Waddell (Nonorganic) Signs and Their Association with Interventional Treatment Outcomes for Low Back Pain

BACKGROUND: The rising use of injections to treat low back pain (LBP) has led to efforts to improve selection. Nonorganic (Waddell) signs have been shown to portend treatment failure for surgery and other therapies but have not been studied for minimally invasive interventions. METHODS: We prospectively evaluated the association between Waddell signs and treatment outcome in 3 cohorts: epidural steroid injections (ESI) for leg pain and sacroiliac joint (SIJ) injections and facet interventions for LBP. Categories of Waddell signs included nonanatomic tenderness, pain during sham stimulation, discrepancy in physical examination, overreaction, and regional disturbances divulging from neuroanatomy. The primary outcome was change in patient-reported "average" numerical rating scale for pain intensity (average NRS-PI), modeled as a function of the number of Waddell signs using simple linear regression. Secondary outcomes included a binary indicator of treatment response. We conducted secondary and sensitivity analyses to account for potential confounders. RESULTS: We enrolled 318 patients: 152 in the ESI cohort, 102 in the facet cohort, and 64 in the SIJ cohort, having sufficient data for primary analysis on 308 patients. Among these, 62% (n = 192) had no Waddell signs, 18% (n = 54) had 1 sign, 11% (n = 33) had 2, 5% (n = 16) had 3, 2% (n = 7) had 4, and about 2% (n = 6) had all 5 signs. The mean change in average NRS-PI in each of these 6 groups was -1.6 ± 2.6, -1.1 ± 2.7, -1.5 ± 2.5, -1.6 ± 2.6, -1 ± 1.5, and 0.7 ± 2.1, respectively, and their corresponding treatment failure rates were 54% (102 of 192), 67% (36 of 54), 70% (23 of 33), 75% (12 of 16), 71% (5 of 7), and 83% (5 of 6). In the primary analysis, an increasing number of Waddell signs were not associated with a significant decrease in average NRS-PI (coefficient [Coef] = 0.19; 95% confidence interval [CI], -0.43 to 0.05; P =.12). A higher number of Waddell signs were associated with treatment failure, with a 1.35 increased odds of treatment failure per cumulative number of signs (P =.008). CONCLUSIONS: Whereas this study found no consistent relationship between Waddell signs and decreased mean pain scores, a significant relationship between the number of Waddell signs and treatment failure was observed.