Time-dependent surgical outcomes following cauda equina syndrome diagnosis: comments on a meta-analysis

Abstract

Study design: In this Journal Club review, we critically examine the methodology and results of a recently published meta-analysis describing the time dependency of surgical outcomes associated with cauda equina syndrome (CES) diagnosis.

Objectives: Our objectives are to clarify the strengths, weaknesses, and appropriate use of statistics regarding the findings of the previous publication.

Summary of background data: Surgical outcomes from published reports are related to both preoperative conditions and the time between surgery and diagnosis of CES.

Methods: We reexamined many of the papers that were used to create the surgical outcomes data set, including the publications that supplied quantitative information for logistic regression. Of special focus were studies that contributed to the "less than 24 hours from diagnosis" data. The use of statistical parameters such as probabilities and odds ratios to infer possible future outcomes is also addressed to clarify the papers' interpretations. To support the overall limitations of their generalized findings, we conducted a power analysis, which highlights the lack of statistical strength in the resulting conclusions.

Results: The results of the initial work state that an advantage exists in treating patients within 48 hours in comparison to treating them during a period greater than 48 hours after the onset of CES symptoms. However, the authors also state that no difference exists between the effects of an early surgery (less than 24 hours) compared with a subsequent period between 24 and 48 hours after CES onset. Key elements of logistic regression and meta-analysis are used to refute specific aspects of their methodology as well as their clinical interpretation regarding acute intervention.

Conclusions: We conclude that a flawed methodology and misinterpretation of results are reported, understating the value of early surgical intervention.