This single center, retrospective, longitudinal study included 48 consecutive patients who underwent mSFR for the surgical repair of lumbar spondylolysis pseudarthrosis, which is recognized as the terminal stage of lumbar spondylolysis, between October 2014 and December 2022.

We conducted this study in compliance with the principles of the Declaration of Helsinki. The study was approved by the Institutional Review Board of Mito Kyodo General Hospital (approval number: 22–11), and written informed consent was obtained from all participants and their legal guardians.

Patients were eligible if they experienced persistent low back pain with or without neurological symptoms in the lower limbs. Lumbar spondylolysis was diagnosed using radiography, computed tomography (CT), and magnetic resonance imaging (MRI). Pseudarthrosis was diagnosed based on the absence of bone marrow edema on MRI in combination with a bone defect on CT. Patients who experienced persistent pain despite at least 3 months of conservative treatment, including medication and physical therapy, and those unable to return to competitive sports were included in the study. Patients were excluded if they were unable to complete the follow-up, presented with recurrence, or had a history of spinal surgery.

Patient background data at admission included age, sex, presence or absence of spina bifida occulta, intervertebral disc degeneration at the lower intervertebral level in lumbar spondylolysis (Pfirrmann classification [15]), spondylolisthesis severity (Meyerding grade [16]), and duration of conservative treatment. Postoperative clinical outcomes included achievement of bone union or nonunion, operation time, length of hospital stay, and time of radiological bone union evaluation. Imaging findings included pre-, and postoperative pars defect length, and the difference in pars defect length pre- and postoperatively.

In this study, “bone union” was defined as the presence of continuity in at least two of the three planes (sagittal, coronal, and axial) on postoperative CT. Nonunion was defined as the absence of continuity at 24 months. If the bone union was confirmed within 24 months, further CT scans were not performed.

For each par defect, pre- and postoperative lengths of the pars defect, as well as the difference between them (Δpars defect length), were evaluated. The measurements were conducted according to a previous study [14]. CT scans were reconstructed to identify the central position of the pars interarticularis in the axial plane. The length of the pars defect was measured on sagittal images passing through this position (Fig. 1). Postoperative measurements were performed using the same method. All measurements were performed independently by two orthopedic surgeons.

The mSFR technique was performed using the bilateral Wiltse approach with 3–4 cm lateral incisions [11]. Pedicle screws were placed at the lateral edge of the pedicle, and a U-shaped rod was inserted subcutaneously through the interspinous ligament to preserve the supraspinous ligament. Autologous iliac crest bone was then used for additional bone grafting in the space of the decorticated pars cleft (Fig. 2). All surgeries were performed by a single surgeon (M.T.), and there were no significant differences in surgical techniques between the groups. Rehabilitation began on the second postoperative day with standing and walking training using semi-rigid braces. Passive and active lower limb exercises were initiated immediately, and isometric trunk exercises and hamstring stretching were introduced after one month. Lower limb strengthening exercises, including static movements and electrical stimulation, were incorporated later. Spinal flexion and extension were avoided, and braces were worn for three months. Activities such as jogging and axial rotation were gradually introduced thereafter.

Statistical analyses were conducted using IBM SPSS ver. 29.0 (IBM Corp., Armonk, NY, USA). Continuous variables are expressed as the mean±standard deviation. The Mann-Whitney U test was used for continuous variables, and the chi-square test was used for categorical variables. Spearman’s correlation was used to assess the relationship between pre- and postoperative pars defect length. Univariate logistic regression analysis was performed to identify the preoperative factors that were significantly associated with bone union. Receiver operating characteristic (ROC) curve analysis and area under the curve (AUC) values were used to evaluate the predictive value of pars defect length for bone union. Statistical significance was set at p<0.05.

Of the 48 patients, 38 were retained for the final analysis after excluding seven who were lost to follow-up, one with recurrence, and two with previous spinal surgeries (Fig. 3). The mean age of the patients was 17.8±6.4 years, consisting of 26 males and 12 females. Among them, 37 had bilateral defects, and one had a unilateral defect, resulting in a total of 75 defects. The majority of the defects were located at L5, accounting for 73 defects.

The mean preoperative pars defect length for all defects was 3.4±1.9 mm, while the mean postoperative length was 1.5±0.9 mm. Defects that achieved bone union had significantly smaller preoperative lengths compared to those that did not (3.0±1.6 mm vs. 5.6±2.2 mm, p=0.0003). Similarly, the mean postoperative pars defect length was significantly smaller in defects with bone union than in those without (1.4±0.8 mm vs. 2.5±0.9 mm, p=0.0002) (Table 2).

The change in pars defect length (Δ) was also significantly smaller in defects with bone union than in those without (1.6±1.1 mm vs. 3.1±1.5 mm, p=0.003). A strong positive correlation (Spearman’s rho=0.76) was observed between the preoperative and postoperative pars defect lengths (Fig. 4).

Among the preoperative factors, the length of the preoperative pars defect was the only variable significantly associated with bone union status. Logistic regression analysis was conducted using preoperative pars defect length as the explanatory variable and bone union status as the dependent variable (Table 3). The analysis revealed that a longer preoperative pars defect length was significantly associated with nonunion (odds ratio, 1.89; 95% confidence interval, 1.29–2.72; p=0.001). ROC analysis revealed that the preoperative pars defect length was a predictive factor for bone union, with an AUC of 0.86 (Fig. 5). The optimal cut-off value, identified using Youden’s J statistic, was 3.8 mm, yielding a sensitivity of 89% and specificity of 75%.