Journal of Pediatric Surgery (2012) 47, 154159

www.elsevier.com/locate/jpedsurg

Magnetic mini-mover procedure for pectus excavatum III:

safety and efficacy in a Food and Drug Administrationsponsored clinical trial
Michael R. Harrison a , Kelly D. Gonzales a,, Barbara J. Bratton a , Darrell Christensen b ,
Patrick F. Curran a , Richard Fechter c , Shinjiro Hirose a
a

Department of Surgery, Division of Pediatric Surgery, University of California, San Francisco, San Francisco,
CA 94143-0570, USA
b
Department of Orthopaedic Surgery, University of California, San Francisco, San Francisco, CA 94143-0570, USA
c
Department of Clinical Engineering, University of California, San Francisco, San Francisco, CA 94143-0570, USA
Received 30 September 2011; accepted 8 October 2011

Key words:
Pectus excavatum;
Sunken chest;
Magnetic Mini-Mover;
Chest wall deformity;
Pediatric device

Abstract
Purpose: The magnetic mini-mover procedure (3MP) uses magnetic force to gradually remodel pectus
excavatum deformity. A magnet is implanted on the sternum and coupled with an external magnetic
brace. Under Investigational Device Exemption and Institutional Review Board approval, we performed
a pilot study of safety, probable efficacy, and cost-effectiveness of this new treatment of an orphan
disease using an implantable pediatric device.
Methods: Ten otherwise healthy patients, ages 8 to 14 years, with severe pectus excavatum (pectus
severity index [PSI] N 3.5) underwent 3MP treatment (mean, 18.8 2.5 months). Safety was assessed by
postimplant and postexplant electrocardiograms and monthly chest x-rays. Efficacy was assessed by
change in pectus severity index as measured using pretreatment and posttreatment computed
tomographic scan. Cost of 3MP was compared with that of standard procedures.
Results: The 3MP device had no detectable ill effect. Device weld failure or malpositioning required
revision in 5 patients. Average wear time was 16 h/d. Pectus severity index improved in patients in the
early or mid puberty but not in patients with noncompliant chest walls. Average cost for 3MP was
$46,859, compared with $81,206 and $81,022 for Nuss and Ravitch, respectively.
Conclusion: The 3MP is a safe, cost-effective, outpatient alternative treatment for pectus excavatum that
achieves good results for patients in early and midpuberty stages.
2012 Elsevier Inc. All rights reserved.

Pectus excavatum, or sunken chest, is the most common

congenital chest wall defect. It is characterized by a posterior
depression of the sternum. The standard surgical interventions
Corresponding author. Tel.: +1 415 476 4086; fax: +1 415 476 2314.
E-mail address: kellydgonzales@comcast.net (K.D. Gonzales).
0022-3468/$ see front matter 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.jpedsurg.2011.10.039

are the Nuss and Ravitch procedures [1,2]. In patients with

pectus excavatum, the magnetic mini-mover procedure (3MP)
applies a sustained outward force on the depressed sternum to
slowly bring the sternum into the correct anatomical position
[3]. The Magnimplant (Texcel, LLC, East Longmeadow,
Mass) consists of a neodymium iron boron rare-earth disc

3MP for PE: safety and efficacy

magnet (1.5 diameter 0.1875 thick) and a ferromagnetic
focusing plate encapsulated in a low-profile titanium shell that
is positioned on the front of the sternum. It is inserted through a
2-cm substernal incision under general anesthesia in an
outpatient procedure. The Magnimplant is secured to the
sternum by screwing it into the threaded stem of a titanium disk
backplate that goes through the sternum and is brought into
position using a flexible monofilament guide wire (Fig. 1A).
When treatment is completed, the device is removed in an
outpatient procedure by making an incision through the prior
implant incision, opening the fibrous capsule around the
device, and then unscrewing the Magnimplant.
The Magnatract is a custom-made, external orthotic brace
with a housing unit containing a second rare-earth magnet
(Fig. 1B). The brace is held in place using the attractive force
between the coupled internal and external magnets, and the
patient is able to adjust the force exerted on the sternum.
Force and temperature are measured every 10 minutes with
an onboard sensor, and the data are downloaded at monthly
visits to monitor brace wear compliance.
We previously reported the design evolution of the
Magnimplant and the Magnatract, as well as our development of the surgical procedure [4,5]. After obtaining
institutional review board approval (H603-28932), Investigational Device Exemption (IDE G050196), and funding
through the Food and Drug Administration (FDA) Office of
Orphan Product Development Grant Program (no.
R01FD00341), we conducted a pilot study to test the safety
and proof of concept of the procedure. We now report the
final results of this small FDA-sponsored pilot study.

1. Materials and methods

Ten otherwise healthy patients (ages 8-14 years) with
severe pectus excavatum (pectus severity index [PSI] N 3.5)
who met all inclusion criteria were enrolled in this single
institution study between July 2007 and April 2011. Two
surgeons (M.H. and S.H.) performed the 3MP procedure in all
10 subjects. Although the FDA limited the duration of

155
treatment to 18 months, treatment ranged up to 25 months
because some patients refused to be explanted to end
treatment. We excluded from the study patients with other
congenital anomalies (e.g., scoliosis); bleeding disorders; heart
disease such as arrhythmia; pacemaker implant or living with
someone with a pacemaker; complicated chest deformity (e.g.,
Poland syndrome); contraindication to foreign body implantation; increased risk of complications with general anesthesia;
or respiratory conditions requiring steroid treatment in the last
3 years, were pregnant, or patients who refused to wear a brace.

1.1. Safety
To determine whether the magnetic field associated with
either the Magnimplant or the Magnatract could have any
effect on cardiac electrical function, we performed an
electrocardiogram (EKG) at the preimplant, 1-month postimplant, and 1-month postexplant time points. We performed
monthly wound checks to monitor for skin changes and
infection and performed monthly chest x-rays to check
device integrity.

1.2. Compliance
A component of the Magnatract is a data logger able to
record force and temperature every 10 minutes. Brace-wear
(interpreted as compliance) was defined as an attractive force
greater than 0.1 N and temperature greater than 25C. Wear
time data were downloaded at each monthly visit to
track compliance.

1.3. Proof of concept

We used PSIs obtained from computed tomographic (CT)
scans pretreatment and posttreatment to objectively measure
efficacy, although respiratory variations made interpretation
of the measurements difficult. We administered quality-oflife questionnaires 1 month postimplant, 1 month postexplant, and 1 year postexplant for subjective assessment of
severity/correction of the pectus defect.

Fig. 1 Magnimplant and Magnatract devices. A, The Magnimplant consists of a neodymium iron boron rare-earth disc magnet (1.5
diameter 0.1875 thick) and a ferromagnetic focusing plate encapsulated in a low-profile titanium shell that is positioned on the front of the
sternum and is secured to the sternum by screwing it into the threaded stem of a titanium disk backplate. B, The Magnatract is a custom-made,
external orthotic brace with a housing unit for a second rare-earth magnet. The brace is held in place using the attractive force between the
coupled internal and external magnets, and the force exerted on the sternum is adjustable.

156

M.R. Harrison et al.

1.4. Cost-effectiveness
We compared patient care costs, including surgical,
anesthesia, and hospital costs, for study patients with those
of patients who underwent either the Nuss or Ravitch
procedure at our institution during the same period. We
excluded costs associated with procedures and tests unique to
the study protocol (e.g., CT scans, monthly visits, chest xrays), as they will not be standard of care if 3MP is used in
the general population.

1.5. Statistical analysis

All statistical analyses were performed using GraphPad
Prism 5.03 (GraphPad Software, La Jolla, Calif) and
included paired t test and Wilcoxon matched pairs signed
rank test. Where appropriate, analyses were 2-tailed, and the
results were considered significant for P b .05.

2. Results
Seven boys and 3 girls with a mean age of 12.7 2 years
participated in the pilot study. Their data are summarized in
Table 1. The average duration of device implant was 18.8
2.5 months. One subject (S0008), who was withdrawn from
the study because of early device failure, elected to continue
treatment with a new device implanted under compassionate
use designation. After device explant, all patients were seen
at 1 month for examination and CT scan and at 1 year for
follow-up visit and questionnaire.

2.1. Safety
None of the subjects experienced a change in electrocardiogram or had any clinical indication of cardiac malfunction. There were no incidents of permanent skin damage or
Table 1

discoloration from wearing the external orthotic device,

although one patient experienced mild erythema that
resolved with brace reconfiguration. There were no infections associated with device implantation or chronic use,
but 3 patients did experience postoperative wound infection
after removal of the Magnimplant. Of these three, one patient
required hospitalization for intravenous antibiotics and
wound exploration for possible osteomyelitis that proved
negative, and the other 2 patients were treated as outpatients
with oral antibiotics.
We previously reported the evolution of the surgical
procedure to implant the magnet on the sternum, including
the use of a flexible monofilament guide wire to guide the
stem of the posterior plate through the hole in the sternum
[5]. Initial problems with retained pleural air were solved by
obtaining a chest x-ray in the operating room before
awakening the patient from anesthesia. Air was evacuated
in 3 patients while under anesthesia. Operating time for both
implant and explant decreased from 105 to 45 minutes as
techniques and instruments evolved.
Outpatient reoperation was required 6 times in 5 patients.
One reoperation was to loosen an overtightened implant that
was causing the patient persistent pain. Two were to either
reattach or replace a device that had become nonfunctional
because a misalignment in cross-threading at the initial
operation caused it to uncouple later. Three reoperations
were to remove devices that had a mechanical failure at the
weld point between the posterior plate and the female
threaded post (the break was found on routine monthly chest
x-ray without symptoms, as the device components were
encased in fibrous tissue that prevented migration).
The one case that raised possible safety concern was the
development of a significant pericardial effusion 16 months
after implantation that required urgent pericardiocentesis. A
thorough search (imaging, echocardiogram, blood, and skin
tests) for any evidence that would link the pericardial
effusion to the implanted sternal magnet (and thus require
urgent removal) revealed no association. The implant was

Data of mean compliance per 24 hours and pectus severity index pretreatment and posttreatment

Subject

Age at
implant, y

Sex

Total months
in study

Mean
compliance, %

Pectus severity index

at enrollment

Final pectus
severity index

Correction,
%

S0001
S0002
S0003
S0004
S0005
S0006
S0007
S0008
S0009
S0010

14
14
14
12
11
8
12
14
14
14

Male
Male
Female
Female
Male
Female
Male
Male
Male
Male

18
17
20
17
25
16
19
19
18
19

48.8
86.2
60.5
72.9
61.2
71.1
67.6
92
74.5
65.6

4.3
3.71
3.6
6.1
3.96
4.19
7.3
4.86
4.1
5.66

3.48
3.55
4.24
6.37
3.01
3.66
8.12
3.76
5
7.65

78.1
34.8
182.9
9.5
133.8
56.4
20.2
68.3
105.9
82.6

Achieved percentage of ideal correction is defined as [(PSIpretreatment PSIcurrent)/(PSIpretreatment PSIcutoff)] 100, where PSIcutoff = 3.25 (cutoff value that
differentiates normal control from pectus excavatum patient population). S0008 was withdrawn from study and overall analysis because of early device
failure. PSI indicates pectus severity index.

3MP for PE: safety and efficacy

157

not removed; the pericardial effusion was aspirated and did

not recur. The consulting cardiologists and immunologists
concluded that the pericardial effusion was viral in origin and
had no relation to the 3MP device.

2.2. Compliance
A data logger residing in the external Magnatract brace
made it possible to record force and temperature over time and
thus provide objective measurements of compliance based on
brace wear. However, recorded force was not precise enough
to correlate force as experienced by the patient to improvement of the pectus defect. Wear time varied significantly
among patients but was surprisingly good overall (average,
16 h/d; range, 11-23 h/d). We also found that wear time
generally increased over the course of the study.

2.3. Proof of concept

For this study, we used the preoperative and postoperative
PSI measured using CT scan as a rough guide for efficacy of
treatment [6]. However, we maintain that PSI is uninformative in assessing the effect of treatment as it is a dynamic
measure that varies with patient position and respiration at
any given measurement, in general worsens (PSI increases)
during the rapid growth spurt in most individuals, and there
is no reliable control data on the magnitude of this increase
to compare with the changes seen in our patients. It may be
that the effect of the 3MP treatment may be primarily in
preventing the expected worsening of the defect during the
pubertal growth spurt. Thus, when the means of the
pretreatment and posttreatment PSI are compared, there is
no statistical difference. When individual patient changes in
PSI are analyzed using the Wilcoxon matched pairs test,
there is no statistical difference (Fig. 2). There was
improvement in patients who were in the early or midpuberty
growth spurt but minimal or no improvement in PSI in
5 older patients who were nearing the end of the pubertal

Pectus Severity Index

S0001

S0002

S0003

S0004
S0005

S0006

S0007

S0009
S0010

2
Before Treatment

After Treatment

Fig. 2 Comparison of pectus severity index from pretreatment to

posttreatment computed tomography. There was no statistical
difference (P = .472; confidence interval, 0.974 to 0.494) when
the means of the pretreatment and posttreatment PSI were
compared. Individual patient changes in PSI analyzed using
the Wilcoxon matched pairs test also showed no statistical
difference (P = .652).

growth spurt. Unfortunately, we cannot know how the PSI

would have changed with time without intervention in
these patients.
A subjective measure of improvement is the subject's
perception of improvement as reflected in brace wear
compliance (hours brace wear per day). Interestingly,
compliance in this study overall was high (67.6% 10.5%
brace wear per day). Furthermore, compliance increased over
the course of treatment for most subjects, suggesting that
they were satisfied with the improving appearance of their
chest over time enough to encourage brace wear. Another
subjective measure comes from the quality-of-life questionnaires. For example, when asked, How satisfied are you
with the correction of your chest?, the average response
from the 9 patients was unsure to satisfied (excluding the
patient who was withdrawn from the study). In addition, 7 of
9 patients also stated they would recommend this treatment
to someone else.

2.4. Cost-effectiveness
Excluding the costs for the research-specific tests and
imaging studies covered by the grant, the average cost for
3MP was $46,859. This amount includes the surgical
procedures to implant and explant, anesthesia, and orthoticrelated expenses (fittings and adjustments). The average cost
for the Nuss or Ravitch performed at our institution during
the same period was $81,206 and $81,022, respectively. The
average billed cost for 3MP was 58% of that for the
standard procedures.

3. Discussion
We learned a great deal from this pilot study. Regarding
safety, the trial showed no detectable ill effect on cardiac
function, the biology of cartilage remodeling, or on wound
repair from the magnetic field or the force applied to the
sternum. There were no infections or significant skin damage
from implantation and brace wear, although 3 skin infections
occurred after explant. The 3 mechanical failures of the
device in which the weld of the threaded post to the posterior
plate broke and the 2 events involving misalignment of the
threaded post led to design improvements to eliminate these
problems in the future. We addressed the initial problems
with retained pleural air with intraoperative chest x-ray and,
if needed, intraoperative evacuation.
We learned about surgical considerations in implanting a
magnet on the sternum. Although 3MP is performed as a
simple outpatient procedure, difficulties encountered during
the implant and explant procedures required evolution in
technique and in the device itself. For example, difficulty in
working in the small space behind the sternum was overcome
using a flexible monofilament guide wire to position the
posterior plate. We also redesigned the attachment

158
mechanism when we learned that the fibrous encapsulation
of the backplate made removal difficult.
We made improvements in the external Magnatract
device over the course of this trial including (1) increasing
the strength of the external magnet, (2) seating the magnet in
a high-permeability steel (mu metal) cup designed to focus
and increase the strength of the magnetic field facing inward
toward the Magnimplant and decrease it facing outward
away from the patient, (3) decreasing the size and increasing
the battery life of the data logger and housing it inside the
Magnatract cup instead of on the brace, (4) reducing the
overall size and profile of the Magnatract brace allowing
access to deep deformities in patients with breast development, and (5) improving the adjustment mechanism
(threaded cup) to allow the patient to make incremental
adjustments in force exerted on the sternum.
Quantitative assessment of pectus severity during the
period of repair proved difficult by any available technique.
The Haller Index or PSI obtained from CT scan could not be
used serially in our study because of the risk of radiation
exposure and scatter created by the metal from the implanted
magnet housing. Instead, we used monthly chest x-rays
during the implant period and performed a CT scan
preimplant and postexplant. We also documented the defect
from 3 angles with monthly photographs.
We learned that changes in the PSI were uninformative
because there is no control data on the natural progression of
PSI during the pubertal growth spurt. So we do not know
what the change in PSI would have been without treatment.
Another interesting observation is that patients who were
younger (8-12 years) when treatment was initiated had
dramatic initial improvement in their pectus deformity.
However, this early correction can be lost when treatment is
stopped (as required by the FDA for this study) before
pubertal growth is complete. This finding suggests that 3MP
is best suited for patients in the early-pubertal or midpubertal
age group [7].
In this study, we learned that the response to 3MP (and
probably other treatments) is directly related to the
compliance of the chest wall, which correlates with skeletal
maturity. Skeletal maturity can be determined by assessing
bone age from a simple handwrist x-ray. Pectus deformities, similar to scoliosis, tend to increase in severity during
the rapid growth spurt, indicating increased susceptibility to
cartilage remodeling during this time [8,9]. It is well-known
that repair of pectus deformities in patients in the late and
postpuberty stage becomes increasingly difficult as the
malformed cartilages ossify and the chest wall becomes more
rigid [10-12]. We suggest that our technique will be best
applied to those patients who are entering the rapid growth
spurt of puberty. We studied the various methods of staging
patients and found that bone age as determined from hand
wrist x-ray is more reliable than the modified TannerWhitehouse III classification [7,13,14].
In this trial, in addition to patient age and chest wall
compliance, the speed and degree of correction is directly

M.R. Harrison et al.

related to the amount of force applied to the sternum and the
length of time it is applied [13,15]. Patients who had longer
periods of brace wear had more rapid correction. The
duration of treatment in this study was limited by the FDA to
18 months, which proved insufficient for some patients who
were demonstrating ongoing improvement at the end of the
18-month trial period. Because there was no demonstrable ill
effect from the treatment, we have FDA approval to extend
the duration of treatment duration in our upcoming
multicenter clinical trial to 24 months. Our current
recommendation is that patients begin treatment within a
year of starting the pubertal growth spurt, continue until
correction is achieved (12-24 months), and then keep the
implant for intermittent treatment (like a retainer in
orthodontics), targeting removal when the growth spurt
is finished.
Finally, we learned that the process of investigatorinitiated device development for orphan pediatric conditions is daunting. The process of getting an approved IDE
from the Committee for Devices and Radiologic Health
required 8 application revisions and responses for the
course of a year and a half. Every innovative improvement
for the course of the study (4 years) required a revision, rereview, and reapproval of the IDE itself, the Office of
Orphan Products grant protocol, and the University of
California, San Francisco Committee on Human Research
protocol. We found the density of regulatory burden
surprising because the supposed advantage of working in
the pediatric orphan market is that the regulatory path for
approval of a Humanitarian Device Exemption is supposed
to be less rigorous than for a standard premarket approval.
The lesser regulatory burden was supposed to compensate
for the much smaller potential market, and thus the lack of
financial support from industry, venture groups, or other
traditional sources of funding. Therefore, investigatorinitiated projects are entirely reliant on grants, in this case
from the Office of Orphan Products division of the FDA,
and the entire process from IDE approval to grant
competition is daunting. Our experience with the conduct
of this trial suggests that most potentially beneficial or even
lifesaving devices aimed at orphan pediatric markets will
make it to patients only after a lengthy, arduous, and
expensive process.

4. Conclusion
This study of 10 patients establishes the 3MP as a safe,
minimally invasive, outpatient procedure that is a costeffective treatment for pectus excavatum. The technique
produces satisfactory results in younger patients in prepuberty and early puberty stage but not in patients who are
older or in late or postpuberty stage. We recommend timing
the treatment to the rapid growth spurt of puberty using bone
age from handwrist x-rays. For older patients, the 3MP may
become feasible if the compliance of the deformed cartilage

3MP for PE: safety and efficacy

can be modified using surgical or pharmacologic means now
being tested in our laboratory.

Acknowledgment
This work was supported by a generous grant from the
FDA Office of Orphan Products Development Grant
Program (1R01 HD003341).

References
[1] Nuss D, Kelly RE, Croitoru DP, et al. A 10-year review of a minimally
invasive technique for the correction of pectus excavatum. J Pediatr
Surg 1998;33:545-52.
[2] Ravitch MM. The operative treatment of pectus excavatum. Ann Surg
1949;129:429-44.
[3] Jamshidi R, Harrison M. Magnet-mediated thoracic remodeling: a new
approach to the sunken chest. Expert Rev Med Devices 2007;4:283-6.
[4] Harrison MR, Estefan-Ventura D, Fechter R, et al. Magnetic minimover procedure for pectus excavatum I: development, design, and
simulations for feasibility and safety. J Pediatr Surg 2007;42:81-5.

159
[5] Harrison MR, Curran PF, Jamshidi R, et al. Magnetic mini-mover
procedure for pectus excavatum II: initial findings of a Food and Drug
Administration-sponsored trial. J Pediatr Surg 2010;45:185-91.
[6] Haller JA, Kramer SS, Lietman SA. Use of CT scans in selection of
patients for pectus excavatum surgery: a preliminary report. J Pediatr
Surg 1987;22:904-6.
[7] Tanner JM. Normal growth and techniques of growth assessment. Clin
Endocrinol Metab 1986;15:411-51.
[8] Sanders JO, Browne RH, McConnell SJ, et al. Maturity assessment
and curve progression in girls with idiopathic scoliosis. J Bone Joint
Surg Am 2007;89:64-73.
[9] Fonkalsrud EW. Current management of pectus excavatum. World J
Surg 2003;27:502-8.
[10] Fonkalsrud EW, Reemtsen B. Force required to elevate the sternum of
pectus excavatum patients. J Am Coll Surg 2002;195:575-7.
[11] Fonkalsrud EW. 912 open pectus excavatum repairs: changing trends,
lessons learned: one surgeon's experience. World J Surg 2009;33:
180-90.
[12] Molik KA, Engum SA, Rescorla FJ, et al. Pectus excavatum repair:
experience with standard and minimal invasive techniques. J Pediatr
Surg 2001;36:324-8.
[13] Cox LA. Tanner-Whitehouse method of assessing skeletal maturity:
problems and common errors. Horm Res 1996;45:53-5.
[14] Gilli G. The assessment of skeletal maturation. Horm Res 1996;45:
49-52.
[15] Mead J, Sly P, Le Souef P, et al. Rib cage mobility in pectus
excavatum. Am Rev Respir Dis 1985;132:1223-8.