Eur Spine J (2013) 22 (Suppl 2):S118–S130

DOI 10.1007/s00586-012-2327-7

REVIEW ARTICLE

Growing rod concepts: state of the art

Muharrem Yazici • Z. Deniz Olgun

Received: 7 February 2012 / Revised: 10 April 2012 / Accepted: 15 April 2012 / Published online: 8 May 2012
Ó Springer-Verlag 2012

Abstract heatedly discussed subject in spinal surgery in the past few

Introduction Early-onset spinal deformities present mul- years. Advances in implant technology and new under-
tiple challenges to the surgeon. They may be rapidly pro- standing of disease processes have led to groundbreaking
gressive and unresponsive to conservative treatment, changes in the treatment of adult and adolescent deformi-
necessitating surgical intervention at an early age. ties, some of which have been transferred to early-onset
Materials and methods This text attempts to provide a disorders as well. However, the large potential of growth in
review of current literature and to summarize the authors’ this special patient population presents a wide spectrum of
opinions. dilemmas and complications in their treatment.
Results This paper attempts to concisely review available Spinal deformities that cannot be controlled by conser-
literature regarding the growing rod’s inception, evolution, vative means require surgical intervention. This statement
technique, results, and complications and answers some of is no less true for early-onset spinal disorders. While no
the controversy still surrounding it. concise blanket treatment, conservative or surgical, exists
Conclusions The growing rod is one of the first, most for the management of the myriad conditions that make up
evolved, most popular and one of the most heatedly dis- early-onset spinal deformities, the growing rod is one of the
cussed technique of fusionless spinal instrumentation. most popular and widely used methods.

Keywords Early-onset scoliosis
Surgery

Growth preservation Conception and history

One of the most powerful tools against progressive defor-

Introduction mity in the armament of the spinal surgeon is fusion. While
one of the most intricately constructed and in health stable
Due to its in many ways central location in the body, the structures of the skeletal system, the dynamic relationship
vertebral column plays a huge role both physiologically between vertebrae becomes the very reason why deformi-
and functionally in the growth and development of the ties after a certain magnitude progress. Fusion abolishes
child’s body. Any interference with the normal growth and this dynamic relationship, turning the spine into a rigid
development of the spinal column, be it deformity or even construct, aiming to establish balance by decreasing or
the treatment thereof, will invariably have consequences correcting deformity. In an adult or an adolescent with very
for the entire body. This is one of the many reasons why little growth remaining, the loss of this mobility is usually
early-onset spinal deformities in children have become a well compensated. In the young child, however, it causes
unique problems, and correction of the deformity is but a
lesser component of it. More essential is the struggle to
M. Yazici (&)
Z. D. Olgun
preserve the growth potential of the spine.
Department of Orthopaedics and Traumatology,
Hacettepe University, Sihhiye, Ankara 60100, Turkey Due to its central, or axial, location in the body, the
e-mail: [email protected] spine is in the center of all growth and development taking

123
Eur Spine J (2013) 22 (Suppl 2):S118–S130 S119

place around it. Its growth, or lack thereof, affects the force to the rib hump and some amount of traction with a
thorax, the abdomen, and the pelvis. Perhaps the most neck piece. It is repeated every 2–3 months over a period
significant relationship, however, is with the thorax and its of 6–9 months and often successful in moderate deformi-
contents: the lungs. ties. Both brace and cast treatment are associated with skin
The growth of these two structures proceeds together irritation and pressure sores, respiratory difficulty and rib
after birth, continuing in a non-linear fashion. Rapid acc- deformations.
eleration of growth occurs before the age of three and If fusion is out of the picture and delaying tactics no
during the pubertal growth spurt [1]. In a child with a longer work, instrumentation becomes essential for the
healthy spine, the volumes of both these structures are control of deformity. The first instrumentation to gain
proportional to height. Lung growth is a complex process, popular application, the Harrington rod, was also the first
with different parts of the lungs growing at different rates implant to be used for fusionless surgery. Reports of
and different times during the life of the child, and of these, Harrington rods used without fusion go back as early as the
alveolar development continues well into the teens. Dis- late 1970s and early 1980s [5]. Moe et al. [6] described a
ruption of growth as occurs in early-onset spinal deformity technique similar to that used today, where the subcuta-
leads to many pulmonary consequences, one of which has neous insertion of a Harrington rod was intended to func-
been defined as the thoracic insufficiency syndrome, the tion like an internal brace, and it was supplemented with
inability of the thorax to support normal lung function [2]. the use of external orthosis postoperatively. No segments
Growth of the deformed spine, whether the deformity be were fused and the rod distracted periodically according to
of congenital or idiopathic in origin, is not normal. This the indication of an increase in deformity by 10°. Patients
may be due to deficient or excessive growth centers as in this series attained 84 % of expected growth in the
found in congenital deformities, or asymmetric compres- instrumented segment, although many required unplanned
sive and distractive forces exerted upon growth plates in surgery due to implant-related complications. This high
all, or inherent, yet-to-be-discovered faults within the complication rate caused reports of a discouraging nature
growth plate itself. It is, however, well known that a fused [7], including Mineiro and Weinstein [8], who questioned
spine does not grow. It has been previously published that the method due to this and less-than-expected growth in the
both vital capacity and lung diffusion capacity are nega- instrumented segment. In their 2002 report, Acaroglu et al.
tively affected by early fusion [3]. While the adage of ‘a published the results of 12 patients, where a significant
short but straight spine is better than a long and crooked increase was observed in rotational deformity, although
one’ has been an undisputed motto for spinal surgery in the coronal plane deformity could be controlled with subcu-
past, these recent developments regarding complications taneous rod insertion without fusion. According to the
after the loss of growth have spurred the research for authors, this situation may have two explanations: one, the
methods to achieve a long and straight spine. subcutaneous rod is not able to control rotational plane
Once the drawbacks of early fusion became apparent, deformity as effectively as that in the coronal plane and
the search for methods to postpone final fusion for as long two, the spontaneous fusion or ankylosis that is thought to
as possible while maintaining an amount of control over take place in the uninstrumented region of the spine
the deformity began. These so-called delaying tactics are accounting for the increase in rigidity at definitive fusion
conservative methods that are combined with close obser- may have caused an increase in rotation with a mechanism
vation of the patient [4]. As it is now known that the greater similar to the crankshaft phenomenon [9].
the age at commencement of growing rod treatment and These and other reports caused a setback in the popu-
therefore the fewer the surgeries the child has to undergo, larity of the method during the 1990s and early 2000s, until
the less chance of complications occurring exists; these the 2005 report by Akbarnia et al. where the technique was
delaying tactics are used for growing rod treatment as well. revised to include two instead of one rod on the concave
Among these are brace treatment, serial casting and halo- side, subperiosteal dissection only at anchor sites to mini-
gravity traction. Bracing, one of the oldest conservative mize the possibility of implant failure and periodic
methods for the control of spinal deformity, can be used in lengthening twice a year regardless of the progression of the
patients with idiopathic or idiopathic-like curves if the curve. They achieved good correction and control of
compressive action of the brace does not interfere with deformity and an average T1–S1 length increase of
respiratory function, as may occur in patients with neuro- 1.21 cm/year. While 48 % of the patients had complica-
muscular disorders and other medical conditions. In some tions of any nature, only 17 % required unplanned surgery.
cases, serial casting may be preferable over a brace as it is The authors concluded that the dual growing rod technique
basically a brace that cannot be removed and therefore was safe and effective, and had an acceptable rate of
negates cooperation problems. Casting in spinal deformity complications compared with previous reports while pre-
attempts to correct the deformity by an indirect molding serving near-normal growth [10]. Consistently better results

123
S120 Eur Spine J (2013) 22 (Suppl 2):S118–S130

have been reported with the use of double rods instead of a The approach is begun with a midline longitudinal
single one [11, 12]. incision. Depending on the preference of the surgeon, two
Changes in implant technology have led to changes in separate incisions or one long incision can be used. Sub-
anchor selection as well. The technique, originally defined muscular dissection is continued through the incision but
for Harrington instrumentation, has over time evolved to only single motion segments at the anchor points are
include hooks, bisegmental claw formations using hooks, exposed subperiosteally. The implants of choice are placed
specifically designed growing (or tandem) connectors and at the previously selected anchor levels. Either hooks,
the use of cantilever torsion maneuvers for deformity claw-constructs using hooks or pedicle screws can be used
correction alongside the original primary distraction forces. as implants; the author prefers pedicle screws at both ends
as previous studies have reported increased stability with
their use and transverse connectors [13]. Only the facet
Technique joints at the anchor levels are excised and fusion per-
formed; inter- and intraspinous ligaments and facet joint
Selection of instrumentation levels is performed by the capsules at adjacent levels are preserved. Once implant
analysis of deformity in standing anteroposterior, lateral placement is complete, four rods, two proximal and two
and especially traction X-rays taken under anesthesia. As distal, are contoured according to the sagittal alignment of
distraction is still the primary method for deformity cor- the spine, taking into account natural kyphosis. The rods
rection, Harrington principles are used and vertebrae within are placed in a submuscular location within grooves
the stable zone are selected as distal and proximal anchor established in the paravertebral muscles and engaged into
points (Figs. 1, 2). the anchor implants. A transverse connector can be used

Fig. 1 The patient in this figure is a 105-month-old girl with a posterior, supine traction under general anesthesia, and lateral standing
diagnosis of congenital scoliosis. She underwent growing rod surgery radiographs at index surgery. b Preoperative clinical appearance of
at a standing height of 135 cm 1 year after the excision of a the patient. c Standing anterior–posterior and lateral X-rays after index
diastometamyelic spur and dural repair, and a ventriculoabdominal surgery. d Patient’s clinical appearance before her fifth lengthening
shunt implantation in the neonatal period. a Standing anterior–

123
Eur Spine J (2013) 22 (Suppl 2):S118–S130 S121

Fig. 2 The patient from Fig. 1 underwent a total of 11 lengthenings. at this stage. c Standing anterior–posterior, right-side bending, left-
No complications requiring unplanned surgery were encountered. Her side bending and lateral X-rays of the patient 6 months after implant
standing height before finalization of growing rod treatment (implant removal. Obvious increase in deformity and loss of trunk balance can
removal) was 153.5 cm, her forced vital capacity 1.90 L (70 % of be appreciated. Patient underwent definitive treatment with posterior
expected) and forced expiratory volume in the first second was 91 % instrumentation and fusion after this follow-up appointment. d Stand-
of expected. a Standing anterior–posterior and lateral views of the ing anterior–posterior and lateral radiographs following definitive
patient before finalization of growing rod treatment. b Patient’s fusion. The deformity has been reasonably corrected and balance re-
standing anterior–posterior and lateral X-rays after removal of established. e Patient’s clinical appearance 2 months after definitive
implants. The CT sections show the extent of her congenital deformity fusion

both proximally and distally to connect the two rods initially more in the halo group, but was surpassed by the
together. A growing connector (or tandem connector) is spinal release group at follow-up. A high rate of device
placed between the two rods and using cantilever torsion, complications was seen in the spinal release group, and
the proximal and distal rods are connected to each other. neurologic complications (8 % total) were more frequent in
Distraction and compression between segments are used to the halo-gravity group.
achieve better spinal balance. Closure proceeds in an
anatomic fashion, with or without the placement of a drain
left to the surgeon’s discretion. Results
An external TLSO brace is used for the first 6 months
following index surgery. Regardless of the evolution of the The first extensive report of the modern dual growing rod
deformity, the rods are distracted every 6 months with pre- technique was the aforementioned 2005 study by Akbarnia
planned lengthening surgery. No external support is used et al. In this study, the authors followed 23 patients oper-
after lengthening sessions, which are usually performed on ated on with the dual growing rod technique and included
an outpatient basis. only patients with follow-up of at least 2 years after index
Several additional interventions have been described for surgery, seven of whom underwent final fusion during the
growing rod treatment including apical fusion to arrest the duration of the study. The patients had an average of 6.6
increase in vertebral rotation and the crankshaft phenom- lengthenings. Mean scoliosis improved from 82° to 36° at
enon, annulotomies and other methods of spinal release. final follow-up while T1–S1 length increased from an
Caubet et al. [14] reported on the results of halo-gravity average of 23.01 to 28.00 cm, averaging a length increase
traction compared with surgical release before the of 1.21 cm/year. They also reported an improved space
implantation of fusionless, expandable spinal devices. available for the lung ratio in their patients, going from
They observed that spinal release resulted in an improved 0.87 to 1 at the conclusion of the study. The authors con-
correction of scoliosis over halo-gravity traction or no cluded that the dual growing rod technique maintains the
release, but halo-gravity traction was superior in the cor- correction achieved at index surgery while allowing growth
rection of kyphosis while thoracic spine height increased to carry on.

123
S122 Eur Spine J (2013) 22 (Suppl 2):S118–S130

Akbarnia et al. followed this report with a 2008 study of are close, and while controversy still exists, making grow-
13 patients who all underwent final fusion and had a ing rod treatment a safe and effective treatment technique in
minimum of 2-year follow-up. None of the patients had congenital scoliosis [17]. Similarly favorable results have
undergone previous surgery and all had non-congenital been achieved with the use of growing rods for the treat-
curve etiologies. Scoliosis correction was similar with the ment of spinal muscular atrophy (SMA) as well [18, 19].
previous report, going from 81° to 27.7° after final fusion The effect of growing rod treatment on the rib cage has
while T1–S1 length increased from 24.4 to 29.3 cm, aver- been studied. Sabourin et al. [20] used low-dose stereora-
aging 1.46 cm/year. The authors also compared patients diographic imaging (EOS) to analyze the changes in tho-
undergoing lengthening procedures at most every 6 months racic geometry following the application of a growing rod.
with those where the period between lengthenings was The authors used a single-rod construct in the study and
longer and determined that more frequent lengthenings reported that with Cobb correction from 50.8° to 26°, a
resulted in increased rates of growth (1.8 vs. 1.0 cm/year) complex three-dimensional effect took place on the rib
and better control of deformity (scoliosis correction 79 vs. cage, improving its structure in both the coronal and sag-
48 %). In a 2010 study by Farooq et al. [15], the results of a ittal planes, although this was less extensive than in the
series of 88 patients treated with a single growing rod and case of a spinal arthrodesis. The authors conclude that
lengthening performed at an average of 9-month intervals, longer follow-up and more patients should be analyzed in
scoliosis was observed to have improved from 73° to 44° order to improve understanding of correction forces exer-
with an annual growth rate of 1.04 cm/year. ted upon the rib cage.
Several factors regarding the nature of the child’s
deformity are thought to affect outcome. The effect of
thoracic kyphosis in growing rod treatment has been Complications
investigated in a study by Schroerlucke et al. [16]. The
authors compared complication rates in patients with nor- Although the logic behind its technique is sound and
mal kyphosis and abnormal thoracic kyphosis. In the 90 innovative, a high complication rate reported in the first
patients they analyzed, the authors reported that patients series regarding its use has caused the growing rod to be met
were with hyperkyphosis were 3.1 times more likely to with skepticism and prejudice in the 1990s and early 2000s.
experience implant-related complications than patients with Among the most common complications are implant-rela-
normal kyphosis, and patients with hypo or hyperkyphosis ted ones, due to instrumentation being unsupported by
were more likely to suffer from general medical compli- strong fusion, and postoperative morbidity due to repeated
cations as well. The authors suggested close monitorization surgeries. In the past few years, as long-term follow-up data
of such patients and family counseling. These findings have become available, other, somewhat unexpected com-
should be taken into account when surgical strategy is plications have turned up as well.
planned and any measure taken to improve fixation in hy- Complication rates in with the original, single-rod tech-
perkyphotic patients. In order to avoid junctional kyphosis nique were significantly higher. In the series of Mineiro and
at the proximal anchor sites, the rods should be contoured Weinstein [8] published in 2002, a retrospective radio-
into kyphosis at the top of the construct while interspinal graphical review of 11 patients undergoing subcutaneous
ligaments should be kept intact. The anchor sites can be rod insertion for early-onset deformity unresponsive to
extended to T2 or even higher for this purpose. conservative treatment, patients were operated on using the
Although growing rod treatment was conceived for idi- method described by Moe et al. using Moe rods in nine and
opathic and idiopathic-like deformities, its use in other Harrington rods in two and followed for 5.1 years. The
indications has become widespread mainly due to the highly authors reported 17 complications, an average of 1.5 per
progressive nature of non-idiopathic deformities. In a 2010 patient, on a total of 53 operative procedures, index and
paper by Elsebai et al., the results of 19 patients with con- distraction procedures combined. Rod failure was the most
genital scoliosis undergoing growing rod treatment were commonly encountered complication in this paper, occur-
retrospectively studied. The patients underwent index sur- ring ten times in eight patients. Hook dislodgement in two
gery at an average of 6.9 years and had a minimum of and infection in two (one deep, one superficial) were also
2 years of follow-up, at the conclusion of which their sco- reported. The authors also noted clinically increased ver-
liosis improved from an average of 66° to 47° while T1–S1 tebral rotation despite adequate control of coronal plane
length increased from a mean of 268.3 to 315.4 mm, deformity and concluded that the limited growth obtained
showing an annual growth rate of 11.7 mm/year. Their during treatment did not justify its preference over fusion at
space available for the lung ratio also improved from 0.81 to an early age.
0.94 at last follow-up. While these numbers are lower than In the 2002 paper by Acaroglu et al. [9, 12] patients
those reported for idiopathic and idiopathic-like series, they undergoing fusionless instrumentation without fusion were

123
Eur Spine J (2013) 22 (Suppl 2):S118–S130 S123

evaluated. The average number of lengthening operations due to poor bone quality. The authors reported a total of 177
per patient was reported to be 4.6 and the overall number of complications in 81 patients (58 %), with a mean number of
surgeries, 7. Six patients in this series developed compli- complications of 2.2, 103 (58 %) of which could be
cations, including dislodgment of hooks, rod breakage, addressed during planned surgery. Patients treated with a
facet or laminar fractures and surgical site infection. The single rod had significantly higher complications than
patients were hospitalized for an average of 101 days dur- patients with a dual rod construct while curve correction was
ing the course of the treatment. The authors also observed also found to be more favorable in the dual rod group.
increased rigidity following distraction treatment at defini- Double rods and subcutaneous placement of the instru-
tive surgery, in some instances requiring the use of extra mentation were found to be associated with higher wound-
interventions to increase spinal flexibility. site problems. The authors noted that due to the protracted
Complications were reported also in the study per- treatment period of early-onset deformity regardless of
formed by Akbarnia et al. [10] and published in 2005. The chosen modality, a high rate of complications are expected
authors revised the technique of subcutaneous rodding and and can be reduced by delaying index surgery as long as
used two rods instead of one and dedicated connectors for possible (13 % decrease of complications for each year
distraction procedures with routine lengthening without increase at initial surgery), using dual rods instead of a
waiting for deformity to increase. They reported compli- single rod and decreasing the number of lengthenings as the
cations in 11 of the 23 patients in the series, but noted that complication risk was found to increase by 24 % for each
most complications could be addressed during planned additional surgical procedure performed.
surgical procedures. Two patients had deep wound infec- Risk factors for rod fracture, a common complication
tions requiring surgical debridement, while four had during growing rod treatment, were analyzed in a 2011
superficial surgical site infections, and had to be addressed report by Yang et al. [22]. The records of 327 patients were
with unplanned surgeries. Implant-related complications analyzed and 86 rod fractures observed in 49 patients, 16 of
consisted of two broken rods, two dislodged hooks and one whom had repeat fractures and 8 having more than 2. The
screw pull-out in five patients and all could be addressed most common breakage location was determined to be
during planned lengthening surgery. The authors also above and below the tandem connectors and near the tho-
reported alignment problems, with the crankshaft phe- racolumbar junction. Ambulatory patients, patients with
nomenon occurring in one patient and one junctional ky- syndromic scoliosis and single rods had more frequent rod
phosis requiring extension of the implant construct. The breakages. The authors concluded that risk factors for rod
rate for unplanned surgeries was only 4 % of total surgical fractures include prior breakage, single rods used, stainless
interventions. The authors concluded that while still rea- steel rods over titanium rods, rods with smaller diameters,
sonably high, the complication rate was not prohibitive of shorter tandem connectors and ambulation. Length of
the technique when its multiple advantages are considered. instrumentation, anchor type and pelvic fixation were not
In another report by Akbarnia et al. [12], 13 patients found to have a significant effect. Other studies to decrease
followed for 3–11 years after final fusion were analyzed. implant-related complications have been published as well
Six patients (46 %) experienced complications, including [23].
the treatment period and follow-up after final fusion. Only Another complication deserving specific mention is
three implant-related complications were observed during junctional kyphosis. When spinal motion is eliminated by
the treatment period, two of which were rod breakages that fusion and forces exerted upon the spine are altered due to
did not require unplanned surgery and were addressed the presence of implants, creating significant difference of
during final fusion. One patient had proximal hook pullout, load between segments, it is impossible to avoid increased
which required unplanned surgery. This patient also stresses at junctional regions. This localized concentration
developed deep wound infection and underwent two of stress often causes adjacent segment degeneration and/or
unplanned surgical operations. Although the complication the development of new deformities. Soft tissue preserva-
rates were still found to be high, the authors concluded that tion during surgery, the localization of the junctional
the technique was a reasonable choice in the treatment of region, the patient’s bone quality, the preoperative severity
early-onset deformity. of deformity and how well the sagittal contours of the spine
In a more recent report, 140 patients regardless of diag- have been restored postoperatively are all factors that
nosis were analyzed especially for complications of growing influence the frequency and severity of adjacent segment
rod treatment [21]. The patients underwent 823 planned problems. This complication has been defined in the
(140 index, 633 lengthenings and 50 final fusions) and 74 pediatric age group after treatment of AIS and Scheuer-
unplanned surgeries, with a mean number of lengthenings of mann’s kyphosis and how frequently it occurs after
4.3 per patient and a mean interval between them of growing rod surgery is a matter of debate. Repetitive dis-
10.4 months. The treatment was aborted in only one patient tractions during growing rod treatment may pose a risk

123
S124 Eur Spine J (2013) 22 (Suppl 2):S118–S130

factor not found in instrumentation and fusion. On the other designed, magnetically controlled growing rod. Five ani-
hand, the lack of fusion in growing rod treatment may mals in the experimental growth underwent weekly spine
allow the spinal implant to be less rigid and therefore distraction, achieving an average of 39 mm after a 7-week
decrease the stresses that occur at junctional areas. This period and were compared to a sham group of three ani-
speculation has been confirmed by the experimental study mals. While the experimental group achieved 80 % of
performed by Yilgor et al. [24] where the authors observed predicted spinal height with the distraction equipment, an
that the quantity of forces exerted upon adjacent segments accelerated increase in vertebral body height was observed
after growing rod application are in fact close to normal after the magnetic implants were removed. No complica-
values. Bess et al. [21] have reported 3 patients with tions related to the implant occurred in this study.
junctional kyphosis in a series of 177. Akbarnia et al. [25]
have suggested proper contouring of the rod with regards to
kyphosis and the careful preservation of ligaments during Controversies
surgery to avoid this complication. In a recent study,
Skaggs et al. [26] have proposed that this complication is While there is some agreement in practice and principle
actually more frequent than has been reported and that this regarding the use of growing rods for curves over 60° in
low incidence is due to an error in measurement. However, patients under the age of ten, no consensus yet exists
this observation is limited to the patients of one center. It is regarding the optimal age for index surgery, the manage-
unknown what the rate of this complication would be in ment of sagittal plane deformity, suitable diagnoses for
series where its incidence was found to be low if the growing rod treatment, the interval between lengthenings,
patients’ data were to be re-measured accordingly to the types of foundations or the placement of rods subcutane-
authors’ method. ously or submuscularly [30, 31].
Neurologic complications in straightforward growing Although the effectiveness and reliability of the
rod surgery are rare. They may be caused by excessive growing rod in early-onset scoliosis has been proven
distraction or significant deformity correction. Sankar et al. numerous times in the past years by reports published in
[27] studied the neurologic risk in growing rod surgery and prestigious journals, the negative psychological and social
questioned the need for neuromonitoring. They reviewed effects of young children having to go through multiple
data from 782 growing rod surgeries, 252 of which were surgical procedures so early in their development on them
index surgery, 168 implant exchanges and 362 were and their families has become a point of discussion. In an
lengthenings, 73 % of which were performed with neuro- attempt to quantify the social consequences of the
monitoring. Only one injury occurred in the series, during growing rod, the data of 265 patients from 16 interna-
an implant exchange (pedicle screw placement), resulting tional centers were studied [32]. More than 90 % of the
in an injury rate of 0.1 %. Neuromonitoring changes patients undergoing growing surgery were found to be
occurred in 0.9 % of index surgeries, 0.9 % of implant \10 years of age, and active treatment to take on average
exchanges, and 0.5 % of lengthenings. The only case 5 years, which meant patients had the potential to
where this change occurred during lengthening was a undergo up to 12 procedures during this time frame. Five
complicated case that had had a change during index sur- of the 16 centers experienced resistance toward regular
gery as well. The authors concluded that the use of neur- lengthening on part of the family. This study underlines
omonitoring is justified when implants are placed for the the necessity of having mutual understanding and sound
first time or exchanged, but lengthenings, especially in cooperation with the family.
patients who have had no problems during index surgery, In order to analyze these issues and determine the
do not seem to require routine neuromonitoring. The severity of it, quality of life studies with validated outcome
authors also note that their evidence is anecdotal and urge measures are required. A questionnaire specific to early-
caution while interpreting their results. onset scoliosis is now available [33]. The application of
The finding that increased number of surgeries causes an this questionnaire in large patient series with a diversity of
increased number of complications has ignited a flurry diagnoses that are from different cultures and having these
of research into the development of remotely controlled results compared with healthy subjects and patients
devices where lengthening can be performed without an requiring hospitalizations or multiple surgeries for pathol-
invasive procedure. Experimental studies of these implants ogies other than spinal deformity will guide the way to a
are scant. Takaso et al. [28] reported good results with their better understanding of the impact of treatment. It is not
remote-controlled growing rod instrumentation on beagle sufficient to measure the success of treatment only with the
dogs with induced scoliotic deformities. More recently, numbers of deformity analysis, trunk growth and pulmon-
Akbarnia et al. [29] published the results of their experi- ary functions; quality of life needs to be assessed, and only
mental study on immature pigs instrumented with a newly then it is possible to speak of success or failure.

123
Eur Spine J (2013) 22 (Suppl 2):S118–S130 S125

What implant is the most favorable one for foundation/ into account when proximal anchor sites are chosen and
anchor sites? rods should be contoured accordingly to avoid junctional
kyphosis. Generally, this places anchor sites in high tho-
There is variation in surgeons’ choice of anchor implants and racic (T2–3) and high lumbar (L2–3) in idiopathic and
no clear consensus on which implant to use. Mahar et al. [13] idiopathic-like deformities. Congenital scoliosis poses a
reported on the biomechanical comparison of various anchors challenge for implant site selection and these patients
and observed that screws at both anchor sites with cross-link should be considered on a case-to-case basis.
constructs demonstrated the greatest load to failure, and all
screw-only constructs were superior to constructs including
hooks. These findings are concurrent with other reports Should the rods be placed in a submuscular
regarding pedicle screw stability. Pedicle screws provide or subcutaneous location?
sound three-column fixation with greater pull-out strengths,
making them the author’s first choice for anchor fixation. Although the technique was initially defined as the ‘sub-
Skaggs et al. compared complications between hooks cutaneous rod’, wound complications and implant promi-
and pedicle screws and found that out of 896 pedicle screws, nence have caused surgeons to search for an alternative
there were 22 (2.4 %) complications directly related to the placement of rods. Submuscular placement offers the best
screw while of 867 hooks studied, there were 60 (6.9 %) alternative in this group of patients, who, due to their
complications observed. None of the complications were conditions and deformity, are generally of small stature
associated with a neurologic or vascular injury. The authors with very little subcutaneous fat. In the 2011 paper by Bess
concluded that pedicle screws in growing rods have sig- et al. [21], it was found that subcutaneous placement in
nificantly less complications than hooks [34]. double rod constructs was associated with a significantly
Pelvic fixation is often used in growing rods applied to higher incidence of wound-site complications, while sub-
patients with long sweeping curves as encountered in neu- muscular placement decreased them.
romuscular and syndromic scoliosis. Sponseller et al. studied
the outcomes and complications in this subgroup of by
reviewing 36 patients with growing rods anchored in the What should be the frequency of lengthening?
pelvis and compared rod breakage rates with patients whose
constructs were not fixed in the pelvis [27]. It was observed The first reported indication for lengthening procedures
that iliac screws achieved better deformity and pelvic obliq- was determined to be a 10° increase in deformity during
uity correction than sacral fixation. The same was found to be follow-up after index surgery. This evolved to routine
true for double rods as well. Rod breakage rate was not found lengthening to catch up and keep pace with growth, with
to be significantly higher in patients with pelvic fixation than variable intervals of time due to reports detailing near-
in patients without pelvic fixation. Iliac screw breakage was normal growth with such practice [10, 12]. It has also been
found to be a complication. The authors concluded that observed that more frequent lengthenings result in a greater
both iliac screws and rods provide satisfactory distal fixation amount of growth achieved during the course of treatment
in growing rod constructs that need to span the pelvis and [12]. However, an increased number of surgeries has been
that lumbar lordosis may be better preserved with these found to be associated with an increased rate of compli-
rather than spine-to-spine constructs as iliac anchors extend cations as well. The current trend is to perform routine
anterior to the center of mass. The more frequent breakages lengthening every 4 months in very small children, every
observed in iliac screws do not seem to affect outcome. 6 months in most children and every 9 months when the
involved segment of the spine is short [25], with compli-
cations occurring during the course of treatment addressed
How many segments should be included during planned surgery whenever possible.
in the construct?

As the growing rod technique still primarily employs dis- Does the underlying diagnosis present an indication
tractive forces for correction, Harrington principles come or contraindication for growing rod treatment? Is
into consideration when selecting anchor sites. Vertebrae the growing rod treatment suitable for idiopathic
within the stable zone as defined by Harrington should be and idiopathic-like deformities only, or can it be used
used. Determination of the stable vertebrae is, in the in other diagnoses as well?
author’s experience, best performed on traction X-rays
obtained under general anesthesia just before the com- While the growing rod was initially defined for idiopathic
mencement of surgery. Thoracic kyphosis should be taken and idiopathic-like deformities, early-onset deformity

123
S126 Eur Spine J (2013) 22 (Suppl 2):S118–S130

includes many diagnoses, including syndromic cases, In a recent study by Yang et al. [36], the results of 16
cerebral palsy, congenital deformities, Marfan syndrome, patients with spinal deformity associated with cerebral
SMA and others. The treatment of these incurs additional palsy that were treated with growing rods were analyzed.
challenges, especially when their generally unresponsive The patients’ mean deformity decreased from 83° to 51°
nature to conservative treatment is considered. and their T1–S1 length increased an average of 9 cm
In the 2011 study by Elsebai et al., the data of 19 during treatment with lengthenings every 9 months. Only
patients with congenital deformities undergoing growing one significant implant-related complication was observed
rod treatment were analyzed. Correction of deformity and while there were six deep wound infections and four
maintenance of this correction and growth obtained during pneumonias. The authors’ results indicate that growing rod
treatment were found to be comparable to results in the treatment can be successfully employed in the treatment of
idiopathic-like group, with improvements in space avail- severe scoliotic curves due to cerebral palsy.
able for the lung ratio. The complication rate was not found
to be different. The authors concluded that the growing rod
is a safe and effective treatment for selected patients with Do we reach our intended result with growing rod
congenital spinal deformities. treatment?
Growing rod treatment has also been used in SMA, a
disorder associated with early-onset, highly progressive The treatment of early-onset scoliosis is a special kind of
curves usually unresponsive to conservative treatment and challenge. It involves a child, often too young to completely
complicated with comorbidities of other systems, such as understand the intricacies of his or her treatment, and a
frequent pulmonary infections and poor nutritional status. family that is worried and anxious. Conservative treatment
McElroy et al. [18] compared 15 SMA patients with 80 in itself is a long-term commitment, with restrictive and
juvenile/infantile idiopathic scoliosis patients undergoing uncomfortable braces or casts worn over long periods of the
growing rod treatment (Figs. 3, 4). Curve correction, pelvic child’s life, often interfering with daily activities, causing
obliquity and space available for the lung ratios in SMA its own set of complications that range from skin problems
patients improved, but the treatment was not found to halt to an inability to control deformity. Its effectiveness also
rib cage collapse. Hospital stays in SMA patients were leaves a lot to be desired.
longer compared with patients with idiopathic deformities, Considering the growing rod as what it is—a sort of
but the rate of major complications was found to be lower. temporary internal brace intended to tide the child over to
Chandran et al. [19] studied 11 patients with SMA types I as close to skeletal maturity as possible before definitive
and II undergoing growing rod treatment and observed fusion is performed, and to let him or her grow in the
good curve correction and a low rate of complications meanwhile as well—it can be said with some confidence
(only postoperative medical complications). Growing rods that it does reach its intended goal. Its complications are
seem to be a viable treatment option in early-onset scoliosis frequent and have many facets, from objective ones such as
associated with SMA. implant failures and wound site infections, to the psycho-
Growing rod treatment in deformities associated with logical and familial effects of having a child forced to
Marfan syndrome has also been studied. Sponseller et al. in undergo surgery, be hospitalized and receive general
their study included ten patients with deformities that anesthesia routinely every 6 months. However, it should
developed before the age of 3 years, who were operated on never be forgotten that these children already struck out on
using the growing rod technique and lengthenings per- the chance of a normal life with their highly progressive,
formed once a year for patients on warfarin treatment. Mean severe curves, and the growing rod provides advantages
curve correction for dual rods (7 of the patients) was that would be impossible to achieve with early fusion.
reported to be 60 %, with an overall length of 11.5 cm
obtained during the course of the study. Two rod breakages,
one anchor dislodgment, and three intraoperative dural Can the spine be lengthened at the same pace
leaks were reported. The authors concluded that growing throughout the duration of treatment?
rods are an effective treatment for early-onset deformity in
Marfan syndrome, helping prevent large infantile curves It has been observed before during the course of growing
from becoming severe and allowing definitive fusion closer rod treatment that the spine stiffens with repeated distrac-
to skeletal maturity. Acute heart failure following growing tion procedures, in some cases going on to ankylosis
rod surgery for Marfan syndrome-related scoliosis has been of uninstrumented (and unexposed) segments. In most
reported, thought to result from mechanical torsion of cor- patients, repeated distractions increasingly require more
onary arteries due to overdistraction. Release of distraction force and achieve less length. This phenomenon has been
reversed the symptoms in this patient [35]. studied.

123
Eur Spine J (2013) 22 (Suppl 2):S118–S130 S127

Fig. 3 The patient in this figure is a 74-month-old girl with infantile capacity 2 L (76 % of expected) and forced expiratory volume in the
idiopathic scoliosis whose standing height at the commencement of first second 94 % of expected. b Standing anterior–posterior and
growing rod treatment was 115 cm. a Preoperative and immedi- lateral X-rays 6 months after the last lengthening. Excellent correc-
ately postoperative standing X-rays after index surgery. The patient tion of deformity and establishment of trunk balance led to the
underwent a total of ten lengthenings and required no unplanned decision of implant removal and observation at the conclusion of
surgery during the treatment period. Her standing height at the growing rod treatment. Immediate postoperative X-rays show good
conclusion of growing rod treatment was 151.5 cm, her forced vital maintenance of correction and balance

Sankar et al. [37] reported on 38 patients with at least 3 T1–S1 distance indicates that despite possible fusion, a
lengthenings and a minimum of 2 years of follow-up and biologically active tissue remains that can still be stimu-
measured T1–S1 lengths after index surgery and each lated to grow.
lengthening. They observed that major curve correction Noordeen et al. [38] measured the forces and the amount
occurred after index surgery (74° to 36°) and the average of distraction required over time in patients treated with the
annual length gain in the T1–S1 segment was 1.76 cm/ growing rod. They prospectively measured the distractive
year. The average T1–S1 gain, however, decreased sig- forces required for 60 lengthenings in 26 patients with a
nificantly with repeated lengthenings and when time was single submuscular rod and observed that the force
considered as a factor, this gain appeared to decrease over required to distract the spine doubled at the fifth length-
time as well. The authors concluded that their findings ening, and it was significantly higher than that required for
might be due to autofusion of the spine or immobilization the fourth lengthening. The mean length that was achieved
by a rigid device. However, the continued gain in the also decreased gradually and became 8 mm or less by the

123
S128 Eur Spine J (2013) 22 (Suppl 2):S118–S130

Fig. 4 a The patient from Fig. 3 returned for a follow-up visit the spine is rigid and immobile despite no formal fusion (or exposure)
6 months after implant removal. Standing anterior–posterior and has been performed on intermediate segments. b Patient’s clinical
lateral X-rays show no increase in deformity and good coronal and appearance at 6 months after implant removal
sagittal alignment. However, bending X-rays clearly demonstrate that

fifth lengthening. It can be concluded, based on this and the fusion takes place late during the course of the treatment,
aforementioned study, that after five or six lengthenings, and even if not, it has been shown in the past years that
more force is needed to distract the spine and even so, less even unsegmented bars show growth when placed under
gain in height is achieved. distraction. There is no doubt that the growing rod causes
the spine to stiffen and lose mobility, but there is also no
doubt that growth is preserved. How does the spine con-
Are there unexpected situations, and if yes, tinue growing if it is fused? If growth continues, how can
how significant are they? we say there is fusion? Is one of these observations wrong?
Quite possibly, they are both right. Cahill et al. noted
Cahill et al. [39] reported on nine patients who had fusion at the last stage, just before definitive fusion. Even if
undergone growing rod treatment and definitive fusion with there is fusion in the spine, it may have occurred very late
a mean of 9.6 years of follow-up. They observed autofu- in the duration of treatment, until which time spinal growth
sion in uninstrumented segments in 89 % of children and is almost complete. Even if spontaneous fusion does take
had to perform on average 7 Smith-Petersen osteotomies to place, this fusion mass is probably a thin, weak layer of
achieve a Cobb angle correction of 44 % at definitive new bone formation that fractures during every lengthening
surgery. The total correction during treatment was 61 %, with distraction. It is also conceivable that this new bone
while T1–S1 length increase was 11.2 cm. The authors formation is a biologically active type of tissue and con-
questioned the fusionless and growth-sparing nature of the tinues growing under distractive forces, much like in the
technique but could not otherwise explain the growth example of the unsegmented bar that grows when it is
achieved during treatment. It is possible that this reported distracted.

123
Eur Spine J (2013) 22 (Suppl 2):S118–S130 S129

Do we preserve growth or stimulate it? Conflict of interest None.

Growth preservation can be achieved with routine distrac-

tion. While this determination is an important milestone in References
the treatment of early-onset scoliosis, it also brings with it
the question of whether growth is preserved or stimulated. 1. Dimeglio A (2001) Growth in pediatric orthopaedics. J Pediatr
Orthop 21(4):549–555
Patients in Akbarnia et al.’s [10, 12] series who were, for 2. Campbell RM Jr, Smith MD, Mayes TC et al (2003) The char-
various reasons, lengthened more frequently than 6 months acteristics of thoracic insufficiency syndrome associated with
showed a greater amount of growth. Yilmaz et al. [40] fused ribs and congenital scoliosis. J Bone Joint Surg Am
showed in their experimental study on 12 immature pigs 85-A(3):399–408
3. Goldberg CJ, Gillic I, Connaughton O et al (2003) Respiratory
that vertebral growth continues during growing rod treat- function and cosmesis at maturity in infantile-onset scoliosis.
ment at an accelerated but statistically not significant rate. Spine 28(20):2397–2406
In a 2012 study performed at the authors’ institution [41], 4. Johnston CE (2011) Delaying tactics: traction, casting and brac-
vertebrae within and outside instrumentation levels in 20 ing. In: Yazici M, EPOS (eds) Non-idiopathic spine deformities
in young children, chap 9. Springer, Heidelberg, pp 109–121
patients with idiopathic-like deformities undergoing grow- 5. Gillespie R, O’Brien J (1981) Harrington instrumentation without
ing rod treatment were measured and compared for growth. fusion. J Bone Joint Surg Br 63:461
Growth taking place in vertebrae within instrumentation 6. Moe JH, Kharrat K, Winter RB, Cummine JL (1984) Harrington
levels was 7.0 mm compared with 5.2 mm in vertebrae instrumentation without fusion plus external orthotic support for
the treatment of difficult curvature problems in young children.
outside instrumentation levels. Considering that the verte- Clin Orthop Relat Res 185:35–45
brae outside instrumentation levels are lumbar and expected 7. Rinsky LA, Gamble JG, Bleck EE (1985) Segmental instru-
to grow even faster than their within-instrumentation mentation without fusion in children with progressive scoliosis.
counterparts, this study indicates that growth may be stim- J Pediatr Orthop 5(6):687–690
8. Mineiro J, Weinstein SL (2002) Subcutaneous rodding for pro-
ulated rather than preserved with growing rod treatment. gressive spinal curvatures: early results. J Pediatr Orthop 22(3):
If the knowledge that complications will occur is con- 290–295
sidered and accepted, it is possible to control deformity and 9. Acaroglu E, Yazici M, Alanay A, Surat A (2002) Three-dimen-
preserve near-normal growth with the growing rod treat- sional evolution of scoliotic curve during instrumentation without
fusion in young children. J Pediatr Orthop 22(4):492–496
ment without causing excessive morbidity on patients. 10. Akbarnia BA, Marks DS, Boachie-Adjei O, Thompson AG,
Even though it is by definition a fusionless technique, the Asher MA (2005) Dual growing rod technique for the treatment
vertebral column is perturbed many times during the course of progressive early-onset scoliosis: a multicenter study. Spine
of treatment, and while it keeps growing, it is impossible to (Phila Pa 1976) 30(17 Suppl):S46–S57
11. Thompson GH, Akbarnia BA, Kostial P et al (2005) Comparison
say that the spine is completely unaffected. This autofusion of single and dual growing rod techniques followed through
may make it possible in selected cases to remove the definitive surgery: a preliminary study. Spine (Phila Pa 1976)
implants and leave the spine be without additional fusion 30(18):2039–2044
and instrumentation procedures. 12. Akbarnia BA, Breakwell LM, Marks DS et al (2008) Dual
growing rod technique followed for three to eleven years until
final fusion: the effect of frequency of lengthening. Spine (Phila
Pa 1976) 33(9):984–990
Conclusion 13. Mahar AT, Bagheri R, Oka R, Kostial P, Akbarnia BA (2008)
Biomechanical comparison of different anchors (foundations) for
the pediatric dual growing rod technique. Spine J 8(6):933–939
A young child with an early-onset deformity requires a 14. Caubet JF, Emans JB (2011) Halo-gravity traction versus surgical
long-term method to control his or her deformity until release before implantation of expandable spinal devices: a
spinal fusion can be performed with the most minimal of comparison of results and complications in early-onset spinal
disadvantages possible. The growing rod method brings deformity. J Spinal Disord Tech 24(2):99–104
15. Farooq N, Garrido E, Altaf F et al (2010) Minimizing compli-
this child to near skeletal maturity with height similar to his cations with single submuscular growing rods: a review of
or her peers, pulmonary functions above 80 % of normal, technique and results on 88 patients with minimum two-year
but, unfortunately, not a healthy spine. follow-up. Spine (Phila Pa 1976) 35(25):2252–2258
In conclusion, the spinal surgeon is no longer forced to 16. Schroerlucke SR, Akbarnia BA, Pawelek JB et al (2011) How
does thoracic kyphosis affect patient outcomes in growing rod
choose between a crooked-but-long or straight-but-short surgery? Spine (Phila Pa 1976) (Epub ahead of print)
spine. We get a long, straight spine, which also happens to 17. Elsebai HB, Yazici M, Thompson GH et al (2011) Safety and
be rigid. No treatment at the moment can completely erase efficacy of growing rod technique for pediatric congenital spinal
the effects of an inherent error in the spine, curing early- deformities. J Pediatr Orthop 31(1):1–5
18. McElroy MJ, Shaner AC, Crawford TO et al (2011) Growing
onset deformity completely, but as alternatives are scant, rods for scoliosis in spinal muscular atrophy: structural effects,
the growing rod treatment remains our best option in the complications, and hospital stays. Spine (Phila Pa 1976)
treatment of early-onset deformity regardless of etiology. 36(16):1305–1311

123
S130 Eur Spine J (2013) 22 (Suppl 2):S118–S130

19. Chandran S, McCarthy J, Noonan K, Mann D, Nemeth B, 30. Yang JS, McElroy MJ, Akbarnia BA et al (2010) Growing rods
Guiliani T (2011) Early treatment of scoliosis with growing rods for spinal deformity: characterizing consensus and variation in
in children with severe spinal muscular atrophy: a preliminary current use. J Pediatr Orthop 30(3):264–270
report. J Pediatr Orthop 31(4):450–454 31. Vitale MG, Gomez JA, Matsumoto H, Roye DP Jr (2011) Vari-
20. Sabourin M, Jolivet E, Miladi L, Wicart P, Rampal V, Skalli W ability of expert opinion in treatment of early-onset scoliosis. Clin
(2010) Three-dimensional stereoradiographic modeling of rib Orthop Relat Res 469(5):1317–1322
cage before and after spinal growing rod procedures in early- 32. Akbarnia BA, Campbell R, Dimeglio A et al (2011) Fusionless
onset scoliosis. Clin Biomech (Bristol, Avon) 25(4):284–291 procedures for the management of early-onset spine deformities
21. Bess S, Akbarnia BA, Thompson GH et al (2010) Complications in 2011: what do we know? J Child Orthop 5(3):159–172
of growing-rod treatment for early-onset scoliosis: analysis of one 33. Corona J, Matsumoto H, Roye DP, Vitale MG (2011) Measuring
hundred and forty patients. J Bone Joint Surg Am 92(15): quality of life in children with early onset scoliosis: development
2533–2543 and initial validation of the early onset scoliosis questionnaire.
22. Yang JS, Sponseller PD, Thompson GH et al (2011) Growing rod J Pediatr Orthop 31(2):180–185
fractures: risk factors and opportunities for prevention. Spine 34. Skaggs DL, Myung KS, Johnston CE, Akbarnia BA, Group GSS
(Phila Pa 1976) 36(20):1639–1644 (2010) Pedicle screws have fewer complications than hooks in
23. Sankar WN, Acevedo DC, Skaggs DL (2010) Comparison of children with growing rods. International Congress on Early
complications among growing spinal implants. Spine (Phila Pa Onset Scoliosis and Growing Spine, Toronto
1976) 35(23):2091–2096 35. Skaggs DL, Bushman G, Grunander T, Wong PC, Sankar WN,
24. Yilgor C, Aritan S, Demirkran G, Daglioglu K, Kosemetmetoglu Tolo VT (2008) Shortening of growing-rod spinal instrumenta-
K, Yazici M (2011) Does growing rod cause degeneration and tion reverses cardiac failure in child with Marfan syndrome and
secondary deformity in the adjacent segment? J Child Orthop scoliosis. A case report. J Bone Joint Surg Am 90(12):2745–2750
5:394–395 36. Sponseller PD, Wild AT, Growing Spine Study Group (2010)
25. Akbarnia BA, Emans JB (2010) Complications of growth-sparing Growing rods for cerebral palsy. J Child Orthop 4:495–496
surgery in early onset scoliosis. Spine (Phila Pa 1976) 35(25): 37. Sankar WN, Skaggs DL, Yazici M et al (2011) Lengthening of
2193–2204 dual growing rods and the law of diminishing returns. Spine
26. Skaggs DL, Myung KS, Lee C (2011) Proximal junctional (Phila Pa 1976) 36(10):806–809
kyphosis in distraction-based growing rods. Free paper, the 5th 38. Noordeen HM, Shah SA, Elsebaie HB, Garrido E, Farooq N, Al
international congress on early onset scoliosis and growing spine Mukhtar M (2011) In vivo distraction force and length mea-
(ICEOS), 18–19 November 2011, Orlando, Florida surements of growing rods: which factors influence the ability to
27. Sankar WN, Skaggs DL, Emans JB et al (2009) Neurologic risk lengthen? Spine (Phila Pa 1976) 36(26):2299–2303
in growing rod spine surgery in early onset scoliosis: is neuro- 39. Cahill PJ, Marvil S, Cuddihy L et al (2010) Autofusion in the
monitoring necessary for all cases? Spine (Phila Pa 1976) 34(18): immature spine treated with growing rods. Spine (Phila Pa 1976)
1952–1955 35(22):E1199–E1203
28. Takaso M, Moriya H, Kitahara H et al (1998) New remote-con- 40. Yilmaz G, Huri G, Demirkran G et al (2010) The effect of posterior
trolled growing-rod spinal instrumentation possibly applicable for distraction on vertebral growth in immature pigs: an experimental
scoliosis in young children. J Orthop Sci 3(6):336–340 simulation of growing rod technique. Spine (Phila Pa 1976)
29. Akbarnia BA, Mundis GM Jr, Salari P, Yaszay B, Pawelek JB 35(7):730–733
(2011) Innovation in growing rod technique: a study of safety and 41. Olgun ZD, Ahmadiadli H, Alanay A, Yazici M (2012) Vertebral
efficacy of a magnetically controlled growing rod in a porcine body growth during growing rod instrumentation: growth pres-
model*. Spine (Phila Pa 1976) (Epub ahead of print) ervation or stimulation? J Pediatr Orthop 32(2):184–189

123