Halo-pelvic traction in the treatment of severe scoliosis: a meta-analysis

To provide better evidence of the efficacy and safety of preoperative halo-pelvic traction on the improvements of deformity and pulmonary functions in patients with severe scoliosis. Electronic database searches were conducted including the Cochrane Library, PubMed, Web of Science and Embase. All studies of halo-pelvic traction for the management of severe spinal deformity were included. We referred to a list of four criteria developed by the Agency for Healthcare Research and Quality (AHRQ) to assess the quality of included studies. The meta-analysis was performed using RevMan 5.4 software. Based on the study selection criteria, a total of eight articles consisting of a total of 210 patients were included. Statistically significant differences were found in coronal Cobb angle (P < 0.001), sagittal Cobb angle (P < 0.001) and height (P < 0.001) between pre- and post-traction. Sensitivity analysis was conducted, and there were substantial changes in heterogeneity with preoperative thoracoplasty subgroup in coronal Cobb angle (P < 0.001). Three trials including 74 subjects reported FVC and FEV1 predicted value between pre- and post-traction. There were statistically significant differences in FVC, FVC%, FEV1 and FEV1% (P < 0.001). The complication rate was 6.6–26.7%, and symptoms disappeared after reasonable traction strategy and intensive care. Preoperative halo-pelvic traction achieved significant improvements in spinal deformity and pulmonary functions, with minor and curable complications. Thus, it is an effective and safe solution before surgery and may be the optimal choice for severe scoliosis. In light of the heterogeneity and limitations, future researches are needed to better determine the long-term efficacy on comprehensive assessment and to explore the appropriate traction system.


Background
Severe scoliosis is a complex three-dimensional spinal deformity, usually accompanied by severe cardiopulmonary function impairment, and a significant increase in mortality due to the aggravation of the natural course in the late stage. There is a lack of unified diagnostic criteria for severe spinal deformity, and it is usually diagnosed when scoliosis or kyphosis Cobb angle is greater than 90° [1]. The need for safe treatment of rigid severe spinal deformity is especially urgent in China, where the major curves of an extensive number of patients can be more than 120° [2]. However, the optimal treatment for patients with a large major curve remains to be determined due to the critical condition, higher incidence of postoperative pulmonary complications, high correction rate expectation and technical difficulty. Spinal deformity surgery was the preferred treatment option, but direct one-stage orthopedic surgery had been recognized as wildly risky and high demanding, which required complex osteotomies and prolonged high-intensity or even multiple surgeries, resulting in serious complications such as pulmonary complications, massive blood loss and irreversible damage to the spinal cord [3]. Therefore, preoperative halo-traction was recommended to increase the safety and correction rate of second-stage surgery [4,5].
In 1959, halo-pelvic traction was first reported by Perry and Nickel to immobilize unstable cervical segments [6]. In 1971, O'Brien et al. [7] reported 118 cases of scoliosis treated with halo-pelvic traction. Since then, the technique was recognized by spinal surgeons in a real sense. Firststage halo-pelvic traction may improve patients' cardiopulmonary and digestive function, reduce the risks of secondstage osteotomy, and increase correction rate [8,9]. As such, earlier postoperative rehabilitation is permitted. Compared to widely used halo-gravity traction, halo-pelvic traction may be more applicable for severe rigid spinal deformity. Indeed, halo-pelvic traction provides effective, continuous, controllable corrective strength for patients with severe spinal deformities. Also, it is reliable and has fewer complications which can be significantly relieved by considerate care, reasonable traction strength and duration. Patients treated with halo-pelvic traction avert long-term bed rests; therefore, the incidence of bed-related complications is significantly reduced, such as bed sores and respiratory infections [2,10].
There were currently clinical studies with modified halopelvic traction devices. However, definite conclusions of halo-pelvic traction on severe scoliosis were still lacking. Thus, we conducted this study to provide better evidence of the efficacy and safety of preoperative halo-pelvic traction on the improvements of deformity and pulmonary functions in patients with severe scoliosis.

Search strategy
Electronic database searches were conducted including the Cochrane Library, PubMed, Web of Science and Embase. Our search strategies were the combinations of "Scoliosis" (MeSH Terms) and the relevant keyword "Halo-pelvic traction." The language was restricted to English or Chinese, with no limitation on subheadings.

Study selection
All studies of halo-pelvic traction for the management of severe spinal deformity were included in the present study. Non-clinical studies, full-text article could not be obtained, and cross-sectional studies, case reports, comments and reviews were excluded.

Data extraction
The data processing was managed by two authors with Endnote X8 software independently, and disagreements were resolved by the third author. Information for each eligible study included ① descriptive statistics such as author information, publication year, study design, country, data sources and sample sizes; ② intervention characteristics such as detailed halo-pelvic traction strategy; and ③ the outcome of interest: coronal Cobb angle, sagittal Cobb angle, SVA, FVC predicted value, FEV1 predicted value, height and adverse events.

Quality assessment and risk of bias
We referred to a list of four criteria developed by the Agency for Healthcare Research and Quality (AHRQ) to assess the quality of included studies [11,12]. The four criteria were: ① defining the source of information; ② listing inclusion and exclusion criteria; ③ indicating the time period used for identifying patients; and ④ explaining any patient exclusions from the analysis. Each criterion was described as "Yes" or "No" or "Unknown" according to each study.

Statistical analysis
The meta-analysis was performed using RevMan 5.4 software. The mean difference (MD) of 95% CIs was used regarding continuous outcomes of halo-pelvic traction. Statistical heterogeneity was evaluated utilizing the I 2 test. I 2 value of less than 25% indicated low heterogeneity and less than 50% indicated moderate heterogeneity. Then, a fixed effects model was adopted. Otherwise, an I 2 value greater than 50% was regarded as significant heterogeneity, and a random effects model was adopted. If there was significant heterogeneity, a subgroup analysis was performed by sequentially removing included study. P < 0.05 were considered statistically significant.

Study characteristics
Based on the study selection criteria, a total of eight articles consisting of a total of 210 patients were included [2,10,[13][14][15][16][17][18]. The study selection process is shown in Fig. 1. Three included studies had a controlled group, while the other did not. The traction strategy was different regarding elongation rate and traction period. Treatment strategies also varied including traction-fusion and traction-surgical release-fusion. The characteristics of the included trials are summarized in Table 1.

Quality assessment
As shown in Table 2, all the studies defined the source of information and indicated time period used for identifying patients; all but two studies did not list clear inclusion and exclusion criteria of patients; only one study explained patient exclusions from analysis.

Radiographic measurement
Eight trials including 210 subjects reported coronal Cobb angle between pre-and post-traction. As shown in Fig. 2, statistically significant difference was found [MD = 57.39 (95%CI 45.57 -69.20), P < 0.001, I 2 = 98%], and a random effects model was utilized due to severe heterogeneity. A sensitivity analysis was conducted to evaluate the effect of studies on the overall outcome by sequentially removing studies. Of note, there was substantial change in heterogeneity when reducing to a subgroup of two studies with preoperative thoracoplasty [MD = 68.47 (95%CI 65.47-71.48), P < 0.001, Eight trials including 210 subjects reported sagittal Cobb angle between pre-and post-traction. As shown in Fig. 3, a statistically significant difference was found [MD = 44.75 (95%CI 35.26-54.23), P < 0.001, I 2 = 97%], and a random effects model was utilized due to severe heterogeneity. Similarly, a sensitivity analysis was conducted, and no substantial change was shown in heterogeneity.
Three trials including 74 subjects reported FEV1 or FEV1% between pre-and post-traction. As shown in Fig. 6, statistically significant differences were found in

Publication bias
Sagittal Cobb angle was the primary outcome of included trials. Therefore, the outcome index was used to make a funnel plot to detect publication bias, as shown in Fig. 7. Visual inspection of the funnel plots showed symmetry, suggesting that there was no publication bias.

Discussion
Since halo-pelvic traction was applied for spinal deformities in 1971, it became the optimal treatment due to desired correction rate [7,19]. However, the utilization of halo-pelvic traction has gradually declined with the rise of internal fixation. Clinically, the long period of wear for 7.5 months leads to intense discomfort and various complications in most patients [20]. Generally, internal fixation or osteotomy is feasible for most spinal deformities. Nevertheless, patients with rigid severe spinal deformity suffer from combination of cardiopulmonary and digestive dysfunction [21,22], the relative poor nutritional status [23]. One-stage orthopedics surgical operation may encounter complex or infeasible osteotomy, long operation time, massive intraoperative and postoperative blood loss [24], and even irreversible spinal cord injury [25]. As such, the incidence of various complications in the perioperative period is significantly increased [26]. Indeed, it is difficult to achieve a satisfactory curative effect after one-stage orthopedic surgery, especially when the Cobb angle exceeds 120º [2]. Preoperative reduction in Cobb angles, improvement of pulmonary function, cardiac function and basic nutritional status significantly reduced the risk of surgery [27]. Thus, preoperative halo-traction was recommended to increase the correction rate and safety of second-stage surgery.
In the present study, halo-pelvic traction achieved significant improvements in preoperative correction. The results showed that the pooled mean change of coronal and sagittal Cobb angle was 63.36º and 49.66º, respectively. In addition, there showed an increase in height with the pooled mean change of 13.92 cm. According to spinal biomechanics, the constant longitudinal force induced persistent fatigue of muscles, slight displacement and rupture of tendons, ligaments, blood vessels and spinal cord cells. As such, tissue cells spontaneously repaired to adapt to new structural  states, and the correction was achieved. Clinically, it had substantial advantages by sharing the orthopedic pressure during surgery, simplifying the procedure and reducing the risk of nerve damage. Moreover, the extension of spine length may significantly increase the volume of the thoracic and abdominal cavity which effectively improves the pulmonary and digestive function, nutritional status and better tolerance for aggressive surgery [28]. In addition, it may also provide a basis for surgical efficacy and prognosis. The principle was that the spinal flexibility can be evaluated by slow and continuous traction, so as to predict the correction degree of scoliosis, which played a crucial role in the prevention of spinal cord and nerve injury caused by excessive correction. Then, we performed sensitivity analyses due to severe heterogeneity. A combination of halo-pelvic traction and thoracoplasty before osteotomy showed better results regarding the coronal Cobb angle with the pooled mean change of 68.47º. However, the difference was not significant in outcomes with or without thoracoplasty. According to the report [2], the average 2-year correction rate was 65% and 64% respectively. Of note, Koller et al. [29] indicated that the change was more observable in flexible curves. We speculated that differences in spine flexibility may be a major reason for severe heterogeneity.
In general, creating a better chest wall or chest volume was a commonly used way to improve lung function in patients with severe spinal deformities [30]. Halo-pelvic traction improved the pulmonary functions regarding FVC and FEV1 in the present study, mainly as it successfully corrected the curvature, stretched the spine, improved the volume of the thoracic cavity and function of the diaphragm [31]. LaMont et al. [32] demonstrated the correlation between pulmonary function tests and thoracic height after halo-gravity traction. Huh et al. [33] reported that the Cobb angle was negatively correlated with FVC. However, one meta-analysis showed that the Cobb angle improved more significantly than lung function in patients, which may be due to the inherent defect of noncontinuous traction by halo-gravity traction [12]. Patients with severe scoliosis often encounter malnutrition due to physical inactivity and metabolic disorders [4]. The nutritional status was not reported in the current study. However, it was noted that growing rod constructs was beneficial to improvement in percentage weight gain, especially for idiopathic and congenital scoliosis [34]. There may be several reasons. One mechanism may be reduced energy expenditure through improved respiratory function [35], in addition to the intensive intervention of the clinical dietitian during hospitalization [32]. Besides, it may also correlate to the correction of gastrointestinal malalignment [36]. Overall, future studies are warranted to fully confirm these conclusions.
Compared to widely used halo-gravity traction, halo-pelvic traction may be more applicable for severe rigid spinal deformity. Theoretically, halo-pelvic traction provides further effective, continuous, controllable corrective strength. It was proved to be an ideal treatment for tuberculous kyphosis represented by rigid severe spinal deformities [22]. On the contrary, the strength of halo-gravity traction may reach a plateau within 2 weeks and subject to small traction weight of up to one-half body weight. In a comparative study, patients treated with halo-pelvic traction exhibited a better correction Fig. 7 Funnel plot to detect publication bias for the studies rate and pulmonary function, with less high-grade osteotomies [10]. Halo-gravity traction was widely used in clinical practice largely due to its convenience and the lower incidence of traction-related complications [28]. However, halopelvic traction was reliable with relatively minor complications which can be significantly relieved by considerate care, reasonable traction strength and duration. In this study, the complication rate was 6.6%-26.7% including pin infections, cervical stiffness and neurological symptoms. More importantly, no permanent neurological deficits or death occurred and symptoms disappeared after prompt traction adjustment and intensive care. In recent years, clinical studies showed the testing and improvement of various details of the traction frame in halo-pelvic traction devices [2,13,17,37]. Indeed, the effect of traction was better when the resultant force line is located on the concave side of the spinal deformity. Therefore, the modified halo-pelvic traction with the rod located anterolateral to the patient is more suitable for severe kyphosis. Nevertheless, three-dimensional printing techniques and biomechanical measurement were necessary to verify the stability and rationality of the innovative design, and to promote the standardized design which would greatly improve the curative effect and popularize halo-pelvic traction treatment.

Limitation
To the best of our knowledge, this was the first meta-analysis to evaluate the effectiveness and safety of halo-pelvic traction on severe scoliosis. However, the study had some limitations. First, the number of studies included was limited, which may influence the overall research conclusions. Second, different treatments, elongation rates and duration of traction may increase heterogeneity. Also, the included studies lacked control groups to draw sufficient conclusions. Third, the present study only analyzed pre-and post-traction data. The long-term correction efficacy, respiratory system function, quality of life, nutritional status and complications remain to be observed. Based on the above, more RCTs are needed in the future to better determine efficacy and facilitate standardized treatment regimens.

Conclusion
In conclusion, preoperative halo-pelvic traction achieved significant improvements in spinal deformity and pulmonary functions, with minor and curable complications. Our systematic review and meta-analysis provided moderate-quality evidence that halo-pelvic traction was an effective and safe solution before surgery, and it may be the optimal choice for severe scoliosis. Future researches are needed to determine the long-term efficacy on comprehensive assessment and explore the appropriate traction strategy.