Electroacupuncture for Treating Chronic Low-Back Pain: Preliminary Research Results

Introduction

Chronic low-back pain (CLBP) is the most-common pain and the second leading reason for missed work in the United States. ¹ Surgery, steroid blocks, and medications often lead to significant side-effects, one of which contributes to the ongoing opioid crisis in the United States. Studies by Haake et al., ² Cherkin et al., ³ and Vickers et al. ⁴ supported using acupuncture for treating CLBP. While this research is encouraging, some concerns still remain regarding the use of acupuncture for treating CLBP. First, most of the previous studies involved manual acupuncture and showed similar clinical effects between intervention and sham-control groups, while animal studies showed that verum electroacupuncture (EA) outperformed sham EA. ⁵ Few studies have examined EA against sham controls directly in human clinical populations. Second, given that not everyone responds to acupuncture, it is important to identify the characteristics that predict clinical responses. The 2 studies described, in which this was carried out, addressed these 2 questions.

Discussion

In 2 articles Kong et al. ⁶ and Edwards et al. ⁷ described the designs and preliminary results of 2 parallel, independent, randomized, controlled clinical trials examining verum and sham EA for treating CLBP. With ∼200 patients recruited from the San Francisco Bay area, Kong and colleagues examined the clinical outcomes in both pain and function, and, more importantly, central pain-regulatory mechanisms of EA approximated by quantitative sensory testing (QST). ⁶ These researchers hypothesized that EA leads to pain relief via changes in central pain-regulatory mechanisms as measured by QST.

Other researchers assessed if multidimensional phenotyping at baseline (QST, expectations, physical examination, psychosocial factors, and Traditional Chinese Medicine (TCM) diagnosis) could be used to predict responses to EA. ⁷ These authors hypothesized that verum EA would be more effective than sham EA.

Two parallel trials were conducted. A neural-mechanisms investigation (N = 99) performed QST at baseline, midtrial, and postintervention. ⁸ The QST parameters included thermal temporal summation (ascending pain facilitation), conditioned pain modulation (descending pain inhibition), and pressure–pain threshold and tolerance. The second study (N = 102) developed and validated a prediction algorithm using reduction techniques from machine learning on a large number of baseline characteristics. ⁹

Inclusion criteria were people ages 21–65 with CLBP lasting >6 months, and no acupuncture within the past 3 years. Both studies^8,9 used the Roland Morris Disability Questionnaire (RMDQ) as the primary functional outcome measure.^10,11 The neural mechanism study ⁸ used a numeric rating scale (0–100) for pain intensity, and the other study ⁹ used the PROMIS^® [Patient-Reported Outcomes Measurement Information System] pain-intensity instrument. Verum- and sham-EA interventions were conducted twice per week for 6–8 weeks. The researchers found statistically and clinically greater reduction of RMDQ in the verum, compared to the sham arm, in the prediction study. ⁹ However, this difference was not statistically significant in the mechanism study. ⁸ The contrasting results between the 2 studies could have been due to differences in the enrollment criteria, the lapses between enrollment and first treatment in the studies, treatment durations, or provider oversights.

Least Absolute Shrinkage and Selection Operator (LASSO) is a statistical method used to improve the accuracy and interpretability of prediction models by shrinking coefficients and variable selections. ¹² Kong and colleagues used this method to predict patient responses to verum and sham EA. ⁸ Even with differences between the studies, preliminary LASSO regression identified a predictive algorithm from the neural mechanisms study ⁸ that was validated in the second study ⁹ and predicted 40% of the variance in the RMDQ from the responder group. Ultimately, this study will help refine an individually adaptable, effective treatment protocol for CLBP, with the ability to identify responders to acupuncture and perform cost-effective triaging. Additional studies will examine the effects of gender and TCM diagnosis, among other factors. Future studies with larger sample sizes will also be planned to validate the above prediction algorithm and improve the accuracy of the models further.

Kong et al conducted 2 clinical trials to: (1) investigate the clinical effectiveness of EA versus sham EA for treating CLBP; (2) develop and validate a prediction model for the differential clinical response of verum EA over sham EA. One of the 2 studies demonstrated a statistically significant clinical effect of verum EA over sham EA in functional improvement. The prediction algorithm developed from the first study was validated in the second study and predicted 40% of the variance in functional outcome.