Secondary Logo

Journal Logo

Focus Paper

Dual Growing Rod Technique for the Treatment of Progressive Early-Onset Scoliosis

A Multicenter Study

Akbarnia, Behrooz A. MD*; Marks, David S. FRCS; Boachie-Adjei, Oheneba MD; Thompson, Alistair G. FRCS; Asher, Marc A. MD§

Author Information
doi: 10.1097/01.brs.0000175190.08134.73
  • Free

Progressive spinal deformity in early life presents significant health risks for the child and a challenge to the treating surgeon.1,2 Many different etiologies such as infantile and juvenile idiopathic scoliosis, congenital vertebral anomalies, neuromuscular conditions, and various syndromes may cause progressive spinal deformities grouped under early-onset scoliosis. Although there are some curves in very young children that do not progress, progress very slowly or even spontaneously resolve, there are others that in spite of nonoperative treatment may show significant deterioration at a young age requiring early intervention.

Significant progression of these curves, if left untreated, may be associated with life-threatening health risks.2,3 Pulmonary development is not complete at birth and thoracic deformity caused by scoliosis may adversely affect lung maturation up to the age of eight.3–5 The primary effect of scoliosis on the developing lung is to inhibit the growth of both alveoli and pulmonary arterioles. This incomplete maturation of the lung and pulmonary vasculature is the primary cause of the ventilation defect seen in patients with early-onset scoliosis.6

Spinal growth is in its peak during the first 5 years of life. The T1–S1 length increases about 10 cm during this time (2 cm per year) and the thoracic and lumbar spine achieves two thirds of its adult height by the age of 5 years. Growth rate then diminishes significantly from 5 to 10 years7,8 before increasing again. Scoliotic spinal segments do not grow normally. Early treatment of progressive curves without arthrodesis allows for spinal growth to continue with different rates based on etiology of the scoliosis and child's age.9 Therefore, treating severe and progressive curves at an early age becomes very important to improve respiratory and visceral development and to normalize spinal growth.

Traditional nonoperative treatment options for early-onset scoliosis include casting, orthotic treatment, or a combination of the two.10 Halo traction may occasionally be helpful in combination with surgery (Johnston C, personal communication). Orthotic treatment, however, is difficult in very young children with severe and progressive curves.11,12 Casting and bracing to control curve progression, especially for long-term use, may be ineffective, since the immature rib cage often deforms before significant correction is transmitted to the spine. Patients who have pulmonary problems may be further compromised by wearing a brace for an extended period of time. In addition, skin problems such as pressure sores may also develop.

Orthotic treatment, however, may be effective in select patients. For infantile scoliosis, orthotic treatment is indicated for curves over 35° and progressive curves.10 A Milwaukee-type orthosis is preferred to an underarm thoraco-lumbo-sacral orthosis (TLSO) to prevent pressure on the thorax, subsequent thoracic deformity, and pulmonary complication. For severe curves, serial casting is an option and is accomplished under general anesthesia until adequate correction is achieved before resuming orthotic treatment. In patients with juvenile curves over 25°, orthotic treatment is continued until skeletal maturity.13 Patients are observed every 6 months thereafter, with standing radiographs, for any progression. Types of orthosis and duration of treatment for curves of other etiologies are based on the diagnosis and rate of progression.

When curves show continued progression and do not respond to nonoperative treatment such as cast or orthosis, surgical intervention is indicated. These curves are usually greater than 50°.12,14 Thoracic curves are treated earlier to prevent pulmonary complications.2 There are two main options available, fusion and nonfusion techniques, when surgery being considered in a very young child. One of the nonfusion options is growing rod instrumentation.

The goal of the growing rod technique is to achieve deformity correction and maintain it during the treatment period while allowing continued spinal growth.12,15,16 Once maximum spinal growth is accomplished, definitive spinal arthrodesis and instrumentation are performed. During the treatment period (from initial surgery to final fusion), the child is brought to the operating room periodically and the instrumentation is lengthened or exchanged. Spinal instrumentation is combined with limited arthrodesis of the anchor sites for added stability. The purpose of this paper is to present the technique, results, and overall clinical experience with the dual growing rod treatment method.

Surgical Technique

Initial Procedure

Surgical technique includes preparation of foundation sites for anchors, rod contouring, subcutaneous or subfascial rod insertion, and insertion of tandem connectors (Figure 1). The skin incision may be one long midline or two separate midline incisions depending on the length of the child's spine.

Figure 1
Figure 1:
Technique of dual rod instrumentation. (Reprinted with permission from Akbarnia B, McCarthy R. In: McCarthy R, ed. Spinal Instrumentation Techniques. Chicago, Scoliosis Research Society, 1998. A, Anteroposterior drawing. B, Lateral view showing the construct contoured to maintain sagittal alignment. Extended tandem connectors are placed in the thoracolumbar spine to minimize profile.


The initial corrective load of the dual rod construct is applied by means of upper and lower foundations. Foundations are defined as assembly of at least two anchors and one or two rods that are stable and strong enough to accept corrective loads and to resist deforming loads without dislodgment of the anchors or plastic deformation of the rod.17 The cephalad and caudal exposures for insertion of anchors is subperiosteal. The remaining area of exposure is subcutaneous or subfascial to prevent premature unwanted fusion.

The selection of the anchor sites is based on the location and type of the curve as well as the child's age and diagnosis. For instance, in patients with neuromuscular conditions, longer instrumentation is preferred. The upper foundation is usually at the T2–T4 levels. Hooks and/or screws can be used for both foundations. We have customarily used hooks for upper foundations, but pedicle screws can also be used, if feasible, based on availability of posterior elements and anatomic variations. Pedicle screws appear to add additional stability to the construct.18 The superior hooks are placed over the transverse process or in a supra-laminar location. The inferior hooks are positioned upward under the facets in a claw fashion. The hooks can be staggered over two levels if there is concern regarding the small size of the spinal canal or if additional stability is required. The caudal foundation levels are usually two or three levels below the scoliosis lower end vertebra. Placing either local bone graft or bone graft extenders at the exposed levels of foundations allows for a limited fusion and added stability at the anchor sites.

Dual Rods

The Pediatric Isola rods are stainless steel and 3/16 inch in diameter. They are cut into two segments, two for each side, and contoured for sagittal alignment. Proper contouring of the rods usually corrects kyphosis by a cantilever maneuver, if the deformity is flexible. At the foundation levels, the rods are connected to the anchors and a transverse connector before all the screws are tightened. There should be two rods connected to each foundation. The proximal and distal rods are then connected by atandem connector on each side. The tandem connectors are attached to the rods by sliding them first cephalad and then caudal and placed at the thoracolumbar junction to have the least adverse effect on the sagittal contour of the spine. They should also be turned appropriately to keep the profile as low as possible.

Postoperative Care

In most patients, a TLSO is used for the first 6 months when the child is upright and until the fusion of the foundation sites is solid. Thereafter, the use of the brace is based on factors such as diagnosis, age, bone quality, and child's level of activity. Fourteen of our 23 study patients (61%) were treated with orthoses for at least 6 months after surgery.

Lengthening Procedure

Lengthening is performed every 6 months as an outpatient or inpatient procedure. The tandem connector is palpated and partially exposed through a small midline incision. The set-screws at one end of the connectors, usually proximal, are loosened, and distraction is accomplished between the two rods by placing a special distractor inside the tandem connector. The set-screws are then tightened. Distraction can also be accomplished between the rod and the tandem connector on either side by using the rod holder and regular distractor. Excessive distraction force should be avoided, especially at the first lengthening. During our preliminary experience with dual growing rods, the timing for the lengthening was decided based on the age, diagnosis, sitting height, and curve progression. It is usually between 5 and 9 months and in most cases around 6 months. Since 1998, the standard has become 6-month intervals between lengthenings. Lengthenings are stopped when no further distraction can be achieved. The patients then undergo the final correction and arthrodesis.

Final Fusion

The final arthrodesis usually necessitates removal of implants, reconstruction, and reinstrumentation. The levels are usually the same as the initial surgical procedure unless progression of the curve either above the fusion, below, or both has occurred.

It is necessary to follow the surgical technique in detail to reduce the rate of complications and to achieve the best long-term results. This is especially important in the initial surgery with regards to proper preparation of the anchor sites and contouring of the rods. It is important to obtain and maintain acceptable coronal and sagittal balance at initial surgery. To avoid proximal junctional kyphosis, the rods should be contoured into kyphosis at the top of the construct and the interspinal ligaments kept intact. The tandem connectors should be placed at the thoracolumbar junction. Short instrumentation, especially in patients with nonidiopathic scoliosis, should be avoided (Figure 2).

Figure 2
Figure 2:
Clinical example of dual growing rod instrumentation in the treatment of a 3 + 11 year-old girl with Marfan's syndrome. A, Early radiograph of the spine taken at the age of 8 months. B, Preoperative anteroposterior radiograph showing significant progression. C and D, Anteroposterior and lateral radiographs at 7 years follow-up (Cobb angle measures 37°). E, Preoperative clinical photograph. F, Postoperative photograph 7 years following initial surgery (age 11 years). The two-incision approach for lengthening is atypical and used only on this patient as a means of alternating the incision site. G, Postoperative lateral photograph (age 11 years).

Materials and Methods

Between 1993 and 2001, 89 consecutive patients from four institutions had posterior single or double “growing rod” instrumentation without fusion for treatment of their progressive early-onset scoliosis. Seventy-two of the 89 patients had dual rod instrumentation using Pediatric Isola implants and a tandem connector long enough to connect proximal and distal rods and accommodate periodic lengthening. We excluded 23 patients because of previous surgical intervention for their scoliosis. In the remaining 49 patients, the procedure was primary. Twenty-three of the 49 patients had greater than 2-year follow-up (range, 2–9.25 years) and are the subject of this study (Table 1).

Table 1
Table 1:
Clinical Data

With the Institutional Review Board's approval, all medical records and radiographs were retrospectively reviewed. The clinical information included age at surgery, sex, diagnosis, prior nonoperative treatment, surgical information (levels of instrumentation, number of lengthenings, lengthening intervals), and surgical complications. Based on the rate of spinal growth described by Dimeglio and Bonnel,7 patients were divided into three groups based on their age: Group 1, from 0 to 5 years (N = 10) with an average age of 3.20 years (range, 1.92–3.92 years); Group 2, 5 to 10 years (N = 12) with an average age of 6.80 years (range, 5.33–9.83 years); and Group 3, consisting of one skeletally immature 12-year-old child. Patients were also divided into two cohorts: those who had finished treatment and reached final fusion and those who are still undergoing treatment.

All upright posteroanterior and lateral radiographs were available before surgery, after surgery (within 1 month after initial surgery), at latest follow-up, and both pre- and post-final fusion. Radiographs were measured in coronal and sagittal planes for scoliosis, kyphosis, lordosis, T1–S1 length, coronal balance, sagittal balance, length of instrumentation, and Campbell's space available for lung ratio (SAL). All postoperative radiographic measurements were calibrated and corrected for magnification to represent actual change. Two observers measured each radiograph independently.

T1–S1 length was measured from the middle of the upper endplate of T1 to the middle of the upper endplate of S1. To calculate Campbell's SAL ratio, the distance from the apex of the T1 rib to the apex of the diaphragm was measured on both the convex and concave sides of the curve. The calculated ratio of the concave to the convex side measurement gave the space available for lung ratio.2

Descriptive statistics were done to determine means and ranges. Changes in radiographic findings between preoperative, postoperative, and follow-up were tested using paired t test and sign rank test when appropriate. Comparisons between groups on variables of interest were done using Student's t test and ANOVA. Correlations were done with the Pearson correlation coefficient.

Surgical Data.

All procedures were primary; no revisions were included in this series. There were 16 girls and 7 boys. Diagnoses included 5 patients with infantile idiopathic scoliosis, 2 juvenile idiopathic, 3 congenital, 2 neuromuscular, 2 Marfan's syndrome, 2 neurofibromatosis, 2 Soto's syndrome, and 1 of each of the following: chromosomal abnormality, congenital hand anomaly, Ulrich syndrome, Beal's syndrome, and Ehlers-Danlos syndrome. Fourteen patients had thoracic curves and 9 had thoracolumbar curves.

Indication for surgery was unsuccessful nonoperative treatment (bracing or casting) with a curve progression of over 10°. The mean age at surgery was 5.43 years (range, 1.92–12.00 years). The average number of vertebral segments instrumented was 13 (range, 11–17). Seven patients underwent final spinal fusion at an average age of 10.24 years (range, 8.08–14.00 years). Of a total of 189 surgical procedures, 23 were initial procedures, 7 were final fusions, and 151 were lengthenings. There were eight unplanned procedures in 4 patients for complications during the treatment period and 3 following final fusion.

The upper foundation anchors were thoracic hooks in all patients. The lower foundation anchors were all hooks in 11, all screws in 2, a combination of hooks and screws in 9, and hook and Galveston fixation to the pelvis in 1 patient. A transverse connector was used at or close to the upper foundation in 6, at or close to the lower foundation in 2, and at both locations in 11 patients. In 4 patients, no transverse connector was used. Bilateral tandem connectors (end-to-end) were used in 20 patients. The 3 remaining patients had side-to-side connectors initially and were switched to tandem connectors later. The location of the dual rods was subcutaneous in 16 and subfascial in 7 patients. The upper level of instrumentation was at T1 in 3, T2 in 5, T3 in 12, and T4 in 3 patients. The lower level of instrumentation ended at L1 in 1, L2 in 12, L3 in 1, L4 in 7, L5 in 1, and the ilium in 1 patient. During the planned lengthening procedures, the rods were changed in 14 patients and a tandem connector was changed in 17 patients.

Six patients also had an anterior annulotomy, to achieve better correction. Two of these 6 patients had nucleotomy in addition to annulotomy. All additional procedures were performed on the same date or before posterior surgery.19


The patients were observed for a minimum of 2 years after initial surgical treatment with an average of 4.75 years follow-up (range, 2.00–9.25 years). During the treatment period (initial surgery to final fusion), which averaged 4.02 years (range, 2.00–6.75 years), the average number of lengthenings was 6.6 (range, 3–11) per patient with an interval of 7.4 months (range, 5.5–21 months). The average number of lengthenings for the first two groups was 6.3 (range, 3–10) and 7 (range, 3–11), respectively. The average interval between lengthenings was 7.4 months (range, 6–12) and 7.5 months (range, 5.5–21), respectively. There was no considerable difference in the number of lengthenings and the interval between lengthenings between Group 1 and Group 2.

Measurement of Curves and Spinal Balance

The scoliosis Cobb angle improved from an average of 82° (range, 50°–130°) pre-initial to 38° (range, 13°–66°) post-initial and 36° (range, 4°–53°) at latest follow-up or post-final fusion (Table 2). The percent change of scoliosis Cobb angle from pre- to post-initial was 53% (9%–80%) and from pre-initial to latest follow-up or post-final was 54% (14%–94%, P < 0.0001). There were 6 patients whose scoliosis correction deteriorated from post-initial to last follow-up or post-final fusion. There was no variable that established an association among these patients, indicating a reason for loss of correction. However, all patients significantly improved from their preoperative deformity.

Table 2
Table 2:
Clinical Data of 23 Patients With 2-Year Follow-up Treated With Primary Dual Growing Rods Without Fusion

Of the 6 patients who had an annulotomy, before or at the time of insertion of posterior instrumentation, the average pre-initial curve was 91° (range, 67°–130°), correctable on side bending radiographs only to 61° (range, 40°–90°). The curves in this group improved to 46° (range, 27°–63°) post-initial and 37° (range, 17°–58°) at latest follow-up or post-final fusion.

Kyphosis was 50° (range, 15°–95°) pre-initial, 35° (range, 10°–80°) post-initial and 45° (range, 20°–105°) at latest follow-up or post-final fusion. Percent change in kyphosis averaged a 25% reduction from pre-initial to post-initial ranging from a 68% reduction to a 73% increase in kyphosis. Progression in kyphosis from post-initial to follow-up or post-final fusion averaged 40%, ranging from a 33% reduction to a 220% increase and was not significant. Lordosis from L1 to S1 measured −45° (range, −78°–17°) pre-initial, −42° (range, −73°–−21°) post-initial and −48° (range, −90°–−17°) at latest follow-up or post-final fusion. The displacement of sagittal alignment between T1 and S1 was 3.72 (range, 0–9.20), 2.33 (range, 3.00–6.80), and 3.92 cm (range, 0.40–8.20 cm), respectively. The coronal balance (deviation from midline) was 2.81 cm (range, 0–15.40 cm) pre-initial, improved to 1.76 cm (range, 0–7.90 cm) post-initial and was 1.96 cm (range, 0–7.30 cm) at latest follow-up or post-final fusion.


The spinal length obtained was calculated from the amount of elongation of T1 to S1 after the initial procedure plus the subsequent growth from post-initial to latest follow-up or post-final fusion (Table 3). This change was also confirmed by the length of the instrumentation from post-initial to pre-final fusion or latest follow-up.

Table 3
Table 3:
Results of 23 Patients With 2-yr Follow-up Treated With Primary Dual Growing Rods Without Fusion

The T1–S1 length increased from 23.01 cm (range, 13.80–31.20 cm) pre-initial to 28.00 cm (range, 19.50–35.50 cm) post-initial (elongation) and to 32.65 cm (range, 25.60–41.00 cm) at latest follow-up or post-final fusion (growth). In addition to initial elongation averaging 5.00 cm (range, 1.30–12.00 cm), the growth over the treatment period was 4.64 cm (range, 0.30–10.70 cm) with an average of 1.21 cm/year (range, 0.13–2.59 cm/year). The growth was 1.19 cm/year (range, 0.13–1.78 cm/year) for Group 1 and 1.13 cm per year (range, 0.17–2.59 cm/year) for Group 2. The mean length increase from pre-initial to post-initial was 23% (range, 6%–62%), from pre-initial to follow-up was 44% (range, 16%–92%) and from post-initial to follow-up was 17% (range, 1%–36%), all statistically significant at P < 0.0001 (Table 4). The length of the instrumentation changed from 23.66 cm (range, 15.00–30.00 cm) post-initial to 27.75 cm (range, 21.20–36.00 cm) at last follow-up or pre-final fusion, an approximate change of 4.67 cm (range, 1.00–10.70 cm). Post-final fusion measurements were not used since the original construct is often replaced at final fusion.

Table 4
Table 4:
T1–S1 Growth From Pre-Initial to Post-Initial Surgery, Pre-Initial to Follow-up, and Post-Initial to Follow-up for the 23 Dual Growing Rod Patients

The amount of initial elongation was also correlated with the initial diagnosis. Three patients with congenital spinal anomalies had a significantly smaller increase in T1–S1 length pre- to post-initial (9%; range, 6%–13%) compared with the rest of the group who had no spinal congenital anomalies (25%; range, 7%–62%). However, there was little difference in T1–S1 growth obtained through serial lengthenings between the congenital and noncongenital groups. There was no difference in growth rate in patients without congenital abnormality regardless of the diagnosis. Also, the number of segments under instrumentation did not correlate with the amount of T1–S1 growth from postoperation to follow-up (r = 0.2; P = 0.3567).

Final Fusion Patient Data

The 7 patients who underwent definitive final fusion were compared with the patients still in active treatment (n = 16). The average age at initial surgery was 6.70 years (range, 2.08–12.00 years) and 4.88 years (range, 1.92–9.83 years), respectively. The average age at final fusion was 10.24 years of age (range, 8.08–14.00 years) with an average treatment period of 3.57 years (range, 2.00–6.00 years). The average treatment period for patients who have not yet undergone definitive fusion is 4.22 years (range, 2.00–6.75 years). The mean number of lengthenings for the definitive spinal fusion group was 6.1 (range, 3–10) at 6-month intervals (range, 5.5–6.7). The patient cohort in active treatment had an average of 6.8 lengthenings (range, 3–11) at an average interval of 8 months (range, 6–21) between lengthenings. The mean scoliosis Cobb angle of the definitive fusion patients was 92° (range, 71°–130°) pre-initial, 39° (range, 15°–62°) post-initial, 33° (range, 4°–53°) pre-final fusion and 26° (range, 4°–53°) post-final fusion. For the remaining patients, mean Cobb angle improved from 78° (range, 50°–125°) pre-initial, to 38° (range, 13°–66°) post-initial, and 40° (range, 18°–53°) at most recent follow-up.

The 7 final fusion patients had an average T1–S1 measurement of 24.59 cm (range, 20.60–31.20 cm) pre-initial, 30.49 cm (range, 26.00–35.50 cm) post-initial, 34.57 cm (range, 31.50–39.00 cm) pre-final fusion, and 36.37 cm (range, 33.10–40.20 cm) post-final fusion. The patients still undergoing lengthening had an average T1-S1 length of 22.33 cm (range, 13.80–29.70 cm) pre-initial, 26.92 cm (range, 19.50–33.00 cm) post-initial, and 31.02 cm (range, 25.60–41.00 cm) at most recent follow-up. Thus far, these patients have a total T1–S1 length increase of 8.69 cm (range, 3.90–14.70 cm) from pre-initial to last follow-up, 53% (range, 18%–96%) of this growth is attributed to initial elongation and 47% (range, 4%–82%) to continued growth. Growth per year for this group averaged 1.01 cm per year (range, 1.30–2.59 cm) after the first surgery. The 7 definitive fusion patients had a total length increase of 11.78 cm (range, 4.60–16.70 cm) from pre-initial to final fusion, 54% (range, 34%–68%) of which was initial elongation and 46% (range, 20%–64%) was growth (1.66 cm per year; range, 0.37–2.35) after the first surgery.


Using Campbell's lung ratio, the SAL2,20 was calculated in 14 patients who had thoracic scoliosis. This ratio was 0.87 (range, 0.7–1.1) preoperation, improved to 0.96 (range, 0.49–1.16) postoperation and 1.00 (range, 0.79–1.23) at follow-up. The mean change was 13% (range, −37%–55%) from preoperation to postoperation, 7% (−11%–64%) from postoperation to follow-up, and 18% (range, −16%–64%) from preoperation to follow-up (P = 0.01).


During Treatment Period.

Eleven of the 23 patients (48%) had 13 complications during the “treatment period” from initial surgery to most recent follow-up or to final fusion. Most complications were able to be addressed during planned lengthening.

There were 2 patients with deep wound infections requiring a total of six unplanned surgical procedures. Both patients were treated by debridement and primary wound closure with 1 patient requiring removal of one of the rods, which was eventually replaced. Four patients had four superficial wound problems, two of which required unplanned surgeries to treat a draining fistula and an incision granuloma. Implant-related complications included 2 rod breakages, 2 hook dislodgements, and 1 screw pull-out in 5 patients. All implant-related complications were addressed during planned lengthening procedures. The two alignment-related complications during treatment period occurred in 2 patients and included 1 crankshaft phenomenon and 1 junctional kyphosis requiring extension of the construct during a planned lengthening.

Following Final Fusion.

Two patients required three extensions of their fusions, two for curve progression and one for lumbosacral pain. One patient's fusion was extended both cephalad and caudal in two separate procedures. The other patient's fusion was extended to the pelvis (Figure 3).

Figure 3
Figure 3:
Venn diagram of the complications during the treatment period and following final fusion in 12 of the 23 study patients.

Unplanned Procedures.

There were eight unplanned procedures in 4 patients during the treatment period, accounting for 4% of total surgical interventions (n = 189). There were an additional three unplanned procedures in 2 patients following final fusion, making the total unplanned procedures 11 in 5 of the 23 patients during the treatment period and following final fusion. As we follow these patients for longer periods of time, it is expected that the complications rates as well as number of unplanned procedures will increase.


Surgical treatment of progressive early onset scoliosis is complex. The decision for choice of surgery is based on the age, diagnosis, type and severity of the curve, and presence of congenital anomalies of the spine.

Spinal Fusion.

Traditional posterior spinal arthrodesis with or without instrumentation in this age group should be supplemented by anterior arthrodesis to prevent crankshaft phenomenon.21,22 Circumferential arthrodesis will halt curve progression; however, at the same time, it prevents future spinal and thoracic growth. Furthermore, studies have shown that fusion surgery at an early age does not improve respiratory function.23

In some patients, such as children with congenital scoliosis, spinal fusion may be indicated. Surgical treatment is often based on the type and location of the anomaly and the age of the patient. If the deformity is limited to a few segments, a limited apex arthrodesis and/or resection may be possible in place of a long fusion and still allow growth of other spinal segments.


Hemiepiphysiodesis of the convex side of the curve with or without instrumentation is another surgical technique used to allow progressive correction and prevent deterioration of the deformity. This technique is most effective in congenital curves and gradual correction may be expected. Hemiepiphysiodesis is not very effective in patients with infantile and juvenile idiopathic scoliosis in achieving the normal spinal growth potentials even with the addition of instrumentation.24 Marks et al reported that simultaneous application of Harrington instrumentation seems to arrest the curve progression and may reduce the thoracic deformity provided that the rod is lengthened.24 However, the procedure does not reverse the established deformity.24–26

Luque trolley instrumentation also has been used in conjunction with or without hemiepiphysiodesis in infantile and juvenile scoliosis.27 Instrumentation alone has not prevented curve progression.28 Additional convex epiphysiodesis has resulted in curve resolution in some patients.28

Anterior Stapling.

Recently, interest has returned to attempts at modulating the growth of the scoliotic spine with anterior spinal growth arrest performed with staples placed endoscopically.29 Historically, this technique had some failures resulting from the loss of staple fixation among other reasons.29 Improved staple designs promise better fixation. Betz et al recently reviewed 21 patients with adolescent idiopathic scoliosis treated with vertebral body stapling.29 However, no patients with infantile scoliosis were included.

Titanium Rib (VEPTR).

The technique of expansion thoracoplasty for the treatment of thoracic insufficiency syndrome with the vertically expandable prosthetic titanium rib (VEPTR) has recently been described.30 This device was designed to treat thoracic deformities resulting from absent and/or fused ribs in congenital and syndromic conditions. This effort has resulted in a broader and deeper understanding of the central role that the spine plays in the architecture and function of the chest wall and thorax. This new awareness emphasizes the importance of evaluating thoracic volume in addition to a standard assessment of anteroposterior and sagittal spinal alignment. Restoring this “fourth dimension” and thereby maximizing the potential for pulmonary development is emerging as an important goal in the treatment of early-onset scoliosis, particularly for those patients with congenital spine and rib anomalies.

Growing Rods.

Harrington originally reported the technique in 1962.15 Moe subsequently developed the technique of “subcutaneous rods” for the treatment of progressive curves in young children and reported it to the Scoliosis Research Society in 1978.31 Several modifications were introduced subsequently to improve the results.32 Luque introduced his technique with sub laminar wires.27 However, long-term follow-up of “Luque Trolley” cases revealed premature auto fusion because of the need for subperiosteal exposure.16,33–35

Few long-term studies of the single growing rod technique method are available to evaluate the growth achieved by these methods. Klemme et al12 reported on a heterogeneous group of 67 children with progressive scoliosis seen between 1973 and 1993. They were treated with a program of incremental-distraction spinal instrumentation without fusion supplemented by full-time external orthotic support. Over the course of treatment, curve magnitude improved from an average of 67° (range, 38°–118°) at initial instrumentation to 47° (range, 19°–88°) at definitive fusion. For all patients, curve response tended to decline with consecutive procedures. The measured growth of the instrumented but unfused spinal segments averaged 3.1 cm (range, 0.0–10.2 cm) over a mean treatment period of 3.1 years (range, 0.5–6.6 years). A total of 30% of patients had arrest of improvement; and in 33%, progression was observed.12 Most of these patients had threaded, modified, or standard Harrington rods, and a few had newer segmental implants.12

The previous studies regarding instrumentation without fusion lack uniformity in data collection as well as the analysis of the results and the actual growth obtained.12,36,37 It is evident, however, that there is a high rate of complications associated with this technique. Blakemore et al reported on 29 patients who underwent single submuscular Isola rod with a 24% complication rate.36 Mineiro and Weinstein studied 11 children who were treated by consecutive distraction of subcutaneous rods, and in 2 patients with rodding and anterior apical fusion.37 At surgery, the average patient age was 5.7 years (range, 2.8–9.0 years), with a mean Cobb angle of 74° (range, 53°–100°) and an average Pedriole angle of 39° (range, 25°–60°). Subcutaneous rodding halted curve progression in all patients. Approximately 5 years after surgery, 1 patient showed no deterioration of the curve and 9 patients showed a 32° (range, 18°–60°) improvement in the magnitude of the original curvature. Eight of these patients had already had definitive surgery performed with segmental spinal instrumentation and fusion. Spinal growth as measured in the instrumented area occurred in all 11 patients and ranged from 0.5 to 4.5 cm (mean, 2.0 cm). It was concluded that subcutaneous rodding with consecutive distraction allows correction of progressive early-onset scoliosis that failed to respond to nonoperative management. Rotational deformity did not deteriorate radiographically, but clinical deformity increased subjectively. There were 1.5 complications per patient. The amount of growth achieved and the number of procedures required to obtain these results raised the question of whether patients would be better served by a single anterior, posterior fusion and instrumentation at a young age.37

A recent study conducted by Acaroglu et al retrospectively evaluated 12 patients who underwent single rod instrumentation without fusion; they noted an improvement in the degree of spinal curvature but with an increase in vertebral rotation.38 As in all previous reports with active distraction technique, a single rod was used to lengthen the concavity of the curve.

Dissatisfaction with outcomes, unpredictability of previous methods of treatment, and high rate of complications has led to further refinement (McCarthy R, personal communication)39–41 and improved methods of surgical technique including the dual rod construct.9,19 Recent reports have also indicated superior results of the dual rod compared with single rod construct in the correction of the adolescent idiopathic scoliosis.42

Initial experience on the dual rod technique and a comparison study of single rod and dual rod techniques in early-onset scoliosis have shown less implant and crankshaft complications compared with previous reports of single rod techniques.14,37,43

Our series include only the patients treated with dual rod instrumentation. There has been no previous report of patients treated exclusively with dual growing rod technique. In our patients, the initial elongation obtained at the first surgery constitutes a major portion of the total increase in sitting height. The growth per year after the initial elongation is very close to the normal growth of the spine. The duration of the treatment has already been longer than previously reported series, and the complication rate is within acceptable limits for these complex group of patients.43

To minimize the development of crankshaft phenomenon and extend the duration of the treatment, the rods are placed bilaterally with transverse connectors attaching them caudal and cephalad for added stability. The lengthening intervals should be timely and bone and soft tissue exposures should be kept to absolute minimum. We now routinely lengthen the construct approximately every 6 months.

The technique is contraindicated in heavy patients and patients who have significantly rigid curves. If the curve magnitude is large and the curve is rigid, a first-stage annulotomy may be indicated before dual rod instrumentation. The technique is also not indicated when there is not enough growth potential to justify multiple surgical procedures. The indications for patients with neuromuscular disorders and patients with significant rib anomalies remain unclear. If the correction is maintained until skeletal maturity without final arthrodesis, it may be possible to remove the implants at the completion of growth, allowing preservation of spinal mobility.


The dual growing rod technique is safe and effective. It maintains correction obtained at initial surgery while allowing spinal growth to continue. It provides adequate stability, increases the duration of treatment period, and has an acceptable rate of complication compared with previous reports using the single rod technique. However, the complications are still considerable, and the family should fully understand the long-term commitment before the treatment is begun. In addition to allowing continued spinal growth, this method of surgical treatment improves the thoracic cage volume in this challenging patient population. This is an ongoing study, and long-term follow-up is needed to confirm our interim findings.

Key Points

  • Dual growing rod technique is a valid surgical alternative for the treatment of progressive early onset scoliosis.
  • Dual growing rod technique provides deformity correction, maintains it throughout the course of treatment, and allows reasonable spinal growth.
  • Early intervention with dual growing rod instrumentation inhibits continued deformation of the thorax, thereby significantly improving the space available for lung ratio.


The authors thank Bruce Gillingham, MD, for his case contribution, Sue Min Lai, PhD, and Andrew Mahar, MS, for statistic analysis, and Pat Kostial, RN, BSN, and Sarah Canale, BS, for data collection and editorial work.


1.Branthwaite MA. Cardiorespiratory consequences of unfused idiopathic scoliosis. Br J Dis Chest 1986;80:360–9.
2.Campbell RM, Smith MD, Mayes TC, et al. The characteristics of thoracic insufficiency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am 2003;85:399–408.
3.Davies G, Reid L. Effect of scoliosis on growth of alveoli and pulmonary arteries and on right ventricle. Arch Dis Child 1971;46:623–32.
4.Dunnill M. Postnatal growth of the lung. Thorax 1962;17:329–33.
5.Emery JL, Mithal A. The number of alveoli in the terminal respiratory unit of man during late intrauterine fife and childhood. Arch Dis Child 1960;35:544.
6.Muirhead A, Conner AN. The assessment of lung function in children with scoliosis. J Bone Joint Surg Br 1985;67:699–702.
7.Dimeglio A, Bonnel F. Le Rachis en Croissance. France: Springer-Verlag, 1990:392–4.
8.Dimeglio A. Growth of the spine before age 5 years. J Pediatr Orthop 1993;1:102–7.
9.Akbarnia BA, Marks DS. Instrumentation with limited arthrodesis for the treatment of progressive early-onset scoliosis. Spine: State of the Art Review 2000;14:181–9.
10.Mehta M, Morel G. In: Zorab P, Siezler D, eds. The Non-Operative Treatment of Infantile Idiopathic Scoliosis. London: Academic Press, 1979.
11.Tello C. Harrington instrumentation without arthrodesis and consecutive distraction program for young children with severe spinal deformities: experience and technical details. Orthop Clin North Am 1994;25:333–51.
12.Klemme WR, Denis F, Winter RB, et al. Spinal instrumentation without fusion for progressive scoliosis in young children. J Pediatr Orthop 1997;17:734–42.
13.Masso P, Meeropol E, Lennon E. Juvenile-onset scoliosis followed up to adulthood: orthopaedic and functional outcomes. J Pediatr Orthop 2002;22:279–84.
14.Tello C, Bersusky E, Francheri A, et al. Severe infantile scoliosis treated with repetitive distractions followed by definitive arthrodesis. Scoliosis Research Society Annual Meeting, Seattle, 2002.
15.Harrington PR. Treatment of scoliosis: correction and internal fixation by spine instrumentation. J Bone Joint Surg (Am) 1962;44:591–610.
16.Moe JH, Kharrat K, Winter RB, et al. Harrington instrumentation without fusion plus external orthotic support for the treatment of difficult curvature problems in young children. Clin Orthop 1984;185:35–45.
17.Yazici M, Asher M, Hardacker J. The safety and efficacy of Isola-Galveston instrumentation and arthrodesis in the treatment of neuromuscular spinal deformities. J Bone Joint Surg Am 2000;82:524–43.
18.Bagheri R, Oka R, Mahar A, et al. Biomechanical comparisons of different anchors (foundations) used in the growing dual rod technique. IMAST 11th Annual Meeting, Southhampton, Bermuda, 2004.
19.Akbarnia BA, Marks DS, Boachie-Adjei O, et al. Dual rod posterior instrumentation without fusion for the treatment of progressive early onset scoliosis: a multicenter study. Scoliosis Research Society 38th Annual Meeting, 2003. Quebec City, Quebec, Canada, 2003.
20.Campbell RM, Smith MD, Mayes TC, et al. The effect of opening wedge thoracostomy on thoracic insufficiency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am 2004;86:1659–74.
21.Dubousset J, Herring JA, Shufflebarger H. The crankshaft phenomenon. J Pediatr Orthop 1989;9:541–50.
22.Roberto R, Lonstein J, Winter RB, et al. Curve progression in Risser stage 0 or 1 patients after posterior spinal fusion for idiopathic scoliosis. J Pediatr Orthop 1997;17:718–25.
23.Goldberg C, Gillic I, Connaughton O, et al. Respiratory function and cosmesis at maturity in infantile-onset scoliosis. Spine 2003;28:2397–406.
24.Marks DS, Iqbal MJ, Thompson AG, et al. Convex spinal epiphysiodesis in the management of progressive infantile idiopathic scoliosis. Spine 1996;21:1884–8.
25.Marks DS, Sayampanathan S, Thompson AG, et al. Long-term results of convex epiphysiodesis for congenital scoliosis. Eur Spine J 1995;4:296–301.
26.Thompson AG, Marks DS, Sayampanathan S, et al. Long-term results of combined anterior and posterior convex epiphysiodesis for congenital scoliosis due to hemivertebrae. Spine 1995;20:1380–5.
27.Luque E, Cardosa A. Segmental spinal instrumentation in growing children. Orthop Trans 1977;1:37.
28.Pratt RK, Webb JK, Burwell RG, et al. Luque trolley and convex epiphysiodesis in the management of infantile and juvenile idiopathic scoliosis. Spine 1999;24:1538–47.
29.Betz RR, Kim J, D'Andrea LP et al. An innovative technique of vertebral body stapling for the treatment of patients with adolescent idiopathic scoliosis: a feasibility, safety, and utility study. Spine 2003;28:S255–65.
30.Campbell RM, Smith M, Hell-Vocke A. Expansion thoracoplasty: the surgical technique of opening-wedge thoracostomy [surgical technique]. J Bone Joint Surg Am 2004;86(suppl 1):51–64.
31.Moe JH, Cummine J, Winter RB, et al. Harrington instrumentation without fusion combined with the Milwaukee brace for difficult scoliosis problems in young children. Scoliosis Research Society, Cambridge, MA, 1978.
32.Marchetti P, Faldini A. ‘End fusions’ in the treatment of some progressive scoliosis in childhood or early adolescence. Orthop Trans 1978;2:271.
33.Mardjetko SM, Hammerberg KW, Lubicky JP, et al. The Luque trolley revisited: review of nine cases requiring revision. Spine 1992;17:582–9.
34.Rinsky LA, Gamble JG, Bleck EE. Segmental instrumentation without fusion in children with progressive scoliosis. J Pediatr Orthop 1985;5:687–90.
35.Sengupta D, Freeman B, Grevitt M, et al. Long term follow-up of Luque trolley growing-rod construct in the surgical treatment of early onset idiopathic scoliosis. Scoliosis Research Society Annual Meeting, Seattle, 2002.
36.Blakemore LC, Scoles PV, Poe-Kochert C, et al. Submuscular Isola rod with or without limited apical fusion in the management of severe spinal deformities in young children: preliminary report. Spine 2001;26:2044–8.
37.Mineiro J, Weinstein S. Subcutaneous rodding for progressive spinal curvatures: early results. J Pediatr Orthop 2002;22:290–5.
38.Acaroglu E, Yazici M, Alanay A, et al. Three-dimensional evolution of scoliotic curve during instrumentation without fusion in young children. J Pediatr Orthop 2002;22:492–6.
39.Asher M. Isola spinal instrumentation system for scoliosis. In: DeWald R, ed. The Textbook of Spinal Surgery. Philadelphia: Lippincott-Raven, 1997:569–609.
40.Akbarnia BA, Marks DS, Perez-Grueso F, et al. Instrumentation with limited fusion for the treatment of progressive early onset scoliosis. Scoliosis Research Society, San Diego, 1999.
41.Akbarnia BA, McCarthy R. Pediatric isola instrumentation without fusion for the treatment of progressive early onset scoliosis. In: McCarthy R, ed. Spinal Instrumentation Techniques. Chicago: Scoliosis Research Society, 1998.
42.Wattenbarger J, Richards B, Herring J. A comparison of single-rod instrumentation with double-rod instrumentation in adolescent idiopathic scoliosis. Spine 2000;25:1680–8.
43.Thompson GH, Akbarnia BA, Kostial P, et al. Comparison of single and dual growing rods techniques followed through definitive spinal fusion. 39th Annual Meeting of Scoliosis Research Society, Buenos Aires, Argentina, 2004. Accepted for publication, Spine, 2005.

scoliosis; early onset; infantile; growing rod; instrumentation without fusion; growing spine

© 2005 Lippincott Williams & Wilkins, Inc.