Pediatric Anesthesia ISSN 1155-5645

ORIGINAL ARTICLE

Quantitative analysis of continuous intravenous infusions in pediatric anesthesia: safety implications of dead volume, ow rates, and uid delivery
Haobo Ma1, Mark A. Lovich2 & Robert A. Peterfreund1
1 Department of Anesthesia, Critical Care, and Pain Medicine, Massachusetts General Hospital, Boston, MA, USA 2 Department of Anesthesiology and Pain Medicine, Caritas Saint Elizabeths Medical Center, Boston, MA, USA

Keywords dead volume; drug infusions; propofol; remifentanil; uid management Correspondence Robert A. Peterfreund, Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Jackson 439, 55 Fruit Street, Boston, MA 02114, USA Email: RPeterfreund@Partners.org Section Editor: Andrew Davidson Accepted 28 October 2010 doi:10.1111/j.1460-9592.2010.03475.x

Summary Objective: Quantitative characterization of continuous pediatric drug infusions. Background: The dynamics of drug delivery by continuous infusion to pediatric patients have not been systematically examined. This study extends previously described analytic models to propofol and remifentanil delivery, focusing on infants and toddlers. We postulated that infusion system dead volume, and drug and carrier ow rates, signicantly inuence drug delivery. Methods: We studied effects of patient weight, infusion system dead volume, drug and carrier ow rates, along with drug stock concentration and dose, on propofol and remifentanil delivery to the circulation. We calculated the drug mass available for inadvertent bolus in the dead volume, the volume of uid supplied by drug infusions, and model-based estimates of the range of lag times to achieve a targeted steady-state rate of drug delivery. Results: The drug mass in the dead volume at steady state increased with dead volume size and drug dose. For infants, this drug mass could exceed 100% of commonly used loading doses. Predicted lag times to steady state depend on patient size, uid ow rates, and the mixing behavior of the drug entering the main uid pathway. Neonates have the longest lag times to achieve steady state. Fluid quantities delivered by drug infusions increase with drug ow rate and can represent a large fraction of estimated maintenance uid requirements. Fluid delivery increases if stock drug concentrations are diluted. These relationships were qualitatively similar for propofol and remifentanil. Conclusions: Traditional studies focus on drug disposition once a drug enters the circulation. Our analysis shows the potential importance of factors inuencing drug delivery to the patients circulation, focusing on propofol and remifentanil administration to small patients. The drug mass available for inadvertent bolus residing in the reservoir of the dead volume at steady state may be large and clinically relevant. Lag times to achieve steady-state delivery are long, depending on the infusion systems architecture and uid ow rates. By themselves, drug infusions can deliver signicant uid loads to children. These observations have practical and perhaps safety implications for infusions of drugs commonly administered to infants and children.

78

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

H. Ma et al.

Quantitative analysis of pediatric intravenous infusions

Introduction Infants and children commonly receive vasopressor, opioid, or sedative/hypnotic medications by continuous intravenous infusion. Infusion pumps drive infusions at specied rates based on weight with adjustments for clinical responses. When the child has only one intravenous access, infusions join the main uid pathway via side ports, stopcocks, or manifolds with the main uid ow serving as a carrier. From their entry points into the carrier ow, the infused drugs must traverse a volume in the uid pathway before reaching the childs blood stream, dened as the infusion system dead volume (Vd) (1). Focusing mainly on adults, our previous work revealed a complex interplay between drug infusion rate, carrier ow rate, and dead volume during vasoactive drug delivery (25). We hypothesized that concepts developed from these investigations would also apply to other drugs commonly used in the care of children such as remifentanil and propofol. Therefore, we explored the effects of dead volume and infusion rates on the delivery of propofol and remifentanil, specically focusing upon infants and toddlers.

Methods We dened the weight range of interest as 350 kg, and we assumed stock concentrations for propofol (10 mgml)1) and remifentanil (50 mcgml)1) for initial analysis. Infusion sets We assumed a typical piggyback drug infusion setup in which drugs enter a carrier uid line by a side port, a manifold, or a stopcock. With calibrated syringes, we measured the dead volumes from the possible entry points in two representative pediatric infusion setups (Figure 1) and used these values in subsequent analyses. Basic calculation formulas The dead volume, Vd, will contain a mass of the infused remifentanil and propofol. The mass of the drug in the Vd during steady-state drug infusion (Mv, units of mass) is proportional to the size of the Vd. It is also inuenced by the drug concentration (Cd, units of mass/volume), the infused dose (Id, units of mass/ time), and the drug (Qd) and carrier (Qc) ow rates

(a)

(b)

5 1

Figure 1 Photographs of two typical pediatric infusion sets. Two infusion sets are commonly used in pediatric practice in our hospitals. They are (a) the 60 drop ARROW IV set (ARROW W21560 Universal 60 Drop IV set; Arrow International, Inc., Reading, PA, USA) and (b) the LifeShield Burette set (LifeShield 150 mL Burette Set 19208-01; Hospira, Inc., Lake Forest, IL, USA). In practice, clinicians usually connect a stopcock and a T-piece connector (LifeShield Microbore Extension Set 11572-18; Hospira, Inc.) at the end

of these sets to provide a less bulky connection for the intravenous catheter and to provide an additional access port close to patient. These sets are representative of other commonly employed clinical infusion sets. (a) 60 drop ARROW IV set with T-connector and a stopcock (distal portion). The dead volume from 1 to 2 is approximately 0.5 ml, and from 1 to 3, it is approximately 2.5 ml. (b) LifeShield Burette set with T-connector and a stopcock. The dead volume from site 4 to site 5 is approximately 1.5 ml.

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

79

Quantitative analysis of pediatric intravenous infusions

H. Ma et al.

(units of volume/time). Formulas 1 and 2 relate these factors (2). Mv Vd Id Qd Qc 1

carrier ow. The drug mixture then travels downstream as a plug or packet. From the Plug-Flow model, one time constant (s) denes the lag time to steady-state infusion (Formula 5) (1,2). s Vd Qd Qc 5

The ow of drug, Qd, is equal to the drug dose divided by the drug concentration. Qd Id Cd 2

Formula 1 can be modied to accommodate weightbased drug dosing where Id is the weight-based infusion dose (e.g., mcgkg)1min)1) (Formula 3). Mv Vd Id0 Wt Qd Qc 3

To account for uid administration adjusted for patient size, we assumed that the carrier ow rate, Qc, equals the weight-based maintenance uid rate dened by the 4-2-1 rule, which is used at many institutions to guide initial uid therapy (see Appendix S1, Table 1) (6,7). We derived three formulas for Mv depending on weight ranges (See Appendix S1, Table 2). Mv is independent of weight when the patient weighs 10 kg or less because drug dose and uid administration increase proportionally [See Appendix S1, Table 2, equation (a)]. If the loading dose of a drug is C Wt [C is a constant (units of drug mass/patient weight), and Wt is the patient weight in kilograms], we can calculate the ratio of Mv, the mass of drug in the dead volume at weight-based steady-state infusion, to the typical loading (induction) dose of the drug (Formula 4). Mv-to-loading dose ratio Mv C Wt 4

In contrast to the Plug-Flow model, the Well-Mixed model assumes instant, uniform, mixing of the drug and carrier uids throughout the dead volume when the drug infusion joins the carrier ow. The time predicted to achieve steady-state infusion according to the Well-Mixed model is approximately three times that predicted by the Plug-Flow model (1,2,4). Data suggest that the real lag time is between one and three time constants, or between s and 3s (1,3,4). Drug infusions also contribute to the total uid load administered, and this is readily calculated from the drug dose and concentration (Formula 2). Formula 2 can then be modied to determine the uid load (see Appendix S1, Formula 7). Results Dead volume measurements The measured Vd from the distal tip of the T-connector (Hospira, Inc., Lake Forest, IL, USA) to the nearest upstream stopcock (including the stopcock dead volume) is 0.5 ml. Neglecting the contribution to Vd from the intravascular catheter, modest for 24- or 22-g devices, this is the minimum dead volume for the uid pathway in this infusion system architecture. Measuring from the distal tip of the T-connector to the nearest side port access in the 60 drop ARROW IV set (Arrow International, Inc., Reading, PA, USA) and the burette set (Hospira, Inc.), the dead volumes were 2.5 and 1.5 ml, respectively. We therefore used a Vd range of 0.52.5 ml in subsequent modeling analyses, i.e., 0.5 ml representing the closest possible junction and 2.5 ml representing a more distant access port for one standard Arrow IV system. Drug mass residing in the dead volume at steady-state delivery (Mv) Effects of Vd. Our analysis reveals that Mv increases with Vd at constant patient weight and infusion dose (Figure 2a,b). For example, in a 3 kg infant, when propofol infuses at a steady-state rate of 250 mcgkg)1min)1 with the IV carrier ow (Qc) set at the maintenance rate of 12 mlh)1 (Figure 2a), a Vd of 0.5 ml contains

The Vd also inuences the time course to establish steady-state delivery when starting or changing the infusion. We previously demonstrated in laboratory studies unexpectedly long lag times to achieve targeted delivery rates for infusions of vasopressors, particularly with central line catheters (3,5). Two conceptual models describe the mixing behavior of an infused drug with the carrier uid when these two ows join in the dead volume. These models assume forward ow only (no retrograde ow). The Plug-Flow model assumes that the streams of drug and carrier mix perfectly just at the point of entry of the drug infusion into the

80

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

H. Ma et al.

Quantitative analysis of pediatric intravenous infusions

1.4 mg of propofol. If the larger Vd of 2.5 ml is used, the uid pathway contains 6.8 mg of propofol. The predictions for remifentanil are qualitatively similar (see supplemental Figure S1). For example, if remifentanil infuses at 0.15 mcgkg)1min)1 into a 3 kg infant, the Vd of 0.5 ml will contain 1.1 mcg of remifentanil; at a Vd of 2.5 ml, the remifentanil Mv is 5.4 mcg. Effects of infusion dose. As the infusion dose increases, Mv also rises (Figures 2b). For a 3 kg infant, increasing the propofol infusion dose to 400 from 250 mcg kg)1min)1 raises the Mv from 6.8 to 9.4 mg in a Vd of 2.5 ml at steady-state infusion (Figure 2b). Assuming a dead volume of 2.5 ml, raising the infusion rate of remifentanil to 0.3 from 0.15 mcgkg)1min)1 increases the Mv from 5.4 to 10.3 mcg (see supplemental Figure S1).
Id = 250 (mcgkg1min1) Cd = 10 mgml1 Qc = maintenance fluid rate Vd = 0.5 ml 8 Vd = 1 ml 6 4 2 0 0 10 20 30 40 50 Vd = 1.5 ml Vd = 2 ml Vd = 2.5 ml

Effects of patient weight. Patient weight also affects the Mv in weight-based propofol and remifentanil infusions. Because dosing is in units of masskg)1 per unit of time, a larger quantity of drug infuses per time unit as patient weight rises. However, carrier ow rates set to the weight-based maintenance level may also rise along with patient weight. As carrier ow rates increase, the drug mass, Mv, is diluted. The two factors, drug mass delivery and carrier ow rate, interplay such that for children weighing 10 kg or less, the Mv will not change with weight. Assuming a Vd of 2.5 ml and the weight-based propofol infusion dose of 250 mcgkg)1 min)1, Mv is the same (6.8 mg) for two infants weighing 3 and 5 kg, respectively (Figure 2a). When the childs weight exceeds 10 kg, Mv rises with weight (Figure 2a). Under the same propofol dosing and dead volume conditions, however, the Mv for a 30 kg child is 9.8 mg, while the Mv for a 50 kg child is 11.4 mg.
Wt = 3 kg Cd = 10 mgml1 Qc = 12 mlh1 Vd = 0.5 ml 6 4 2 0 0
50

(a)
Propofol mass in dead volume (mg)

12 10

(b)
Propofol mass in dead volume (mg)

10 8

Vd = 1 ml Vd = 1.5 ml Vd = 2 ml Vd = 2.5 ml

100

150

200

250

300

350

400

Patient weight (kg)

Id = 250 (mcgkg1min1) Cd = 10 mgml1 Qc = maintenance fluid rate Propofol loading dose = 2 mgkg1

Propofol infusion dose (mcgkg1min1)

Wt = 3 kg Cd = 10 mgml1 Qc = 12 mlh1 Propofol loading dose = 2 mgkg1

(c) 120
Ratio of propofol Mv to loading dose (%)
100 80

(d) 160
Vd = 0.5 ml 140

Vd = 0.5 m Vd = 1 ml Vd = 1.5 m Vd = 2 ml Vd = 2.5 m

Ratio of propofol Mv To loading dose (%)

Vd = 1 ml Vd = 1.5 ml 60 40 20 0 0 10 20 30 40 50 Vd = 2 ml Vd = 2.5 ml

120 100 80 60 40 20 0 0
50

100

150

200

250

300

350

400

Patient weight (kg)

Figure 2 Propofol in the dead volume. (a) The mass of propofol in the dead volume (Mv) at steady-state total intravenous anesthesia infusion as a function of patient weight. The propofol infusion rate (Id) is assumed to be 250 mcgkg)1min)1. Mv is independent of patient weight when the weight is 10 kg, if carrier ow (Qc) is set to the maintenance uid rate based on the 4-2-1 rule (see text for details). (b) Mv as a function of propofol infusion dose. The patient weight (Wt) is assumed to be 3 kg. Qc is the maintenance uid rate (12 mlh)1). (c) The ratio of the propofol Mv to the loading

Propofol infusion dose (mcgkg1min1)

dose, as a function of patient weight. The weight-based propofol infusion dose (Id) is assumed to be 250 mcgkg)1min)1. Qc is the maintenance uid rate. (d) The ratio of the propofol Mv to the loading dose, as a function of propofol infusion dose. The patient weight (Wt) is assumed to be 3 kg. Qc is the maintenance uid rate (12 mlh)1). Conditions: Dead volumes used in calculations are 0.5, 1, 1.5, 2, and 2.5 ml. The propofol stock drug concentration (Cd) is 10 mgml)1. The propofol loading dose is dened as 2 mgkg)1.

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

81

Quantitative analysis of pediatric intravenous infusions

H. Ma et al.

Implications of Mv in relation to patient weight The drug mass in the Vd is available for inadvertent bolus. The clinical impact of any particular bolus depends on the childs weight. To account for the patients weight, we derived and modeled a new parameter, the ratio of Mv to the drugs weight-based loading dose (Formula 4). We specied the propofol loading dose as 2 mgkg)1 and the remifentanil loading dose as 1 mcgkg)1. Converting the absolute mass in the dead volume to the ratio of Mv to weight-based loading dose reveals the potential clinical impact of an inadvertent bolus of the Vd contents. At a steady-state propofol infusion of 250 mcgkg)1min)1, a Vd of 2.5 ml contains 6.8 mg of the drug, which represents 114% of loading dose for a 3 kg patient (Figure 2c). For a 50 kg child receiving the same infusion dose of 250 mcgkg)1 min)1, the same 2.5 ml Vd contains propofol, 11.4 mg, which is only 11.4% of a loading dose for that size child (Figure 2c). Mv-to-loading dose ratios are even larger for higher-infusion doses (Figure 2d). Similarly, a 2.5 ml Vd contains remifentanil, 5.4 mcg, during steady-state infusion of 0.15 mcgkg)1min)1 to a 3 kg infant; this represents 180% of the loading dose for that child. The same Vd contains 11.4 mcg remifentanil when a 50 kg child receives the infusion; the potential inadvertent bolus represents 23% of the loading dose (see supplementary Figure S1). Lag time to achieve steady-state drug delivery When starting drug infusions or changing the infusion dose, there is an obligate lag to achieve steady-state drug delivery because of dead volume effects. From the Plug-Flow model, the lower limit of time required to achieve steady-state delivery is one time constant. The upper limit, derived from the Well-Mixed model, is approximately three time constants to achieve 95% of steady-state delivery. As predicted by Formula 5, the time constant lengthens with larger dead volumes (Figure 3) or slower ow rates. Carrier and drug infusion ows are relatively slow for smaller compared to larger children; thus, the time constant will be longer in infants (Figure 3a,b). For example, in a 3 kg infant with carrier owing at 12 mlh)1 and an infusion Vd of 2.5 ml, up to 27 min would be required to reach steadystate propofol delivery as predicted by the Well-Mixed model when starting a new infusion at 250 mcgkg)1 min)1 (Figure 3b); achieving a steady-state delivery of a new remifentanil infusion (0.15 mcgkg)1min)1) requires up to 36 min (see supplementary Figure S2). At higher drug infusion doses, the ow of drug will be higher, thus shortening the time constant (Figure 3c,d).

Fluid delivery by drug infusions During any drug infusion, there is an obligate uid delivery to the child from the drug infusions (Figures 4 and 5). Volume delivery increases with patient weight and infusion dose; diluting the stock drug will also increase volume delivery for the same dose (Formula 2). A propofol infusion at 250 mcgkg)1min)1 necessitates a minimum 75 mlh)1 of uid to a 50 kg child. This represents 83% of the childs predicted maintenance uid requirement. For a 3 kg infant, the same propofol infusion rate requires 4.5 mlh)1 of uid, or 37.5% of the childs 12 mlh)1 maintenance uid requirement. Remifentanil (stock concentration 50 mcgml)1) infusing at 0.15 mcgkg)1min)1 will obligate delivery of 9 ml of uid in 1 h to a 50 kg child, which is 10% of the childs maintenance uid requirement. The same remifentanil infusion will provide a 3 kg infant 0.54 mlh)1 of uid, about 4.5% of the childs maintenance uid requirement. Some clinicians choose to infuse dilute solutions of remifentanil (e.g., 5 mcgml)1) in infants. This results in a reduced Mv and a faster achievement of steadystate levels of drug delivery (shorter time constant) because the drug ow rate is relatively high. However, there is an obligate increase in uid from the drug infusion (Figure 5c,d). For example, if remifentanil (5 mcgml)1) infuses at 0.15 mcgkg)1min)1 into a 3 kg infant, the infusion will provide 5.4 mlh)1 of uid, or 45% of the infants maintenance uid requirement. Discussion The clinical implications of the interactions of various types of intravenous systems, drug concentrations, carrier rates, and patient sizes have not been adequately investigated in children. To our knowledge, this is the rst application of the principles derived from our adult-focused previous research (2) to pediatric anesthesia. An important factor for children is the overall uid load from the combination of the carrier and the drug ows. We studied remifentanil and propofol as model drugs because they are so commonly administered for sedation or total intravenous anesthesia (TIVA). Our ndings demonstrate that the rapidity of response to initiation or changes of infusions and the potential for clinically signicant intermittent drug boluses are greatly inuenced by the rate of carrier ow, the drug concentration, and where the infusion joins the main uid pathway (dead volume factor). This information may be useful to clinicians in training and clinicians who do not routinely provide pediatric anesthetic care.

82

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

H. Ma et al.

Quantitative analysis of pediatric intravenous infusions

(a) 600
Propofol: tau predicted by plug-flow model (s)
500

Propofol: tau predicted by well-mixed model (s)

Id = 250 (mcgkg1min1) Cd = 10 mgml1 Qc = maintenance fluid rate Vd = 0.5 ml

(b)

1800 1600 1400 1200 1000 800 600 400 200 0 0

Id = 250 (mcgkg1min1) Cd = 10 mgml1 Qc = maintenance fluid rate

Vd = 0.5 ml Vd = 1 ml Vd = 1.5 ml Vd = 2 ml Vd = 2.5 ml

400 300 200 100 0

Vd = 1 ml Vd = 1.5 ml Vd = 2 ml Vd = 2.5 ml

10

20

30

40

50

10

20

30

40

50

Patient weight (kg)

Patient weight (kg)

(c)
Propofol: tau predicted by plug-flow model (s)
800 700 600 500 400 300 200 100 0

Propofol: tau predicted by well-mixed model (s)

Wt = 3 kg Cd = 10 mgml1 Qc = 12 mlh1 Vd = 0.5 ml Vd = 1 ml Vd = 1.5 ml Vd = 2 ml Vd = 2.5 ml

(d)
2500 2000

Wt = 3 kg Cd = 10 mgml1 Qc = 12 mlh1 Vd = 0.5 ml

1500 1000 500 0

Vd = 1 ml Vd = 1.5 ml Vd = 2 ml Vd = 2.5 ml

50

100

150

200

250

300

350

400

50

100

150

200

250

300

350

400

Propofol infusion dose (mcgkg1min1)

Figure 3 Time to achieve steady-state propofol infusion after the initiation of an infusion. (a) Predictions based on the Plug-Flow model: effect of patient weight. This time predicted by the PlugFlow model is dened as one time constant (tau or s). The carrier ow (Qc) is the maintenance uid rate. (b) Predictions based on the Well-Mixed model: effect of patient weight. This time predicted by the Well-Mixed model is approximately three time constants (3 s). The carrier ow (Qc) is the maintenance uid rate. (c) Predictions based on the Plug-Flow model: effect of propofol infusion

Propofol infusion dose (mcgkg1min1)

dose. The patient weight (Wt) is dened as 3 kg; the carrier ow (Qc) is the maintenance uid rate (12 mlh)1). (d) Predictions based on the Well-Mixed model: effect of propofol infusion dose. The patient weight (Wt) is dened as 3 kg; the carrier ow (Qc) is the maintenance uid rate (12 mlh)1). Conditions: Dead volumes used in calculations are 0.5, 1, 1.5, 2, and 2.5 ml. The propofol infusion rate (Id) is 250 mcgkg)1min)1, and the stock drug concentration (Cd) is 10 mgml)1.

(a)
120

(b)

Ratio of propofol infusion fluid load to maintenance fluid flow (%)

Propofol infusion fluid load (mlh1)

Cd = 10 mgml1 Wt = 3 Kg Wt = 5 Kg Wt = 10 Kg Wt = 20 Kg Wt = 30 Kg Wt = 40 Kg Wt = 50 Kg

140 120 100 80 60 40 20 0 0

Cd = 10 mgml1 Qc = maintenance fluid rate

100 80 60 40 20 0 0 50 100 150 200 250 300 350 400

Wt10 Kg Wt = 20 Kg Wt = 30 Kg Wt = 40 Kg Wt = 50 Kg

50

100 150 200 250 300 350 400

Propofol infusion dose

(mcgkg1min1)

Propofol infusion dose (mcgkg1min1)

different patient weights. The carrier ow (Qc) is the maintenance uid rate. The ratio is independent of weight when the weight is 10 kg so that the curves are superimposed. Conditions: The stock propofol drug concentration (Cd) is 10 mgml)1.

Figure 4 Propofol infusion uid loads. (a) Hourly uid load from propofol infusion, as a function of infusion dose, for different patient weights. (b) The ratio of propofol infusion uid load to patient maintenance uid ow, as a function of infusion dose, for

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

83

Quantitative analysis of pediatric intravenous infusions

H. Ma et al.

18 16

(a)

Cd = 50 mcgml1

20

14 12 10 8 6 4 2 0 0 0.05 0.1 0.15 0.2 0.25 0.3

Ratio of remifentanil infusion fluid load to maintenance fluid flow (%)

Remifentanil infusion fluid load (mlh1)

18 16 14 12 10 8 6 4 2 0 0

(b)

Cd = 50 mcgml1 Qc = maintenance fluid rate

Wt = 3 Kg Wt = 5 Kg Wt = 10 Kg Wt = 20 Kg Wt = 30 kg Wt = 40 Kg Wt = 50 Kg

Wt 10 Kg Wt = 20 Kg Wt = 30 Kg Wt = 40 Kg Wt = 50 Kg

0.05

0.1

0.15

0.2

0.25

0.3

Remifentanil infusion dose (mcgkg1min1)

180

Remifentanil infusion dose (mcgkg1min1)

200

Remifentanil infusion fluid load (mlh1)

160 140 120 100 80 60 40 20 0 0

Ratio of remifentanil infusion fluid load to maintenance fluid flow (%)

(c)

Cd = 5 mcgml1

180 160 140 120 100 80 60 40 20 0 0

(d)

Cd = 5 mcgml1 Qc = maintenance fluid rate

Wt = 3 Kg Wt = 5 Kg Wt = 10 Kg Wt = 20 Kg Wt = 30 kg Wt = 40 Kg Wt = 50 Kg

Wt 10 Kg Wt = 20 Kg
Wt = 30 Kg Wt = 40 Kg Wt = 50 Kg

0.05

0.1

0.15

0.2

0.25

0.3

0.05

0.1

0.15

0.2

0.25

0.3

Remifentanil infusion dose (mcgkg1min1)

Remifentanil infusion dose (mcgkg1min1)

Figure 5 Remifentanil infusion uid loads. (a) Hourly uid load from remifentanil infusion, as a function of infusion dose, for different patient weights. The drug concentration (Cd) is 50 mcgml)1. (b) The ratio of remifentanil infusion uid load to patient maintenance uid ow, as a function of infusion dose, for different patient weights. The carrier ow (Qc) is the maintenance uid rate, and the drug concentration (Cd) is 50 mcgml)1. The ratio is independent of weight when the weight is 10 kg so that the curves are

superimposed. (c) Hourly uid load from remifentanil infusion, as a function of remifentanil infusion dose, for different patient weights. The drug concentration (Cd) is 5 mcgml)1. (d) The ratio of remifentanil infusion uid load to patient maintenance uid ow, as a function of infusion dose, for different patient weights. The carrier ow (Qc) is the maintenance uid rate, and the drug concentration (Cd) is 5 mcgml)1. The ratio is independent of weight when the weight is 10 kg so that the curves are superimposed.

A safety issue is that during infusion, the system dead volume will accumulate a mass of drug (Mv) at steady state. This drug mass is potentially available for delivery as an inadvertent bolus should a clinician give an upstream push of uid or another drug (e.g., antibiotic or muscle relaxant), or increase the carrier ow rate. Untoward physiologic effects may result from the bolus, depending on the particular agent and the amount given. The magnitude of a bolus is potentially most signicant in infants. This is because the Mv-toloading dose ratio is relatively greater at low body weights when other variables remain constant. A second intravenous line dedicated to pushed drugs and replacement of blood or third space losses can avoid the problem of inadvertent bolusing of infused drugs. Multiple factors affect Mv and the Mv-to-loading dose ratio (Formulas 3 and 4). In clinical practice, patient weight is xed and the drug infusion dose is
84

determined by clinical requirements. In theory, clinicians can either increase the carrier ow or dilute the drug concentration to decrease Mv or the Mv-to-loading dose ratio. However, both maneuvers will increase obligate uid loads. These uid loads may be undesirable, depending on clinical conditions and comorbidities, e.g., infants on uid restriction because of renal or cardiac disease. The most practical method to minimize Mv and the Mv-to-loading dose ratio is to reduce the infusion system dead volume. This can be accomplished by connecting the drug infusion at an entry point close to the patient in the infusion system. However, even the closest available port may still provide a relatively large infusion system dead volume to a neonate. To join an infusion to the main uid path, some clinicians interpose a Y-piece between the hub of the vascular access catheter and the infusion tubing. Others introduce a
Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

H. Ma et al.

Quantitative analysis of pediatric intravenous infusions

needle at the port of the T-connector and deliver the infused drug through the needle. These strategies may be useful to reduce dead volume and thereby minimize time lags and bolus risks for infusion of only a single drug. When patients receive two or more infused drugs, upstream access points are likely to be required. This obligates that the ow pathway of some drugs will have a larger dead volume. Clinicians may also encounter situations where drug infusions need to be added, but access to the limb with an IV is limited. Upstream junctions in the uid path may be used to connect new infusions to the delivery system, with obligate dead volumes larger than a desired minimum. Our analysis further demonstrate that the lag time to achieve a steady state of drug delivery after starting or changing an infusion can be very long for small children in whom both carrier and drug ow rates are slow relative to the size of the dead volume. This increases the time required to traverse the dead volume. To reduce the lag time, clinicians can increase the carrier ow or dilute the stock drug concentration. Again, these changes will increase obligate uid loads. For an individual drug, reducing the dead volume of the uid pathway will shorten lag times. Optimally, an infusion system with a minimal Vd might be designed that can accommodate multiple separate drug infusions so that drug delivery lag times can be reduced without elevating uid ows. Some anesthesiologists will transiently raise the carrier ow rate to quickly achieve an increase in drug delivery and then reduce the carrier ow to steady-state levels. However, when multiple drugs infuse through one line, drug delivery kinetics become complex because changing one drug infusion will, at least transiently, impact the delivery of the other drugs. This is because each drug infusion contributes to the overall carrier ow for other medications in the same uid pathway. Therefore, a transient increase in carrier ow to promote the delivery of one drug will push other drugs into the patient, delivering boluses of those medications. Conversely, after returning the carrier to a baseline rate, delivery of all infused drugs will temporarily fall until a new steady-state delivery is restored (1). When analyzing uid delivery from drug infusions, we calculated volume loads from propofol and remifentanil individually. However, children can receive uids from concurrent infusions, such as during TIVA. A 3 kg infant treated with propofol at 250 mcgkg)1 min)1 and remifentanil at 0.15 mcg kg)1min)1 receives a combined uid load of at least 5.04 mlh)1 (42% of maintenance uid) depending on the initial dilution of remifentanil. In a 50 kg child, the same propofol and remifentanil combination proPediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd

duces a uid load of 84 mlh)1 just from the TIVA drugs (93% of hourly maintenance). Consequently, clinicians neglecting all sources of uid entry may inadvertently subject their patients to some degree of volume loading. These principals are well recognized in the care of NICU patients but could be overlooked in the operating room. In conclusion, managing continuous drug infusions requires not only an understanding of pharmacokinetics once a drug enters the patient, but also the dynamics of systems delivering the drug to the circulation. Dead volume and uid ow rates are the two most important factors in drug delivery to the circulation. A large dead volume combined with a slow uid ow provides the conditions for a drug mass of signicant size in the dead volume. If pushed into a child as a bolus, this drug mass could produce signicant effects, especially in infants. Large dead volumes and slow uid ow rates also create the conditions for surprisingly long lag times before infusions reach targeted steady-state delivery rates. Increasing uid ow rates can overcome some of these problems at the expense of delivering a larger uid volume. We emphasize that the infusion system architecture affects how anesthetic drugs reach the child and thus the safety and efcacy of drug infusion techniques. In general, infused drugs should be administered as close to the vascular access catheter as possible (minimizing dead volume). The ndings support the recommendation that clinicians consider placing more than one IV access, dedicating one access for infusions plus carrier and an additional IV access for administering drug boluses and replacement uids (blood products, colloid, crystalloid). Although the experienced pediatric anesthesiologist may be familiar with the basic principles described in this report, we aim to convey these concepts in quantitative fashion to trainees and to practitioners who infrequently provide pediatric anesthesia care. Appreciating these principles can help clinicians provide safer and more effective anesthetics to children. Acknowledgments We thank Dr. Charles Cote and Dr. Lucinda Everett for helpful comments during the preparation of this manuscript. Supported by institutional funds from the Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital. Conict of interest None of the authors has any nancial disclosures or conicts of interest relevant to this report.
85

Quantitative analysis of pediatric intravenous infusions

H. Ma et al.

Supporting Information Additional Supporting Information may be found in the online version of this article: Figure S1 Remifentanil in the dead volume. Figure S2 Time to achieve steady-state remifentanil infusion after the initiation of an infusion. Appendix S1 Weight-based formulas for calculating uid ows, the mass of infused drug in the dead References
1 Lovich MA, Doles J, Peterfreund RA. The impact of carrier ow rate and infusion set dead-volume on the dynamics of intravenous drug delivery. Anesth Analg 2005; 100: 1048 1055. 2 Lovich MA, Kinnealley ME, Sims NM et al. The delivery of drugs to patients by continuous intravenous infusion: modeling predicts potential dose uctuations depending on ow rates and infusion system dead volume. Anesth Analg 2006; 102: 11471153. 3 Lovich MA, Peterfreund GL, Sims NM et al. Central venous catheter infusions: a labora-

volume, the time constant, and the volume uid delivered to the child. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

tory model shows large differences in drug delivery dynamics related to catheter dead volume*. Crit Care Med 2007; 35: 2792 2798. 4 Moss DR, Bartels K, Peterfreund GL et al. An in vitro analysis of central venous drug delivery by continuous infusion: the effect of manifold design and port selection. Anesth Analg 2009; 109: 15241529. 5 Bartels K, Moss DR, Peterfreund RA. An analysis of drug delivery dynamics via a pediatric central venous infusion system: quantication

of delays in achieving intended doses. Anesth Analg 2009; 109: 11561161. 6 Holliday MA, Segar WE. The maintenance need for water in parenteral uid therapy. Pediatrics 1957; 19: 823832. 7 Bailey AG, McNaull PP, Jooste E et al. Perioperative crystalloid and colloid uid management in children: where are we and how did we get here? Anesth Analg 2010; 110: 375390.

86

Pediatric Anesthesia 21 (2011) 7886 2010 Blackwell Publishing Ltd