Discussion of Certain Limitations in the Design of a Randomized Controlled Trial of Acupuncture and Strategies for Their Improvement [Response to Letter]

Dear editor

We sincerely thank Dr. Rui Chen et al for their interest in and thoughtful comments on our article.¹ We also appreciate the opportunity to respond to their comments. Below, we provide point-by-point responses to the key issues raised.

Considerations on Deqi Sensation and Electroacupuncture Stimulation

Quantitative control in clinical trials of acupuncture has always been a point of concern. “Deqi”, as a subjective, composite perception experienced by patients, is often difficult to standardise consistently.² In our trial design, to eliminate human factors in the administration of acupuncture and electroacupuncture, acupoints, depth, manipulation, and all parameters of electrical stimulation were standardised. However, the drawback of this standardized approach is that it cannot guarantee a sustained and consistent “deqi” sensation for every participant. This is because acupuncture is inherently a personalized treatment modality, and the perception of needling varies from person to person.³ In clinical practice, an electrical stimulation intensity of 2 mA is tolerable for almost all patients, but some individuals may experience no sensation at all. We did not adjust the stimulation intensity based on individual tolerance levels. The stimulation intensity needed to be kept consistent. However, we will add patient-reported records of the sensation of electrical stimulation and the “deqi” sensation for use in subsequent analyses.

Limitations of the Sham Electroacupuncture Control

Blinding in acupuncture remains an issue without a perfect solution.⁴ There are currently many different forms of sham acupuncture, including no manipulation at all, direct placement of a blunt needle, bracket-assisted placement of a blunt needle, superficial needling, and needling at non-acupoints.⁵ With the exception of the first method, each of the other methods provides some degree of pressure and pain sensation to the participant’s skin, and all can stimulate mechanoreceptors. We believe that the placement of a bracket on the skin itself is more likely to generate mechanical stimulation. Moreover, the irregular shape of the knee joint makes it difficult to stabilise the bracket. Additionally, if the blunt needle does not create a sensation of skin penetration, participants can easily detect that they are not being acupunctured. This will undoubtedly eliminate the placebo effect and the expectancy effect, thereby further influencing the participants’ central nervous signals.

The focus of our planned brain imaging analysis is to determine the most specific effects of real acupuncture, as compared with the microstimulation and expectancy effects of a blunt needle. Therefore, the cutaneous microstimulation produced by sham acupuncture will not affect our analysis.

Concerns About Unblinding and Its Psychological Contamination of the “Skin-Brain Axis”

As noted above, blinding in acupuncture clinical trials is a problem that is difficult to solve perfectly. We have done our best to minimise the risk of unblinding in this trial. Nevertheless, 100% blinding is unattainable. We will use Bang’s Blinding Index to ultimately assess the success of blinding.

We will include participants’ Bang’s Blinding Index and Patients’ Expectancy Scale for Acupuncture scores as covariates in the regression model to control for the substantial psychological effects of unintended unblinding.

We would like to express our sincere gratitude again to Dr. Rui Chen et al for their suggestions on our trial, as well as for this valuable opportunity for communication. We believe these suggestions will contribute to the improvement of our research. We will subsequently strive to complete this trial.