We thank Dr. Pijak and Dr. Muramatsu for their comments and for discussing different possibilities to explain our results. In our paper, we mentioned the following possible reasons for the observed efficacy of acupuncture: effects of needling, more intensive provider interaction, and differences in expectation (placebo effect in the strict sense). Like Dr. Pijak, we do not assume that the superiority of acupuncture can be ascribed entirely to a placebo effect, but we assume that all 3 reasons contribute to the observed effect. We agree that more intensive TCA schemes and less invasive sham acupuncture schemes exist as those chosen in our study, and we agree that the definition of appropriate sham acupuncture is an unsolved problem. Our results suggest that our sham scheme is not a placebo scheme but is probably a real, minimal acupuncture. Furthermore, the definition of an optimal acupuncture concept is still in discussion and whether an optimal point selection exists is questionable. We agree with Dr. Muramatsu that deep needling does not seem to be important for specific acupuncture effects. Specific and nonspecific effects of acupuncture as a complex therapeutic intervention are intertwined (1). Hence, as Dr. Pijak states, investigations with other schemes and devices are necessary to clarify whether there are specific effects of traditional acupuncture techniques. Our study could not answer these questions because it was designed as a pragmatic approach to clarify whether acupuncture as commonly used in German health care is justified as a treatment component for knee pain.

Dr. Wettig is right that we have published our study protocol in advance. He correctly cited articles that informed that our study compared TCA and sham acupuncture. We remain convinced that publication of study protocols in advance is good scientific practice. In addition, we should remind readers that, by ethical reasons in randomized trials, study physicians are obliged to inform the patient about all possible therapies that could be allocated. We doubt that many patients studied the techniques described in the protocol publication in detail to become aware about what type of acupuncture they have received. Our careful analysis showed that patient blinding of acupuncture was successful. Even about half of those patients who were convinced to know the type of acupuncture that they received had incorrect guesses. Moreover, the knowledge of receiving sham acupuncture would not favor the effect in this group but would reduce a placebo effect evoked by the expectation. The constructed conjecture of Dr. Wettig about many unreported additional therapies in the sham acupuncture group is extremely unlikely. The additional therapies, including medication, reported to the blinded interviewer were similar in both acupuncture groups.