Original Contribution

Hemoperfusion using dual pulsatile pump in paraquat poisoning

Gil J. Suh MD^a, Christopher C. Lee MD^b,?, Ik J. Jo MD^c, Sang D. Shin MD^a,

Jung C. Lee MD^d, Byung G. Min MD^d, Adam J. Singer MD^b

^aDepartment of Emergency Medicine, Seoul National University College of Medicine, Seoul, Korea

^bDepartment of Emergency Medicine, Center for International Emergency Medicine, Stony Brook University Medical Center,

Stony Brook, NY, USA

^cDepartment of Emergency Medicine, Samsung Medical Center, Sungkyunkwan University, Seoul, Korea

^dDepartment of Biomedical Engineering, Seoul National University College of Medicine, Seoul, Korea

Received 5 May 2007; revised 26 September 2007; accepted 27 September 2007

Abstract

Introduction: Hemoperfusion is an effective method for removing paraquat from the body. However, the need for special equipment and personnel to perform hemoperfusion limits its applicability in the emergency department. A portable, user-friendly extracorporeal life support system (Twin Pulse Life Support [T-PLS]) capable of producing dual pulsatile flow by an electromechanical blood pump has recently been developed. We compared the effects of hemoperfusion using traditional and dual-pulsatile pumps in a paraquat-intoxicated Canine model. We hypothesized that T-PLS would be as effective as nonpulsatile hemoperfusion in reducing plasma and tissue paraquat levels.

Methods: Twelve adult male mongrel dogs were randomly assigned to hemoperfusion using a standard nonpulsatile pump (n = 6) or the T-PLS (n = 6). paraquat intoxication was induced by Intramuscular injection of paraquat (30 mg/kg). One hour after paraquat administration, hemoperfusion was performed for 4 hours at a flow rate of 125 mL/min. Periodic hemodynamic, chemical, and hematologic parameters as well as paraquat blood levels were obtained before and during the experiment. All animals were euthanized after completing 4 hours of hemoperfusion, and tissue levels of paraquat were determined. Results: During hemoperfusion, hemodynamic parameters including aortic blood pressure, heart rate, rectal temperature, and SaO2 showed no significant difference between the T-PLS group and the nonpulsatile pump groups. Chemical and hematologic parameters such as Serum electrolytes, platelet level, creatinine concentration, the ratio of Blood urea nitrogen to creatinine, alanine aminotransferase/aspartate aminotransferase (ALS/AST) levels, and plasma hemoglobin concentration as an indicator of hemolysis also showed no between-group differences. Plasma paraquat concentrations and lung and kidney tissue paraquat concentrations were also similar in both groups.

Conclusions: In this experimental canine study, T-PLS was as effective as traditional nonpulsatile hemoperfusion in reducing plasma and tissues paraquat levels. There were also no differences in hemodynamic, chemical, and hematologic parameters between the groups. Hemoperfusion using T-PLS may play a role as an alternative method for treating patients with paraquat poisoning.

(C) 2008

* Corresponding author. Department of Emergency Medicine, Center for International Emergency Medicine, Stony Brook University Medical Center, School of Medicine, Stony Brook, NY 11794, USA. Tel.: +1 631 444 3880; fax: +1 631 444 3919.

E-mail address: [email protected] (C.C. Lee).

0735-6757/$ – see front matter (C) 2008 doi:10.1016/j.ajem.2007.09.022

Introduction

Paraquat is a contact herbicide that is widely used throughout the world. In Korea, paraquat is available over the counter, and its use does not require any licensing. As a result, paraquat poisoning is quite common in Korea. Accidental and/or intentional poisoning with paraquat has a very high mortality rate. Proudfoot et al [1] reported that the prognosis of patients with paraquat poisoning is related to the plasma paraquat concentration and time after ingestion. The absorbed paraquat distributes rapidly to most tissues, with the highest concentrations found in the kidneys and the lungs. Greater than 90% of the absorbed dose of paraquat is eliminated by the kidneys within the first 12 to 24 hours after the ingestion [2]. However, due to the paraquat-associated nephrotoxicity, clearance of paraquat falls. In this situation, methods such as hemodialysis, hemofiltration, or hemoper- fusion should be considered to eliminate the absorbed paraquat from the body.

Hemoperfusion is a method for removing toxic materials from the blood by contact with an adsorbent system. It plays an important role in the management of patients with paraquat poisoning [3]. However, there are many factors that may lead to a delay in performing hemoperfusion in the emergent situation. In most hospitals, hemoperfusion is performed in a hemodialysis unit by a specially trained technician or “perfusionist” under the direction of a nephrologist. The need for specialized personnel and equipment is a significant barrier to the performance of hemoperfusion in an emergent situation paraquat poisoning. In addition, the hemodialysis unit is rarely available outside of business hours. Until recently, hemoperfusion has been performed using a standard roller pump that produces nonpulsatile blood flow. This roller

pump has been known to result in damage to the blood cells because the circuit tubes must be squeezed through the rotating rollers with minute small gaps between the tube’s

2 surfaces [4]. Generally, the duration of hemoperfusion using the roller pump in paraquat intoxication is at least 2 or 3 hours because of the limitation of the roller pump in increasing blood flow. As a result, it is difficult to remove paraquat from the blood within a short time frame.

Recently, we developed a portable Extracorporeal life support system that uses a dual pulsatile electromechanical blood pump (Twin Pulse Life Support [T-PLS]). In contrast to nonpulsatile flow, pulsatile flow can maintain circuit pressures with a 20% lower pump flow than the standard nonpulsatile pump flow [5-12]. Previous animal studies have shown that T-PLS can be used to treat animals with congestive heart failure with good hemodynamic and hematologic results [13,14]. The advantages of T-PLS include an easily adjustable pump rate, an autopriming mode, a suction detection mode, and an alarm. These features allow for easy estimation of the flow rate, and as a result, T-PLS can be easily operated by any Healthcare personnel (such as nurses or physicians) with minimal training. These characteristics of the T-PLS offer the opportunity to use this system to perform emergency hemoperfusion in the emer- gency department (ED) employing readily available ED staff. The current study was designed to compare the ability of standard nonpulsatile hemoperfusion and T-PLS to remove paraquat from the blood and the body tissues and to investigate the hemodynamic and hematologic consequences of these 2 methods in a paraquat-intoxicated canine model. We hypothesized that T-PLS would be as effective at reducing plasma and tissue levels of paraquat as standard

nonpulsatile hemoperfusion.

Fig. 1 The exterior view of the T-PLS.

structure and function of T-PLS”>Methods

Table 1 Hemodynamic parameters between the T-PLS group and the nonpulsatile group
		T-PLS group	Nonpulsatile group	P
	Systolic BP (mm Hg)	87.95 +- 20.01	81.68 +- 13.52	.452
	Diastolic BP (mm Hg)	57.57 +- 19.85	54.54 +- 15.53	.712
	Heart rate (beat/min)	103.84 +- 25.54	99.68 +- 14.96	.747
	Rectal temperature (?C)	35.37 +- 1.60	35.60 +- 1.75	.720
	Oxygen saturation (%)	98.70 +- 1.45	98.83 +- 1.36	.935
	BP indicates blood pressure.

Study design

A prospective, controlled experimental design was used to compare blood and tissue levels of paraquat in paraquat- intoxicated dogs treated with standard hemoperfusion and T-PLS. The study was approved by the Institutional Animal Care Committee and was in accordance with the Institutional and National Guidelines for handling laboratory animals.

Structure and function of T-PLS

The weight and dimensions of T-PLS are 25 kg and 380 x 280 x 600 cm, respectively. It also has 4 wheels and handles to control movement. The T-PLS consists of an electrical power source, blood pump, membrane oxygenator, and a controller (Fig. 1). The maximum flow rate of T-PLS is

7 L/min. The priming and stroke volumes of T-PLS are

400 and 70 mL, respectively. The T-PLS features auto- priming and suction protection modes. A 50-W brushless DC motor and 4-hour lithium-ion battery are used to operate the T-PLS. There also is a suction detection mode that sounds an alarm in case of poor inflow and reduces the pump rate automatically.

The twin silicon tube structure of T-PLS is similar to that of other hemoperfusion pumps such as the Anyheart (Korea) [15,16]. In the T-PLS circuit, while the actuator moves to the left side, the left silicon tube is contracted and the right silicon tube is simultaneously expanded, which leads to a positive pressure gradient from the left silicon tube to the right silicon tube. This mechanism results in pulsatile flow (Fig. 2). This mechanism may also effectively reduce the high membrane oxygenator Inlet pressure and subsequently result in less hemolysis. Moreover, less hemolysis can also be expected because of tube compres- sion with a solid actuator.

Fig. 2 Mechanism of T-PLS pump. PIP indicates pump inlet pressure; POP, pump outlet pressure; MOIP, membrane oxygenator inlet pressure; MOOP, membrane oxygenator outlet pressure.

Interventions

Twelve adult male mongrel dogs weighing approximately 30 kg were used in the study. The animals were housed in a controlled environment with free access to food and water before the experiment. The dogs were sedated with an intramuscular injection of ketamine (6 mg/kg) and xylazine (1 mg/kg), followed by endotracheal intubation with a 7F endotracheal tube. Anesthesia was induced with N₂O:O₂ (1:1) plus enflorane 1.5% to 4%. Intermittent mechanical positive-pressure ventilation was provided with a tidal vol- ume of 15 mL/kg. The respiratory rate was adjusted to maintain an end-tidal carbon dioxide of 35 to 40 mm Hg. The dogs were placed in the supine position and restrained at the 4 limbs. lactated Ringer solution (10 mL/kg per hour) was infused through a peripheral intravenous line located in a forelimb. Neuromuscular paralysis was achieved with pancu- ronium (0.2 mg/kg) and repeated as needed. Continuous electrocardiogram (lead II) monitoring was performed, and a rectal temperature probe and urinary catheter were placed. After shaving the cervical, thoracic, and Groin areas, the right Common carotid artery was surgically exposed. Using the Seldinger technique, 2 central venous catheters (5F; Arrow international Inc, USA) were placed via introducer sheaths (7.5F, Arrow international Inc) in the right common carotid artery and the Right internal jugular vein for hemodynamic monitoring and hematologic measurements, respectively.

After dissection of the right groin area, a double-lumen catheter (11F; Medcomp, PA, USA) was inserted into the right femoral vein using the Seldinger technique and connected to the T-PLS or nonpulsatile pump. Ten minutes before connecting the pump, an Intravenous heparin bolus (250 unit/kg) was given followed by a continuous heparin infusion (50 unit/hr) for systemic anticoagulation.

The animals were randomly assigned in 1 of 2 treatments. Randomization was performed using a computerized random numbers table. The nonpulsatile group (n = 6) had hemoperfusion with a nonpulsatile roller pump (AK-90S, Gambro Lundia AB, Sweden). The T-PLS group (n = 6) received hemoperfusion with the T-PLS (Newheartbio Co, Korea). In both groups, the same activated charcoal filters (Adsorba, Gambro Dialysatoren GmbH & Co KG, Germany) were used. Both systEMS used similar oxygen concentrations.

Fig. 3 Serum electrolyte concentrations. A, sodium; B, potassium; C, chloride; D, bicarbonate. BL indicates basal level; HP0, before start of hemoperfusion; HP30, 60, 120, 180, 240: 30, 60, 120, 180, and 240 minutes after the start of hemoperfusion.

Paraquat (30 mg/kg) was administered via intramuscular injection in the thigh area. This route of administration was chosen to minimize variability in bioavailability. The dose chosen is similar to a prior study [3]. The dogs were observed

for 1 hour and then were administered hemoperfusion at a flow rate of 125 mL/min for 4 hours. During hemoperfusion, blood was drawn for measurement of hemodynamic and hematologic parameters.

Fig. 4 Serum Platelet levels. Fig. 5 Serum plasma Hemoglobin levels.

Fig. 6 Serum creatinine concentration and BUN/Cr. A, BUN; B, BUN/creatinine ratio.

During the experiment, hemodynamic parameters includ- ing aortic blood pressure, heart rate, body temperature, and SaO2 were measured. Chemical and hematologic parameters included CBC, plasma electrolytes, AST/ALT, BUN, creatinine, plasma hemoglobin, and plasma paraquat con- centrations. Blood was withdrawn 30 minutes before and 60 minutes after paraquat injection and at 30, 90, 150, 210, and 270 minutes after hemoperfusion. After the experiment, the dogs were euthanized with an intravenous injection of 20 mL of KCl, and kidney and lung tissue concentrations of paraquat were determined.

Measurement of paraquat concentration

The levels of paraquat were measured colorimetrically as described previously [17-19].

Statistical analysis

For statistical analysis, SPSS software package 11.0 was used. A Mann-Whitney test was used to compare the hemodynamic and hematologic parameters between the T- PLS group and the nonpulsatile group. The data were expressed as means +- SDs. A P value less than .05 was considered to be statistically significant.

Results

Basic hemodynamic parameters such as aortic blood pressure, heart rate, body temperature, oxygen saturation,

Fig. 7 Serum glutamic-oxaloacetic transaminase (GOT) and glutamic-pyravic transaminase (GPT) levels. A, GOT; B, GPT.

Fig. 8 Serum paraquat levels.

and electrocardiogram were continuously monitored during the experiment. Although aortic blood pressures and heart rates in the T-PLS group were higher than those of the nonpulsatile group, these differences were not significant (Table 1). The changes in plasma sodium, potassium, chloride, and bicarbonate were similar between groups (Fig. 3). The number of platelets in the plasma dropped similarly in both groups (Fig. 4), whereas there was a similar rise in the hemoglobin levels in both groups (Fig. 5). The changes in renal and Hepatic function also did not differ between the groups as reflected in Figs. 6 and 7.

In both groups, there was an early rise in paraquat levels followed by a rapid decline over the next 4 hours. The changes in paraquat levels were similar in both groups (Fig. 8). Similarly, tissues paraquat concentrations in the lungs and kidneys also demonstrated no significant differ- ences between the 2 groups (Fig. 9).

Discussion

This study failed to show significant differences in plasma and tissue concentrations of paraquat between paraquat- intoxicated dogs treated with T-PLS and those treated with standard nonpulsatile hemoperfusion. This suggests that the novel pulsatile pump within the T-PLS is as effective at eliminating paraquat from both the blood and body tissues as the more standard yet cumbersome nonpulsatile hemoperfu- sion technique.

It is well known that pulsatile flow results in increased tissue perfusion compared with nonpulsatile flow [20-22]. Therefore, one might have expected to find a more rapid decline in blood and tissue levels of paraquat using the T-PLS system. However, our results showed that there were

no significant differences in the rate of decline in plasma and tissue paraquat concentrations between T-PLS and nonpul- satile pump. Although this finding cannot be completely explained, it is possible that the main factor leading to the removal of paraquat from the body may not be the type of blood flow produced by the hemoperfusion pump (pulsatile or nonpulsatile) but by the ability of charcoal filter to adsorb paraquat. Further study is needed to elucidate the extent of adsorption of paraquat in the charcoal filter by the pulsatile flow.

Many problems, including hemolysis and thrombocyto- penia, can occur as a result of extracorporeal circulation [4]. In this study, the measurement of plasma hemoglobin as an indicator of red blood cell destruction (or hemolysis) failed to show a significant difference between T-PLS and nonpulsa- tile pump, although hemoperfusion using T-PLS demon- strated a trend toward a decrease in plasma hemoglobin

Fig. 9 Distribution plot of tissue paraquat concentrations in the lung (P = .142 by Mann-Whitney test) and kidney (P = .902 by Mann-Whitney test). KD indicates kidney; Roller, nonpulsatile group; TPLS, T-PLS group.

concentrations. platelet counts were also lower in the T-PLS group compared with that of the nonpulsatile pump group. These finding suggest that T-PLS may result in a slight increase in the incidence of adverse events such as hemolysis or thrombocytopenia when compared with standard non- pulsatile hemoperfusion. Although the plasma creatinine concentration and the ratio of BUN to creatinine as indicators of nephrotoxicity failed to show significant differences between the groups, there was a trend toward reduced levels of BUN and a reduced BUN/creatinine ratio in the T-PLS group, and the differences between the groups seemed to grow over the course of the study. As a result, it is possible that had the duration of the experiment been extended, a significant difference would have been noted between the groups. Furthermore, because the number of animals was low, this study may have lacked the power to detect significant differences between the groups in renal function. Because of its rapid absorption and toxicity, hemoperfu- sion should not be delayed in paraquat poisoning. As mentioned earlier, hemoperfusion generally is performed by a perfusionist or specially trained technician in a specialized area such as a hemodialysis unit, making the use of hemoperfusion in the emergent situation very difficult. Thus, having an easy method for performing hemoperfusion in the ED by the ED staff would be very beneficial. Because

of its ease of use, the T-PLS meets these requirements.

The T-PLS is the first commercial dual pulsatile electromechanical extracorporeal life support device. The advantages of T-PLS include an easily adjustable pump rate, an autopriming mode, a suction detection mode, an alarm, and a flow estimation that can help medical personnel operate the system without specially trained pump techni- cians. For example, when the inflow is poor, the T-PLS sounds an alarm and automatically adjusts the pump rate. Because of its relatively compact size and weight, the T-PLS is portable and can even be loaded on an ambulance. The T-PLS also is equipped with a 4-hour lithium-ion battery and wheels allowing it to be easily moved from one area of the ED to another.

Limitations

Our study has several limitations that are worth noting. First, we used a canine model of paraquat poising. It is unclear how well this model reflects paraquat poisoning in humans. Therefore, the clinical application of the hemoper- fusion using T-PLS is relatively unknown. We have utilized hemoperfusion using T-PLS in 2 patients with paraquat poisoning. One patient survived, and the other died. However, it is difficult to evaluate the effects of hemoperfu- sion using T-PLS based on these 2 patients because many factors, including the dose of ingested paraquat, the time from ingestion to hemoperfusion, the frequency of hemo- perfusion, the patient’s underlying condition, and so on, might affect the prognosis of patients with paraquat

poisoning. Before widespread clinical application, further studies of hemoperfusion using the T-PLS in paraquat poisoning are needed. Second, paraquat poisoning was induced via intramuscular injection of paraquat, which can cause rapid distribution of paraquat into the various organs of body. In contrast, most paraquat poisonings in the ED are caused by oral ingestion of paraquat. In this situation, the absorption of paraquat from the gastrointestinal tract into the blood and tissues can be delayed by several factors such as the presence of food material, the rate of gastric emptying, gastrointestinal decontamination, and so on. Therefore, animal models that better mimic the clinical scenario should also be developed to evaluate the effects of hemoperfusion by T-PLS. Finally, our study did not include an untreated control group that also received paraquat. However, a prior animal study has already compared standard hemoperfusion to a control group that was not treated and found that hemoperfusion reduced mortality [3].

Conclusion

In this study, T-PLS was as effective at reducing serum and tissue paraquat levels as the standard nonpulsatile hemoperfusion machine. The T-PLS also produced similar hemodynamic and hematologic effects as the nonpulsatile pump. Because of its ease of use, hemoperfusion using T-PLS should be considered as an alternative method for the management of patients with paraquat poisoning.

References

1. Proudfoot AT, Stewart MJ, Levitt T, et al. Paraquat poisoning: significance of plasma-paraquat concentration. Lancet 1979;2:330-2.
2. Bismuth C, Sherrmann JM, Garnier R, et al. Elimination of paraquat. Hum Toxicol 1987;6:63-7.
3. Yang TS, Chang YL, Yen CK. Hemoperfusion treatment in pigs experimentally intoxicated by paraquat. Hum Exp Toxicol 1997;16: 709-15.
4. Qian KX, Zeng P, Ru WM, Yuan HY, Feng ZG, Li I. How to produce a pulsatile flow with low hemolysis. J Med Eng Technol 2000;24:227-9.
5. Sanderson JM, Wright G, Sims FW. Brain damage in dogs immediately following pulsatile and non-pulsatile blood flows in extracorporeal circulation. Thorax 1972;27:225-86.
6. Taylor KM, Bain HB, Maxted KJ, et al. Comparative studies of pulsatile and nonpulsatile flow during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1978;75:569-73.
7. Trinkle JK, Helton NE, Bryant LR, Griffen WO. Pulsatile cardio- pulmonary bypass: clinical evaluation. Surgery 1970;68:1074-7.
8. Trinkle JK, Helton NE, Wood RE, Bryant LR. Metabolic comparison of a new pulsatile pump and a roller pump for cardiopulmonary bypass. J Thorac Cardiovasc Surg 1969;58:562-7.
9. Chiu IS, Chu SH, Hung CR. Pulsatile flow during routine cardiopulmonary bypass. J Cardiovasc Surg 1984;25:530-6.
10. Anstadt MP, Tedder M, Hegde SS, et al. Pulsatile versus nonpulsatile reperfusion improves cerebral blood flow after cardiac arrest. Ann Thorac Surg 1993;56:453-61.
11. Anstadt MP, Stonnington MJ, Tedder M, et al. Pulsatile reperfusion after cardiac arrest improves neurologic outcome. Ann Surg 1991;214: 478-88.
12. Undar A, Masai T, Yang SQ, et al. pulsatile perfusion improves regional Myocardial blood flow during and after hypothermic cardiopulmonary bypass in a neonatal piglet model. ASAIO J 2002;48:90-5.
13. Baek KJ, Kim JS, Sun K, Son HS, Ki W, Han SB, et al. Effects of an extra-corporeal life support system using a dual pulsatile pump. J Korean Soc Emerg Med 2002;13:489-96.
14. Roh YR, Choi H, Lee JC, et al. Application of the pulsatile flow versatile ECLS: in vivo studies. Int J Artif Organs 2003;26(5): 428-35.
15. Choi WW, Kim HC, Min BG. A new automatic cardiac output control algorithm for moving actuator total artificial heart by motor current waveform analysis. Int J Artificial Organs 1996;19:189-96.
16. Min BG, Kim HC, Choi JW, Ryu GH, Seo KP, Rho JR, et al. A moving-actuator type electromechanical total artificial heart–part II:

circular type and Animal experiment. IEEE Trans Biomed Eng 1990; 37:1195-200.

1. Tsunenari S. The determination of paraquat using chromatography. Forensic Sci 1975;5:61-7.
2. Tsunoda N. Selective extraction of paraquat using SepPak C18 cartridge. Eisei Kagaku 1983;29:206-11.
3. Maruyama H, et al. Simple ten-channel elutor for chromatograph procedures. Clin Chem 1986;32:1988.
4. Jacobs LA, Klopp EH, Seamone W, et al. Improved Organ function during cardiac bypass with a roller pump to deliver pulsatile flow. J Thorac Cardiovasc Surg 1969;58:703-12.
5. Shepard RB, Kirklin JW. Relation of pulsatile flow to oxygen consumption and other variables during cardiopulmonary bypass. J Thorac Cardiovasc Surgv 1969;58:694-702.
6. Nakano T, Tominaga R, Nagano I, et al. Pulsatile flow enhances endothelium-derived nitric oxide release in the peripheral vasculature. Am J Physiol Heart Circ Physiol 2000;278:H1098-104.