J Neurosurg (Pediatrics 2) 101:64–68, 2004

Performance and complications associated with the

Synchromed 10-ml infusion pump for intrathecal baclofen
administration in children

A. LELAND ALBRIGHT, M.D., YASSER AWAAD, M.D., MICHAEL MUHONEN, M.D.,

WILLIAM R. BOYDSTON, M.D., RICHARD GILMARTIN, M.D., LINDA E. KRACH, M.D.,
MICHAEL TURNER, M.D., KATHRYN A. ZIDEK, M.D., ED WRIGHT, M.D., DALE SWIFT, M.D.,
AND KAREN BLOOM, M.D.

Departments of Pediatric Neurosurgery, Children’s Hospital of Pittsburgh; and Department of

Neurosurgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

Object. The objectives of this multicenter study were to monitor the performance of a 10-ml pump infusing intrathecal
baclofen to treat 100 children with cerebral spasticity, to monitor complications associated with the pump, and to correlate
pump-related complications with body habitus.
Methods. Age at implantation of the pump ranged from 1.4 to 16.8 years (mean 8.1 years). The effects of ITB on spas-
ticity in the upper and lower extremities were evaluated using the Ashworth Scale. Data were collected regarding implant
site, infection, complication, and body mass index (BMI). Ashworth Scale scores decreased significantly in the upper and
lower extremities at 6 and 12 months after pump implantation (p  0.001). There were four serious system-related com-
plications, all specific to catheters. There were 32 serious procedure-related complications in 21 patients: 11 complications
were infections that occurred in nine patients. Four of nine pump-induced infections were treated with pump removal and
antibiotic therapy; five infections were treated successfully with antibiotic therapy alone, without pump removal. In chil-
dren younger than 8 years of age there was a significantly higher incidence of serious procedure-related adverse events
than in older children. There was no significant correlation between BMI and the incidence of pump pocket–related com-
plications or infections.
Conclusions. The 10-ml pump can be used therapeutically in small children, particularly those weighing less than 40
lbs, with greater ease and less wound tension, than the conventional 18-ml pump. The incidence of complications associ-
ated with the 10-ml pump in younger children appears to be similar to that previously reported with the 18-ml pump in
larger-sized children.

KEY WORDS • baclofen • infusion pump • spasticity • cerebral palsy • complications •

pediatric neurosurgery

URING the past decade, several groups have docu-
D mented the effectiveness of implanted pumps to
infuse intrathecal baclofen for the treatment of
patients with severe spasticity or dystonia.1–3,5,7 The most
commonly used pump has been the programmable 18-ml
one made by Medtronic, Inc. (Minneapolis, MN). The
dimensions of that pump are 65
28 mm, approximately
the size of a hockey puck. A pump of that thickness, how-
ever, is sometimes difficult to implant into small or thin
children.
In July 1996, the Center for Devices and Radiological
Health approved a smaller Medtronic pump, with a reser-
voir of 10 ml and a thickness of 22 mm. That pump was
intended for patients considered to be too small for the
Abbreviations used in this paper: BMI = body mass index;
CAP = catheter access port; CSF = cerebrospinal fluid; FDA =
Food and Drug Administration; ITB = intrathecal baclofen; SD =
standard deviation.
conventional 18-ml pump. The FDA approved the use of
the 10-ml pump with the stipulation that clinical data be
collected for 1 year on 100 pediatric patients in whom the
smaller pump was implanted.
This study was conducted to fulfill the FDA require-
ment and to provide additional information correlating
body habitus to pump-related complications. The objec-
tives of the study were to monitor the performance of the
10-ml pump in treating spasticity, to monitor complica-
tions associated with the pump, and to correlate any com-
plications with body habitus.
Clinical Material and Methods
This was an open-label, prospective, nonblinded study,
conducted in 11 US medical centers (Appendix 1). Each
center obtained institutional review board approval for the
study. Patients enrolled in the study consisted of two
groups: 1) 14 children who had been enrolled in the

64 J. Neurosurg: Pediatrics / Volume 101 / August, 2004

Brought to you by Universidad De Antioquia | Unauthenticated | Downloaded 10/17/20 07:30 PM UTC
Synchromed baclofen pump
pre–market approval evaluation of the 10-ml pump; and 2)
86 children who were treated with ITB after approval of the
device. The study was closed on December 10, 2001.
Pumps were implanted after induction of general anes-
thesia. Techniques of pump implantation and choices of
intrathecal catheters were left to the discretion of neuro-
surgeons, as was the prevention and management of com-
plications.
Standardized data forms were completed at enrollment
and at follow-up visits 6 and 12 months postoperatively.
Evidence of baclofen effect on spasticity was evaluated
using the Ashworth Scale to grade muscle tone in the upper
and lower extremities. Data were collected concerning the
pump implant site, including the abdominal quadrant of
implantation and any evidence of inflammation, infection,
hematoma/seroma, CSF accumulation, skin erosion, or
wound dehiscence.
A patient’s height and weight were measured during the
course of treatment. As an estimate of nutritional status,
height, and weight were combined in the BMI (BMI =
weight in kilograms/height in meters). The associations be-
tween BMI, infections, and complications were evaluated.
To monitor drug delivery by the pump, at the time of
each refill, investigators compared the difference between
the programmer-calculated residual volume and the actu-
al residual volume aspirated from the reservoir. Adverse
events were recorded on an adverse event case report form
used to collect data about signs and symptoms, complica-
tions, interventions, possible cause, and outcome. Organ-
isms causing infections were recorded.
Adverse events were graded by the medical director of
Medtronic Drug Delivery rather than by participating in-
vestigators in an attempt to achieve uniformity in grading.
Events were classified as serious if they were life threat-
ening, resulted in patient death, resulted in permanent or
significant disability or incapacity, necessitated invasive
intervention, prolonged hospitalization for 24 hours or
more, or resulted in an additional hospitalization. Centers
were notified whenever an event was designated as seri-
ous; data were also provided to the FDA. Adverse events
not meeting the aforementioned criteria were classified as
nonserious and included events such as vomiting, pneu-
monia, wound hematoma, seroma, inflammation, and uri-
nary retention.
Complications were classified as the following: a) sys-
tem related (those related to the pump, catheter, or pump
access port); or b) procedure related (those related to pump
implantation, pump refills, or reprogramming). Special at-
tention was given to any complication that may have been
an infection related to the implanted infusion system.
Height and weight were correlated with the frequency
of complications by using the Fisher exact test. Changes
in baseline Ashworth Scale scores at 6 and 12 months
were compared using paired t-tests. Lower-extremity, up-
per-extremity, and total scores were analyzed. Data were
analyzed with statistical software (SAS version 8.2; SAS
INstitute, Inc., Cary, NC) .
Results
Of 102 patients enrolled in the study, two were subse-
quently excluded because they were older than 16 years of
age at the time of pump implantation. Spasticity was asso-
J. Neurosurg: Pediatrics / Volume 101 / August, 2004
Brought to you by Universidad De Antioquia | Unauthenticated | Downloaded 10/17/20 07:30 PM UTC
ciated with cerebral palsy in 90%, traumatic brain injury
in 5%, anoxia in 3%, holoprosencephaly in 1%, and
Huntington disease in 1%. The male/female ratio was
65:35. The mean age at first implantation was 8.1 years
(range 1.4–16.8 years); seven patients were younger than
4 years of age when the pump was implanted. Ninety-two
patients completed the study; five pumps were explanted
before completion of the study, and three patients were
lost to follow up when they moved to other states. The
mean duration of therapy was 11.8 months (range 0.5–17
months). In all patients a 10-ml pump was implanted; 53
pumps were equipped with catheter access ports.
Most pumps (55%) were inserted in the right lower
quadrant; 32% were located in the right upper quadrant,
3% in the left upper quadrant, and 10% in the left lower
quadrant. There was no association between pump
implantation site and risk of complications.
Data pertaining to refilling of the pumps were available
from 399 instances. The mean difference between the pro-
grammer-calculated residual volume and the actual resid-
ual volume was 0.59  0.67 ml ( SD), with the pro-
grammer calculating that 0.59 ml more remained in the
pump than was removed. These results were within the
pump specifications of  15%.
Changes in Ashworth Scale scores are shown in Table
1. Spasticity in the upper and lower extremities was sig-
nificantly decreased at 6 and 12 months compared with
baseline (p  0.001).
There were four serious system-related complications,
all specific to the catheters. One catheter broke, one dis-
connected at the lumbar site, one migrated out of the
intrathecal space, and one disconnected from the pump.
Two of these catheter-related complications occurred in
cases involving catheters not made by Medtronic; their
dimensions may not have fit well with the Medtronic
components. There were no nonserious system-related
complications.
Thirty-two procedure-related complications, occurring
in 21 patients, were graded as serious (Table 2). The com-
plications were observed during an overall follow-up
duration of 1178 months; thus, the 1-year cumulative rate
of freedom from serious procedure-related complications
was 79%. There were 22 nonserious procedure-related
complications that occurred in 18 patients. None of the
adverse events associated with the 10-ml pump was unan-
ticipated.
The relationship between the presence of a side port
(CAP) and pump pocket–related complications was eval-
TABLE 1
Summary of Ashworth Scale scores after treatment with ITB
Change From Baseline No. of Cases Mean  SD p Value
upper extremity
to 6 mos 70 0.34  0.87 0.002
to 12 mos 70 0.39  0.88 0.001
lower extremity
to 6 mos 79 0.93  0.94 0.001
to 12 mos 81 0.97  0.88 0.001
total score
to 6 mos 81 0.72  0.89 0.001
to 12 mos 83 0.75  0.85 0.001
65
 A. L. Albright, et al.

TABLE 2 TABLE 4
Serious procedure-related complications during the Serious procedure-related adverse events
12-month follow-up period* stratified by age group*
No. of No. of Age Group
Component Adverse Event Events Cases
Factor 8 Yrs 8 Yrs p Value
implant inversion 2 2
catheter breakage 1 1 no. of patients 55 45 NA
dislodgment from intrathecal space 1 1 no. of patient yrs 53.34 44.91 NA
disconnection at pump 1 1 no. of serious procedure AEs 25 7 0.007†
disconnection at lumbar site 1 1 no. of patients w/ 1 AE 14 7 0.32‡
pump pocket infection 3 3 no. of infections 8 3 0.36†
hematoma/seroma 6 6 no. of patients w/ 1 infection 6 3 0.51‡
CSF accumulation 4 3 * AE = adverse event; NA = not applicable.
wound dehiscence 2 2 † Based on Poisson rate per person-year of exposure.
lumbar infection 1 1 ‡ Based on Fisher exact test.
lumbar hematoma/seroma 1 1
lumbar CSF accumulation 1 1
lumbar wound dehiscence 1 1
CSF leak 3 3
meningitis 1 1 neously and connected to the pump during a later oper-
spinal headache 1 1 ation.
other event Of 23 adverse events reported, none was considered to
dural leak CSF accumulation 1 1 be caused by the implanted system. These events were
catheter lumbar CSF accumulation 1 1 usually associated with an intercurrent illness, injury, or
insertion
activity.
* There were a total of 32 adverse events occurring in 21 patients. The We correlated serious procedure-related adverse events
1-year complication-free rate was 79% (95% confidence limit) with age by comparing those occurring in children
younger than 8 years with those at least 8 years of age
(Table 4). In the former group significantly more adverse
uated. Pump pocket complications occurred in 11 of 46 events were demonstrated than in their older counterparts,
patients without a CAP compared with 12 of 54 patients although there was no significant intergroup difference in
with a CAP, obviously an insignificant difference. In the incidence of infections.
terms of overall complications, 16 were demonstrated in Although investigators had been requested to obtain
46 without a CAP and in 13 of the 54 with a CAP. weight and height measurements on three serial occasions,
Eleven infections occurred in 9 patients during an over- these data were often incomplete. Measurement of height
all follow-up period of 1178 months (Table 3). The result- was difficult in some children because of their contrac-
ing 1-year cumulative infection-free rate was 92%. Ten of tures (deformities) or because they could not stand up to
the 11 infections were classified as procedure related be measured. Baseline height, weight, and BMI data are
because they occurred within 2 months of pump implan- summarized in Table 5. The median BMI was at approxi-
tation. The infecting organism was identified in only three mately the 25th percentile for an 8-year-old child, based
of 11 infections; two infections were caused by Staphylo- on the previously established criteria.6 For example, a 4-
coccus aureus and one by a nonspecified Staphylococcus ft-tall child with a BMI of 14.96 (the mean in this study)
species. Of the nine patients with infections, four under- would weigh only 48 lb. Table 6 provides data that docu-
went explantation of the pumps; five with either pump ment the changes in height, weight, and BMI during the
site– or lumbar-related infections were treated effectively study. The increases in height and weight between base-
with antibiotic agents and completed the study. Data were line and 12-month evaluation (7.95 cm [3.1 in] and 3.3 kg
not available on the use of intra- or perioperative antibiot- [7.3 lb], respectively) were statistically significant. The
ic drugs. Three of the infections were associated with the difference in BMI between baseline and 12 months was
use of a nonstandard catheter that was initially external- not statistically significant.
ized for the screening trial and then implanted subcuta- To evaluate the relationship between BMI and infec-
tions or complications, the occurrence of those events was
TABLE 3 compared between children with a BMI below the group
Infections documented during the 12-month follow-up period*
median and those with a BMI above the median. In pa-
tients in whom BMI data were available, five infections
Infection No. of Events No. of Cases were documented in four patients with a BMI lower than
the group median, and four infections occurred in three
inflammation 1 1 patients with a BMI greater than the median. The inci-
pump pocket infection 5 5
wound dehiscence 1 1
dences of infection in the two groups was not statistically
lumbar infection 2 2 different; however, it is important to note that this study
lumbar wound dehiscence 1 1 had only 12% power to detect a doubling of the percent-
meningitis 1 1 age with infections in the low-BMI group. When the num-
* There were a total of 11 events that occurred in nine patients (at least ber of pump pocket–related complications was correlated
one event). The 1-year cumulative infection-free rate was 91.9% (95% with BMI, the association neared significance (p = 0.09
confidence limits 86.5–97.3%). Fisher exact test); in 12 of 33 patients with a BMI below

66 J. Neurosurg: Pediatrics / Volume 101 / August, 2004

Brought to you by Universidad De Antioquia | Unauthenticated | Downloaded 10/17/20 07:30 PM UTC
Synchromed baclofen pump

TABLE 5 In a recent publication investigators reported that

Summary of baseline height, weight, and BMI data
Factor No. of Cases Mean  SD Median
height (cm) 65 113.33  16.21 111.80
weight (kg) 100 19.58  6.33 17.70
BMI 65 14.96  2.44 14.71
the median 13 complications occurred, whereas in five of
32 patients with a BMI above the median nine complica-
tions were observed. None of the complications could be
attributed to patient growth.
Discussion
The 10-ml pump was shown to perform within manu-
facturer’s specifications for delivery of  15% of the pro-
grammed volume. Delivery accuracy was evaluated by
comparing the measured residual volume against the
pump-calculated volume. The mean difference of 0.59 ml
was within the 95% confidence interval for the mean
(0.52–0.66 ml). Syringe volume measurements are inher-
ently less accurate than device-estimated calculations of
residual volume because of error in syringe marking,
human error in reading the syringe volume at the time of
both emptying and refilling, and error due to the inability
to empty the pump contents completely.
The changes in Ashworth Scale scores were consistent
with those reported in other studies.3,5 Our findings demon-
strated again that chronic ITB infusion decreases upper-
and lower-extremity spasticity and that this improvement
occurred in children whose median age (8 years) was lower
than that in previous multicenter studies. The efficacy of
ITB in younger children is not surprising; the first reported
successful use of ITB was observed in a 4-year-old child.4
In these authors’ experience, ITB was successful in treating
spasticity in a 9-month-old child who suffered from spas-
ticity after a near-drowning incident.
The incidences of adverse events and complications
reported in association with the 10-ml pump were similar
to those reported in other series involving evaluation of
the 18-ml pumps manufactured by Medtronic.3,5 None of
the complications was unexpected.
pumps equipped with a side port were associated with a
higher incidence of complications that those without one.
Range Based on the data in our larger multicenter study, no such
79.09–160.00 relationship was found. We would not have predicted that
9.00–43.20 pumps with side ports were associated with a higher inci-
10.82–23.60 dence of complications, because the incidence was only
slightly higher than those without side ports, and we be-
lieve that they offer distinct advantages, particularly the
ability to allow for dye studies to evaluate the catheters, as
well as the ability to withdraw CSF to evaluate culture or
baclofen levels.
The infection rate (11 infections in nine children) was
similar to that reported in other multicenter studies.5 In our
opinion, it was somewhat surprising that five infections,
including four pump site–related infections, were apparent-
ly eradicated by antibiotic therapy without requiring pump
removal. Cerebrospinal fluid cultures were not obtained in
these cases. At our center, when a pump infection develops,
we aspirate and culture fluid around the pump and also
aspirate and culture CSF from the pump side port. If bacte-
ria grow in both sites, we remove the entire system in near-
ly all cases. When bacteria arise from fluid around the
pump and the CSF, we have successfully eradicated the
infection in only one case (of ~12 attempts), even when
intravenous antibiotic agents were supplemented with
those administered into the CSF via the pump.
The incidence of infections did not appear to be related
to nutritional status as reflected in the BMI; however,
because of our small sample size we cannot draw any con-
clusion about a correlation; the statistical beta error in the
calculations was high. In most of the patients the BMI was
less than the 25th percentile, and it is possible that the rate
of infections was higher than if the series had involved
children with a normal BMI.
The male/female ratio enrolled was approximately 2:1.
Cerebral palsy occurs with equal frequency in the sexes,
and spasticity would be expected to occur with equal fre-
quency as well. The cause of the sex discrepancy in ITB
treatment is unknown. Parents of girls might theoretically
be more hesitant to sanction the 3-in abdominal incision
required for either the 10-ml or the 18-ml pump than par-
ents of boys. Alternatively, parents of boys might pursue
more aggressive treatments in hopes of increasing their

TABLE 6
Changes in height, weight, and BMI during the 12-month follow-up period
Mean Value
Mean %
Change From Baseline No. of Cases Baseline Follow Up Gain SD of Gain Increase

height gain (cm)

to 6 mos 27 108.55 114.67 6.11* 3.63 5.83
to 12 mos 33 110.90 118.85 7.95* 3.94 7.41
weight gain (kg)
to 6 mos 65 17.83 19.95 2.12* 1.69 12.15
to 12 mos 85 19.05 22.36 3.30* 3.43 18.19
BMI
to 6 mos 27 14.69 14.74 0.05† 1.20 0.42
to 12 mos 33 14.61 14.91 0.30† 1.19 2.26
* Statistically significant change (p  0.001).
† Not statistically significant.

J. Neurosurg: Pediatrics / Volume 101 / August, 2004 67

Brought to you by Universidad De Antioquia | Unauthenticated | Downloaded 10/17/20 07:30 PM UTC

son’s activities. The sex of children who presented to the
individual institutions for treatment of spasticity, whether
for ITB or another therapy, was not recorded.
It was predicted that the 10-ml pump, which is approx-
imately one-third thinner than the standard 18-ml pump,
would be associated with a lower incidence of complica-
tions than the larger pump because the former would
cause fewer pump insertion site–related complications.
The finding that the rate of complications was similar to
that demonstrated in studies involving the 18-ml pump
implanted in older children might be extrapolated to indi-
cate a lower complication rate than if the larger pump had
been inserted in these younger, smaller children. This pos-
tulate cannot be answered.
The smaller reservoir capacity of the 10-ml pump is a
drawback when used to treat children requiring large daily
ITB doses; in some cases the reservoir must be refilled
every 3 to 4 weeks or more often. The frequent refills have
been speculated as causing a higher infection rate. In fact,
only five infections in the present study occurred later
than 2 months after pump insertion, during chronic refills.
SynchroMed II, the second-generation Medtronic pump,
is now available. It is a more rounded pump than the pre-
sent one; it is one third the height of the present 18-ml
pump and its reservoir capacity is slightly larger (20 ml
compared with 18 ml). In situations in which Synchromed
II is not available, the 10-ml pump can be inserted in small
children, particularly those weighing less than 40 lb, with
greater ease and less wound tension than the 18-ml pump.
Disclaimer
Dr. Albright is a consultant for Medtronic, Inc.
Appendix
Participating Centers and Investigators
Children’s Hospital of Pittsburgh, Pennsylvania
A. Leland Albright, M.D.
Oakwood Healthcare Systems, Dearborn, Michigan
Yasser Awaad, M.D.
Rehab. Associates, Louisville, Kentucky
Karen Bloom, M.D.
68
Brought to you by Universidad De Antioquia | Unauthenticated | Downloaded 10/17/20 07:30 PM UTC
A. L. Albright, et al.
Children’s Healthcare of Atlanta, Georgia
William R. Boydston, M.D., Ph.D.
Wesley Medical Center, Wichita, Kansas
Richard Gilmartin, M.D.
Gillette Children’s Specialty Healthcare, St. Paul, Minnesota
Linda E. Krach, M.D.
Children’s Hospital of Orange County, California
Michael Muhonen, M.D.
Medical City Dallas Hospital, Dallas, Texas
Dale Swift, M.D.
Indianapolis Neurosurgical, Indianapolis, Indiana
Michael Turner, M.D.
University of Missouri, Columbia, Missouri
Ed Wright, M.D.
The Institute for Rehab and Research, Houston, Texas
Kathryn A. Zidek, M.D.
References
1. Albright AL, Barry MJ, Shafton DH, et al: Intrathecal baclofen
for generalized dystonia. Dev Med Child Neurol 43:652–657,
2001
2. Albright AL, Cervi A, Singletary J: Intrathecal baclofen for
spasticity in cerebral palsy. JAMA 265:1418–1422, 1991
3. Albright AL, Gilmartin R, Swift D, et al: Long-term intrathecal
baclofen therapy for severe spasticitiy of cerebral origin. J Neu-
rosurg 98:291–295, 2003
4. Dralle D, Muller H, Zierski J, et al: Intrathecal baclofen for
spasticity. Lancet 2:1003, 1985
5. Gilmartin R, Bruce D, Storrs BB, et al: Intrathecal baclofen for
management of spastic cerebral palsy: multicenter trial. J Child
Neurol 15:71–77, 2000
6. Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al: CDC
Growth Charts: United States. Advance Data From Vital and
Health Statistics; No. 314. Hyattsville, MD: National Center
for Health Statistics, 2000
7. Sgouros S, Seri S: The effect of intrathecal baclofen on muscle
co-contraction in children with spasticity of cerebral origin. Pe-
diatr Neurosurg 37:225–230, 2002
Manuscript received September 15, 2003.
Accepted in final form April 2, 2004.
Address reprint requests to: A. Leland Albright, M.D., Depart-
ment of Neurosurgery, Children’s Hospital of Pittsburgh, 3705 Fifth
Avenue, Pittsburgh, Pennsylvania 15213. email: Leland.Albright@
chp.edu.
J. Neurosurg: Pediatrics / Volume 101 / August, 2004