Anaesthesia 2014, 69, 111–117
doi:10.1111/anae.12463
Original Article The properties of an improvised piston pump for the rapid delivery of intravenous fluids* C. M. Smart,1 C. W. Primrose,2 A. L. Peters2 and E. J. Speirits2 1 Anaesthetic Trainee, Ayr Hospital, Ayr, UK 2 Medical Student, Glasgow University, Glasgow, UK
Summary To maximise the effect of a small fluid load, it is occasionally desirable to bolus manually with multiple depressions of a large-capacity syringe. This is usually achieved by placing the syringe on the side port of a three-way tap. We modified this technique by placing two-one-way valves in line with the three-way tap, effectively creating a piston pump, the infusion rates via which we compared with those achieved by an inflatable pressure-infuser in a simulated resuscitation. Fluid flow was faster using the piston pump than with the pressure-infuser (mean (SD) time to infuse 2000 ml saline 0.9% via a 16-G cannula 352 (10) s vs 495 (19) s, respectively, p < 0.0001). The piston pump appears to have potential for both tight control of fluid delivery and major high-volume resuscitation. The lightweight nature of the pump and its lack of reliance on gravity may also make it suitable for the pre-hospital setting. .................................................................................................................................................................
Correspondence to: C. Smart Email: [email protected] *Presented in part at the European Society of Emergency Medicine, Antalya, Turkey; October 2012. The Intensive Care Society State of the Art Meeting, London, UK; December 2012. and the Scottish Intensive Care Society, St Andrews, UK; January 2013. Accepted: 7 September 2013
Introduction The use of a large-capacity syringe to deliver controlled boluses of fluid is a well-described technique. Manual syringing of intravenous fluid may be desirable in three settings: first, if there is a need to control the volume tightly (e.g. in paediatric resuscitation) [1, 2], secondly, if there is a need to use higher pressures to drive flow (e.g. with intra-osseous devices) [3, 4] and thirdly, if it is necessary to introduce a small volume rapidly (e.g. to assess left ventricular volume responsiveness) [5]. Manual syringing is usually performed by attaching the syringe to the side port of a three-way tap connected in series with an intravenous giving set [1]. © 2014 The Association of Anaesthetists of Great Britain and Ireland
This practice is attention consuming, requiring constant alteration of the three-way tap position, and is made inefficient by refluxing of fluid into the syringe under back pressure. As a result, rapid, high volume resuscitation is onerous, and this severely limits the utility of the technique. Instead, increasing intravenous flow during high-volume resuscitation is normally facilitated by an alternative method, most commonly using an inflatable pressure-infuser (‘bag squeezer’), or a mechanical pressure-infuser, if available [6, 7]. The expanding availability of cardiac output monitoring, and the growing evidence of benefit in certain surgery [8, 9], would seem to make manual syringing 111
Anaesthesia 2014, 69, 111–117
an appealing option. By delivering a small volume rapidly, the fluid has less time to redistribute [10, 11], and the cardiac responsiveness can be more accurately assessed. If the fluid is delivered in small, rapidly titrated aliquots, one can avoid over-resuscitating the patient, whilst responding quickly to any drops in the cardiac index. However, as outlined above, the current syringing technique is limited in its versatility, as it is unsuitable for high-volume fluid resuscitation. We propose a modification of the standard syringing technique, in which a one-way valve is attached both upstream and downstream of the three-way tap. The one-way valves effectively convert the syringe into a simple piston pump. Because no alteration of the three-way tap position is now necessary, the device becomes capable of very rapid infusion. We characterised the properties of the device, and compared it in a simulated resuscitation with the most commonly available alternative, the inflatable pressure cuff.
Smart et al. | Piston pump delivery of intravenous fluids
Figure 1 Image of the piston pump assembly before insertion into circuit.
Circuit: (1)
Circuit: (2)
1
1
Methods The piston pump was constructed around a 50-ml syringe (Plastipak Luer-LokTM; BD Medical, Franklin Lakes, NJ, USA). For clarity, note that a ‘50-ml’ syringe actually has a full capacity of 60 ml. The syringe was connected to the side port of a three-way tap (Discofixâ; Braun Medical, Sheffield, UK). On the other two ports of the tap were attached one-way valves (r-lockTM 165250; Codan, Lensahn, Germany). The position of the tap was adjusted so that all its lumens were open (Fig. 1). Two infusion circuits were then created and primed with saline 0.9%. The total height drop from the top of the bag of infusion fluid to the outflow was 100 cm for both circuits. Circuit (1) comprised a 500-ml bag (sodium chloride 0.9%, FKE 1323, Viaflow; Baxter, Newbury, UK) spiked with a blood-giving set (MMC2071B; Baxter). The piston pump assembly was inserted downstream of the giving set, via Luer-lock connections. The outflow of the pump was attached to a blood warmer cartridge and extension line (Rangerâ Standard Flow Disposable Set with Extension/Model 24250; Arizant Healthcare, Trittau, Germany), inserted into a blood warmer (Ranger blood and fluid warmer; Arizant 112
5 1m 4 2
2 3
3
Figure 2 Diagram of the constructions of the two circuits. 1, fluid bag with giving set; 2, fluid warmer with insert and extension; 3, intravenous cannula; 4, piston pump; 5, inflatable pressure-infuser. Healthcare). The distal end of the extension line was connected to a 16-G intravenous cannula (BD VenflonTM Pro Safety; BD Medical) emptying into a bucket (Fig. 2). Circuit (2) was identical in every respect to circuit (1), but lacking the piston pump. Instead, an inflatable pressure-infuser (Medex C-FusorTM; Smiths Medical, St Paul, MN, USA) was used to drive flow (Fig. 2). © 2014 The Association of Anaesthetists of Great Britain and Ireland
Smart et al. | Piston pump delivery of intravenous fluids
Anaesthesia 2014, 69, 111–117
Infusion times The time to infuse a given fluid load was compared between the two circuits. The times to administer a single 500-ml bag of saline, to exchange bags once the previous one had finished and to administer a 2000-ml fluid load (4 9 500-ml bags including changeovers) were recorded. In circuit (1), one operator pumped the piston as quickly as possible to generate flow, whilst a second operator rapidly changed fluid bags when the previous one had finished. In circuit (2), one operator inflated the pressure bag as quickly as possible to a pressure of 300 mmHg, and maintained this pressure by intermittent insufflation of air throughout the infusion, whilst a second operator changed fluid bags, and detached/ re-attached the pressure-infuser. The circuits were pre-primed with saline in both cases. The timer was started with the spiking of a fresh bag of saline, and the infusion begun using the methods outlined above. The time to run the bag to visible completion was recorded, the bag was removed, and a fresh bag was spiked. The time taken to changeover fluid bags was recorded. This was defined as the time from noting completion of the previous bag, to the moment the next bag was spiked. With each method, the process was repeated as fast as possible until four 500-ml bags had been run through. The total time to complete the full 2000-ml fluid load was then documented for both techniques. The experiment was repeated four times (16 bags of saline per group).
Pressure/time curves A pressure transducer (Truwave disposable; Edwards Lifesciences, Irvine, CA, USA) was connected to the side port of the warmer extension line immediately upstream of, and level with, the cannula (Fig. 3). The intraluminal pressure at this point represents the trans-cannula driving pressure. Four 500-ml bags of saline were run through as rapidly as possible using both methods. Mean pressure measurements (determined by an internal algorithm) were recorded from a digital monitor (IntelliVue MP90; Phillips Healthcare, Best, Netherlands) every 5–10 s until the end of the infusion. © 2014 The Association of Anaesthetists of Great Britain and Ireland
Transduced circuit
1
1m
4
2 5
3
Figure 3 Diagram of the position of the transducer in circuit (1). 1, fluid bag with giving set; 2, fluid warmer with insert and extension; 3, intravenous cannula; 4, piston pump; 5, pressure transducer level with cannula.
Peak pressure recording The circuit was assembled as in Fig. 3, but with a peak pressure sensor (DPM3 Biometer; Fluke Biomedical, Everett, WA, USA) attached to the side port of the warmer circuit. The syringe was filled to its 60-ml capacity. The operator then applied a steady force to the plunger with two thumbs, discharging the contents of the syringe into the circuit at a relatively constant rate. The time taken to empty the barrel of the syringe and the peak pressure generated downstream were recorded. This process was repeated 35 times with various steady forces applied to the syringe plunger and the results were plotted graphically.
Haemolysis A final experiment was designed to assess whether turbulence within the piston pump would haemolyse red cells. A bag of expired packed red blood cells (39 days post-venesection date) was run passively through 113
Anaesthesia 2014, 69, 111–117
Smart et al. | Piston pump delivery of intravenous fluids
Mean pressure (mmHg)
300 250 200 150 100 50 0
0
20
40
60
80
100
120
Time (s)
Figure 4 Algorithm-derived mean intraluminal pressure vs time from start of infusion. Solid lines represent piston pump circuit (4 9 500-ml bags). Dotted lines represent pressure-infuser circuit (4 9 500-ml bags).
circuit (2) until it flowed out of the cannula. The pressure-infuser was then inflated to 300 mmHg with the circuit clamped. The priming volume of the circuit (65 ml), and an additional 20 ml, were then run through under pressure and discarded. Three samples were then taken at the cannula, under conditions of driven flow, at approximately 5-ml intervals. These samples represented blood driven through the entire circuit by the pressure-infuser. Flow was then stopped, the pressure-infuser removed and the piston pump installed as previously described. Eighty-five millilitres were then pumped through rapidly and discarded. Three samples were immediately taken from the cannula (running under pressure) at approximately 5-ml intervals. These represented blood that had run the entire way through the circuit driven by the piston pump. The samples were centrifuged, and the supernatant fluid analysed by spectral absorption (Modular Analytics Evo Solution; Roche Diagnostics, Penzberg, Germany) for the presence of free haemoglobin. A haemolysis index of 0–1000 was generated. The results were analysed by paired or unpaired two-tailed t-tests (Prism Software; GraphPad, La Jolla, CA, USA), with a value of p < 0.05 indicating statistical significance.
Results The piston pump was faster at infusing 500 ml saline (mean (SD) 80 (2) s) and 2000 ml saline in four 114
sequential 500-ml bags (352 (10) s than the pressureinfuser (113 (7) s, p < 0.0001 and 495 (19) s, respectively, p < 0.0001). Compared with the pressure-infuser, the time taken to replace an empty fluid bag with the piston pump system was significantly shorter (mean (SD) 10 (3) s vs 13 (3) s, p = 0.0036). The pressure-time curves for the piston pump and pressure-infuser are shown in Fig. 4. A rectangular hyperbola resulted when the peak intraluminal pressure was recorded against the time taken to empty a 60-ml syringe via the piston pump assembly (Fig. 5), from which a straight line of best fit with zero intercept of pressure/timereciprocal was plotted to derive a function relating peak pressure to time (pressure = 2318/time). This function is consistent with laminar flow in accordance with Poiseuille’s Law [12]. Compared with the pressure-infuser, there was no significant difference in the haemolysis index using the piston pump (mean (SD) 61.3 (1.2) vs 63.3 (1.5), p = 0.15).
Discussion The improvised piston pump proved effective in our simulated resuscitation scenario, and significantly outperformed the pressure-infuser. We noted time savings in the infusion of single intravenous bags, the time to changeover bags and cumulative savings over the full 2000-ml resuscitation. These results can be understood by considering the pressure driving the two infusions (Fig. 4). For © 2014 The Association of Anaesthetists of Great Britain and Ireland
Smart et al. | Piston pump delivery of intravenous fluids
Anaesthesia 2014, 69, 111–117
Peak pressure (mmHg)
700 600 500 400 300 200 100 0
0
5
10
15
20
Time to depress plunger (s)
Figure 5 Intraluminal peak pressure vs time to depress 60-ml plunger. clarity, the piston pump generates saw-tooth pressure waves, so the mean pressure is analysed. At the start of infusion, the piston pump assembly responds rapidly, with a sharp rise to plateau pressure. The pressure-infuser is inefficient by contrast, with a slow rise-time, and a tail-off (caused by the shape of the internal bladder). The piston pump’s steady pressure profile means that flow is driven consistently throughout. This efficiency produces disproportionately large time savings, with only a minimal increase in plateau pressure. Although the mean pressure of the cycle dictates flow, the peak pressure affects safety. This is discussed separately below. Peak pressure analysis can be used to study the fluid dynamics of the pump. By observing a simple reciprocal relationship between pressure and time (Fig. 5), we conclude that flow through the cannula remained consistently laminar according to Poiseuille’s Law [12]; this is desirable in reducing shear stress and haemolysis [13]. Ergonomically, the arrangement of the piston pump means no manipulation of the three-way tap is
necessary for either active or passive flow. When flow is passive, the syringe will automatically refill. The plunger is pushed back by hydrostatic pressure until arrested by the internal lip in the barrel. The full syringe is then bypassed, diverting all flow to the patient (Fig. 6). Several other features may be of particular interest to anaesthetists. The 60-ml syringe exceeds the volume of the basilic vein of the arm [14], so one plunger depression should flush intravenous drugs into the central circulation, even with sluggish venous return. Furthermore, as the syringe refills automatically, the anaesthetist is seldom without means to flush in drugs, even if the current intravenous bag is empty. The piston pump can generate enough pressure to cause barotrauma [15]. Intraluminal pressures as high as 635 mmHg resulted when a rapid push was attempted (Fig. 5). It should be understood, however, that the intraluminal pressure is not the same as the intravenous pressure. Intravenous pressure will always be lower, being downstream of the resistance of the cannula. Furthermore, the peak pressure of the pump cycle will be damped by the compliance of the venous plexus [12]. Evidence to back these assumptions comes from an experiment that used a similar cyclical pump (Manual Bulb Pump; Tuta laboratories, Lane cove, NSW, Australia) infusing via a 16-G cannula into a forearm vein [15]. Transducing via a second cannula inserted proximally, Goodie et al. measured the intravenous pressure generated by the pump. They found it to be consistently lower than 300 mmHg despite (in a previous experiment) having registered pressures exceeding 1000 mmHg within the intravenous line itself [16]. They noted
Figure 6 Schematic illustrating self-refilling under hydrostatic pressure. © 2014 The Association of Anaesthetists of Great Britain and Ireland
115
Anaesthesia 2014, 69, 111–117
that occluding the proximal run-off from the vein (by inflating a blood pressure cuff on the arm) caused a linear increase in venous pressure, in proportion to the pressure within the cuff. The maximum intravenous pressure they recorded was around 200 mmHg, with the blood pressure cuff inflated to 150 mmHg. We are reasonably reassured, therefore, that the syringe pump will not cause barotrauma as long as the venous run-off remains sufficient. There have been case reports of extravasation and compartment syndrome caused by bulb pumps, with sawtooth pressure profiles [17]. However, there are similar reports relating to inflatable pressure cuffs [18– 20] and even rapid infusers with occlusion alarms [7]. Cannula site vigilance is thus advised irrespective of pressurisation method. We suggest that any cannula attached to a piston pump should be placed in a large superficial vein (not a central vein) on the non-blood pressure cuff arm, and checked if resistance is felt during syringe plunger depression. Furthermore, the piston pump should only be used once passive flow under gravity has been demonstrated. This should ensure that there is no serious obstruction of venous return, and that the one-way valves are correctly oriented. Approximately 40–100 ml air is contained within a 1000-ml intravenous bag [7, 21]. This will not advance under gravity, but may do so under pressure. Serious air embolism can be caused by both inflatable and automated pressure-infusers [21, 22], and may occur when using a piston pump. As with any pressure infuser, ideally all air should be aspirated from the intravenous bag prior to infusion, although it has been noted that this is poorly complied with in practice [21]. We conclude that the piston pump is a useful adjunct to resuscitation, and could be employed in an emergency if a rapid infuser is unavailable. Its ability to deliver rapid, titrated boluses is suited to cardiac output monitoring. Being compact, simple, and only 39 g in weight, this device could also be employed in the pre-hospital setting, entrapment, or battlefield, where there may be limited space around the patient [23]. We commend its use to anaesthetists and would encourage colleagues to conduct further studies of its clinical safety and efficacy. 116
Smart et al. | Piston pump delivery of intravenous fluids
Acknowledgements The authors thank the anaesthesia department at Ayr Hospital for the kind donation of the materials necessary to conduct the research.
Competing interests No external funding and no competing interests declared.
References 1. Bree S, Nordmann GR, Wood K, McNicholas J. The paediatric transfusion challenge on deployed operations. Journal of the Royal Army Medical Corps 2010; 156: 361–4. 2. Holden HB, Maher JJ. Some aspects of blood loss and fluid balance in paediatric adenotonsillectomy. British Medical Journal 1965; 5474: 1349–51. 3. LaRocco BS, Henry E, Wang MD, Brian G. Intraosseous infusion. Prehospital Emergency Care 2003; 7: 280–5. 4. Macnab A, Christenson J, Findlay J, et al. A new system for sternal intraosseous infusion in adults. Prehospital Emergency Care 2000; 4: 173–7. 5. Hood JA, Wilson RJ. Pleth variability index to predict fluid responsiveness in colorectal surgery. Anesthesia and Analgesia 2011; 113: 1058–63. 6. Comunale ME. A laboratory evaluation of the Level 1 rapid infuser (H1025) and the Belmont Instrument fluid management system (FMS 2000) for rapid transfusion. Anesthesia and Analgesia 2003; 97: 1064–9. 7. Smith CE, Kabbara A, Kramer RP, Gill I. Evaluation of a new IV fluid and blood warming system to prevent air embolism. Trauma Care 2001; 11: 78–82. 8. Gan TJ, Soppit A, Maroof M, et al. Goal-directed intraoperative fluid administration reduces length of hospital stay after major surgery. Anesthesiology 2002; 97: 820–6. 9. Pearse R, Dawson D, Fawcett J, Rhodes A, Grounds RM, Bennett ED. Early goal-directed therapy after major surgery reduces complications and duration of hospital stay. A randomised, controlled trial. Critical Care 2005; 9: 687–93. 10. Drobin D, Hahn RG. Volume kinetics of Ringer’s solution in hypovolemic volunteers. Anesthesiology 1999; 90: 81–91. n C. Perioperative fluid management. Anes11. Prough DS, Svense thesia and Analgesia 2002; 10: 84–91. 12. Acheson DJ. Elementary Fluid Dynamics. Oxford: Clarendon Press, 1990. 13. Kamaneva MV, Burgreen GW, Kono K, Repko B, Antaki JF, Umezu M. Effects of turbulent stresses upon mechanical hemolysis: experimental and computational analysis. American Society for Artificial Internal Organs Journal 2004; 50: 418– 23. 14. Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium enhancement at CT. Part I. Prediction with a computer model. Radiology 1998; 207: 647–55. 15. Goodie DB. The effect of manual bulb pump infusion systems on venous luminal pressure and vein wall integrity. Anesthesia and Analgesia 1995; 80: 552–6. 16. Goodie DB, Morris RW. Bulb pump infusions and extravasation injury. Anesthesia and Analgesia 1994; 78: 812. 17. Sneyd JR, Lau W, McLaren ID. Forearm compartment syndrome following intravenous infusion with a manual ‘‘bulb’’ pump. Anesthesia and Analgesia 1993; 76: 1160–1. © 2014 The Association of Anaesthetists of Great Britain and Ireland
Smart et al. | Piston pump delivery of intravenous fluids 18. Seiler JG, Valadie AL, Drvaric DM, Frederick RW, Whiteside TE. Perioperative compartment syndrome. A report of four cases. Journal of Bone and Joint Surgery 1996; 78: 600–2. 19. Sung CY, Chung RK, Ra YS, Lee HS, Lee GY. Impending compartment syndrome of the forearm and hand after a pressurized infusion in a patient under general anesthesia-A case report. Korean Journal of Anesthesiology 2011; 60: 60–3. 20. Willsey DB, Peterfreund RA. Compartment syndrome of the upper arm after pressurized infiltration of intravenous fluid. Journal of Clinical Anesthesia 1997; 9: 428–30.
© 2014 The Association of Anaesthetists of Great Britain and Ireland
Anaesthesia 2014, 69, 111–117 21. Adhikary GS, Massey SR. Massive air embolism: a case report. Journal of Clinical Anesthesia 1998; 10: 70–2. 22. Sviri S, Woods WP, Van Heerden P. Air embolism – a case series and review. Critical Care and Resuscitation 2004; 6: 271– 6. 23. Pearce FJ, Lee J, Van Albert S. Closing the Loop on Critical Care Life Support for Military En-Route Care Environments. Silver Spring, MD: Walter Reed Army Institute of Research, 2004.
117
doi:10.1111/anae.12463
Original Article The properties of an improvised piston pump for the rapid delivery of intravenous fluids* C. M. Smart,1 C. W. Primrose,2 A. L. Peters2 and E. J. Speirits2 1 Anaesthetic Trainee, Ayr Hospital, Ayr, UK 2 Medical Student, Glasgow University, Glasgow, UK
Summary To maximise the effect of a small fluid load, it is occasionally desirable to bolus manually with multiple depressions of a large-capacity syringe. This is usually achieved by placing the syringe on the side port of a three-way tap. We modified this technique by placing two-one-way valves in line with the three-way tap, effectively creating a piston pump, the infusion rates via which we compared with those achieved by an inflatable pressure-infuser in a simulated resuscitation. Fluid flow was faster using the piston pump than with the pressure-infuser (mean (SD) time to infuse 2000 ml saline 0.9% via a 16-G cannula 352 (10) s vs 495 (19) s, respectively, p < 0.0001). The piston pump appears to have potential for both tight control of fluid delivery and major high-volume resuscitation. The lightweight nature of the pump and its lack of reliance on gravity may also make it suitable for the pre-hospital setting. .................................................................................................................................................................
Correspondence to: C. Smart Email: [email protected] *Presented in part at the European Society of Emergency Medicine, Antalya, Turkey; October 2012. The Intensive Care Society State of the Art Meeting, London, UK; December 2012. and the Scottish Intensive Care Society, St Andrews, UK; January 2013. Accepted: 7 September 2013
Introduction The use of a large-capacity syringe to deliver controlled boluses of fluid is a well-described technique. Manual syringing of intravenous fluid may be desirable in three settings: first, if there is a need to control the volume tightly (e.g. in paediatric resuscitation) [1, 2], secondly, if there is a need to use higher pressures to drive flow (e.g. with intra-osseous devices) [3, 4] and thirdly, if it is necessary to introduce a small volume rapidly (e.g. to assess left ventricular volume responsiveness) [5]. Manual syringing is usually performed by attaching the syringe to the side port of a three-way tap connected in series with an intravenous giving set [1]. © 2014 The Association of Anaesthetists of Great Britain and Ireland
This practice is attention consuming, requiring constant alteration of the three-way tap position, and is made inefficient by refluxing of fluid into the syringe under back pressure. As a result, rapid, high volume resuscitation is onerous, and this severely limits the utility of the technique. Instead, increasing intravenous flow during high-volume resuscitation is normally facilitated by an alternative method, most commonly using an inflatable pressure-infuser (‘bag squeezer’), or a mechanical pressure-infuser, if available [6, 7]. The expanding availability of cardiac output monitoring, and the growing evidence of benefit in certain surgery [8, 9], would seem to make manual syringing 111
Anaesthesia 2014, 69, 111–117
an appealing option. By delivering a small volume rapidly, the fluid has less time to redistribute [10, 11], and the cardiac responsiveness can be more accurately assessed. If the fluid is delivered in small, rapidly titrated aliquots, one can avoid over-resuscitating the patient, whilst responding quickly to any drops in the cardiac index. However, as outlined above, the current syringing technique is limited in its versatility, as it is unsuitable for high-volume fluid resuscitation. We propose a modification of the standard syringing technique, in which a one-way valve is attached both upstream and downstream of the three-way tap. The one-way valves effectively convert the syringe into a simple piston pump. Because no alteration of the three-way tap position is now necessary, the device becomes capable of very rapid infusion. We characterised the properties of the device, and compared it in a simulated resuscitation with the most commonly available alternative, the inflatable pressure cuff.
Smart et al. | Piston pump delivery of intravenous fluids
Figure 1 Image of the piston pump assembly before insertion into circuit.
Circuit: (1)
Circuit: (2)
1
1
Methods The piston pump was constructed around a 50-ml syringe (Plastipak Luer-LokTM; BD Medical, Franklin Lakes, NJ, USA). For clarity, note that a ‘50-ml’ syringe actually has a full capacity of 60 ml. The syringe was connected to the side port of a three-way tap (Discofixâ; Braun Medical, Sheffield, UK). On the other two ports of the tap were attached one-way valves (r-lockTM 165250; Codan, Lensahn, Germany). The position of the tap was adjusted so that all its lumens were open (Fig. 1). Two infusion circuits were then created and primed with saline 0.9%. The total height drop from the top of the bag of infusion fluid to the outflow was 100 cm for both circuits. Circuit (1) comprised a 500-ml bag (sodium chloride 0.9%, FKE 1323, Viaflow; Baxter, Newbury, UK) spiked with a blood-giving set (MMC2071B; Baxter). The piston pump assembly was inserted downstream of the giving set, via Luer-lock connections. The outflow of the pump was attached to a blood warmer cartridge and extension line (Rangerâ Standard Flow Disposable Set with Extension/Model 24250; Arizant Healthcare, Trittau, Germany), inserted into a blood warmer (Ranger blood and fluid warmer; Arizant 112
5 1m 4 2
2 3
3
Figure 2 Diagram of the constructions of the two circuits. 1, fluid bag with giving set; 2, fluid warmer with insert and extension; 3, intravenous cannula; 4, piston pump; 5, inflatable pressure-infuser. Healthcare). The distal end of the extension line was connected to a 16-G intravenous cannula (BD VenflonTM Pro Safety; BD Medical) emptying into a bucket (Fig. 2). Circuit (2) was identical in every respect to circuit (1), but lacking the piston pump. Instead, an inflatable pressure-infuser (Medex C-FusorTM; Smiths Medical, St Paul, MN, USA) was used to drive flow (Fig. 2). © 2014 The Association of Anaesthetists of Great Britain and Ireland
Smart et al. | Piston pump delivery of intravenous fluids
Anaesthesia 2014, 69, 111–117
Infusion times The time to infuse a given fluid load was compared between the two circuits. The times to administer a single 500-ml bag of saline, to exchange bags once the previous one had finished and to administer a 2000-ml fluid load (4 9 500-ml bags including changeovers) were recorded. In circuit (1), one operator pumped the piston as quickly as possible to generate flow, whilst a second operator rapidly changed fluid bags when the previous one had finished. In circuit (2), one operator inflated the pressure bag as quickly as possible to a pressure of 300 mmHg, and maintained this pressure by intermittent insufflation of air throughout the infusion, whilst a second operator changed fluid bags, and detached/ re-attached the pressure-infuser. The circuits were pre-primed with saline in both cases. The timer was started with the spiking of a fresh bag of saline, and the infusion begun using the methods outlined above. The time to run the bag to visible completion was recorded, the bag was removed, and a fresh bag was spiked. The time taken to changeover fluid bags was recorded. This was defined as the time from noting completion of the previous bag, to the moment the next bag was spiked. With each method, the process was repeated as fast as possible until four 500-ml bags had been run through. The total time to complete the full 2000-ml fluid load was then documented for both techniques. The experiment was repeated four times (16 bags of saline per group).
Pressure/time curves A pressure transducer (Truwave disposable; Edwards Lifesciences, Irvine, CA, USA) was connected to the side port of the warmer extension line immediately upstream of, and level with, the cannula (Fig. 3). The intraluminal pressure at this point represents the trans-cannula driving pressure. Four 500-ml bags of saline were run through as rapidly as possible using both methods. Mean pressure measurements (determined by an internal algorithm) were recorded from a digital monitor (IntelliVue MP90; Phillips Healthcare, Best, Netherlands) every 5–10 s until the end of the infusion. © 2014 The Association of Anaesthetists of Great Britain and Ireland
Transduced circuit
1
1m
4
2 5
3
Figure 3 Diagram of the position of the transducer in circuit (1). 1, fluid bag with giving set; 2, fluid warmer with insert and extension; 3, intravenous cannula; 4, piston pump; 5, pressure transducer level with cannula.
Peak pressure recording The circuit was assembled as in Fig. 3, but with a peak pressure sensor (DPM3 Biometer; Fluke Biomedical, Everett, WA, USA) attached to the side port of the warmer circuit. The syringe was filled to its 60-ml capacity. The operator then applied a steady force to the plunger with two thumbs, discharging the contents of the syringe into the circuit at a relatively constant rate. The time taken to empty the barrel of the syringe and the peak pressure generated downstream were recorded. This process was repeated 35 times with various steady forces applied to the syringe plunger and the results were plotted graphically.
Haemolysis A final experiment was designed to assess whether turbulence within the piston pump would haemolyse red cells. A bag of expired packed red blood cells (39 days post-venesection date) was run passively through 113
Anaesthesia 2014, 69, 111–117
Smart et al. | Piston pump delivery of intravenous fluids
Mean pressure (mmHg)
300 250 200 150 100 50 0
0
20
40
60
80
100
120
Time (s)
Figure 4 Algorithm-derived mean intraluminal pressure vs time from start of infusion. Solid lines represent piston pump circuit (4 9 500-ml bags). Dotted lines represent pressure-infuser circuit (4 9 500-ml bags).
circuit (2) until it flowed out of the cannula. The pressure-infuser was then inflated to 300 mmHg with the circuit clamped. The priming volume of the circuit (65 ml), and an additional 20 ml, were then run through under pressure and discarded. Three samples were then taken at the cannula, under conditions of driven flow, at approximately 5-ml intervals. These samples represented blood driven through the entire circuit by the pressure-infuser. Flow was then stopped, the pressure-infuser removed and the piston pump installed as previously described. Eighty-five millilitres were then pumped through rapidly and discarded. Three samples were immediately taken from the cannula (running under pressure) at approximately 5-ml intervals. These represented blood that had run the entire way through the circuit driven by the piston pump. The samples were centrifuged, and the supernatant fluid analysed by spectral absorption (Modular Analytics Evo Solution; Roche Diagnostics, Penzberg, Germany) for the presence of free haemoglobin. A haemolysis index of 0–1000 was generated. The results were analysed by paired or unpaired two-tailed t-tests (Prism Software; GraphPad, La Jolla, CA, USA), with a value of p < 0.05 indicating statistical significance.
Results The piston pump was faster at infusing 500 ml saline (mean (SD) 80 (2) s) and 2000 ml saline in four 114
sequential 500-ml bags (352 (10) s than the pressureinfuser (113 (7) s, p < 0.0001 and 495 (19) s, respectively, p < 0.0001). Compared with the pressure-infuser, the time taken to replace an empty fluid bag with the piston pump system was significantly shorter (mean (SD) 10 (3) s vs 13 (3) s, p = 0.0036). The pressure-time curves for the piston pump and pressure-infuser are shown in Fig. 4. A rectangular hyperbola resulted when the peak intraluminal pressure was recorded against the time taken to empty a 60-ml syringe via the piston pump assembly (Fig. 5), from which a straight line of best fit with zero intercept of pressure/timereciprocal was plotted to derive a function relating peak pressure to time (pressure = 2318/time). This function is consistent with laminar flow in accordance with Poiseuille’s Law [12]. Compared with the pressure-infuser, there was no significant difference in the haemolysis index using the piston pump (mean (SD) 61.3 (1.2) vs 63.3 (1.5), p = 0.15).
Discussion The improvised piston pump proved effective in our simulated resuscitation scenario, and significantly outperformed the pressure-infuser. We noted time savings in the infusion of single intravenous bags, the time to changeover bags and cumulative savings over the full 2000-ml resuscitation. These results can be understood by considering the pressure driving the two infusions (Fig. 4). For © 2014 The Association of Anaesthetists of Great Britain and Ireland
Smart et al. | Piston pump delivery of intravenous fluids
Anaesthesia 2014, 69, 111–117
Peak pressure (mmHg)
700 600 500 400 300 200 100 0
0
5
10
15
20
Time to depress plunger (s)
Figure 5 Intraluminal peak pressure vs time to depress 60-ml plunger. clarity, the piston pump generates saw-tooth pressure waves, so the mean pressure is analysed. At the start of infusion, the piston pump assembly responds rapidly, with a sharp rise to plateau pressure. The pressure-infuser is inefficient by contrast, with a slow rise-time, and a tail-off (caused by the shape of the internal bladder). The piston pump’s steady pressure profile means that flow is driven consistently throughout. This efficiency produces disproportionately large time savings, with only a minimal increase in plateau pressure. Although the mean pressure of the cycle dictates flow, the peak pressure affects safety. This is discussed separately below. Peak pressure analysis can be used to study the fluid dynamics of the pump. By observing a simple reciprocal relationship between pressure and time (Fig. 5), we conclude that flow through the cannula remained consistently laminar according to Poiseuille’s Law [12]; this is desirable in reducing shear stress and haemolysis [13]. Ergonomically, the arrangement of the piston pump means no manipulation of the three-way tap is
necessary for either active or passive flow. When flow is passive, the syringe will automatically refill. The plunger is pushed back by hydrostatic pressure until arrested by the internal lip in the barrel. The full syringe is then bypassed, diverting all flow to the patient (Fig. 6). Several other features may be of particular interest to anaesthetists. The 60-ml syringe exceeds the volume of the basilic vein of the arm [14], so one plunger depression should flush intravenous drugs into the central circulation, even with sluggish venous return. Furthermore, as the syringe refills automatically, the anaesthetist is seldom without means to flush in drugs, even if the current intravenous bag is empty. The piston pump can generate enough pressure to cause barotrauma [15]. Intraluminal pressures as high as 635 mmHg resulted when a rapid push was attempted (Fig. 5). It should be understood, however, that the intraluminal pressure is not the same as the intravenous pressure. Intravenous pressure will always be lower, being downstream of the resistance of the cannula. Furthermore, the peak pressure of the pump cycle will be damped by the compliance of the venous plexus [12]. Evidence to back these assumptions comes from an experiment that used a similar cyclical pump (Manual Bulb Pump; Tuta laboratories, Lane cove, NSW, Australia) infusing via a 16-G cannula into a forearm vein [15]. Transducing via a second cannula inserted proximally, Goodie et al. measured the intravenous pressure generated by the pump. They found it to be consistently lower than 300 mmHg despite (in a previous experiment) having registered pressures exceeding 1000 mmHg within the intravenous line itself [16]. They noted
Figure 6 Schematic illustrating self-refilling under hydrostatic pressure. © 2014 The Association of Anaesthetists of Great Britain and Ireland
115
Anaesthesia 2014, 69, 111–117
that occluding the proximal run-off from the vein (by inflating a blood pressure cuff on the arm) caused a linear increase in venous pressure, in proportion to the pressure within the cuff. The maximum intravenous pressure they recorded was around 200 mmHg, with the blood pressure cuff inflated to 150 mmHg. We are reasonably reassured, therefore, that the syringe pump will not cause barotrauma as long as the venous run-off remains sufficient. There have been case reports of extravasation and compartment syndrome caused by bulb pumps, with sawtooth pressure profiles [17]. However, there are similar reports relating to inflatable pressure cuffs [18– 20] and even rapid infusers with occlusion alarms [7]. Cannula site vigilance is thus advised irrespective of pressurisation method. We suggest that any cannula attached to a piston pump should be placed in a large superficial vein (not a central vein) on the non-blood pressure cuff arm, and checked if resistance is felt during syringe plunger depression. Furthermore, the piston pump should only be used once passive flow under gravity has been demonstrated. This should ensure that there is no serious obstruction of venous return, and that the one-way valves are correctly oriented. Approximately 40–100 ml air is contained within a 1000-ml intravenous bag [7, 21]. This will not advance under gravity, but may do so under pressure. Serious air embolism can be caused by both inflatable and automated pressure-infusers [21, 22], and may occur when using a piston pump. As with any pressure infuser, ideally all air should be aspirated from the intravenous bag prior to infusion, although it has been noted that this is poorly complied with in practice [21]. We conclude that the piston pump is a useful adjunct to resuscitation, and could be employed in an emergency if a rapid infuser is unavailable. Its ability to deliver rapid, titrated boluses is suited to cardiac output monitoring. Being compact, simple, and only 39 g in weight, this device could also be employed in the pre-hospital setting, entrapment, or battlefield, where there may be limited space around the patient [23]. We commend its use to anaesthetists and would encourage colleagues to conduct further studies of its clinical safety and efficacy. 116
Smart et al. | Piston pump delivery of intravenous fluids
Acknowledgements The authors thank the anaesthesia department at Ayr Hospital for the kind donation of the materials necessary to conduct the research.
Competing interests No external funding and no competing interests declared.
References 1. Bree S, Nordmann GR, Wood K, McNicholas J. The paediatric transfusion challenge on deployed operations. Journal of the Royal Army Medical Corps 2010; 156: 361–4. 2. Holden HB, Maher JJ. Some aspects of blood loss and fluid balance in paediatric adenotonsillectomy. British Medical Journal 1965; 5474: 1349–51. 3. LaRocco BS, Henry E, Wang MD, Brian G. Intraosseous infusion. Prehospital Emergency Care 2003; 7: 280–5. 4. Macnab A, Christenson J, Findlay J, et al. A new system for sternal intraosseous infusion in adults. Prehospital Emergency Care 2000; 4: 173–7. 5. Hood JA, Wilson RJ. Pleth variability index to predict fluid responsiveness in colorectal surgery. Anesthesia and Analgesia 2011; 113: 1058–63. 6. Comunale ME. A laboratory evaluation of the Level 1 rapid infuser (H1025) and the Belmont Instrument fluid management system (FMS 2000) for rapid transfusion. Anesthesia and Analgesia 2003; 97: 1064–9. 7. Smith CE, Kabbara A, Kramer RP, Gill I. Evaluation of a new IV fluid and blood warming system to prevent air embolism. Trauma Care 2001; 11: 78–82. 8. Gan TJ, Soppit A, Maroof M, et al. Goal-directed intraoperative fluid administration reduces length of hospital stay after major surgery. Anesthesiology 2002; 97: 820–6. 9. Pearse R, Dawson D, Fawcett J, Rhodes A, Grounds RM, Bennett ED. Early goal-directed therapy after major surgery reduces complications and duration of hospital stay. A randomised, controlled trial. Critical Care 2005; 9: 687–93. 10. Drobin D, Hahn RG. Volume kinetics of Ringer’s solution in hypovolemic volunteers. Anesthesiology 1999; 90: 81–91. n C. Perioperative fluid management. Anes11. Prough DS, Svense thesia and Analgesia 2002; 10: 84–91. 12. Acheson DJ. Elementary Fluid Dynamics. Oxford: Clarendon Press, 1990. 13. Kamaneva MV, Burgreen GW, Kono K, Repko B, Antaki JF, Umezu M. Effects of turbulent stresses upon mechanical hemolysis: experimental and computational analysis. American Society for Artificial Internal Organs Journal 2004; 50: 418– 23. 14. Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium enhancement at CT. Part I. Prediction with a computer model. Radiology 1998; 207: 647–55. 15. Goodie DB. The effect of manual bulb pump infusion systems on venous luminal pressure and vein wall integrity. Anesthesia and Analgesia 1995; 80: 552–6. 16. Goodie DB, Morris RW. Bulb pump infusions and extravasation injury. Anesthesia and Analgesia 1994; 78: 812. 17. Sneyd JR, Lau W, McLaren ID. Forearm compartment syndrome following intravenous infusion with a manual ‘‘bulb’’ pump. Anesthesia and Analgesia 1993; 76: 1160–1. © 2014 The Association of Anaesthetists of Great Britain and Ireland
Smart et al. | Piston pump delivery of intravenous fluids 18. Seiler JG, Valadie AL, Drvaric DM, Frederick RW, Whiteside TE. Perioperative compartment syndrome. A report of four cases. Journal of Bone and Joint Surgery 1996; 78: 600–2. 19. Sung CY, Chung RK, Ra YS, Lee HS, Lee GY. Impending compartment syndrome of the forearm and hand after a pressurized infusion in a patient under general anesthesia-A case report. Korean Journal of Anesthesiology 2011; 60: 60–3. 20. Willsey DB, Peterfreund RA. Compartment syndrome of the upper arm after pressurized infiltration of intravenous fluid. Journal of Clinical Anesthesia 1997; 9: 428–30.
© 2014 The Association of Anaesthetists of Great Britain and Ireland
Anaesthesia 2014, 69, 111–117 21. Adhikary GS, Massey SR. Massive air embolism: a case report. Journal of Clinical Anesthesia 1998; 10: 70–2. 22. Sviri S, Woods WP, Van Heerden P. Air embolism – a case series and review. Critical Care and Resuscitation 2004; 6: 271– 6. 23. Pearce FJ, Lee J, Van Albert S. Closing the Loop on Critical Care Life Support for Military En-Route Care Environments. Silver Spring, MD: Walter Reed Army Institute of Research, 2004.
117
