Clinical predictors of screening lumbar zygapophyseal joint blocks: development of clinical prediction rules

BackgroundOnly controlled intra-articular zygapophyseal joint (ZJ) injections or medial branch blocks can diagnose ZJ-mediated low back pain. The low prevalence of ZJ pain implies that identification of clinical predictors of a positive response to a screening block is needed.PurposeTo estimate the predictive power of clinical findings in relation to pain reduction after screening ZJ blocks.Study designAs part of a wider prospective blinded study investigating diagnostic accuracy of clinical variables, a secondary analysis was carried out to seek evidence of variables potentially valuable as predictors of screening ZJ block outcomes.Patient sampleChronic low back pain patients received screening ZJ blocks (n=151) with 120 patients included in the analysis after exclusions.Outcome measuresPain intensity was measured using a 100-mm visual analog scale, and responses were categorized according to 75% through 95% or more pain reduction in 5% increments.MethodsPatients completed pain drawings, questionnaires, and a clinical examination before screening lumbar ZJ blocks. History, demographic and clinical variables were evaluated in cross tabulation and regression analyses with diagnostic accuracy values calculated for variables and variable clusters in relation to different pain reduction standards.ResultsAt the 75% pain reduction standard, 24.5% responded to screening ZJ blocks and 10.8% responded at the 95% standard. The centralization phenomenon is not associated with pain reduction using any standard. No variables were useful predictors of post–ZJ block pain reduction of less than 90%. Seven clinical findings were associated with 95% pain reduction after blocks. Five useful clinical prediction rules (CPRs) were found for ruling out a 95% pain reduction (100% sensitivity), and one CPR had a likelihood ratio of 9.7, producing a fivefold improvement in posttest probability.ConclusionsA negative extension rotation test, the centralization phenomenon, and four CPRs effectively rule out pain ablation after screening ZJ block. One CPR generates a fivefold improvement in posttest probability of a negative or positive response to ZJ block.