Which Way Do We Go?: Commentary on an article by So Kato, MD, et al.

    loading  Checking for direct PDF access through Ovid

Excerpt

Kato et al. performed a subset analysis of prospectively collected data on 757 patients with degenerative cervical myelopathy. Those data were derived from 2 related multicenter studies with sites in Asia, Europe, Latin America, and North America, thus representing a worldwide diversity of patients. The focus of the current study was to compare anterior and posterior approaches in terms of their safety and effectiveness. The authors sought to achieve a higher level of evidence than was obtained in previous cohort studies or meta-analyses of cohort studies while avoiding the ethical dilemmas and substantial challenges with regard to patient volume and equipoise inherent in truly randomized trials. They appropriately eliminated from the analysis patients whose pathophysiology clearly indicated a specific approach. (One rational summary of how to decide on an approach is presented in a recent review by Lebl and Bono1.) They then applied a propensity-score-matched analysis and identified 80 pairs consisting of 1 patient treated with the anterior approach and the other treated with the posterior approach. Those who had combined approaches were excluded. Propensity-score-matched analysis, with which few surgeons are familiar, estimates the probability of the use of a certain treatment based on patient characteristics and is calculated with a logistic regression model. Two patients with the same score but different treatments are selected, enabling paired comparison of the 2 treatments performed in clinically similar circumstances. Propensity scoring performed on prospectively collected data can provide a level of evidence approaching, but not equal to, that of a randomized controlled trial and is being used more frequently in medical research.
Prior investigators comparing anterior and posterior approaches did not include the extent of radiographic assessment provided by Kato et al. When calculating their propensity scores, Kato et al. used 8 characteristics assessed with cervical magnetic resonance imaging (MRI) as well as age, sex, body mass index, comorbidity score, smoking status, and preoperative modified Japanese Orthopaedic Association (mJOA) score. They appropriately pointed out the limitations of propensity score analysis and in particular noted that the results cannot be generalized to patients with the highest and lowest propensity scores. This suggests that there is a relatively small group of patients for whom an anterior or posterior approach is clearly appropriate, with a large majority in between for whom either approach can be used.
Kato et al. evaluated impairment (mJOA score), disability (Neck Disability Index [NDI]), and quality of life (Short Form-36 [SF-36] Physical and Mental Component Summary [PCS and MCS] scores) at 2 years but limited their complications reporting to 30 days. This may have limited their ability to definitively address the finding of increased reoperation rates following anterior approaches in the meta-analysis by Luo et al.2, as reoperations may certainly occur beyond 30 days. In the study by Kato et al., radiculopathy was observed only in the posterior group, which is in contrast to the systematic review by Lawrence et al.3, who found the prevalence of that complication to be similar between the anterior and posterior approaches. The low prevalence of this complication makes it difficult to draw definitive conclusions regarding its association with either the anterior or the posterior approach.
Lessons for the surgeon serving patients with degenerative cervical myelopathy are (1) when it is clear that a particular approach is more appropriate (such as substantial kyphosis suggesting anterior or multilevel involvement trending toward posterior), use that approach; and (2) when it is not a clear that a particular approach is preferable, discuss with the patient his/her individual preferences and select the appropriate approach on that basis.
    loading  Loading Related Articles