Early onset scoliosis (EOS) is defined as curvature of the spine in children seen before 10 years of age which might be depending on various etiologies [14]. Definitive surgical treatments with a large fusion performed on young children with severe spinal deformity may result in severe pulmonary insufficiency, short body stature and growth retardation [21]. In order to prevent these problems, physicians are in search of a treatment modality that allows the thorax to grow and at the same time correcting the deformity [21]. Many techniques have been proposed to prevent early spinal fusion [4,6,10,13]. The goal of these methods is to keep spinal deformity under control and to provide the optimum space for the lungs, until the child reaches skeletal maturity [23].

Traditional growing rod (TGR) technique has been shown to be effective in achieving body lengthening and maturation while spinal growth continues in early-onset scoliosis [4,23]. However, there are concerns that TGR use may cause recurrent surgical intervention and associated complications, as well as leading to worsened sagittal plane deformity of posterior-based distraction systems and increasing the risk of junction problems [3]. As a result of these problems encountered with the conventional methods, a less invasive method which do not require repetitive surgical interventions and has a lower risk of autofusion development, magnetically controlled growing rod (MCGR) systems have been developed [2,24].

Little attention has been paid to the sagittal plane and spinopelvic parameters, focusing on existing GR studies, coronal plane deformity, spine height and lung capacity [23]. The importance of spinopelvic parameters on the patient’s clinical results is well known in the literature and, as a result, is becoming increasingly important in the spinal deformity research [23]. As the changes in the sagittal plane began to gain importance in spinal deformity surgery, studies on alterations in sagittal plane after TGR and MCGR were performed [7,23]. However, a web-based search using databases PubMed and Google scholar regarding the comparison of the effects of TGR and MCGR on sagittal plane did not yield a positive result. Therefore, we aimed to contribute to the literature data by comparing the effects of TGR and MCGR on sagittal plane in treatment of EOS.

The study was approved in advance by the Institutional Review Board of University of Kyrenia (RY-2019-10) and all patients signed an informed consent form. We evaluated patients who were operated in two centers due to EOS with dual TGR or dual MCGR techniques, retrospectively. Inclusion criteria in our study were determined as having a major curvature of 30 degrees or more with at least 3 previous lengthening procedures, previous experience of an extended fusion surgery or an appointment for surgery and fulfilling the follow-up period of at least two years after the first operation. The etiology of all 27 patients included in our study was idiopathic EOS. Twelve patients from one center were operated with dual MCGR technique, while 15 patients were operated with dual TGR technique in another center.

Age and gender distributions were similar in both groups. The baseline data and preoperative values of the patients are given in Table 1. The preoperative, early postoperative and final control radiological data of the patients who were operated with MCGR technique are given in Table 2 (Figs. 1 and 2), and those operated with TGR technique are given in Table 3 (Figs. 3 and 4). The comparison of the radiological findings of the two groups on the postoperative and final control visits is given in Table 4.

No correlation was observed between TK, PJA, and DJA in the postoperative period and last postoperative visit in both patient groups. Correlation was not detected between postoperative LL and DJA in both patients groups. In patients who underwent MCGR, the number of lengthening was positively correlated with postoperative PJA and negatively correlated with postoperative DJA (r=0.674, p=0.016; r=-0.766, p=0.004, respectively). There was a negative correlation detected between PJA and DJA in the postoperative period in patients who underwent MCGR (r=-0.687, p=0.014).

In patients who underwent TGR, the number of lengthening was detected positively correlated with postoperative and last postoperative PJA (r=0.683, p=0.010; r=0.687, p=0.005, respectively), and negatively correlated with postoperative and last postoperative DJA (r=-0.818, p<0.001; r=-0.542, p=0.037, respectively). In patients who underwent TGR operation, a negative correlation was detected between PJA and DJA in the postoperative period and last postoperative period (r=-0.528, p=0.043; r=0.520, p=0.047).

Superficial wound infection in one patient and screw pull out complication in three patients were encountered in the MCGR group. Superficial wound infection in three patients, deep tissue infection in one patient, 6-times rod breakages in four patients and 7-times screw pull out complications in six patients were occurred in TGR group.

In the TGR application, the requirement of periodic lengthening in the operation room every 6 to 9 months, and the requirement of a total of 6 to 18 scheduled surgical procedures from the beginning of treatment are the disadvantages of this technique. In addition, there is an increased risk of complications such as surgical wound complications, the termination of the growth by stiff spine development (autofusion), deteriorated chest development, and thoracic insufficiency syndrome [3,8]. In MCGR application, it is thought that the increase in the frequency of lengthening may reduce the development of the autofusion by meeting the spine growth more closely [12]. There is an opinion that, with the more frequent lengthening in MCGR, which is possible as outpatient, the rods would more closely match up with spinal growth and possibly stress may be reduced throughout the implants [1]. In our study, in the TGR technique, rod breakage in four patients (27%), screw pullout complication in six patients (40%) were encountered, while screw pullout complication in only three patients (25%) were encountered in patients who underwent surgery with MCGR technique.

Akbarnia et al. [4] and Shah et al. [23] have observed a decrease in TK after the first implantation, however, they have observed an increase in the last follow-up. Similarly, LL has initially decreased, and increased at the latest follow-up. In their study comparing TGR and MCGR, Akbarnia et al. [5] reported a decrease in postoperative value of TK compared to the preoperative value in both groups, and a tendency to increase in the last follow up. In our study, in line with the literature, we have observed a decrease in the early postop period in TK and LL values, however, in the last follow up, there was an increase in the patient group treated with TGR, while the decrease was continuing in the patients treated with MCGR. Since the kyphosis could not be given to the magnetic rod actuator part and this part coincided with the patient’s kyphosis region, we think that hypokyphosis might have been developed in the MCGR group.

In order to prevent proximal junctional kyphosis in patients whom they applied TGR, Akbarnia and Emans [3] suggested that rods should be countered in the shape of kyphosis and interspinal ligaments should be kept as strong as possible, upper level should be limited at T2 level and the upper level should rarely be reached, and mentioned that in the presence of thoracic hyperkyphosis, the excessive correction of the rods may lead to implant failure after surgery. in their study in which they have applied MCGR and evaluated the proximal junctional kyphosis, Inaparthy et al. [15] have emphasized that male, syndromic, hyperkyphotic and younger patients should be closely followed-up in terms of proximal junctional kyphosis (PJK) development. In our study, we found that the increase in PJA was higher in the MCGR group in the last follow up, although it did not reach statistical significance. It is evident that the PJA values increased as the TK values of the patients decreased. Although there was an increase in PJA values in the early postoperative period and last postoperative period in patients who underwent TGR operation, there was no correlation between TK and PJA. Since inadequate kyphosis was given to the actuator part in the MCGR group, patients have less kyphosis angle than the TGR group in postoperative and future postoperative follow-up. We think that this situation may be effective in the increase of PJA in MCGR group in the follow-ups. Besides, we think that further lengthening in the MCGR group compared to the TGR group may be the reason for distractive forces to increase more in the adjacent segments and higher increase rate in PJA. It should be kept in mind that repeated distraction below the existing upper thoracic kyphosis may worsen the upper deformity. Due to the small number of patients, we could not subgroup patients according to preoperative kyphosis values and this is one of the limitations of our study. Another limitation of our study was that the PJK values were higher in the MCGR patient group at the last follow-up compared to the TGR patient group, but we could not find a statistically significant value due to the insufficient number of patients and follow-up time.

Lowe et al. [20] have reported that preoperative thoracolumbar kyphosis was a risk factor for postoperative DJK development. Shah et al. [23] have indicated that DJA showed the same pattern as LL and correlated with the change in LL and that DJA closely reflected the change in LL. In our study, we observed that both of our groups had the same correlative pattern between LL and DJA, however, this change was not statistically significant.

Some studies have shown that coronal and sagittal plane deformity can develop after initial surgery in both single and dual GR techniques [25]. In a study, which included 67 patients and examined the effect of serial lengthening on sagittal balance (SB) in TGR, it was reported that the balance tended to come to neutral in the early postoperative period and it was found to be statistically significant, however, this significance was reported to be lost in future follow-up period [19]. We observe in our study that the SB and coronal balance change tended to come to neutral in the MCGR group but may not provide information about the superiority of methods to each other due to small number of patients in the groups and short follow-up periods.

Pelvic incidence (PI) is defined independently of position and has been shown to play an important role in maintaining sagittal alignment [9,18]. The presence of abnormal PI is a risk factor for sagittal imbalance and failure following adult scoliosis surgery [11,17]. In their recent study, Atici et al. [7] stated that pelvic parameters (PI, pelvic tilt, and sacral slope) remained unchanged in patients with EOS treated with GR. Shah et al. [23] have found a 5-degree increase in PI compared to preoperative value, however, they have reported that this was not statistically significant. In our study, although we observed a statistically significant increase in the sacral slope in our follow-up in the MCGR group, we did not find any significant changes in other spinopelvic parameters and spinopelvic parameters in the TGR patient group. As a result, we have not determined statistically significant difference between the MCGR group and the TGR group in terms of spinal and spinopelvic parameters in the short-term follow-up.

Sankar et al. [22] have shown that there was an inconsistency between the expected and the obtained net T1–S1 length in the case of repeated lengthening. They attributed this to the hardening of the spine due to autofusion. The effects of this phenomenon on the sagittal profile are still unclear. We provided statistically significant lengthening of T1–S1 distance in both groups in the follow-up period. When we compared the surgical technique, we could not find a statistically significant difference between the groups. Both methods are effective and successful in providing increase in length. In both methods, it should be considered that the planned and actual lengthening will not be the same.

Limitations of our study were; limited number of patients, being retrospective in the nature, short-term follow-up periods and performance of the two procedures by two different surgical teams and clinics.

Studies on TGR procedures have shown that wound and implant complication rates are as high as 58% [8,16]. In their multicentric study, Choi et al. [12] have shown that MCGR results in a lower rate of infections compared to TGR (11% in TGR versus 4% in MCGR). Rod breakage and screw pull out are frequently observed complications with TGR use and it was seen that the rod type or corset usage did not prevent these problems [8]. It is thought that this is a problem regarding instrumentation, which cannot be overcome by any type of implant, and is caused by spanning unfused spinal segments [12]. On the other hand, the MCGR is considered not eliminating implant-related common complications such as rod breakage or screw pull out [12]. The absence of lengthening after some lengthening procedures is specific to MCGR, however, does not always indicate a problem with the device [12]. Also in our study, implant-related complications were higher in the TGR group. However, planning and follow-up of the treatment (TGR or MCGR) according to the normal pediatric spinal sagittal parameters may help in preventing the complications.

When evaluating MCGR and TGR techniques in their own, we can obtain significant changes in Cobb angle, and T1–S1 values when we look at preoperative, early postoperative and last follow up values. In the evaluation of sagittal pelvic parameters, we observed the lack of significant difference between the two groups. MCGR can cause hypokyphosis and PJK in a minimum 2-year follow-up period. In order to prevent proximal junctional kyphosis, magnetic rods with different kyphotic degrees might be manufactured by the company and compatible magnetic rods with pelvic parameter and lumbar lordosis can be placed. The implant-related complications were less in the MCGR group. However, larger case groups and longer follow-up periods, may aid in the understanding the superiority of one method to other in terms of complications.