Infusions of critical or noncritical medications often use multiaccess infusion sets, especially in the field of anesthesia and intensive care therapy. For example, total IV anesthesia and plasma or effect-site target-control infusion system often use simultaneous infusions of hypnotics, opioids, and muscle relaxants via a single IV access.1–8 IV infusions of anesthetics (i.e., remifentanil or propofol) or life-support drugs, such as vasoactive drugs require a stable infusion rate to maintain a stable effect. The crucial performance aspect of an infusion system is that the patients receive reliable, accurate drug delivery. The rate of infusion of individual drugs is often altered abruptly with the goal of rapid change in drug delivery to the patient. To be effective, the specific design of infusion sets must be considered to minimize a lag in response time from the change in drug flow. Many factors contributing to the inability of gravity-fed infusion devices to maintain a constant flow rate have been described.9,10 Three factors can affect the flow rate: volume of drop, impact of the decrease of the volume in the carrier infusion bag, and continuation change of the internal diameter of the tubing at the flow control clamp location (so called “creep” or “cold flow”).
Higher infusion set dead space volume may affect the time required to reach the desired concentrations. Lovich et al.11,12 precisely demonstrated the impact of carrier flow rate and infusion dead volume on the dynamics of IV drug delivery. Multiaccess infusion sets allow multiple simultaneous infusions but significant variations in the delivery rate of potent drugs are possible.
The aim of this study was to understand the influence of multiaccess infusion device properties, specifically dead space volume and the presence of an antireflux valve (ARV), on individual drug delivery during multi-infusion therapy. The main objective was to investigate the ability of an optimized multiaccess infusion system of pump and tubing to prevent drug delivery inaccuracies (lag time, backflow and bolus) induced by multi-infusion therapy.
Three drugs were infused simultaneously: noradrenalin (A) (Noradrénaline®, Merck Génerique, Lyon, France), midazolam (B), (Hypnovel®, Roche, Neuilly-sur-Seine, France), and isosorbide dinitrate (C) (Risordan®, Sanofi-Aventis, Paris, France). All drugs were used directly without further purification. An isotonic saline solution (NaCl 0.9%, Maco Pharma, France) was used to prepare dilutions. Final solutions of drugs were prepared in 50 mL syringes (Pentaferte, Villeparisis, France) as 276, 125, and 1000 μg/mL, respectively. There were no documented drug–drug interactions among these three drugs. These drugs were chosen based on the ability to analyze the drug concentrations in the combined infusion. They are used as a marker of drug delivery and could be replaced by other products. The saline solution used as carrier fluid was 0.9% NaCl (1 L bag, Maco Pharma, Tourcoing, France).
The three drugs were infused using syringe pumps (Pilote A2, Fresenius Vial, France; flow rate accuracy ±1% on drive mechanism and ±2% on syringes) connected to a multi-infusion device with a carrier fluid. This assessment focuses on the influence of two different gravity-fed infusion sets for the carrier fluid: one standard (M2) and one optimized (M5) (both from Doran International, Toussieu, France). These two infusion sets with similar internal diameter (2.8 mm) are composed of the same materials. M2 and M5 differ in overall length, presence, and position of the ARV and dead space volume. Dead space volume is defined as the internal volume of the administration set tubing from the point of drug infusion to the end of the tubing. M2 components were a sharp piercing spike with a bacterium-tight closable vent, a drip chamber, three-gang-manifolds at 130 cm from distal side (patient’s side) (V = 8.01 mL), and two three-way stopcocks at 100 cm from the distal end (V = 6.16 mL) (Fig. 1A). M5 components were a sharp piercing spike with a bacterium-tight closable vent, a drip chamber, two three-way stopcocks at 110 cm from the distal end (V = 6.39 mL), an ARV, and two flexible low dead volume connectors at the distal end (V = 0.046 mL) (Fig. 1B).
The important characteristics of these devices are listed in Table 1. For each experiment, a new 1 L NaCl 0.9% carrier fluid was used. It was hung 80 cm above the outlet from the infusion set. Special attention was brought to the absence of unwanted obstruction of the infusion line. An 18-gauge polyurethane angiocatheter (Sendal, Saint-Genis-Laval, France) was added at the distal end of each device.
Drug concentrations delivered to the end of the infusion set were determined using UV spectrophotometric analysis (Lambda 25, Perkin-Elmer, Cortaboeuf, France). All information from the spectrophotometer was collected with UV WinLab software (Perkin-Elmer, Cortaboeuf, France). A partial least square (PLS) method on UV spectra was used to simultaneously determine the three drug concentrations at the egress from the angiocatheter. PLS regression was obtained using a PLS module of XLStat software (Addinsoft, Paris, France). The 220- to 300-nm spectral zone was used to obtain the best model. The percentage of recovery was in the 99.5%–101.0% range. The limit of detection and limit of quantification of analytes in mixtures were 0.36 μg/mL limit of detection and 1.11 μg/mL (limit of quantification), 1.19 μg/mL and 3.61 μg/mL, 0.07 μg/mL and 0.22 μg/mL for isosorbide dinitrate, midazolam and noradrenaline, respectively. Samples (1 mL) were collected from the angiocatheter outflow using a fraction collector. Samples exceeding the spectrophotometer’s linear range were diluted in saline. This sensitive, fast, and simple dosage technique allows sample analysis collected at the angiocatheter egress every 30 s.
Four parameters were studied.
- The mass flow rate (expressed as μg/min) is defined as the amount of drug delivered to the patient per unit of time. This parameter is calculated from the drug concentrations measured in the samples collected at the end of the catheter.
- The mass flow rate plateau (expressed as μg/min) is defined as the mean amount of drug delivered to the patient per unit of time during the steady state (concentration equilibrium). For example, with a drug concentration of 10 mg/mL, and a flow rate of 5 mL/min, the theoretical mass flow rate plateau will be 50 (10 × 5) mg/min at steady state.
- To assess the evolution of the amount of drug delivered to the patient after start or change of flow rate, the flow change efficiency (FCE, in percentage) was calculated from the ratio of the area under the experimental mass flow rate curve to the area corresponding to theoretical instant mass flow rate curve after different durations beginning at the start of infusion and lasting 5, 10, or, 15 min thereafter (T = 5, 10, or 15 min) (Fig. 2). The FCE illustrates the ability of all the infusion material used to deliver the expected amount of drug to the patient.
- To assess the impact of infusion devices on mass flow rate plateau, the ratio of experimental mass flow rate plateau to theoretical mass flow rate plateau (RPe/Pt, in percentage) was calculated (Fig. 2). RPe/Pt illustrates the ability with which the infusion device is able to reach the theoretical value at drug mass flow rate plateau. A ratio <100% means that the quantity of drug delivered to the patient is less than the expected one highlighting the likelihood of a backflow through the infusion device. Therefore, 100% RPe/Pt means that the infusion device prevents backflow of the drug.
The experimental protocol was designed to assess the impact of multiaccess infusion devices on mass flow rate during simultaneously infused drugs. Two steps were performed; the first one to quantify possible differences in mass flow rate between M2 and M5 devices, and the second to observe if these differences were constant when outputs varied over time.
First Step: Evolution of Drug Delivery During Multi-Infusion Therapy
This experiment assessed the efficacy of M2 and M5 designs on drug mass flow rate and FCE of new and previous drug administration (Fig. 3A). Multi-infusions of the three drugs with an isotonic saline solution (carrier fluid) were performed using M2 and M5. Carrier fluid infusion rates were fixed at 90 mL/h throughout the experiment. The carrier flow was verified every 15 min by drop count. At T0, the first drug (C) (1000 μg/mL) was infused at a rate of 15 mL/h for 15 min, then at 10 mL/h for 10 min, and at 7 mL/h until the end. At T0 + 95 min, the second drug (B) administration (125 μg/mL) was started at an infusion rate of 7 mL/h until the end. At T0 + 155 min, the third drug (A) (276 μg/mL) was initiated at an infusion rate of 7 mL/h until the end. C was infused from the proximal access, B from the median access, and A from the distal access (patient’s side). Samples were collected at the catheter egress every 30 s during the first 10 min after the start of each drug infusion and then every 2 min. The total experiment time was 220 min. Five trials were performed for each infusion set, M2 and M5.
Second Step: Influence of Multi-Access Infusion Devices Properties on Drug Delivery After Flow Rate Change During Multi-Infusion Therapy
This setup compared M2 and M5 devices-related drug mass flow rate and FCE for Drug A flow rate changes when other drug administration regimens remained stable (Fig. 3B). Infusion of Drug A started (T0) 30 min after the beginning of the carrier fluid (90 mL/h), Drugs C (7 mL/h) and B (7 mL/h) delivery. At T0, Drug A (276 μg/mL) infusion started at a rate of 7 mL/h; at T0 + 10min, it increased to 11 mL/h and at T0 + 20 min, to 14 mL/h. At T0 + 30 min, Drug A infusion rate decreased to 11 mL/h and at T0 + 40min, to 7 mL/h. Samples were collected at the angiocatheter egress every 30 s. Five trials were performed for each infusion set, M2 and M5.
For both steps, samples were immediately analyzed using the UV analytical method previously described.
Mann-Whitney’s test was used to compare drug delivery values, RPe/Pt values, and FCE. P values ≤0.05 were considered statistically significant. The results are expressed as mean values ± sd.
There was no significant difference in mass flow rate for Drugs C and B between infusion devices (Figs. 4 and 5). Start of Drugs B and A infusions induced perceptible transient perturbations of Drug C delivery with M2 and M5 devices (Fig. 4). Start of Drug A infusion induced an unwanted bolus of Drugs B and C with M2 but not with M5. Analysis of Drug A mass flow rate showed significant differences between the two infusion devices during the first 6 min (Fig. 6). The time to reach the Drug A mass flow rate plateau was longer when using the M2 infusion set compared with M5 infusion set. Analysis of Drug A FCE indicated significant differences between the M2 and M5 sets during the first 10 min after the start of the Drug A infusion (Fig. 7). M5 gave a higher FCE than M2 (7.9-fold after 5 min; P = 0.008), 1.47-fold after 10 min (P = 0.016); this difference was no longer significant after 15 min of perfusion (P = 0.841). These results showed that the low dead space volume connectors on M5 achieved a mass flow rate plateau faster after the start of drug infusion when compared with an infusion set having larger dead space volume.
In the second step, variations in Drug A mass flow rate were more rapid with M5 than with M2. Using M5, Drug A reached the mass flow rate plateau more quickly and the level of the plateau was higher than with M2. At each stage of flow rate change, analysis of Drug A delivery showed significant differences between M2 and M5 for about the first 5 min. The results were similar whether the output was increased or decreased (Fig. 8). Mean RPe/Pt values from collected data during the first 10 min after flow rate change were significantly higher with M5 than with M2 (99.6% vs 92.4%; P < 0.0001) (Fig. 9).
To our knowledge, this is the first study designed to quantify the impact of the use of multiaccess infusion devices on in vitro drug infusion mass flow rate during multi-infusion therapy.
We found that drug mass flow rate is not stable in infusion outflow when using multiaccess infusion devices. Many studies have investigated the accuracy of gravity-flow IV infusion. Ziser et al.10 and Flack and Whyte13 identified many factors that affect the size of the drop in gravity infusion. Crass and Vance9 identified 22 factors that may alter flow rates in IV infusion systems and proposed a classification in four categories: factors related to the IV set, factors related to the IV fluid, patient-related factors, and others. They showed a mean difference of 33.6% between desired and measured volume of fluid delivered at 150 mL/h over a considered period. The originality of our work was to elucidate variations of drug delivery with pump syringes (i.e., constant outflows) when used simultaneously throughout multiaccess infusion devices with varying dead volumes. According to our experimental conditions, a sampling rate of one sample every 30 s seems to be adequate to detect rapid changes. Carrier fluid was administered by gravity in this study. Initiation of fluid via stopcock or sideport may induce transient alteration of the speed of infusion of carrier fluid. Nevertheless, the experimental protocol based on verification of carrier fluid flow every 15 min did not allow any conclusion about transient changes in carrier fluid flow. We observed a bolus effect and mass flow rate perturbations at the start of a new drug infusion. When adding a drug infusion to an IV fluid line, distal occlusion may lead to a backflow of the drug into the main line and subsequent administration of a potentially dangerous bolus dose or interaction with another drug. The lack of unwanted bolus after start of drug infusion at distal access with M5 devices may be due to the presence of ARV on this access that prevents the backflow effect. This observation supports the use of an ARV.
The second point to consider is that the dead space volume has considerable consequences on the evolution of drugs mass flow rate. Analysis of FCE showed significant differences between infusion devices. Optimized infusion devices with low dead volume connectors gave higher FCE with next to eightfold better efficiency versus nonoptimized ones having larger dead volume after 5 min. This result could have clinical consequences. When anesthesia is maintained using plasma or an effect-site target infusion system, target changes need to induce changes of the amount of drug delivered as soon as possible. Our results highlighted that low dead volume infusion devices allowed more rapid drug delivery variations compared with larger dead volume infusion sets. This issue seems very important to shorten the delay between target changes on the infusion pump and the intended clinical effects. Indeed, what could be the benefit of even the best model of target-controlled infusion if the quality of the infusion set is so poor that the drug delivery adjustment is delayed? Optimized infusion devices with low dead space volume connectors took less time to reach the mass flow rate plateau with approximately 5 min saved compared with nonoptimized devices. These results are in accordance with previous findings. Hutton and Thornberry14 suggested that drugs travel through extension tubing as a plug with dispersion caused by diffusive and shear forces. Lovich et al.12 suggested two models: a plug-flow model postulating that drug and carrier streams mix instantaneously and perfectly at their meeting point and a well-mixed model postulating that the concentration of a drug within the dead volume is always uniform. These models predicted a lag time in response to changes in carrier or drug flow, which is proportional to the dead volume and inversely related to the total flow rate.11,12 These results confirm the hypothesis of a link between variations in mass flow rate and dead volume. Modeling this relation for efficiency ≤100% may contribute to predicting efficiency according to dead volume and time. The model of a pharmacokinetic system with the assumption of an ideal infusion system may be corrupted by the lag time in response to changes in drug outflow. Application of control theory to the observed effects could establish the sensitivity of pharmacokinetic parameters to lag time and backflow and indirectly to dead space volume of infusion systems.
However, at the present time, our results cannot be extrapolated to other infusion conditions. Our results and conclusion need to be tested according to other flow rate and multi-infusion protocols.
The third point to consider is the impact of the ARV on drugs mass flow rate. ARV is used to prevent backflow of drug into the tubing of the gravity-fed infusion device when the downstream tube has become obstructed. After releasing an occlusion, in the absence of an ARV, a large stored volume of drug solution could be delivered to the patient. In our work, mean mass flow rate at the plateau was significantly higher with an ARV than without it. In our conditions of hydration (90 mL/h) and drug outflow (7 mL/h), the presence of an ARV in the direct upstream of the stopcock led to a transient increase of 8% in mean mass flow rate values at the plateau. This finding may have clinical issues. Optimized devices may prevent backflow of drugs that affects transient delivery of drugs. The lag between the targeted and obtained concentration could be increased when resistance in the distal part of the infusion line increases in some situations, such as obstruction of the venous catheter or infusion set collapse. In these cases, an increase in backflow occurs and can be efficiently prevented by the presence of an ARV, which is clearly mandatory when drugs are given using an infusion pump.
In conclusion, the design of multiaccess infusion devices can have a significant impact on drug delivery during multi-infusion therapy. Small dead volumes and the presence of an ARV improve the accuracy of drug delivery. The expected results of optimized devices are reduction of backflow, shorter response in drug delivery at distal access after flow rate change, and absence of an unwanted bolus of drugs infused at the proximal access at the start of drug infusion by distal access. Care providers must consider dead volume and the presence of an ARV when choosing their infusion devices.
The authors thank Prof. Alain D’Hollander (Department of Anesthesia and Intensive Care, Hôpital Universitaire de Genève, Switzerland) for reviewing the manuscript and Doran International (Toussieu, France) for providing scientific and technological expertise as well as infusion sets (Edelvaiss ™®).
1. Lauven PM, Stoeckel H, Schwilden H. A microprocessor controlled infusion scheme for midazolam to achieve constant plasma levels. Anaesthesist 1982;31:15–20
2. White PF. Clinical uses of intravenous anesthetic and analgesic infusions. Anesth Analg 1989;68:161–71
3. Shafer SL, Gregg KM. Algorithms to rapidly achieve and maintain stable drug concentrations at the site of drug effect with a computer-controlled infusion pump. J Pharmacokinet Biopharm 1992;20:147–69
4. Bailey JM. A technique for approximately maintaining constant plasma levels of intravenous drugs. Anesthesiology 1993;78: 116–23
5. Miller DR. Intravenous infusion anaesthesia and delivery devices. Can J Anaesth 1994;41:639–51
6. Gepts E. Pharmacokinetic concepts for TCI anaesthesia. Anaesthesia 1998;53:4–12
7. Russell D. Intravenous anaesthesia: manual infusion schemes versus TCI systems. Anaesthesia 1998;53:42–5
8. Van Poucke GE, Bravo LJ, Shafer SL. Target controlled infusions: targeting the effect site while limiting peak plasma concentration. IEEE Trans Biomed Eng 2004;51:1869–75
9. Crass RE, Vance JR. In vivo accuracy of gravity-flow i.v. infusion systems. Am J Hosp Pharm 1985;42:328–31
10. Ziser M, Feezor M, Skolaut MW. Regulating intravenous fluid flow: controller versus clamps. Am J Hosp Pharm 1979;36: 1090–4
11. 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
12. Lovich MA, Kinnealley ME, Sims NM, Peterfreund RA. The delivery of drugs to patients by continuous intravenous infusion: modelling predicts potential dose fluctuations depending on flow rates and infusion system dead volume. Anesth Analg 2006;102:1147–53
13. Flack FC, Whyte TD. Variations of drop size in disposable administration sets used for intravenous infusion. J Clin Pathol 1975;28:510–2
© 2009 International Anesthesia Research Society
14. Hutton P, Thornberry EA. Factors affecting delivery of drug through extension tubing. Br J Anaesth 1986;58:1141–8