Nonoperative treatment of spondylolysis and grade I spondylolisthesis in children and young adults: a meta-analysis of observational studies.

BACKGROUND The incidence of spondylolysis is at least 6% by the end of childhood, and painful lesions are not infrequent. The most common treatments for spondylolysis are nonoperative in nature and include bracing, activity restriction, and therapeutic exercises. These treatments have been used either alone or in concert. The aim of this meta-analysis was to identify and summarize the evidence from the literature on the effectiveness of nonoperative treatment for spondylolysis (including those with up to 25% spondylolisthesis) in children and young adults. METHODS A comprehensive literature search identified articles meeting the following inclusion criteria: (1) the target population was children and young adults with spondylolysis (including those with up to 25% spondylolisthesis); (2) the treatment intervention was nonoperative; (3) minimum follow-up was 1 year in studies using clinical parameters as the primary outcome; and (4) the studies included at least 10 subjects. Outcome data from eligible studies were pooled into 1 of 2 groups: clinical outcome or radiographic evidence of a union of the pars defects. RESULTS Fifteen observational studies measuring the clinical outcome had a weighted and pooled success rate of 83.9% in 665 patients. A subgroup analysis comparing the clinical outcome of patients treated with a brace to patients treated without a brace was not significantly different (P=0.75). Ten studies evaluating radiographic healing of the defects had a pooled success rate of 28.0% (n=847). A subgroup analysis showed that unilateral defects healed at a pooled and weighted rate of 71% (n=92), significantly more than bilateral defects at 18.1% (n=446, P<0.0001). An additional subgroup analysis showed acute defects healed at a rate of 68.1% (n=236), significantly more than progressive lesions (28.3%, n=224, P<0.0001) and terminal lesions (n=217, P<0.0001), of which not one defect healed. CONCLUSIONS A meta-analysis of observational studies suggests that 83.9% of patients treated nonoperatively will have a successful clinical outcome after at least 1 year. Bracing does not seem to influence this outcome. In contrast to the high rate of success with clinical parameters, most defects did not heal with nonoperative treatment suggesting that a successful clinical outcome does not depend on healing of the lesion. Lesions diagnosed at the acute stage were more likely to heal after nonoperative treatment as were unilateral defects when compared with bilateral defects. LEVEL OF EVIDENCE Meta-analysis of level IV studies. Therapeutic level IV.