2. Factors Affecting Drug Distribution Through Infusion2. Factors Affecting Drug Distribution Through Infusion by The Infusion Physics Study Group* 05-18-09 Convection Enhanced Delivery

2. Factors Affecting Drug Distribution Through Infusion

by The Infusion Physics Study Group*

05-18-09

Convection Enhanced Delivery (CED) is a technique used to distribute drugs inside the

brain parenchyma using pressure to cause the movement of infused fluid.1 While this

technique does distribute large molecules much further than diffusion alone could do, its

application has been limited because the extent and shape of distribution are variable.

Understanding and reducing the causes of such variability was the purpose of this study.

Previous in vivo experience has shown that the most common departure from ideal

infusion distribution was backflow along the outer surface cannula. (Figure 1a shows an

image of a good infusion and Figure 1b shows an image of an infusion with significant

backflow.) Backflow takes place whenever it is easier for the fluid to travel along an

annular space created between the outer surface of the catheter and the surrounding

medium than out through the pores of the media.

We systematically studied the physics of infusion in gels and attempted to determine the

conditions that contributed to variability. We then tested the applicability of these

findings in vivo. Each experiment was performed at least three times, maintaining the

same conditions, in order to evaluate reproducibility. In total, over 300 experiments were

performed.

This study was based on the hypotheses that

1. Infusion into a uniform and isotropic medium leads to a spherical infusate

distribution,

and

2. Whenever the medium is either non-uniform or non-isotropic, the infusion will

depart from the spherical distribution, often in unpredictable ways.

No single technique ensures reproducibility. Instead some of the techniques increased the

margin for error in the system.

The experiments focused on five key subjects:

(1) Effect of cannula insertion techniques

(2) Effect of using a stepped cannula design

(3) Effect of a prior cannula track in the surrounding medium

(4) Effect of using pulsatile flow

(5) Infusion pressure as a real-time monitor

Experimental methods are described in the Appendix.

Effect of cannula placement technique:

The method of insertion itself can have a significant effect on variability of the infusion

due to the seal of the gel around the catheter. We evaluated various techniques that have

been used in prior work. Figure 2 shows the three insertion modes that were evaluated.

Figure 2a shows the results obtained when the catheter was inserted into the gel while it

was still in liquid form and then allowed to solidify around the catheter. Good spherical

distribution with no backflow was observed in 5 or 6 experiments. In contrast, if after

allowing the gel to solidify around the catheter, and then lifting the catheter 3mm and

then “re seating” it to its original position (to break the seal between the gel and the

catheter), we observed backflow of approximately 17 mm every time. (see Figure 2B)

The third method used to insert the catheter consisted of first allowing the gel to solidify

and then inserting the catheter into the solidified gel. As can be seen in Figure 2c, this

method provided the most variability. Nevertheless, this mode was chosen for most

experiments since it is the most realistic insertion mode for in vivo experiments.

Smooth insertion was an important factor in minimizing backflow. Even a slight lateral

movement of the catheter could provide a low resistance path for backflow. Sometimes

these movements were too small to be visible by the naked eye, necessitating the use of a

video monitor for observation. For most experiments, a mechanical introducer was used

to insert the catheter in a consistent manner.

Inserting the catheter into the gel often causes small cracks to from around the catheter.

Figure 3a shows the typical backflow at room temperature when the gel is constrained in

a plastic cube. This problem can be minimized by raising the temperature of the gel to

body temperature by suspending it in a heated bath and removing the gel from the plastic

cube. This set up is called “unconstrained gel at temperature”. Backflow in such cases is

more typical of backflow in tissue and is shown in Figure 3b.

Effect of using a stepped design: A catheter disturbs the surrounding medium by displacing some of it while being

inserted. Catheters with a smaller outside diameter will cause less displacement than

those with a larger gauge.

A cannula has to be rigid enough to allow successful in vivo insertion. A step design can

give most of the catheter sufficient structural rigidity to allow it to be inserted into a gel

or the brain tissue while minimizing the displacement of the material in the target region

by the use of a step. The step can also help hinder any backflow starting at the cannula

tip. Such a step is illustrated in Figure 4.

Figure 2b shows 15-17mm of backflow which was consistently observed using a

standard straight catheter. Figure 5 shows results obtained using the step catheter. The

step completely stops the backflow in 3 of the 4 cases and impedes the backflow in the

other case at this flow rate.

Figure 6 shows the effect of the step at different flow rates. At low flow rates the

backflow does not reach the step. At intermediate flow rates the step impedes backflow,

but if the flow rate is high enough, the backflow overcomes the step.

Figure 7 shows the effect of the step in vivo with no backflow observed above the step,

either early on or later in the course of the infusion.

The effect of a prior catheter track in the surrounding medium:

In-vivo experiments had shown that tissue damage from a previous cannula track can

provide a low resistance path for the infusate. The top image in Figure 8 shows the

previous catheter track as well as the current catheter. The bottom of Figure 8 shows the

infusion cloud as it is diverted by the prior catheter track.

This condition was replicated in a gel experiment by inserting and then removing a

cannula adjacent to the current catheter. As shown in Figure 9, the irregularity in the

medium presents less resistance to flow and results in a preferential flow path for the

infusate.

Effect of Pulsatile Flow:

Backflow may be looked at as a result of competition between two paths for the fluid

flow: one, through the pores of the surrounding medium, and the other, along an annulus

surrounding the catheter. If these two paths had different elastic behaviors, they would

respond differently to pulsed flow: one path may open more rapidly than the other when

the fluid pressure is increased in a step-wise manner.

We tested the dynamic (time-dependent) characteristics of the two paths by employing

periodic pulsing of the pressure. Typically, the pressure was on for 1-2 seconds and off

for 8- 9, and this cycle was repeated every 10 seconds.

Gel results are shown in Figure 10 where the average flow rate of the pulsed flow is the

same as in the case of steady flow (i.e. 20% duty cycle with a peak flow of 100uL/min is

the same average rate as a steady flow of 20uL/min).. In the steady infusion, backflow

begins at a total infused volume of 50mL, while in the pulsed case, it does not appear

even at 190mL. This suggests that the annulus opens up more slowly in response to a

step increase in pressure than do the pores. Consequently, we are likely to encounter less

backflow.

Figure 11 shows the affect of increasing the duty cycle while keeping the flow rate

constant. In each case, the total Vi is 50uL, but the average flow rate is different. The

10% duty cycle results in a perfect sphere; 20% duty cycle results in backflow at 10uL

and an elliptical shape at 50uL; 50% duty cycle results in backflow at 10uL and an

oblong shape at 50uL. As the pulsed flow approaches steady flow, the backflow tends

toward the behavior expected in steady flow.

Figure 12 shows an indication of this phenomenon in vivo. The infusion on left side of

the putamen was performed at 1uL/min steady flow; the right side used pulsatile flow 1

second on/9 off with with a peak flow of 10uL/min. The right side has a more spherical

distribution and is distributed further into the tissue.

More experiments are planned to evaluate this phenomena in vivo under inter-operative

MRI

Infusion pressure as a real-time indicator:

The pressure required to maintain a given flow rate is a potentially valuable external

indicator of what is happening within the cannula and in the surrounding medium.

Figure 13 shows a typical pressure profile. The pressure rises to the required level and

then remains stable until the infusion is completed. We often saw a similar pressure

profile even in the presence of backflow, as seen in Figure 14.

However, Figure 15 shows a markedly different pressure profile. The pressure rises to a

peak of 43mm Hg and then drops to approximately 25mm Hg in about 30 seconds before

decaying to a steady state pressure of 18mm Hg. Examination of the video of the

infusion showed that an occlusion had initially blocked the flow. As the pressure was

rising, the occlusion was forced out of the catheter, followed by the pressure dropping.

Figure 16 shows a somewhat lower peak of 37mm Hg and a longer decay to about 25mm

Hg. Examination of the video indicated a partial occlusion which was gradually

removed.2

Pressure profiles corresponding to both complete and partial occlusion have been

observed in vivo as well. Figure 17 shows an in vivo example.

Due to the probable important role in causing variability of infusion results, further

studies of occlusion in gels and in vivo will be the subject of another report by the

Infusion Physics Study Group.

*The Infusion Physics Study Group:

Research Contributor

Senior investigators and in

vivo/ex vivo experiments Dr. Krystof Bankiewicz, University of

California at San Francisco

Dr. Marina E. Emborg, University of

Wisconsin, Madison

Fluid physics Dr. Raghu Ragavan, Therataxis

Dr. Martin Brady, Therataxis

Simulations/engineering Chris Ross, Engineering Resources

Group, Inc.

MRI physics Dr. Andrew Alexander, University of

Wisconsin, Madison

Dr. Tracy McKnight, University of

California at San Francisco

Technical contributors

(in alphabetical order) Janine Beyer, University of California at

San Francisco

John Bringas, University of California at

San Francisco

Dr. Kevin Brunner, University of

Wisconsin, Madison

Michael Dobbert, University of

Wisconsin, Madison

Ronald Fisher, University of Wisconsin,

Madison

Valerie Joers, University of Wisconsin,

Madison

Philip Pivirotto, University of California

at San Francisco

James J. Raschke, University of

Wisconsin, Madison

Dr. Dali Yin, University of California at

San Francisco

Elizabeth Zakszewski, University of

Wisconsin, Madison

Project management Ken Kubota, Kinetics Foundation

Tom Dunlap, Kinetics Foundation

References:

1. Convection Enhanced Delivery is described in the companion report, “What is CED?”

by the Infusion Study Group, May, 2009

2. This phenomenon was observed in “Convection-enhanced delivery of macromolecules

in the brain”, R. Hunt Bobo, Douglas W. Laske, Aytac Akbasak, Paul F. Morrison,

Robert L. Dedrick, and Edward H. Oldfield, Proc. Natl. Acad. Sci. USA, Vol. 91, pp.

2076-2080, 1994

3. Krauze MT, Saito R, Noble C, Bringas J, Forsayeth J, McKnight TR, Park J,

Bankiewicz KS. Effects of the perivascular space on convection-enhanced delivery of

liposomes in primate putamen. Exp Neurol. 2005 Nov;196(1):104-11

APPENDIX

05-18-09

Gel experiments:

Initially, experiments were performed at room temperature with the gel (normally 0.2%

or 0.6% agarose) “constrained” in a 2 ¼ x 2 ¼ x 3 inch clear plastic container. Later

experiments were performed in a heated water bath and the gel was removed from the

plastic container and placed in the bath. This set up is called “unconstrained” gel at

temperature and provides an experimental set up that is closer to the in vivo set up. The

water is heated with a recirculator/heater system and the temperature is measured with

a digital pyrometer with its probe in a sacrificial gel in the bath. The combination of

unconstrained gel and body temperature reduces the cracking of the gel that occurs

during insertion into gels. A key component of this system is a custom built computer

which controls the pump, monitors the volume infused and the infusion line pressure.

The pressure data are presented graphically and digitally in real time. The purpose of this

subsystem was to measure the variations in pressure while keeping a constant flow rate.

A high definition video camera was set up to monitor the flow of the infusate into the gel

and it was synchronized with the pressure data acquisition. Still images were selected at

time intervals of interest to correlate the infusion cloud with the pressure trace.

Most experiments used a single port catheter with the port at the tip of the catheter. The

catheter was constructed using fused silica or stainless steel for the main body and

polyimide for the stepped tip in various sizes dictated by the requirements of the

experiment.

The infusion line is a fused silica tube that connects to a transition line through the ‘Zero

Air Chamber’. The ‘Zero Air Chamber’ allows an in-dwelling stylet to be manipulated

without the risk of introducing air into the infusion stream. The transition line is flexible

and guides the stylet through the catheter. It is attached to the catheter. The infusion

system is shown in

Figure A

In Vivo:

Comparisons between infusions require that the parameters are sufficiently controlled

allowing replication of an experiment. This includes the ability to perform accurate

targeting of the desired structure to allow controlling for anatomical variation.

In vivo experiments needed to be coordinated between the gel lab (that tests the delivery

system pre and post infusion), the surgical suite, and the MRI facility. During the surgical

and MRI procedures, vital signs are constantly monitored. Previous research has found

that changes in blood pressure may affect infusate distribution. 3

The in vivo infusion system includes the following equipment:

Pressure monitoring and infusion pump controller system (Engineering Resources

Group)

MRI-compatible syringe pump (Harvard)

step catheter

Silica infusion lines and disposable syringes

Catheter targeting system (Navigus® , a MRI-compatible trajectory guide for

intracerebral biopsies and placement of electrodes for deep brain stimulation that is

part of the Medtronic StealthStation® Navigation system, was modified for catheter

placement.)

MRI compatible stereotactic frame. The frame ensures that the animals’ head is

secured (minimizing movement during MRI and surgery), placed in a similar position

during all imaging recordings and surgical placement, and facilitates positioning of

the navigus base

Solution of Gadolinium DTPA using degassed-sterile-distilled water for in vivo

(MRI) visualization of the infusate

3.0T MRI scanner

Targeting Method:

The placement of the Navigus system is performed in sterile surgical conditions. Using

stereotaxic methods and guided by baseline T1 MRI in the coronal, axial and sagittal

planes, an entry burr hole in the skull is created and the Navigus system is anchored on

top of it. Multiple T1 MRIs are performed before and during the cannula insertion to

ensure that the tip of the canula reaches the target at the desired angle and depth in one

single tissue pass. When this is achieved, the infusion is started.

Figure A

Figures

5-18-09

Figure 1a In vivo coronal MRI; “good infusion

”.

Figure 1b In vivo coronal MRI; with backflow.

Figure 2 The effect of cannula insertion techniques.

a. Gel grown around catheter

b. Same, with gel-to-catheter seal broken

c. Insertion into a solidified gel

All at 5μL/min

a. Grow gel around catheter

b. Grow gel around catheter and re-seat

c. Insert catheter into solidified gel

• March 2008• Room temp• Constrained

• February 2009• 37 degrees C• Un-constrained

a. b.

Figure 3 Effect of ambient temperature and constraint on the flow

Figure 4 Stepped catheter design;10mm long polyimide catheter with 1.5mm step

Figure 5 The effect of the step on the flow; 5μL/min; 0.5mm; polyimide catheters

Figure 6 The effect of the step at different flow rates. Flow rates varied

from 0.1, to 10 uL/min at different time periods.

Figure 7 Step catheter inhibits backflow in vivo, at an early and later point

in the infusion.

Figure 8 The infusion cloud is diverted by a cannula track left from a previous

infusion.

Figure 9 The infusion cloud reaches a channel which had been formed earlier in

the gel. The channel diverts the flow. Steady infusion at 10 uL/min.

Run H03 20µl/m Steady Run Ho5 100µl/m Pulses 2s

ON, 8s Off

0.5minutes, 10µl

2.5minutes, 50µl

5.5minutes, 110µl

9.5minutes, 190µl

Figure 10 The effect of pulsatile flow: no backflow appears even at Vi=190 uL/min;

in an equivalent steady infusion, backflow appears at Vi=50 uL/min. Average flow rate

is the same in both cases.

Figure 11 The effect of duty cycle: as the duty cycle is increased, the effect of

pulsed flow begins to resemble the effect of steady flow.

Figure 12 The effect of pulsatile flow in vivo. Left: steady flow, Right: pulsatile flow.

The average flow rate is the same in the two cases

Figure 13 Pressure profile for a steady infusion at 20μL/min.

Figure 14 Pressure profile in the presence of backflow.

Figure 15 Pressure profile in the case of an occluded catheter; complete occlusion.

Occlusion visually verified.

Figure 16 Pressure profile in the case of a partially occluded catheter.

Occlusion visually verified.

Figure 17 Pressure profile in the case of an occluded catheter, in vivo.