Share this article on:

Drug Infusion System Manifold Dead-Volume Impacts the Delivery Response Time to Changes in Infused Medication Doses In Vitro and Also In Vivo in Anesthetized Swine

Lovich, Mark A. MD, PhD*; Wakim, Matthew G. BS*; Wei, Abraham BS*; Parker, Michael J. MD; Maslov, Mikhail Y. MD, PhD*; Pezone, Matthew J. BA*; Tsukada, Hisashi MD; Peterfreund, Robert A. MD, PhD§

doi: 10.1213/ANE.0b013e3182a76f3b
Technology, Computing, and Simulation: Research Report

BACKGROUND: IV infusion systems can be configured with manifolds connecting multiple drug infusion lines to transcutaneous catheters. Prior in vitro studies suggest that there may be significant lag times for drug delivery to reflect changes in infusion rates set at the pump, especially with low drug and carrier flows and larger infusion system dead-volumes. Drug manifolds allow multiple infusions to connect to a single catheter port but add dead-volume. We hypothesized that the time course of physiological responses to drug infusion in vivo reflects the impact of dead-volume on drug delivery.

METHODS: The kinetic response to starting and stopping epinephrine infusion ([3 mL/h] with constant carrier flow [10 mL/h]) was compared for high- and low-dead-volume manifolds in vitro and in vivo. A manifold consisting of 4 sequential stopcocks with drug entering at the most upstream port was contrasted with a novel design comprising a tube with separate coaxial channels meeting at the downstream connector to the catheter, which virtually eliminates the manifold contribution to the dead-volume. The time to 50% (T50) and 90% (T90) increase or decrease in drug delivery in vitro or contractile response in a swine model in vivo were calculated for initiation and cessation of drug infusion.

RESULTS: The time to steady state after initiation and cessation of drug infusion both in vitro and in vivo was much less with the coaxial low-dead-volume manifold than with the high-volume design. Drug delivery after initiation in vitro reached 50% and 90% of steady state in 1.4 ± 0.12 and 2.2 ± 0.42 minutes with the low-dead-volume manifold and in 7.1 ± 0.58 and 9.8 ± 1.6 minutes with the high-dead-volume manifold, respectively. The contractility in vivo reached 50% and 90% of the full response after drug initiation in 4.3 ± 1.3 and 9.9 ± 3.9 minutes with the low-dead-volume manifold and 11 ± 1.2 and 17 ± 2.6 minutes with the high-dead-volume manifold, respectively. Drug delivery in vitro decreased by 50% and 90% after drug cessation in 1.9 ± 0.17 and 3.5 ± 0.61 minutes with the low-dead-volume manifold and 10.0 ± 1.0 and 17.0 ± 2.8 minutes with the high-dead-volume manifold, respectively. The contractility in vivo decreased by 50% and 90% with drug cessation in 4.1 ± 1.1 and 14 ± 5.2 with the low-dead-volume manifold and 12 ± 2.7 and 23 ± 5.6 minutes with the high-dead-volume manifold, respectively.

CONCLUSIONS: The architecture of the manifold impacts the in vivo biologic response, and the drug delivery rate, to changes in drug infusion rate set at the pump.

From the *Department of Anesthesiology, Critical Care and Pain Medicine, Steward St. Elizabeth’s Medical Center; Harvard Medical School, Division of Pulmonary, Critical Care, and Sleep Medicine, Department of Medicine, Beth Israel Deaconess Medical Center; Division of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, Steward St. Elizabeth’s Medical Center; and §Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, Massachusetts.

Funding: In part by Doran International, American Heart Association Scientists Development Grant, and the Department of Anesthesiology, Critical Care and Pain Medicine, Steward St. Elizabeth’s Medical Center, Boston, MA.

Conflict of Interest: See Disclosures at the end of the article.

Reprints will not be available from the authors.

Address correspondence to Mark A. Lovich, MD, PhD, Department of Anesthesiology, Critical Care and Pain Medicine, Steward St. Elizabeth’s Medical Center, 736 Cambridge St., Boston, MA 02135-2997. Address e-mail to mark.lovich@steward.org.

Surgical and critical care patients commonly receive IV infusions of vasoactive, inotrophic, or sedative-analgesic compounds. Precise control of drug delivery is essential for maintaining the desired pharmacologic effect.1 Many drug infusions delivered by syringe or volumetric pump are configured with a crystalloid carrier flowing through a manifold that allows connection of multiple drug infusions to a single port of an indwelling percutaneous catheter. Adjustment of the dose rate of a drug infusion is accomplished by changing the settings of an infusion pump. We previously demonstrated that the amount of drug exiting the catheter does not immediately reflect the clinician’s intent when adjusting the infusion pump’s settings.2–4 This delivery response lag time results from delays in propagating the new dose through the infusion system. These propagation delays depend on the combined flow rates of carrier fluid plus drug solution, and the infusion system dead-volume measured from the point where drug and carrier streams meet to the patient’s blood.2 The infusion system dead-volume has contributions from the catheter and part of the manifold and is therefore sensitive to manifold design.4

Drug infusion manifolds typically consist of a number of ports for infused drugs to enter the carrier stream. Manifolds may take the form of several connected 3-way stopcock ports (Fig. 1A) allowing infusions to enter the carrier flow. This type of manifold structure contributes significantly to the infusion system dead-volume. Other manifold designs add modest amounts of dead-volume to the infusion system.4 One novel manifold design has 8 ports feeding channels running in parallel (Fig. 1B), such that all drug flows meet downstream at the connection to the percutaneous catheter (Fig. 1C).5 Since all fluid streams remain separate until merging at the tip of the manifold, the additional dead-volume from this device should be negligible.

Figure 1

Figure 1

We sought to confirm that infusion systems configured with large dead-volume manifolds will have longer response times to changes in set drug dosing than found in a novel, small-dead-volume manifold system. In a laboratory model, we compared the kinetics of drug delivery of a simple linear array of stopcocks (Fig. 1A) with this new coaxial manifold (Figs. 1, B and C). We measured the drug delivery rate while starting, stabilizing, and terminating an epinephrine infusion entering the main fluid pathway via each manifold. But infused drugs must circulate from their point of entry into the bloodstream to reach receptors and exert their effects. If the kinetics of drug circulation and action were slow relative to the lag times in delivering drug through the infusion system, the impact of the dead-volume may be less meaningful. We therefore extended the comparison of the effects of manifold design to a porcine model, measuring the kinetics of the myocardial contractile response to an epinephrine infusion. We demonstrate in vitro the impact of infusion system dead-volume on the rate at which drugs are administered into the circulation and show in vivo that such propagation delays have physiologic significance. Recognizing that the infusion system delays are minimized in clinical arenas where carrier flows are high, such as in the operating room, we chose to model typical carrier flows used in intensive care units (10 mL/h) where minimizing fluid becomes a paramount concern for patients with intravascular volume-limiting pathologies managed over days.

Back to Top | Article Outline

METHODS

A 16-gauge 20-cm single-lumen catheter (Arrow CS-04300, priming volume 0.22 mL, Research Triangle Park, NC) was used both in vitro and in vivo. A preconnected bank of 4 stopcocks (Codan in vitro Corporation, W20041, Santa Ana, CA) was used as the large-dead-volume manifold (Fig. 1A). The drug infusion line was connected to the most upstream stopcock; the total system dead-volume was approximately 1.27 mL. An Edelvaiss Multiline® drug infusion system (D2006140, Doran International, Toussieu, France) was used for the low-dead-volume manifold, in which all ports meet at the downstream Luer connector tip (Fig. 1B) adding no dead-volume to that of the catheter. One port of this manifold connects to a conduit formed from one of the 8 peripheral channels and the central lumen, creating a lower resistance pathway designed for gravity-driven carrier infusions. A normal saline carrier was driven by syringe pump (Medfusion 3500, Smiths Medical, Lower Pemberton, UK) at 10 mL/h through both manifold-catheter combinations. Epinephrine (430 µg/mL, Amphastar Pharmaceuticals, Rancho Cucamonga, CA) in normal saline was delivered at 3 mL/h by a syringe pump.

Back to Top | Article Outline

Delivery In Vitro

The drug infusion port of both manifolds was primed and vented in a manner that removes startup delays from the syringe pump drive mechanism and the syringe stopper and tubing compliances.6 Briefly, after priming of the port, the flow from the pump was established but diverted from the manifold by a vent consisting of stopcock attached to the port (Fig. 1A). At the time of drug initiation, this venting stopcock was turned to allow flow into the manifold. Samples were taken every minute from the distal tip of the catheter using a turntable fraction collector. After 30 minutes, while the carrier flowed uninterrupted, epinephrine was discontinued by venting drug away from the manifold. Samples were collected for an additional 30 minutes. These in vitro infusion experiments were repeated 8 times with each manifold.

Back to Top | Article Outline

Epinephrine Concentration Assay

The concentration of epinephrine in each sample was determined by spectrophotometric methods.7 Metaperiodate (6 µL of 2% NaIO4 in ddH2O, Sigma-Aldrich #S1878, St. Louis, MO) and ethanol (9 µL, 100%) were added to 60 µL taken from the samples and the absorbance at 490 nm measured. This concentration multiplied by the total flow rate was the drug delivery rate leaving the catheter. A standard curve was constructed by diluting epinephrine from the same stock used in the experiments.

Back to Top | Article Outline

Drug Delivery In Vivo: Surgical Preparation

All surgical procedures were approved by our Institutional Animal Care and Use Committee. Four adolescent Yorkshire swine (36–41 kg) were sedated, anesthetized, and catheterized as described in the Appendix.

The kinetics of pharmacologic response to drug infusion through high and low-dead-volume manifolds was compared 8 times in 4 animals. Power calculations assuming a 2-fold difference in response times with a standard deviation (SD) of 50% of the high-dead-volume response time suggested that 7 repetitions would be sufficient assuming a 5% chance of a type I error and 90% power.

The priming and venting procedures were identical to the in vitro measurements as were both drug and carrier flow rates. Epinephrine was diluted in normal saline in a 60-cc syringe (Becton-Dickenson) so that at 3 mL/h, the delivered dose rate would be 0.1 µg/kg/min. Drug infusion continued until the heart rate, mean arterial blood pressures, and max dP/dt signals came to steady state, approximately 21 and 30 minutes with the low and high-dead-volume manifolds, respectively. The epinephrine flow was then diverted away from the manifold, and data collection continued until another steady state was reached. In 2 animals, the sequence of testing was low-dead-volume manifold followed by high, high, and then low. In 2 other animals, the sequence was high-dead-volume manifold, followed by low, low, and then high. The animal was rested for 15 minutes between each infusion experiment. Each animal was killed with Euthasol™ 0.1 mL/kg after 4 measurements.

Back to Top | Article Outline

Analysis and Statistics

The time delay from the start of epinephrine flow until drug delivery (in vitro) or contractility (in vivo) started to increase was measured (T0), as was the time to 50% of maximal effect (T50), and time to 90% of maximal effect (T90). The differences between T0, T50, and T90 measurement for low- and high-dead-volume manifolds were compared using an unpaired 1-tailed t tests with Welch correction after verifying that all datasets were normal and that none of these times correlated with the order in which it was acquired in each animal (Prism 5.0, GraphPad Software, Inc., La Jolla, CA). The confidence intervals for the ratios of the T0, T50, and T90 between high- and low-dead-volume manifiolds were calculated by Prism 5.0 which uses Fieller's theorem on nonpaired data-sets that are normally distributed. The threshold for T0 was considered to be 5% of full scale effect for the initiation of drug infusion. Linear interpolation between 1 minute data points was used for in vitro drug delivery measurements. The temporal resolution of in vivo contractility data were much finer, 1 heartbeat, and therefore, no interpolation was needed to measure T0, T50, and T90. For the cessation of drug infusion, drug delivery or contractile response immediately decreased due to a reduction in total flow through the infusion system.2,3 A new plateau may be reached, and the time to a 5% reduction in this new value was considered to be T0 for cessation of drug flow. T50 and T90 reflected a 50% and 90% decrease in drug delivery rate (in vitro) or contractility (in vivo) from peak levels.

Back to Top | Article Outline

RESULTS

In Vitro Studies

The rates of drug exiting the distal tip of the catheter for both low and high-dead-volume manifolds are shown (Fig. 2). For the low-dead-volume manifold, epinephrine starts to exit the distal tip of the catheter in 1 minute and comes to steady drug delivery in 2 minutes (Table 1). In contrast, with the large manifold, drug does not exit the catheter until the 5th minute and does not reach steady state until up to 11 minutes. With the low-dead-volume manifold, when the drug flow is stopped, drug delivery rate out of the distal end of the catheter starts to decrease immediately and completely stops within 4 minutes. With the large manifold, on cessation of the drug flow, delivery decreases instantaneously to a new plateau that lasts for 5 minutes before it slowly tapers off over the next 20 minutes. The T0, T50, and T90 are shown for both initiation and cessation of syringe pump drug flow (Table 1). These findings are consistent with our previously published work.3 All of these times for the low-dead-volume manifold are shorter than for the high-dead-volume design.

Table 1

Table 1

Figure 2

Figure 2

Back to Top | Article Outline

In Vivo Studies

The kinetics of contractility measurements are similar to the kinetics of in vitro drug delivery data but with increased latency (Fig. 3). With the low-dead-volume manifold, contractility (max dP/dt) starts to increase within 2 minutes after the start of drug flow (Fig. 3A). By contrast, with the high-dead-volume manifold, 7 minutes pass before contractility starts to increase, and by this time, the contractility with the low-dead-volume manifold has almost reached peak effect. With both manifolds, contractility decreases to 86% of the peak effect within minutes of discontinuing the drug infusion (Fig. 3B). This early reduction in drug effect is due to the sudden decrease in total flow from 13 to 10 mL/h. With the low-dead-volume manifold, contractility continues to decrease, whereas with the high-dead-volume manifold, max dP/dt plateaus for 8 minutes before slowly decreasing back to baseline. T0, T50, and T90 data show that the times for pharmacologic response with the low-dead-volume manifold were significantly shorter than with the high volume design for both initiation and cessation of drug infusion (Table 2).

Table 2

Table 2

Figure 3

Figure 3

Back to Top | Article Outline

DISCUSSION

Prior analytic models of drug flow in infusion systems show that the time for drug administration to reach steady state after a change in drug or carrier flow rate is proportional to the dead-volume and inversely proportional to the total flow rate.2 In the present study, the flow rate of drug solution and crystalloid carrier was identical for all experiments. Thus, the response time should be proportional to the dead-volumes tested. The same 16-gauge single-lumen catheter was used in both infusion systems and has a priming volume of 0.22 mL. Infusion through the most upstream port of a 4-gang of stopcocks added 1.05 mL of dead-volume for a total of 1.27 mL, whereas the low-dead-volume manifold added essentially no additional dead-volume to that of the catheter. Thus, the dead-volumes tested were approximately 5-fold different. Indeed, the T50 and T90 for initiation of epinephrine infusion in vitro were similarly 4.6 to 5.5-fold (95% CI) and 3.7 to 5.4-fold (95% CI) less for the low-dead-volume manifold than for the high-volume design (Fig. 2, Table 1). The delays to changes in drug pump settings with the larger dead-volume infusion system design come from the prolonged time to propagate a new drug concentration down the manifold and catheter from the point where the carrier dilutes the drug.2

Under the conditions of these experiments, when drug infusion is terminated in vitro, drug delivery instantly decreases to 74% of the steady-state level for both low and high infusion system dead-volumes. This reflects a nearly identical instantaneous decrease in total flow. A plateau is maintained for about 5 minutes with the high-dead-volume manifold only, time needed for carrier fluid to flush through the infusion system. Overall, cessation of drug administration in vitro for the large-dead-volume manifold is approximately 4.8 to 5.8 (95% CI) and 4.0 to 5.8 (95% CI)-fold slower than for the low-dead-volume manifold, in terms of T50 and T90, respectively, and in approximate proportion to the infusion system dead-volumes studied (Table 1).

The biologic response in vivo with drug initiation or cessation is much faster for the low-dead-volume manifold than the high-dead-volume design (Fig. 3A, Table 2). The delay in the time to first observed biologic effect (T0) is 3.0 to 5.0 (95% CI)-fold less for the low-dead-volume than for the high-dead-volume infusion system. Likewise, the T50 and T90 are 2.1 to 3.2 and 1.3 to 2.4 (95% CI)-fold less with the low-dead-volume design, respectively. Cessation of drug leads to similar distinctions with the T0, T50, and T90 being 2.7 to 5.1 (95% CI), 2.3 to 3.7 (95% CI), and 1.2 to 2.3 (95% CI)-fold less with the low-dead-volume manifold, respectively. The ratios of these response times in vivo (between high/low dead-volume manifolds) are less than in vitro, indicating an additional delay to the biologic response other than propagation through the infusion system.

The fact that the in vivo response data are sensitive to manifold dead-volume has significant implications. The time course of pharmacologic response to catecholamine infusion is the combination of propagation through the infusion system and circulatory pharmacokinetics. Once the drug enters the circulation, we anticipate little difference in pharmacologic contractile response between the delivery conditions because the circulatory pharmacokinetics and receptor activation should be similar. If the response to the drug once in the circulation is very slow, circulatory pharmacokinetics would be the rate-limiting step in determining the time course of the pharmacologic response, obscuring the effect of propagation delays through the infusion system. Our data demonstrate that the T50s measured in vivo were 2 to 4 minutes more than in vitro. This difference can be interpreted as the approximate time for drug circulation and effect. This time interval is not long enough to overshadow the impact of delays in drug delivery to the circulation caused by drug traversing the infusion system dead volume.

Overall, our findings demonstrate that results of infusion experiments obtained in an in vitro model predict responses to drug infusions in vivo. The temporal delay for epinephrine propagation through the infusion system in response to a change in pump settings is a significant factor governing the time course of contractility responses. These data demonstrate how optimizing the architecture of a drug infusion system can lead to faster pharmacologic responses to changes in pump settings for drugs delivered by infusion.

Back to Top | Article Outline

APPENDIX

Surgical Preparation

All surgical procedures were approved by our Institutional Animal Care and Use Committee. Four adolescent Yorkshire swine (36–41 kg) were sedated with an IM injection of telazol 250 mg, ketamine 125 mg, and xylazine 125 mg before tracheal intubation with a 6.0-mm cuffed endotracheal tube. The animal was ventilated with 500 mL tidal volumes at 10 breaths per minute. The respiratory rate was adjusted to keep end-tidal carbon dioxide between 35 and 40 mm Hg (Poet-IQ gas monitor, Criticare Systems, INC, Waukesha, WI). Anesthesia was maintained with isoflurane 1% to 2% during catheterization. A femoral artery catheter (16-gauge, CS-04300, Arrow, Research Triangle Park, NC) was placed for monitoring and blood sampling. Introducer sheaths (9 FR) were placed in the contralateral femoral artery and right internal jugular vein. A pressure–volume conductance catheter (Ventri-cath 507, Millar Instruments, Houston, TX) was advanced retrograde into the left ventricle from the femoral artery and used to measure indices of myocardial contractility, max dP/dt. Calibration of this catheter was per the manufacturer’s recommendations. Real-time ventricular pressure volume loops were acquired to a computer for later analysis (Powerlab 16/30, ADI Instruments, Colorado Springs, CO). A continuous cardiac output pulmonary artery catheter was inserted (746HF8, Edwards Life Sciences, Irving, CA) via the internal jugular vein. A 16-gauge 20-cm single lumen catheter (Arrow, CS-04300) was placed in one of the femoral veins, flushed with heparinized saline (2 U/mL), and capped until connected to the high or low-dead-volume manifold.

Following catheterization, the isoflurane anesthetic was transitioned to IV anesthesia with midazolam (0.25 mg/kg/h), fentanyl (12–25 µg/kg/h), and ketamine (5 mg/kg/h). This anesthetic was allowed to stabilize for 30 minutes before initiating epinephrine infusion.

Back to Top | Article Outline

DISCLOSURES

Name: Mark A. Lovich, MD, PhD.

Contribution: This author helped design and conduct the study, analysed the data, and wrote the manuscript.

Attestation: Mark A. Lovich has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Conflicts of Interest: Mark A. Lovich received research funding from Doran International.

Name: Matthew G. Wakim, BS.

Contribution: This author helped conduct the study and analysed the data.

Attestation: Matthew G. Wakim has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

Name: Abraham Wei, BS.

Contribution: This author helped conduct the study and analysed the data.

Attestation: Abraham Wei has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

Name: Michael J. Parker, MD.

Contribution: This author helped design the study and write the manuscript.

Attestation: Michael J. Parker has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

Name: Mikhail Y. Maslov, MD, PhD.

Contribution: This author helped conduct the study.

Attestation: Mikhail Y. Maslov has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

Name: Matthew J. Pezone, BA.

Contribution: This author acquired in vitro data and assisted in statistical analyses.

Attestation: Matthew J. Pezone has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

Name: Hisashi Tsukada, MD.

Contribution: This author helped conduct the study.

Attestation: Hisashi Tsukada has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

Name: Robert A. Peterfreund, MD, PhD.

Contribution: This author helped design the study and write the manuscript.

Attestation: Robert A. Peterfreund has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Conflicts of Interest: The author has no conflicts of interest to declare.

This manuscript was handled by: Steven L. Shafer, MD.

Back to Top | Article Outline

REFERENCES

1. Cook RI. Syringe pump assemblies and the natural history of clinical technology. Can J Anaesth. 2000;47:929–35
2. Lovich MA, Doles J, Peterfreund RA. The impact of carrier flow rate and infusion set dead-volume on the dynamics of intravenous drug delivery. Anesth Analg. 2005;100:1048–55
3. Lovich MA, Peterfreund GL, Sims NM, Peterfreund RA. Central venous catheter infusions: a laboratory model shows large differences in drug delivery dynamics related to catheter dead volume. Crit Care Med. 2007;35:2792–8
4. Moss DR, Bartels K, Peterfreund GL, Lovich MA, Sims NM, Peterfreund RA. An in vitro analysis of central venous drug delivery by continuous infusion: the effect of manifold design and port selection. Anesth Analg. 2009;109:1524–9
5. Foinard A, Décaudin B, Barthélémy C, Debaene B, Odou P. The impact of multilumen infusion devices on the occurrence of known physical drug incompatibility: a controlled in vitro study. Anesth Analg. 2013;116:101–6
6. Tsao AC, Lovich MA, Parker MJ, Zheng H, Peterfreund RA. Delivery interaction between co-infused medications: an in vitro modeling study of microinfusion. Paediatr Anaesth. 2013;23:33–9
7. el-Kommos ME, Mohamed FA, Khedr AS. Spectrophotometric determination of some catecholamine drugs using metaperiodate. J Assoc Off Anal Chem. 1990;73:516–20
© 2013 International Anesthesia Research Society