Computational and experimental study of proximal flow in ventricular catheters

Technical note

Full access

✓ The treatment of hydrocephalus with shunt insertion is fraught with high failure rates. Evidence indicates that the proximal holes in a catheter are the primary sites of blockage. The authors have studied ventricular catheter designs by using computational fluid dynamics (CFD), two-dimensional water table experiments, and a three-dimensional (3D) automated testing apparatus together with an actual catheter. With the CFD model, the authors calculated that 58% of the total fluid mass flows into the catheter's most proximal holes and that greater than 80% flows into the two most proximal sets of holes within an eight-hole catheter. In fact, most of the holes in the catheters were ineffective. These findings were experimentally verified using two completely different methodologies: a water table model of a shunt catheter and a 3D automated testing apparatus with an actual catheter to visualize flow patterns with the aid of ink. Because the majority of flow enters the catheter's most proximal holes, blockages typically occur at this position, and unlike blockages at distal holes, occlusion of proximal holes results in complete catheter failure. Given this finding, new designs that incorporated varying hole pattern distributions and size dimensions of the ventricular catheter were conceived and tested using two models. These changes in the geometrical features significantly changed the entering mass flow rate distribution. In conclusion, new designs in proximal ventricular catheters with variable hole diameters along the catheter tip allowed fluid to enter the catheter more uniformly along its length, thereby reducing the probability of its becoming occluded.

Abstract

✓ The treatment of hydrocephalus with shunt insertion is fraught with high failure rates. Evidence indicates that the proximal holes in a catheter are the primary sites of blockage. The authors have studied ventricular catheter designs by using computational fluid dynamics (CFD), two-dimensional water table experiments, and a three-dimensional (3D) automated testing apparatus together with an actual catheter. With the CFD model, the authors calculated that 58% of the total fluid mass flows into the catheter's most proximal holes and that greater than 80% flows into the two most proximal sets of holes within an eight-hole catheter. In fact, most of the holes in the catheters were ineffective. These findings were experimentally verified using two completely different methodologies: a water table model of a shunt catheter and a 3D automated testing apparatus with an actual catheter to visualize flow patterns with the aid of ink. Because the majority of flow enters the catheter's most proximal holes, blockages typically occur at this position, and unlike blockages at distal holes, occlusion of proximal holes results in complete catheter failure. Given this finding, new designs that incorporated varying hole pattern distributions and size dimensions of the ventricular catheter were conceived and tested using two models. These changes in the geometrical features significantly changed the entering mass flow rate distribution. In conclusion, new designs in proximal ventricular catheters with variable hole diameters along the catheter tip allowed fluid to enter the catheter more uniformly along its length, thereby reducing the probability of its becoming occluded.

The treatment of hydrocephalus with the insertion of a shunt is fraught with high failure rates. Evidence indicates that the proximal end of a catheter is the primary site of blockage.1,2,4,9 Obstructing agents generally include debris such as blood clots, cell clusters, and normal tissues such as choroid plexus and ependymal tissues.2,3,10 Other entrance conditions such as the CSF environment, ventricular size, and catheter tip location are also identified to be important factors.4,6,9,11 In a recent randomized trial, researchers found that a catheter tip surrounded by CSF has a better chance of success.4,11 Controversies also exist regarding ventricular sizes in relation to shunt success and failure. Although some blame collapsed ventricles for increasing the risk of proximal shunt malfunction, others try to optimize the neurodevelopmental function of patients with hydrocephalus by maximizing their cerebral mantle. Regarding catheter tip location, the goal is to avoid obstructing agents such as choroid plexus, ependymal, glial, and brain tissues. Given that the mean inlet size of a ventricular catheter is 1.6 to 2 cm and that the mean intraventricular distance available for frontal horn placement is less than 1.6 cm, it is very difficult to place a catheter so that the inlets are directed away from the choroid plexus.6 In theory, the best location is farthest from the choroid plexus. In clinical practice, it is difficult to place a catheter precisely and to predict the alteration in location as ventricular size decreases. Finally, data from a recent randomized trial of endoscopically guided catheter placement showed no statistical correlation between placement and survival. Our interdisciplinary team studied the fluid dynamics of ventricular catheters in an attempt to answer questions surrounding the mechanisms of proximal obstruction and flow occurring at the inlets.

Materials and Methods

We studied ventricular catheter designs by using CFD, 2D water table experiments, and 3D automated testing apparatus together with an actual catheter. The 2D CFD models produced a visual image of the flow field and results were integrated to evaluate mass flow rates. The 2D water table experiments were used to visualize the flow field and measurements were made to evaluate mass flow rates. The automated 3D test apparatus was used for flow visualization in an actual ventricular catheter. Most commercially available ventricular catheters consist of four rows of eight holes with an inner catheter diameter of 1.2 mm; thus, we began our investigations with this basic geometrical design.

Computational Fluid Dynamics

The general procedure for the development of a CFD model involves incorporating the physical dimensions of the system to be studied into a virtual wire-frame model. The shape and features of the actual physical model are transformed into coordinates for the virtual space of the computer and a CFD computational grid (mesh) is generated. The fluid properties and motion are calculated at each of these grid points. After grid generation, flow field boundary conditions are applied and the fluid's thermodynamic and transport properties are included. Finally, a system of strongly coupled, nonlinear, partial differential conservation equations governing the motion of the flow field are numerically solved. This numerical solution describes the fluid motion and properties.

Computer simulations were performed using a fluid dynamic software package (FIDAP, version 5.0; Fluid Dynamics International, Inc., Evanston, IL). This package, loaded on a Sun Microsystem Ultra 10 workstation, uses a finite element method to solve Navier-Stokes equations. Input files to the software containing both the catheter and the boundary conditions were generated, and Reynolds numbers were matched. In each case, the generated mesh was confined to use a minimum number of iterations for final convergence. This was a 2D model, assuming steady, incompressible, laminar flow in newtonian fluid. The geometrical features of the catheter consisted of two rows of eight holes and an inner diameter of 1.2 mm.

Water Table Experiments

A proximal catheter water table model with a scale factor of × 100 and matched Reynolds number was constructed with the aid of a wooden model. Using frame-grabbing techniques, we calculated flow velocities into each catheter hole. This involved making several time-lapse photographs by overlapping consecutive frames from a video of flow patterns. Flow visualization through each hole was made possible by using ground wax flakes floating on the water. The water entered the table at the distal end of the catheter. Several analyses were then performed: the first involved a model that had holes proportionally the same size as those in the most commonly used commercially available proximal ventricular catheters; the second used a progressive hole pattern to generate an even flow distribution in the catheter as discovered in our CFD procedure. We measured the average velocity of particles moving through each hole and calculated the mass flow rates in each hole.

Automated Experimental Apparatus

This experiment was developed to validate our results further by measuring the mass flow rate and critical pressures of commercial shunt systems. The ventricular pressure is adjusted by changing the elevation of the supply reservoir, which changes available hydrostatic pressure. The catheter is mounted in a model ventricular reservoir. The fluid enters the shunt system through the proximal holes of the catheter. From the proximal holes of the catheter, the fluid drains through the shunt valve and into the distal holes of the catheter. Drainage matter is collected and the mass is continuously measured. The pressure of the draining fluid is measured in the model ventricular reservoir with the aid of a precision pressure transducer. The transparent model ventricular reservoir permits flow visualization at the inlets to the catheter. The output of the apparatus is monitored continuously using a computer-based data acquisition system. The experiment pressure ranged from approximately 50 cm of water positive pressure to −15 cm of water siphoning pressure. We designed a computer-based data acquisition system by using LabView software. With this software we reliably and continuously monitored the performance of a shunt system over a range of operating conditions. This part of the study was performed to visualize flow in an actual ventricular catheter by injecting india ink and allowing the dye to flow into the inlets of the ventricular catheter.

Results

A velocity vector plot of a standard ventricular catheter from 2D CFD modeling is shown in Fig. 1 left. The scale on the right side of each plot demonstrates that most flow vectors are concentrated in the most proximal inlets. Flow velocities can then be calculated into mass flow rates and plotted in percent flow distribution rates (Fig. 2). Specifically, using the CFD model, we calculated that 58% of the total fluid mass flows into the most proximal holes of the catheter and that more than 80% flows into the two most proximal sets of holes within an eight-hole catheter. Most of the holes in the proximal catheters were not used.

Fig. 1.
Fig. 1.

Left: Plot demonstrating results of CFD modeling with a standard catheter. Most if not all of the flow velocity vectors occur proximally in the first two holes. Right: Plot demonstrating results of CFD modeling with a modified catheter with variable sizes of holes. Because the largest holes are situated distally, the flow velocity vectors are more evenly distributed.

Fig. 2.
Fig. 2.

Bar graph showing flow distribution in catheter holes, which was derived from the CFD modeling. Black bars represent the standard catheter, whereas the white bars represent the modified catheter with variable hole sizes, the largest of which was located distally or near the catheter tip.

These findings were experimentally verified in the water table experiment. Figure 3 shows the water table model and results. As fluid flowed into the catheter, flakes of wax aided in creating a visual image of the flow and allowed us to calculate the mass flow rate into each individual inlet, which we obtained using time-lapse photographs of individual flakes and measuring flow velocity. The time-lapse sequences showed that the flow into the catheter begins at Inlet 6 and that the majority flows through Inlets 7 and 8. The gray bars in Fig. 4 show that approximately 90% of flow courses through Inlets 7 and 8 in the water table experiment.

Fig. 3.
Fig. 3.

Upper Left: Diagram depicting the view of the water table model and the direction of flow. Upper Right and Lower: Time-lapse photography sequences obtained in the water table model, visualizing flow with the aid of wax flakes. The earliest sequences show that there was no flow through Inlets 1 through 4. The majority of the flow occurs through Inlets 7 and 8 and, to a lesser degree, Inlet 6.

Fig. 4.
Fig. 4.

Bar graph showing flow distribution through catheter holes, which were derived from the water table modeling. Gray bars represent the standard catheter, whereas black bars represent the modified catheter with variable hole sizes.

Given these findings, we created new catheter designs by varying hole pattern distributions and size dimensions of the proximal end of the ventricular catheter and tested them by using 2D models. A computational result for a catheter with holes that become progressively larger was obtained. Our findings showed that fluid enters the catheter more uniformly along the length of the catheter and significantly improved the flow field distribution within the catheters (Figs. 1 right and 2). Similar results were obtained in the water table experiments (Fig. 4).

The 3D automated experiment was set up to simulate flow in a standard ventricular catheter in a physiological setting. The main thrust of the 3D experiment was to visualize flow; that is, mass flow rate into each inlet was not calculated or measured. Figure 5 shows the time-lapse sequence photographs taken as the ink flowed into only the most proximal inlet.

Fig. 5.
Fig. 5.

Time-lapse photographs demonstrating the flow of india ink injected into the model ventricular reservoir. Note that all of the ink appears to flow into the most proximal inlet.

Discussion

Obstruction of the proximal end of a catheter is one of the most common causes of shunt malfunction. There are various commercially available ventricular catheters designed to improve on the success of the catheter such as the addition of flanges and slots, but none of these features has clinically proven to alter proximal shunt revision rates.8 The exact mechanism of ventricular catheter occlusion is unclear. In our experiments we evaluated the fluid dynamics occurring at the proximal inlets of catheters to evaluate factors of proximal shunt malfunction.

Both the computational and experimental data in this study showed that more than 80% of the flow into the catheter occurs at the two most proximal sets of holes. Thus, most of the holes in the proximal catheter were ineffective because of the difference between the area of the inlets relative to the area of the catheter lumen. Fluid flows to the regions of least resistance, that is, the most proximal inlets. One way to distribute the flow evenly is to increase the resistance of flow proximally by decreasing the area of the proximal inlets and to decrease the resistance distally by increasing the area of the distal inlets. Such a design in proximal ventricular catheters together with variable-hole diameters along the catheter tip would allow fluid to enter the catheter more uniformly along its length, theoretically reducing the probability of occlusion.

Because the majority of flow enters the most proximal catheter holes, blockages will occur at this position. Note that unlike blockages at distal holes, those at proximal holes may result in complete catheter failure when debris or tissue migrates internally and occludes the lumen of the catheter. This theory was examined during proximal shunt revisions performed at our institution. Photographs of occluded proximal catheters were obtained during a 6-month period in which we did not have to insert a metal stylet into the lumen of the blocked catheter to coagulate the occluding tissue. Figure 6 shows these occluded ventricular catheters. All of these catheters contained tissue at their most proximal holes, results which support our laboratory findings. Two drawbacks to our clinical correlation of the occluded catheters are as follows: 1) we did not split the occluded catheters in half lengthwise to verify visually where exactly luminal blockage occurred; and 2) we excluded blocked catheters that required stylet coagulation of the lumen given that this would shrink the tissue and alter the appearance. Therefore the denominators of the examined catheters were not known. Nevertheless, one may still theorize that when the majority of the flow courses through the most proximal catheter hole, it carries debris and tissue such as choroid plexus into the same hole, thus allowing these tissues to grow into and block the lumen of the catheter and causing shunt malfunction. Our findings also explain why it takes only one or two patent holes for the shunt to function and that percutaneous recanalization of an occluded proximal catheter is effective.5,7 Placing the largest hole distally or near the tip would decrease the probability of occluding the entire catheter when only its tip is blocked.

Fig. 6.
Fig. 6.

Photographs illustrating occluded catheters removed during recent shunt revision surgeries. All catheters are occluded at their most proximal holes.

In conclusion, more than 80% of total fluid mass flows into the two most proximal holes. Catheters with variable sized holes and its largest one situated at the catheter tip would redistribute the flow more evenly along the entire length of the catheter. Therefore, favorable changes in the geometry of the proximal catheter can significantly alter the fluid dynamics of the catheter, which may ultimately lead to a decrease in the rate of proximal catheter obstruction.

Acknowledgment

We acknowledge Dr. Robert Hurt for his work on the CFD model, which was performed at Bradley University during the summer of 2001 and funded by a Children's Miracle Network grant.

References

  • 1.

    Christens-Barry WAGuarnieri MCarson BS Sr: Fiberoptic delivery of laser energy to remove occlusions from ventricular shunts: technical report. Neurosurgery 44:3453501999Neurosurgery 44:

  • 2.

    Collins PHockley ADWoollam DH: Surface ultrastructure of tissues occluding ventricular catheters. J Neurosurg 48:6096131978J Neurosurg 48:

  • 3.

    Del Bigio MR: Biological reactions to cerebrospinal fluid shunt devices: a review of the cellular pathology. Neurosurgery 42:3193261998Del Bigio MR: Biological reactions to cerebrospinal fluid shunt devices: a review of the cellular pathology. Neurosurgery 42:

  • 4.

    Drake JMKestle JRWMilner Ret al: Randomized trial of cerebrospinal fluid shunt valve design in pediatric hydrocephalus. Neurosurgery 43:2943051998Neurosurgery 43:

  • 5.

    Ginsberg HJSum ADrake JMet al: Ventriculoperitoneal shunt flow dependency on the number of patent holes in a ventricular catheter. Pediatr Neurosurg 33:7112000Pediatr Neurosurg 33:

  • 6.

    Kaufman BAPark TS: Ventricular anatomy and shunt catheters. Pediatr Neurosurg 31:161999Pediatr Neurosurg 31:

  • 7.

    Pattisapu JVTrumble ERTaylor KRet al: Percutaneous endoscopic recanalization of the catheter: a new technique of proximal shunt revision. Neurosurgery 45:136113671999Neurosurgery 45:

  • 8.

    Portnoy HD: New ventricular catheter for hydrocephalic shunt. Technical note. J Neurosurg 34:7027031971Portnoy HD: New ventricular catheter for hydrocephalic shunt. Technical note. J Neurosurg 34:

  • 9.

    Sainte-Rose C: Shunt obstruction: a preventable complication? Pediatr Neurosurg 19:1561641993Sainte-Rose C: Shunt obstruction: a preventable complication? Pediatr Neurosurg 19:

  • 10.

    Sekhar LNMoossy JGuthkelch AN: Malfunctioning verticuloperitoneal shunts. Clinical and pathological features. J Neurosurg 56:4114161982J Neurosurg 56:

  • 11.

    Tuli SO'Hayon BDrake Jet al: Change in ventricular size and effect of ventricular catheter placement in pediatric patients with shunted hydrocephalus. Neurosurgery 45:132913351999Neurosurgery 45:

Article Information

Address reprint requests to: Julian Lin, M.D., Department of Neurosurgery, 530 Northeast Glen Oak, #3461, Peoria, Illinois 61637–2642. email: jchu888@hotmail.com.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Left: Plot demonstrating results of CFD modeling with a standard catheter. Most if not all of the flow velocity vectors occur proximally in the first two holes. Right: Plot demonstrating results of CFD modeling with a modified catheter with variable sizes of holes. Because the largest holes are situated distally, the flow velocity vectors are more evenly distributed.

  • View in gallery

    Bar graph showing flow distribution in catheter holes, which was derived from the CFD modeling. Black bars represent the standard catheter, whereas the white bars represent the modified catheter with variable hole sizes, the largest of which was located distally or near the catheter tip.

  • View in gallery

    Upper Left: Diagram depicting the view of the water table model and the direction of flow. Upper Right and Lower: Time-lapse photography sequences obtained in the water table model, visualizing flow with the aid of wax flakes. The earliest sequences show that there was no flow through Inlets 1 through 4. The majority of the flow occurs through Inlets 7 and 8 and, to a lesser degree, Inlet 6.

  • View in gallery

    Bar graph showing flow distribution through catheter holes, which were derived from the water table modeling. Gray bars represent the standard catheter, whereas black bars represent the modified catheter with variable hole sizes.

  • View in gallery

    Time-lapse photographs demonstrating the flow of india ink injected into the model ventricular reservoir. Note that all of the ink appears to flow into the most proximal inlet.

  • View in gallery

    Photographs illustrating occluded catheters removed during recent shunt revision surgeries. All catheters are occluded at their most proximal holes.

References

1.

Christens-Barry WAGuarnieri MCarson BS Sr: Fiberoptic delivery of laser energy to remove occlusions from ventricular shunts: technical report. Neurosurgery 44:3453501999Neurosurgery 44:

2.

Collins PHockley ADWoollam DH: Surface ultrastructure of tissues occluding ventricular catheters. J Neurosurg 48:6096131978J Neurosurg 48:

3.

Del Bigio MR: Biological reactions to cerebrospinal fluid shunt devices: a review of the cellular pathology. Neurosurgery 42:3193261998Del Bigio MR: Biological reactions to cerebrospinal fluid shunt devices: a review of the cellular pathology. Neurosurgery 42:

4.

Drake JMKestle JRWMilner Ret al: Randomized trial of cerebrospinal fluid shunt valve design in pediatric hydrocephalus. Neurosurgery 43:2943051998Neurosurgery 43:

5.

Ginsberg HJSum ADrake JMet al: Ventriculoperitoneal shunt flow dependency on the number of patent holes in a ventricular catheter. Pediatr Neurosurg 33:7112000Pediatr Neurosurg 33:

6.

Kaufman BAPark TS: Ventricular anatomy and shunt catheters. Pediatr Neurosurg 31:161999Pediatr Neurosurg 31:

7.

Pattisapu JVTrumble ERTaylor KRet al: Percutaneous endoscopic recanalization of the catheter: a new technique of proximal shunt revision. Neurosurgery 45:136113671999Neurosurgery 45:

8.

Portnoy HD: New ventricular catheter for hydrocephalic shunt. Technical note. J Neurosurg 34:7027031971Portnoy HD: New ventricular catheter for hydrocephalic shunt. Technical note. J Neurosurg 34:

9.

Sainte-Rose C: Shunt obstruction: a preventable complication? Pediatr Neurosurg 19:1561641993Sainte-Rose C: Shunt obstruction: a preventable complication? Pediatr Neurosurg 19:

10.

Sekhar LNMoossy JGuthkelch AN: Malfunctioning verticuloperitoneal shunts. Clinical and pathological features. J Neurosurg 56:4114161982J Neurosurg 56:

11.

Tuli SO'Hayon BDrake Jet al: Change in ventricular size and effect of ventricular catheter placement in pediatric patients with shunted hydrocephalus. Neurosurgery 45:132913351999Neurosurgery 45:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 61 61 15
PDF Downloads 62 62 38
EPUB Downloads 0 0 0

PubMed

Google Scholar