Subanalysis by sex showed no difference in age (males, 16.4 years vs females, 18.7 years; P = .073), slip percentage (10.4% vs 13.4%; P = .168), or presence or absence of slip (25 vs 26; P > .99) (Table 4).
Discussion
Increasing MRI resolution combined with increasing concern about unnecessary radiation exposure has added to the attractiveness of MRI in the diagnosis of spondylolysis. Spondylolysis progresses on a continuum, starting with a stress reaction (early or active defect) and ending with either healing or nonunion of the pars defect (terminal defect) (Figure 4). Although risk factors for progression are not clearly defined, Fujii and colleagues15 showed that the reaction around the defect is the most important factor for osseous union. It would then make sense that the earlier the spondylolytic defect is identified, the higher the likelihood for union, especially with nonoperative treatment such as rest, activity restriction, and bracing.12,17
There is a lack of consensus regarding MRI use in the diagnosis of spondylolysis. Masci and colleagues18 prospectively evaluated 50 defects in 39 patients using a 1.5-Tesla MRI scanner, concluded MRI is inferior to SPECT/CT, and recommended that SPECT remain the first-line advanced imaging modality. Conversely, Campbell and colleagues4 prospectively evaluated 40 defects in 22 patients using a 1.0-Tesla magnet and concluded that MRI can be used as an effective and reliable first-line advanced imaging modality. These are the only 2 prospective studies conducted within the past decade. Both were underpowered and used outdated technology (newer MRI scanners use 3.0-Tesla magnets). In addition, specific imaging characteristics (eg, edema in pars or pedicle on fluid-specific sequences) that suggest a positive finding—versus overt fracture on T1 MRI—have been recently emphasized. Neither Masci and colleagues18 nor Campbell and colleagues4 detailed what constituted a positive MRI finding. Although an adequately powered prospective study will provide a better analysis of the utility of MRI versus SPECT, such a study is costly and time-consuming. It is important to identify patient and lesion characteristics to help optimize the usefulness of MRI. It is also important to identify the subset of patients most likely to experience osseous healing of active defects,16 as this is the same subset of patients most likely to respond to nonoperative treatment.
We conducted the present study to identify any clinical or radiographic characteristics associated with the diagnosis of early or active spondylolysis. Almost equal numbers of active and inactive defects (49, 59) were identified. There were no differences in patient characteristics, including age, body mass index, and symptom duration. However, there was a significant sex difference—a relatively high proportion of males with active spondylolysis. This finding, which had been reported before,16,19,20 is probably the result of several factors, including males’ lower lumbar spine bone mineral density21; their relatively less spinal flexibility, which affects the distribution of torsional loads on the spine22; and their relatively greater participation in sports, especially sports involving high-velocity, torsional loading of the lumbar spine.23 Studies are needed to delineate the extent to which sex influences the development and persistence of active spondylolytic lesions. Alternatively, a subanalysis revealed an age difference, between our female and male cohorts (18.7 vs 16.4 years), that may have contributed to the high proportion of males with active spondylolysis.
Although the groups’ difference in symptom duration was not significant, it was trending toward significance. As discussed, it could be explained that, along the continuum of disease, earlier defects are more active and either achieve fibrous or osseous union or become chronic and “burn out” to inactive lesions, potentially leading to a listhesis.24 The listhesis association was higher in the inactive group than in the active group (P = .006). The difference in numbers of active and inactive defects at each stage (early, progressive, late) confirms this finding, with no inactive lesions in the early and progressive stages and many fewer active lesions in the terminal stage. Overall, presence of a spondylolisthesis on plain radiographs may obviate the need for SPECT or MRI, as it indicates an inactive chronic lesion—unless new symptoms are suspicious for reactivation or development of previously described adjacent-level pars defects.
No other radiographic parameters were found to be significant—consistent with findings of other studies.2,5,16 Pelvic incidence has been shown to predict progression of spondylisthesis, but under our study parameters it appears not to be associated with development of a slip.
This study had several weaknesses. First, it was retrospective, and imaging parameters were inconsistent, as we included patients who underwent imaging at other facilities. Second, the timing of imaging was inconsistent. Ideally, the same sequence protocol would be used, and all imaging studies (MRI, SPECT, CT) would be performed within a specific period after the initial concern for a spondylolysis was raised. Last, not all patients underwent all 3 advanced imaging modalities; having all 3 would have allowed for a retrospective comparison of MRI and SPECT sensitivity in detecting spondylolysis. Such a comparison would have been interesting, though it was not the goal of this study.