Background: Facet joint radiofrequency (RF) ablation is characterized by a high failure rate, which is partly due to the fact that pain relief after diagnostic blocks is inherently subjective. An area that has yet to be explored is whether more objective measures, such as changes in vital signs after blocks, might be used to predict treatment outcomes.
Methods: A multicenter, prospective study was performed in 223 patients who underwent diagnostic lumbar medial branch blocks, of whom 87 proceeded to RF denervation. Blood pressure (BP), heart rate (HR), and pain scores were recorded preblock and 20 minutes postblock. A positive vital sign response was designated as a decrease of less than 7.5 units in BP or HR, and a positive facet block as pain relief of 50% or greater based on 6-hour pain diary scores.
Results: Overall, 125 subjects (56.1%; 95% confidence interval, 49.3%-62.6%) experienced a positive facet block, and 71 had 3-month follow-up information after denervation. Correlations between changes in NRS scores and HR (r = -0.01, P = 0.893), systolic BP (r = 0.05, P = 0.47), diastolic BP (DBP) (r = 0.08, P = 0.22), and mean arterial pressure (r = 0.08, P = 0.21) were weak and nonsignificant. No associations were found between facet block results and any vital sign. Six (85.7%) of 7 patients who experienced a decrease in DBP of greater than 7.5 mm Hg after facet block had a positive RF denervation outcome at 3 months, compared with 43.8% who did not (odds ratio, 7.52; 95% confidence interval, 0.84-363.8; P = 0.049). A classification tree based on significant decrease in DBP, pain duration, and baseline NRS pain score showed a 76.7% (range, 65.8%-86.3%) accuracy rate.
Conclusions: Although a decrease in DBP of more than 7.5 mm Hg had 97.3% specificity and 85.7% positive predictive value for predicting positive RF ablation outcomes, the low negative predictive value (56.3%) precludes its use as a solitary screening tool. An algorithm based on age, baseline NRS pain score, and a significant decrease in DBP was able to predict 76.7% (range, 65.8%-86.3%) of RF denervation outcomes.