Adequate muscle relaxation was maintained with incremental doses of rocuronium 0.3 mg/kg every 30 min. Tidal volume and respiratory rate were adjusted to maintain end-tidal CO 2 between 32 and 35 mmHg.
Signs of inadequate anesthesia [e.g. increases in blood pressure greater than the targeted MAP (³65 mmHg), tearing, or sweating] were treated with additional boluses of 1 mg/kg of fentanyl.
During the course of induced hypotension, reflex tachycardia, defined as constant increase in HR exceeding 110 beats/min for more than 10 min, was treated with a bolus dose of propranolol (Mayestrotense; Alex, Alexandria Company for Pharmaceuticals and Chemical Industries) 0.2 mg. If the MAP decreased below 50 mmHg, gradual decrease of the studied drug was initiated together with a decrease in inspired isoflurane concentration and 200 ml fluid bolus; if no response to this measures, phenylephrine 100 mg/kg intravenous boluses was given progressed to stopping the study drug if needed and patients were excluded from the study.
Bradycardia (HR <45 beats/min) was treated with 0.2 mg/kg anticholinergic intravenously, and in case of recurrence the dose was repeated; if bradycardia persisted, epinephrine in a dose of 0.01 mg/kg (1/10 000) was given and patients were excluded from the study.
(1) HR, invasive blood pressure, and peripheral oxygen saturation were recorded, initially at 5-min intervals from the induction of anesthesia till the reduction of MAP 30% below baseline value (value taken after premedication), and then every 15 min till the end of surgery.
(2) End-tidal isoflurane concentration was recorded every 15 min till the end of surgery.
(3) Arterial blood gas samples were drawn before surgery to determine baseline values of pH, HCO 3 , and base excess, then every 30 min till the end of the surgery, and finally the results were recorded.
(4) The quality of surgical field was estimated by the surgeon every 15 min during surgery using the scale adapted from that of Fromme et al. , on which a reading of 0 represented 'no bleeding' (bloodless field), 1 represented 'mild bleeding' (not even a surgical nuisance), 2 represented 'moderate bleeding' (a nuisance but without interference with accurate dissection), 3 represented 'moderate bleeding' (moderately compromised surgical dissection), 4 represented 'heavy bleeding' (controllable, and which significantly interfered with dissection), and 5 represented 'massive bleeding' (uncontrollable bleeding). The surgeon was blinded to the studied drugs used.
Hypotensive infusions in both groups were stopped 10 min before the end of surgery to allow rise in blood pressure for effective hemostasis. After skin closure, inhalational anesthetics were closed and the patients were kept to inhale 100% oxygen. Residual nondepolarizing muscle relaxant was reversed with neostigmine 0.05 mg/kg and atropine 0.02 mg/kg given intravenously, after which the patients were extubated and transferred to the recovery room. After a modified Aldrete score  higher than 9, patients were discharged from the recovery room to the wards.
A power analysis indicated that 35 patients were required per group. The difference in quality of the surgical field and the percentage of inhaled anesthesia was not less than 30% (50-80%) with 95% confidence limit, and with 80% power of the study and group to group ratio of 1 : 1. The collected data were analyzed using SPSS software (version 16; SPSS Inc., Chicago, Illinois, USA). For quantitative data, comparison between two groups and more was performed using c2 -test. For comparison between means of two groups of parametric data, Student's t-test was used. Intergroup comparisons were performed using repeated measures of t-test.
| Results|| |
A total of 80 patients were enrolled in the study; 10 were excluded from statistical analysis because they met one or more of the exclusion criteria. In both groups the age range of patients was 20-48 years [Table 1]. The study included 44 (62.8%) men and 26 (37.1%) women; 61 (87.1%) patients suffered from ASA I and the remaining nine (12.8%) patients suffered from ASA II. In comparison between both groups, there were no statistically significant differences as regards age, weight, sex, ASA physical status, and duration of surgery (P > 0.05). Fracture mandible was the most commonly performed surgery with 25 (35.7%) patients, fracture maxilla with 12 (17.1%) patients, fracture zygoma with 10 (14.3%) patients, rhinoplasty with seven (10%) patients, condylectomy with eight (11.4%) patients, and oronasal fistulectomy and maxillary cyst, the least performed surgeries, with four (5.7%) patients each.
Mean HR was significantly lower in the DEX group compared with the SNP group from the induction of drug till 15 min after its discontinuation (P < 0.05) [Figure 1]. MAP was not significantly different (P > 0.05) in both groups at all times of measurements; the two groups reached the desired MAP (50-65 mmHg) with no intergroup significant difference [Figure 2]. End-tidal isoflurane concentration (%) was significantly different between the two groups (P < 0.05) 15 min after induction till the end of the surgery; the DEX group was lower ranging from 0.4 ± 0.1 to 0.9 ± 0.2 than the SNP group ranging from 0.9 ± 0.1 to 1.4 ± 0.2 [Figure 3].
|Figure 1: Heart rate was statistically lower in the dexmedetomidine (DEX) group (P < 0.05) at all times of measurements except baseline. SNP, sodium nitroprusside|
Click here to view
|Figure 2: Mean arterial pressure was no statistically different (P > 0.05) in both groups at all times of measurements. DEX, dexmedetomidine; SNP, sodium nitroprusside|
Click here to view
|Figure 3: End-tidal concentration of isoflurane percentage was statistically lower in the dexmedetomidine (DEX group) (P < 0.05). SNP, sodium nitroprusside|
Click here to view
Arterial blood gases analyses in both groups as regards pH, HCO 3 , and base excess showed no statistically significant difference (P > 0.05) at all times of measurements. Quality of the surgical field, according to Fromme et al. , was better in the DEX group (score range: 0-2) compared with the SNP group (score range: 1-4) at all times of assessment [Figure 4].
|Figure 4: Quality of surgical field was statistically better in the dexmedetomidine (DEX) group (P < 0.05). SNP, sodium nitroprusside|
Click here to view
Intraoperative data, the time to achieve the target hypotensive level (MAP: 50-65 mmHg), was significantly longer in the DEX group (mean: 12.5 ± 2 min) compared with the SNP group (mean: 7.5 ± 1.6 min) [Table 2]. Only 11.4% (4/35) of patients in the DEX group and all (100%, 35/35) patients in the SNP group received intraoperative fentanyl supplementation. Fentanyl requirements were significantly lower in the DEX group (4 ± 12.7 mg) compared with the SNP group (72 ± 16.7 mg). Propranolol administration was necessary in 23 (65.7%) of patients in the SNP group to control HR (propranolol mean value: 0.3 ± 0.3 mg); in comparison, no (0%) patient in the DEX group received propranolol therapy. Time for recovery, which was defined as the duration of time from administration of reversal of neuromuscular blockade to sustained eye opening (for >5 s) on command, was significantly longer in the DEX group (22.9 ± 3.1 min) compared with the SNP group (10.4 ± 1.5 min). There were no episodes of hypotension or bradycardia requiring treatment, and no anticholinergic drug or phenylephrine was given. No patient in our study required an additional hypotensive agent to achieve controlled hypotension.
| Discussion|| |
The main purposes of hypotensive anesthesia during maxillofacial surgery are to decrease blood loss, improve quality of surgical field, and decrease duration of surgery, all related to improving visualization . In the present study comprising patients undergoing elective maxillofacial surgeries, an intraoperative initial loading dose of DEX of 1 mg/kg followed by infusion at 0.2-0.5 mg/kg/h decreased HR, anesthetic and fentanyl requirement intraoperatively, and also achieved better quality of surgical field in comparison with the group that received SNP intraoperatively.
Dexmedetomidine is a highly selective a2-adrenoceptor agonist and is 8-10 times more selective toward a2-adrenoceptor than clonidine. In the heart, the dominant action of a2-adrenoceptor agonists is a decrease in tachycardia (through blocking cardioaccelerator nerve) and bradycardia (through a vagomimetic action). In the peripheral vascular system, there is sympatholysis-mediated vasodilatation and smooth muscle cells receptor-mediated vasoconstriction .
Hemodynamic data of our study were consistent with data by other researchers. Gao et al. , who studied the effect of DEX as an adjuvant of sevoflurane for controlled hypotension in endoscopic sinus surgery, found that the HR values were lower in the DEX group compared with the control group, and they concluded that DEX is safe and effective in assisting controlled hypotension during endoscopic sinus surgery.
On the basis of clinical practice, SNP was administered at infusion rate of 0.25 mg/kg/min to the control group. SNP is a direct-acting, nonselective peripheral vasodilator. Peripheral vasodilation results in baroreceptor-mediated sympathetic responses with tachycardia and increased myocardial contractility .
In the SNP group there was significant increase in HR during constant hypotension compared both with baseline and the DEX group. Propranolol administration was necessary in 13 (65%) patients in the SNP group to control increased HR; this is consistent with data by other researchers. Yosry and Othman  compared both MgSO 4 and SNP for induction of controlled hypotension in choroidal melanoma resection surgery and revealed that MgSO 4 was associated with significant decrease of HR from baseline compared with SNP, which was associated with significant increase of HR. Degoute et al.  studied the effect of remifentanil in comparison with nitroprusside or esmolol for controlled hypotension during tympanoplasty, and found significant increase in the HR with SNP in comparison with remifentanil or esmolol.
The isoflurane concentration was titrated to maintain adequate depth of anesthesia in both groups; isoflurane concentration was less in the DEX group than in the SPN group, which could be explained by the anesthetic sparing effects of DEX, which decrease the intraoperative consumption of volatile anesthetic .
Rescue fentanyl were significantly lower in the DEX group compared with the SNP group. The reduction in fentanyl consumption could be explained by the opiate sparing properties of DEX, which help in decreasing opioid requirements by up to 90% .
It was evident that patients receiving DEX had a better surgical field as compared with patients receiving SNP. These findings can be attributed to the fact that DEX reduces sympathetic activity, resulting in lower blood pressure and reduced HR, thereby decreasing blood loss at the surgical site and improving the quality of surgical field . Poor quality of surgical field was found with nitroprusside. As a result of the direct effect of nitroprusside on vascular smooth muscle, reflex tachycardia, and increased cardiac output, the flow through the tissue would be increased, the resultant thus increasing the flow to and through the tissue . Our findings were in agreement with the study conducted by Nasreen et al. , who used DEX to provide hypotensive anesthesia during middle ear surgery, and observed that patients receiving DEX had a better surgical field compared with patients receiving placebo. Our results were also supported by Guven et al. , who reported that intraoperative bleeding was less in DEX group compared with placebo in functional endoscopic sinus surgery. Our results were in agreement with other researchers who reported that mild nitroprusside-induced hypotension produced fewer optimal surgical conditions .
In both groups, oxygen saturation remained 97% or greater in all patients, with an inspired 100% O 2 indicating no clinically significant intrapulmonary shunting.
pH, HCO 3 , and base excess were within normal ranges, and there was no significant difference in both groups, indicating trivial anaerobic metabolism or tissue hypoxia because of hypoperfusion caused by hypotensive anesthesia. Our findings were in agreement with other researchers .
The time for recovery was significantly longer in the DEX group compared with the SNP group. Similar to our results, Bulow et al.  observed that there was a prolonged recovery time in patients receiving DEX when compared with those receiving remifentanil during gynecologic laparoscopic surgery. Recovery in patients receiving DEX was prolonged most probably because of its sedative effects through central actions in the locus coeruleus .
This study had several limitations including lack of a control group, which was because of the certain benefit of controlled hypotension. The anesthetist team was not blinded to the study drug because the dosing requirements differed. We assessed the depth of anesthesia subjectively/clinically as we lacked a bispectral index monitor. In addition, the quality of the surgical field was subjectively estimated by surgeons, which may predispose to error in collecting data.
| Conclusion|| |
We concluded that DEX is an effective and safe drug with anesthetic benefits for controlled hypotension. Compared with SNP, DEX offers the advantage of better quality of the surgical field and decreased anesthetic and fentanyl requirements intraoperatively. However, DEX was associated with significantly longer time to reach the target MAP and delayed recovery from anesthesia compared with SNP.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Farah GJ, de Moraes M, Filho LI, Pavan AJ, Camarini ET, Previdelli IT, Coelho L. Induced hypotension in orthognathic surgery: a comparative study of 2 pharmacological protocols. J Oral Maxillofac Surg 2008; 66:2261-2269.
Degoute CS. Controlled hypotension: a guide to drug choice. Drugs 2007; 67:1053-1076.
Elsharnouby NM, Elsharnouby MM. Magnesium sulphate as a technique of hypotensive anaesthesia. Br J Anaesth 2006; 96:727-731.
Richa F, Yazigi A, Sleilaty G, Yazbeck P. Comparison between dexmedetomidine and remifentanil for controlled hypotension during tympanoplasty. Eur J Anaesthesiol 2008; 25:369-374.
Paris A, Tonner PH. Dexmedetomidine in anesthesia. Curr Opin Anaesthesiol 2005; 18:412-418.
Hessen M, Dietrich GV, Boldt J, Zeiler D, Alberts CH, Hempelmann G. Beta 2-adrenoceptor density of human lymphocytes after nitroprusside-induced hypotension. Anesth Analg 1995; 81:1250-1254.
Guler G, Akin A, Tosun Z, Eskitascoglu E, Mizrak A, Boyaci A. Single-dose dexmedetomidine attenuates airway and circulatory reflexes during extubation. Acta Anaesthesiol Scand 2005; 49:1088-1091.
Tufanogullari B, White PF, Peixoto MP, Kianpour D, Lacour T, Griffin J, et al
. Dexmedetomidine infusion during laparoscopic bariatric surgery: the effect on recovery outcome variables. Anesth Analg 2008; 106: 1741-1748.
Fromme GA, MacKenzie RA, Gould AB Jr, Lund BA, Offord KP. Controlled hypotension for orthognathic surgery. Anesth Analg 1986; 65:683-686.
Aldrete JA. The post-anesthesia recovery score revisited. J Clin Anesth 1995; 7:89-91.
Posnick JC. Anesthesia techniques, blood loss/fluid replacement, airway management & convalescence in the treatment of dentofacial deformities. In: Posnick JC, editor. Orthoenathic surgery: principles and practice
. Elsevier Health Sciences, ISBN 0323226892; 2014; 308-336.
Jones CR. Perioperative uses of dexmedetomidine. Int Anesthesiol Clin 2013; 51:81-96.
Gao GJ, Xu YY, Wang B, Yang WY, Shang Y. Feasibility of dexmedetomidine assisting sevoflurane for controlled hypotension in endoscopic sinus surgery. Med J Chin PLA 2012; 37:45-48.
Amr YM, Amin SM. Effects of preoperative beta-blocker on blood loss and blood transfusion during spinal surgeries with sodium nitroprusside-controlled hypotension. Saudi J Anaesth 2012; 6:263-267.
Yosry M, Othman IS. Controlled hypotension in adults undergoing choroidal melanoma resection: Comparison between the efficacy of nitroprusside and magnesium sulphate. Eur J Anaesthesiol 2008; 25: 891-896.
Degoute CS, Ray MJ, Manchon M, Dubreuil C, Banssillon V. Remifentanil and controlled hypotension; comparison with nitroprusside or esmolol during tympanoplasty. Can J Anaesth 2001; 48:20-27.
Mizrak A, Ganidagli S, Cengiz MT, Oner U, Saricicek V. The effects of dex premedication on volatile induction of mask anesthesia (vima)and sevoflurane requirements. J Clin Monit Comput 2013; 27:329-334.
Paranjpe J. Dexmedetomidine: Expanding role in anesthesia. Med J DY Patil Univ 2013; 6:5-13.
Boezaart AP, van der Merwe J, Coetzee A. Comparison of sodium nitroprusside- and esmolol-induced controlled hypotension for functional endoscopic sinus surgery. Can J Anaesth 1995; 42:373-376.
Nasreen F, Bano S, Khan RM, Hasan SA. Dexmedetomidine used to provide hypotensive anesthesia during middle ear surgery. Indian J Otolaryngol Head Neck Surg 2009; 61:205-207.
Guven DG, Demiraran Y, Sezen G, Kepek O, Iskender A. Evaluation of outcomes in patients given dexmedetomidine in functional endoscopic sinus surgery. Ann Otol Rhinol Laryngol 2011; 120:586-592.
Bulow NM, Barbosa NV, Rocha JB. Opioid consumption in total intravenous anesthesia is reduced with dexmedetomidine: a comparative study with remifentanil in gynecologic videolaparoscopic surgery. J Clin Anesth 2007; 19:280-285.
Patel CR, Engineer SR, Shah BJ, Madhu S. Effect of intravenous infusion of dexmedetomidine on perioperative haemodynamic changes and postoperative recovery: a study with entropy analysis. Indian J Anaesth 2012; 56:542-546.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2]