Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
ADC Online First, published on January 30, 2014 as 10.1136/archdischild-2013-304887 Review
Intravenous drug delivery in neonates: lessons learnt Catherine M T Sherwin,1 Natalie J Medlicott,2 David M Reith,3 Roland S Broadbent3 1
Division of Clinical Pharmacology & Clinical Trials Office, Department of Pediatrics, University of Utah School of Medicine, Salt Lake City, Utah, USA 2 New Zealand’s National School of Pharmacy, University of Otago, Dunedin, New Zealand 3 Department of Women’s and Children’s Health, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand Correspondence to Dr Catherine M T Sherwin, Division of Clinical Pharmacology & Clinical Trials Office, Department of Pediatrics, University of Utah School of Medicine, 295 Chipeta Way, 2S010, Salt Lake City, UT 84108, USA; catherine.sherwin@hsc. utah.edu
To cite: Sherwin CMT, Medlicott NJ, Reith DM, et al. Arch Dis Child Published Online First: [please include Day Month Year] doi:10.1136/ archdischild-2013-304887
INTRODUCTION The intravenous route is important for the delivery of drugs to neonates.1 Although seemingly a simple and reliable method of drug administration, there are aspects of neonatal infusion set-ups that lead to potential problems for the smallest patients. Table 1 outlines problems related to intravenous drug administration in neonates.2–6 These include slow intravenous flow rates, significant dead space volumes, small drug dose volumes and limitations on the flush volume that can be safely given. Medications with narrow therapeutic indices (such as dopamine, aminoglycosides and insulin) present the greatest potential concern. Unaccounted variability in drug delivery through neonatal intravenous lines may be difficult to interpret in the clinical setting. Problems related to erratic, delayed or extended drug delivery are anticipated to be greatest in extremely low birth weight (ELBW, 75 min) using an infusion set-up for ELBW infants.5 This was due to the small dose volume (0.2 mL) added into a slow primary infusion (3.8 mL/h). However, other factors related to the intravenous line configuration may also be important. If gentamicin delivery continues for longer than expected, the distribution phase may not be completed by 60 min, time to maximum plasma concentration (Tmax) will be increased and peak (Cmax) concentrations may be unreliable because of inappropriate sampling times. The volume of tubing between the point of drug administration and end of the patient catheter is an important variable and primarily determined by component geometry. This fluid volume has been
Sherwin CMT, et al. Arch Dis Child(or 2014;0:1–5. doi:10.1136/archdischild-2013-304887 1 Copyright Article author their employer) 2014. Produced by BMJ Publishing Group Ltd (& RCPCH) under licence.
Drug therapy
Received 19 July 2013 Revised 18 December 2013 Accepted 13 January 2014
ABSTRACT Intravenous drug administration presents a series of challenges that relate to the pathophysiology of the neonate and intravenous infusion systems in neonates. These challenges arise from slow intravenous flow rates, small drug volume, dead space volume and limitations on the flush volume in neonates. While there is a reasonable understanding of newborn pharmacokinetics, an appreciation of the substantial delay and variability in the rate of drug delivery from the intravenous line is often lacking. This can lead to difficulties in accurately determining the pharmacokinetic and pharmacodynamic relationship of drugs in the smallest patients. The physical variables that affect the passage of drugs through neonatal lines need to be further explored in order to improve our understanding of their impact on the delivery of drugs by this route in neonates. Through careful investigation, the underlying causes of delayed drug delivery may be identified and administration protocols can then be modified to ensure predictable, appropriate drug input kinetics.
Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
Review Table 1 Problems and consequences of intravenous drug administration in neonates Problem Drug related Drug preservatives and contaminants Drug solution density Drug solution pH Drug–drug or drug–intravenous fluid incompatibility Layering (differences in specific gravity) Multiple intravenous drug therapy Small dosage volumes
Drug therapy
Delivery related Dead space at injection site Fluid flow dynamics Inline filters Inadequate mixing of additives Injection site location in intravenous line Intraluminal tubing diameter Phototherapy/light exposure Plastic intravenous line components Slow or stop/start infusion rates
Consequence
Risk of toxic reactions or intoxication Countercurrent flow Incompatibility issues Drug precipitation and/or inactivation Drug trapping and delay in delivery Possible fluid overload/incompatibility Measurement inaccuracy—‘volume of dose not measured correctly’ Drug trapping and potential overdose Layering and/or drug trapping Adsorption of drugs to membrane Possibility of overdosage Delay in drug delivery Pooling, minimal laminar flow recovery Possible photo-degradation of the drug Possible drug adsorption to plastic Delay in delivery, unpredictable effects
solutions along the infusion line. This includes the drug solution itself and any flush solutions that follow. The usual drug infusion time in neonates ranges from short (3–5 min) with a slow push of a small dose to longer (30 min— continuous infusion) durations when a syringe pump is used. Delays in drug effects (both on-effects and offeffects) are reported in anaesthetics as the pharmacological responses can be readily observed by the clinician.11 14 21 22 For other drugs where pharmacodynamics are not so easily observed (eg, aminoglycosides), any unintended delays or alteration in drug administration may be missed. In a study by Medlicott et al,6 retrograde flow of drug and/or flush solution was reported to have the potential to lead to erratic drug input kinetics. When this is recognised, it may be reduced by modifying drug administration protocols including strategic placement of one-way (or antireflux) valves in the line.18 Start-up times (times between start of infusion and effective administration) of infusion pumps are affected by dead space and low infusion rates. Start-up times vary widely between models, ranging anywhere from 7 min to 1 h at a 1 mL/h infusion rate.23 24 However, priming the pump by delivering a 2 mL fluid bolus prior to connection can significantly reduce the start-up delay regardless of the pump model.24 Additionally, start-up time can be improved by increasing flow rate (even from 0.1 to 1 mL/h) using a smaller syringe or a syringe plunger with reduced compressibility. However, altering these factors should be considered in light of the specific therapy being administered.25
WHAT ARE THE VARIATIONS SEEN IN INTRAVENOUS INFUSION SET-UPS? termed the infusion system ‘dead volume’ by Lovich et al.14 15 Neonatal intravenous line components have been designed with reduced dead volumes (eg, narrow lumen tubing) and these are common in neonatal units.16 17 A very low dead space volume (0.046 mL) has been shown to achieve the desired flow rate more rapidly compared with systems with a dead space volume ≥1.85 mL.18 19 These studies were conducted with primary fluid flow rates of 7 and 35 mL/h. Additionally, Lannoy et al20 demonstrated that very low dead space volumes also minimised the effects of stopping and restarting carrier fluid flow. Further to normal tubing volume, any variables that introduce additional volume into the intravenous line can contribute to the fluid bulk flow within the line and potentially drive administered drug
Table 2 Factors recognised to affect drug delivery (rate and extent) through neonatal intravenous infusion lines
2
Factor
Reference
Primary fluid flow rate Primary fluid composition (eg, dextrose 10%) Drug solution infusion rates Syringe pump start-up delay Flush solution volume Tubing volume between drug administration point and end of the catheter Loss of drug by adsorption onto inline filters, for example, insulin Geometry of drug administration ports Placement of antireflux valves into the infusion line
5 6 27 28 6 27 28 5 6 23 24 14 15 22 31 42 14–17
6 27 34–36 38 54
26 18
Injection ports, inline filters and other components within the infusion line can potentially add large volumes through which the drug solution must travel to reach the patient.16 17 The length of tubing between the injection port and the patient can be minimised if syringe drivers delivering slow infusions are placed close to the patient within the incubator. Different geometries are available for intravenous line components (eg, Y- or T-shaped injection ports) but little comparative information is reported about the efficiency of drug administration through these or their relative potential for dose sequestration or ‘trapping’.26 Some information is available regarding the component dead space volumes but these have not been systematically compared with to determine optimal combinations. If drug solutions become trapped within the recesses of an infusion line, there is a risk that it may be inadvertently administered at a later time. Three studies have demonstrated that differences in the specific gravity between an infused drug solution and the primary intravenous solution can significantly alter the time course of dose delivery, particularly at the slow infusion rates typically used for preterm newborns.6 27 28 Talon and Meiburg have described a fluid flow model for miscible layers of different viscosities showing the disruption to normal laminar flow conditions at low Reynolds values.29 We showed that in the absence of one-way flow valves in the primary line, gravity has a significant effect on the direction of travel of a drug solution infused into a slow (4 mL/h) primary line containing 10% dextrose.6 This suggests that the position of the infusion tubing in neonatal incubators may be an important but often overlooked parameter in the delivery of medications to neonates, especially when there are differences in the specific gravity of the two solutions. There is considerable variability in intravenous line configurations used at different hospitals, adding to the potential for variation in drug infusion characteristics. What we do know from Sherwin CMT, et al. Arch Dis Child 2014;0:1–5. doi:10.1136/archdischild-2013-304887
Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
Review the literature2 7 15 21 22 28 30–33 and from our own experimental work5 6 is that it is not possible to make assumptions regarding the drug delivery rate in neonates when the primary intravenous lines have very low flow rates.
WHAT ARE THE EFFECTS OF INLINE FILTERS?
WHAT IS THE RELEVANCE OF FLUID DYNAMICS WITHIN THE INTRAVENOUS LINE? Optimal drug delivery through intravenous lines requires infused drug solutions to move predictably through the intravenous line. When the appearance of a drug in the blood compartment is not predictable it may be important for the clinician to consider the impact of drug delivery on its therapeutic effectiveness before adjusting dosage regimens. For example, variability in aminoglycoside plasma concentrations is commonly attributed to inter-patient differences in drug clearance, sample collection or drug analysis.41 However, unintended delays in drug input may also contribute to this and affect peak plasma concentrations. Administering flush solutions after drug solutions is a usual practice to clear drug from the infusion port into the main stream of fluid. The flush volume may be chosen with guidance from studies that recommended at least 1 mL is needed to obtain full delivery of high-density drug solutions and to avoid their separation in the infusion stream into potential sequestration sites.31 42 However, when the combined volumes of the drug and flush solutions exceed the fluid tolerance for a neonate, then lower volumes must be used. Conversely, larger flush volumes, if tolerated, may be required if there is a clinical need to ensure that a drug enters the circulation within a defined time period. Although standard drug infusion arrangements are widely advocated,43 there are no systematic studies on the minimum flush volumes required to effectively flush infusion lines with differing configurations.
WHAT ARE THE DELIVERY KINETICS AND FLUID DYNAMICS THAT MATTER? Variability in drug delivery can be influenced by the placement of the infusion tubing. The practice of placing an infusion pump within the incubator to minimise tubing length also means that Sherwin CMT, et al. Arch Dis Child 2014;0:1–5. doi:10.1136/archdischild-2013-304887
CONTINUOUS INFUSIONS: WHAT ARE THE SPECIAL CONSIDERATIONS IN THE NEWBORN? Continuous infusions also have practical difficulties in the smallest patients. For example, vasoactive drug infusions such as dopamine require precise infusion rates and patients do not tolerate small interruptions of drug delivery or variations in the delivery rate. Continuous infusion pumps are connected to the main infusion as close to the patient as possible, which reduces drug delivery. The drug delivery rate to the patient is determined primarily by the drug solution infusion rate if adequate mixing into the primary fluid occurs. Perhaps the most important considerations for continuous infusions are (i) understanding factors that affect the time to onset of drug effect, (ii) time to transition to new infusion rates and (iii) time to offset of effect once the infusion is stopped. How to determine or even manipulate these intervals has not been well studied. There are some studies reported for delivery of anaesthetic agents through paediatric21 and neonatal45 infusion systems which suggest that intervals can be considerably long when the primary fluid flow rate is low. Syringe size has been shown to have an effect on the accuracy of low rate infusions,46 affect the time taken to achieve the target flow rate47 and also the impact of syringe ‘stiction’ (the static friction that can cause a syringe driver in an infusion pump system to stick). Stiction is one of the major sources of start-up delay, particularly at low flow rates.48 49 This delay can be reduced through the use of a smaller syringe, with micropiston syringes (0.8 mL pulse size). This has demonstrated smaller initial flow deviations on an infusion pump trumpet curve, which is used for accuracy determination.49 50
HEPARIN- OR SALINE-LOCKED LINES USED FOR INTRAVENOUS DRUG DOSING: WHAT ARE THE SPECIAL CONSIDERATIONS IN THE NEWBORN? When intravenous fluids are not required, intravenous drugs are often delivered using a heparin- or saline-locked line with no fluid being infused between drug doses. A recent study by Cook et al51 suggests that the use of heparin-locked lines is not necessary and should not be used in neonates as it leads to unnecessary exposure to heparin. The main consideration for this mode of drug delivery is to ensure complete drug delivery occurs during the infusion. The flush volume should reliably clear all drug solution out of the infusion system, so attention to the dead space volume is important. For incompletely flushed lines, the remaining drug may be infused the next time the line is used producing an unexpected bolus of the previous drug. 3
Drug therapy
Inline filters are known to adsorb insulin, hence insulin infusions are typically added below any inline filter.34–36 Inline filters are typically used to remove bacteria, endotoxins and any other particulates associated with intravenous therapy. Adsorption takes place rapidly and the drug is lost from solution until the available absorbing surface becomes saturated with drug.37 For potent low dose drugs, a significant proportion of the dose can be lost onto the filter. Some literature describes the binding of antimicrobials, particularly gentamicin, to inline filters.38 However, this information is likely to be filter-specific and is therefore difficult to generalise. We have shown that gentamicin does not bind significantly to Poisdyne Neo (PALL Corp.).6 Nazeravich and Otten27 have shown no loss with other filters available in the 1980s, but reported that inline filter position (horizontal of vertical) affected gentamicin recovery. The type of coating of the intravenous tubing can also play a significant role in drug infusions. Hooymans et al39 demonstrated that there was a considerable amount of adsorption of clonazepam to polyvinylchloride (PVC) tubing (∼60%) while there was no drug loss in a polyethylene tubing system. While few studies look at drug absorption in drugs relevant to neonates, there is evidence that insulin can adsorb to PVC tubing unless the system is properly preconditioned and flushed.40
infusions may be administered into tubing placed on inclined surfaces. While a difference in drug delivery delay has been observed between horizontally or vertically placed infusion tubing,27 a more subtle effect was observed by Medlicott et al6 for infusions administered into lines placed on a 15° inclined surface. In this study, there was a fluid density difference between the infused drug solution (density=1.00 g/cm3) and the 10% dextrose primary infusion (density=1.04 g/cm3). When the primary fluid flow rates were low (4 mL/h), a layer of the lower density fluid formed, which moved upwards (away from the patient) against the direction of the primary fluid and mixing of the layers occurred slowly.6 The flow characteristics of similar coaxially miscible layers in channels are described in fluid mechanics, which is used to enhance mixing and reduce the unwanted effect of layering29 44 but application to neonatal infusion lines has not been reported. Introduction of one-way (or antireflux) valves into the lines may reduce this problem.18
Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
Review HOW CAN INCOMPATIBILITY REACTIONS BE PREVENTED? There is a need to prevent incompatibility reactions, particularly at lower flow rates. Drug compatibility is a major issue with infusions in neonatal intensive care units with only 4% of coadministered drugs having known unrestricted compatibility.52 Incompatibility of co-infused drugs can lead to precipitation, variations in pH, drug degradation or even formation of toxic products, making it crucial to avoid.53 Foinard et al53 showed that at a minimum carrier flow rate of 30 mL/h, and use of a multilumen infusion device can reduce or prevent precipitation. While these results represent a step forward in prevention of co-infusion drug incompatibility, the required flow rates are currently too high for neonates.
Acknowledgements The authors would like to acknowledge Dr Michael Spigarelli of the Division of Clinical Pharmacology & Clinical Trials Office, Department of Pediatrics, University of Utah School of Medicine for informative discussions related to this work. Contributors All authors have made substantial contributions to the following. Conception and review of the literature: CMTS, NJM, DMR, RSB. Analysis and interpretation of data: CMTS, NJM, DMR. Drafting the article or revising it critically for important intellectual content: CMTS, NJM, DMR, RSB. Final approval of the version that is being submitted: CMTS, NJM, DMR, RSB. Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES 1
Drug therapy
WHAT NEEDS TO BE DONE TO REMOVE/REDUCE VARIABILITY IN DRUG DELIVERY? Low fluid tolerances, the associated inability to administer medications into infusion lines with high flow rates or use of large flush volumes presents a unique set of problems for reliably delivering intravenously administered medications to neonatal patients. Evaluating infusion protocols and practices in neonatal settings may give insight into unexpected therapeutic outcomes in clinical practice. Additionally, closer attention should be given to drug infusion rates for improved interpretation of neonatal pharmacokinetics and pharmacodynamic studies. Box 1 outlines evidence-based recommendations for the minimisation of adverse effects related to neonatal intravenous infusions. Future investigations should seek to examine delivery kinetics of aminoglycosides and other drugs such inotropes, insulin and caffeine where loss of drug in the infusion line or delays in the delivery of the drug may be misinterpreted as treatment failure. In addition to scientific investigation and consensus development, attention needs to be given to education of the neonatal workforce regarding the special hazards and difficulties associated with intravenous drug administration to neonates.
2 3 4 5
6 7 8 9 10 11 12 13
14
Box 1 Evidence-based recommendations for the minimisation of adverse effects related to neonatal intravenous infusions
15
16
▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ 4
Seek to minimise drug delivery interruptions Minimise variations in drug delivery rates Locate infusion pumps as close to the patient as possible When vasoactive medications are being infused, bolus doses of other medications must be given in another line. Use one-way valves to prevent retrograde drug flow When administering small molecular weight drugs at concentrations
ADC Online First, published on January 30, 2014 as 10.1136/archdischild-2013-304887 Review
Intravenous drug delivery in neonates: lessons learnt Catherine M T Sherwin,1 Natalie J Medlicott,2 David M Reith,3 Roland S Broadbent3 1
Division of Clinical Pharmacology & Clinical Trials Office, Department of Pediatrics, University of Utah School of Medicine, Salt Lake City, Utah, USA 2 New Zealand’s National School of Pharmacy, University of Otago, Dunedin, New Zealand 3 Department of Women’s and Children’s Health, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand Correspondence to Dr Catherine M T Sherwin, Division of Clinical Pharmacology & Clinical Trials Office, Department of Pediatrics, University of Utah School of Medicine, 295 Chipeta Way, 2S010, Salt Lake City, UT 84108, USA; catherine.sherwin@hsc. utah.edu
To cite: Sherwin CMT, Medlicott NJ, Reith DM, et al. Arch Dis Child Published Online First: [please include Day Month Year] doi:10.1136/ archdischild-2013-304887
INTRODUCTION The intravenous route is important for the delivery of drugs to neonates.1 Although seemingly a simple and reliable method of drug administration, there are aspects of neonatal infusion set-ups that lead to potential problems for the smallest patients. Table 1 outlines problems related to intravenous drug administration in neonates.2–6 These include slow intravenous flow rates, significant dead space volumes, small drug dose volumes and limitations on the flush volume that can be safely given. Medications with narrow therapeutic indices (such as dopamine, aminoglycosides and insulin) present the greatest potential concern. Unaccounted variability in drug delivery through neonatal intravenous lines may be difficult to interpret in the clinical setting. Problems related to erratic, delayed or extended drug delivery are anticipated to be greatest in extremely low birth weight (ELBW, 75 min) using an infusion set-up for ELBW infants.5 This was due to the small dose volume (0.2 mL) added into a slow primary infusion (3.8 mL/h). However, other factors related to the intravenous line configuration may also be important. If gentamicin delivery continues for longer than expected, the distribution phase may not be completed by 60 min, time to maximum plasma concentration (Tmax) will be increased and peak (Cmax) concentrations may be unreliable because of inappropriate sampling times. The volume of tubing between the point of drug administration and end of the patient catheter is an important variable and primarily determined by component geometry. This fluid volume has been
Sherwin CMT, et al. Arch Dis Child(or 2014;0:1–5. doi:10.1136/archdischild-2013-304887 1 Copyright Article author their employer) 2014. Produced by BMJ Publishing Group Ltd (& RCPCH) under licence.
Drug therapy
Received 19 July 2013 Revised 18 December 2013 Accepted 13 January 2014
ABSTRACT Intravenous drug administration presents a series of challenges that relate to the pathophysiology of the neonate and intravenous infusion systems in neonates. These challenges arise from slow intravenous flow rates, small drug volume, dead space volume and limitations on the flush volume in neonates. While there is a reasonable understanding of newborn pharmacokinetics, an appreciation of the substantial delay and variability in the rate of drug delivery from the intravenous line is often lacking. This can lead to difficulties in accurately determining the pharmacokinetic and pharmacodynamic relationship of drugs in the smallest patients. The physical variables that affect the passage of drugs through neonatal lines need to be further explored in order to improve our understanding of their impact on the delivery of drugs by this route in neonates. Through careful investigation, the underlying causes of delayed drug delivery may be identified and administration protocols can then be modified to ensure predictable, appropriate drug input kinetics.
Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
Review Table 1 Problems and consequences of intravenous drug administration in neonates Problem Drug related Drug preservatives and contaminants Drug solution density Drug solution pH Drug–drug or drug–intravenous fluid incompatibility Layering (differences in specific gravity) Multiple intravenous drug therapy Small dosage volumes
Drug therapy
Delivery related Dead space at injection site Fluid flow dynamics Inline filters Inadequate mixing of additives Injection site location in intravenous line Intraluminal tubing diameter Phototherapy/light exposure Plastic intravenous line components Slow or stop/start infusion rates
Consequence
Risk of toxic reactions or intoxication Countercurrent flow Incompatibility issues Drug precipitation and/or inactivation Drug trapping and delay in delivery Possible fluid overload/incompatibility Measurement inaccuracy—‘volume of dose not measured correctly’ Drug trapping and potential overdose Layering and/or drug trapping Adsorption of drugs to membrane Possibility of overdosage Delay in drug delivery Pooling, minimal laminar flow recovery Possible photo-degradation of the drug Possible drug adsorption to plastic Delay in delivery, unpredictable effects
solutions along the infusion line. This includes the drug solution itself and any flush solutions that follow. The usual drug infusion time in neonates ranges from short (3–5 min) with a slow push of a small dose to longer (30 min— continuous infusion) durations when a syringe pump is used. Delays in drug effects (both on-effects and offeffects) are reported in anaesthetics as the pharmacological responses can be readily observed by the clinician.11 14 21 22 For other drugs where pharmacodynamics are not so easily observed (eg, aminoglycosides), any unintended delays or alteration in drug administration may be missed. In a study by Medlicott et al,6 retrograde flow of drug and/or flush solution was reported to have the potential to lead to erratic drug input kinetics. When this is recognised, it may be reduced by modifying drug administration protocols including strategic placement of one-way (or antireflux) valves in the line.18 Start-up times (times between start of infusion and effective administration) of infusion pumps are affected by dead space and low infusion rates. Start-up times vary widely between models, ranging anywhere from 7 min to 1 h at a 1 mL/h infusion rate.23 24 However, priming the pump by delivering a 2 mL fluid bolus prior to connection can significantly reduce the start-up delay regardless of the pump model.24 Additionally, start-up time can be improved by increasing flow rate (even from 0.1 to 1 mL/h) using a smaller syringe or a syringe plunger with reduced compressibility. However, altering these factors should be considered in light of the specific therapy being administered.25
WHAT ARE THE VARIATIONS SEEN IN INTRAVENOUS INFUSION SET-UPS? termed the infusion system ‘dead volume’ by Lovich et al.14 15 Neonatal intravenous line components have been designed with reduced dead volumes (eg, narrow lumen tubing) and these are common in neonatal units.16 17 A very low dead space volume (0.046 mL) has been shown to achieve the desired flow rate more rapidly compared with systems with a dead space volume ≥1.85 mL.18 19 These studies were conducted with primary fluid flow rates of 7 and 35 mL/h. Additionally, Lannoy et al20 demonstrated that very low dead space volumes also minimised the effects of stopping and restarting carrier fluid flow. Further to normal tubing volume, any variables that introduce additional volume into the intravenous line can contribute to the fluid bulk flow within the line and potentially drive administered drug
Table 2 Factors recognised to affect drug delivery (rate and extent) through neonatal intravenous infusion lines
2
Factor
Reference
Primary fluid flow rate Primary fluid composition (eg, dextrose 10%) Drug solution infusion rates Syringe pump start-up delay Flush solution volume Tubing volume between drug administration point and end of the catheter Loss of drug by adsorption onto inline filters, for example, insulin Geometry of drug administration ports Placement of antireflux valves into the infusion line
5 6 27 28 6 27 28 5 6 23 24 14 15 22 31 42 14–17
6 27 34–36 38 54
26 18
Injection ports, inline filters and other components within the infusion line can potentially add large volumes through which the drug solution must travel to reach the patient.16 17 The length of tubing between the injection port and the patient can be minimised if syringe drivers delivering slow infusions are placed close to the patient within the incubator. Different geometries are available for intravenous line components (eg, Y- or T-shaped injection ports) but little comparative information is reported about the efficiency of drug administration through these or their relative potential for dose sequestration or ‘trapping’.26 Some information is available regarding the component dead space volumes but these have not been systematically compared with to determine optimal combinations. If drug solutions become trapped within the recesses of an infusion line, there is a risk that it may be inadvertently administered at a later time. Three studies have demonstrated that differences in the specific gravity between an infused drug solution and the primary intravenous solution can significantly alter the time course of dose delivery, particularly at the slow infusion rates typically used for preterm newborns.6 27 28 Talon and Meiburg have described a fluid flow model for miscible layers of different viscosities showing the disruption to normal laminar flow conditions at low Reynolds values.29 We showed that in the absence of one-way flow valves in the primary line, gravity has a significant effect on the direction of travel of a drug solution infused into a slow (4 mL/h) primary line containing 10% dextrose.6 This suggests that the position of the infusion tubing in neonatal incubators may be an important but often overlooked parameter in the delivery of medications to neonates, especially when there are differences in the specific gravity of the two solutions. There is considerable variability in intravenous line configurations used at different hospitals, adding to the potential for variation in drug infusion characteristics. What we do know from Sherwin CMT, et al. Arch Dis Child 2014;0:1–5. doi:10.1136/archdischild-2013-304887
Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
Review the literature2 7 15 21 22 28 30–33 and from our own experimental work5 6 is that it is not possible to make assumptions regarding the drug delivery rate in neonates when the primary intravenous lines have very low flow rates.
WHAT ARE THE EFFECTS OF INLINE FILTERS?
WHAT IS THE RELEVANCE OF FLUID DYNAMICS WITHIN THE INTRAVENOUS LINE? Optimal drug delivery through intravenous lines requires infused drug solutions to move predictably through the intravenous line. When the appearance of a drug in the blood compartment is not predictable it may be important for the clinician to consider the impact of drug delivery on its therapeutic effectiveness before adjusting dosage regimens. For example, variability in aminoglycoside plasma concentrations is commonly attributed to inter-patient differences in drug clearance, sample collection or drug analysis.41 However, unintended delays in drug input may also contribute to this and affect peak plasma concentrations. Administering flush solutions after drug solutions is a usual practice to clear drug from the infusion port into the main stream of fluid. The flush volume may be chosen with guidance from studies that recommended at least 1 mL is needed to obtain full delivery of high-density drug solutions and to avoid their separation in the infusion stream into potential sequestration sites.31 42 However, when the combined volumes of the drug and flush solutions exceed the fluid tolerance for a neonate, then lower volumes must be used. Conversely, larger flush volumes, if tolerated, may be required if there is a clinical need to ensure that a drug enters the circulation within a defined time period. Although standard drug infusion arrangements are widely advocated,43 there are no systematic studies on the minimum flush volumes required to effectively flush infusion lines with differing configurations.
WHAT ARE THE DELIVERY KINETICS AND FLUID DYNAMICS THAT MATTER? Variability in drug delivery can be influenced by the placement of the infusion tubing. The practice of placing an infusion pump within the incubator to minimise tubing length also means that Sherwin CMT, et al. Arch Dis Child 2014;0:1–5. doi:10.1136/archdischild-2013-304887
CONTINUOUS INFUSIONS: WHAT ARE THE SPECIAL CONSIDERATIONS IN THE NEWBORN? Continuous infusions also have practical difficulties in the smallest patients. For example, vasoactive drug infusions such as dopamine require precise infusion rates and patients do not tolerate small interruptions of drug delivery or variations in the delivery rate. Continuous infusion pumps are connected to the main infusion as close to the patient as possible, which reduces drug delivery. The drug delivery rate to the patient is determined primarily by the drug solution infusion rate if adequate mixing into the primary fluid occurs. Perhaps the most important considerations for continuous infusions are (i) understanding factors that affect the time to onset of drug effect, (ii) time to transition to new infusion rates and (iii) time to offset of effect once the infusion is stopped. How to determine or even manipulate these intervals has not been well studied. There are some studies reported for delivery of anaesthetic agents through paediatric21 and neonatal45 infusion systems which suggest that intervals can be considerably long when the primary fluid flow rate is low. Syringe size has been shown to have an effect on the accuracy of low rate infusions,46 affect the time taken to achieve the target flow rate47 and also the impact of syringe ‘stiction’ (the static friction that can cause a syringe driver in an infusion pump system to stick). Stiction is one of the major sources of start-up delay, particularly at low flow rates.48 49 This delay can be reduced through the use of a smaller syringe, with micropiston syringes (0.8 mL pulse size). This has demonstrated smaller initial flow deviations on an infusion pump trumpet curve, which is used for accuracy determination.49 50
HEPARIN- OR SALINE-LOCKED LINES USED FOR INTRAVENOUS DRUG DOSING: WHAT ARE THE SPECIAL CONSIDERATIONS IN THE NEWBORN? When intravenous fluids are not required, intravenous drugs are often delivered using a heparin- or saline-locked line with no fluid being infused between drug doses. A recent study by Cook et al51 suggests that the use of heparin-locked lines is not necessary and should not be used in neonates as it leads to unnecessary exposure to heparin. The main consideration for this mode of drug delivery is to ensure complete drug delivery occurs during the infusion. The flush volume should reliably clear all drug solution out of the infusion system, so attention to the dead space volume is important. For incompletely flushed lines, the remaining drug may be infused the next time the line is used producing an unexpected bolus of the previous drug. 3
Drug therapy
Inline filters are known to adsorb insulin, hence insulin infusions are typically added below any inline filter.34–36 Inline filters are typically used to remove bacteria, endotoxins and any other particulates associated with intravenous therapy. Adsorption takes place rapidly and the drug is lost from solution until the available absorbing surface becomes saturated with drug.37 For potent low dose drugs, a significant proportion of the dose can be lost onto the filter. Some literature describes the binding of antimicrobials, particularly gentamicin, to inline filters.38 However, this information is likely to be filter-specific and is therefore difficult to generalise. We have shown that gentamicin does not bind significantly to Poisdyne Neo (PALL Corp.).6 Nazeravich and Otten27 have shown no loss with other filters available in the 1980s, but reported that inline filter position (horizontal of vertical) affected gentamicin recovery. The type of coating of the intravenous tubing can also play a significant role in drug infusions. Hooymans et al39 demonstrated that there was a considerable amount of adsorption of clonazepam to polyvinylchloride (PVC) tubing (∼60%) while there was no drug loss in a polyethylene tubing system. While few studies look at drug absorption in drugs relevant to neonates, there is evidence that insulin can adsorb to PVC tubing unless the system is properly preconditioned and flushed.40
infusions may be administered into tubing placed on inclined surfaces. While a difference in drug delivery delay has been observed between horizontally or vertically placed infusion tubing,27 a more subtle effect was observed by Medlicott et al6 for infusions administered into lines placed on a 15° inclined surface. In this study, there was a fluid density difference between the infused drug solution (density=1.00 g/cm3) and the 10% dextrose primary infusion (density=1.04 g/cm3). When the primary fluid flow rates were low (4 mL/h), a layer of the lower density fluid formed, which moved upwards (away from the patient) against the direction of the primary fluid and mixing of the layers occurred slowly.6 The flow characteristics of similar coaxially miscible layers in channels are described in fluid mechanics, which is used to enhance mixing and reduce the unwanted effect of layering29 44 but application to neonatal infusion lines has not been reported. Introduction of one-way (or antireflux) valves into the lines may reduce this problem.18
Downloaded from adc.bmj.com on May 6, 2014 - Published by group.bmj.com
Review HOW CAN INCOMPATIBILITY REACTIONS BE PREVENTED? There is a need to prevent incompatibility reactions, particularly at lower flow rates. Drug compatibility is a major issue with infusions in neonatal intensive care units with only 4% of coadministered drugs having known unrestricted compatibility.52 Incompatibility of co-infused drugs can lead to precipitation, variations in pH, drug degradation or even formation of toxic products, making it crucial to avoid.53 Foinard et al53 showed that at a minimum carrier flow rate of 30 mL/h, and use of a multilumen infusion device can reduce or prevent precipitation. While these results represent a step forward in prevention of co-infusion drug incompatibility, the required flow rates are currently too high for neonates.
Acknowledgements The authors would like to acknowledge Dr Michael Spigarelli of the Division of Clinical Pharmacology & Clinical Trials Office, Department of Pediatrics, University of Utah School of Medicine for informative discussions related to this work. Contributors All authors have made substantial contributions to the following. Conception and review of the literature: CMTS, NJM, DMR, RSB. Analysis and interpretation of data: CMTS, NJM, DMR. Drafting the article or revising it critically for important intellectual content: CMTS, NJM, DMR, RSB. Final approval of the version that is being submitted: CMTS, NJM, DMR, RSB. Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES 1
Drug therapy
WHAT NEEDS TO BE DONE TO REMOVE/REDUCE VARIABILITY IN DRUG DELIVERY? Low fluid tolerances, the associated inability to administer medications into infusion lines with high flow rates or use of large flush volumes presents a unique set of problems for reliably delivering intravenously administered medications to neonatal patients. Evaluating infusion protocols and practices in neonatal settings may give insight into unexpected therapeutic outcomes in clinical practice. Additionally, closer attention should be given to drug infusion rates for improved interpretation of neonatal pharmacokinetics and pharmacodynamic studies. Box 1 outlines evidence-based recommendations for the minimisation of adverse effects related to neonatal intravenous infusions. Future investigations should seek to examine delivery kinetics of aminoglycosides and other drugs such inotropes, insulin and caffeine where loss of drug in the infusion line or delays in the delivery of the drug may be misinterpreted as treatment failure. In addition to scientific investigation and consensus development, attention needs to be given to education of the neonatal workforce regarding the special hazards and difficulties associated with intravenous drug administration to neonates.
2 3 4 5
6 7 8 9 10 11 12 13
14
Box 1 Evidence-based recommendations for the minimisation of adverse effects related to neonatal intravenous infusions
15
16
▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ ▸ 4
Seek to minimise drug delivery interruptions Minimise variations in drug delivery rates Locate infusion pumps as close to the patient as possible When vasoactive medications are being infused, bolus doses of other medications must be given in another line. Use one-way valves to prevent retrograde drug flow When administering small molecular weight drugs at concentrations
