Is Deqi an Indicator of Clinical Efficacy of Acupuncture? A Systematic Review

Table 7

Summary of the findings table for the evidence of comparative effects of AWD versus AOD for pain.

Acupuncture with Deqi versus acupuncture without Deqi for pain
Study ID and descriptions	Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Quality of the evidence (grade)
		AOD (assumed risk)	AWD (corresponding risk)
Lund et al. 2006 [66]; Subject: acupuncture for pelvic pain in late pregnancy; Setting: two maternity healthcare departments in Sweden	Effective rate for change in morning pain intensity after treatment assessed with change in VAS scores, grouped into “lower,” “unchanged,” and “higher”	Study population		RR 0.99  (0.69 to 1.41)	47 (1 study)	⊕⊕⊝⊝; low^1,2,3,4,5
		727 per 1000	720 per 1000   (502 to 1000)
	Effective rate for change in evening pain intensity after treatment assessed with change in VAS scores, grouped into “lower,” “unchanged,” and “higher”	Study population		RR 1.06  (0.73 to 1.54)	47 (1 study)	⊕⊕⊝⊝; low^1,2,3,4
		682 per 1000	723 per 1000    (498 to 1000)
Haker and Lundeberg 1990 [67]; Subject: acupuncture for epicondylalgia; Setting: outpatients in Sweden	Effective rate for patient reported recovery after treatment assessed with a scale from “unchanged”/“worse” to “excellent” recovery	Study population		RR 1.35  (1.05 to 1.73)	82 (1 study)	⊕⊕⊝⊝; low^2,4,6,7
		658 per 1000	888 per 1000    (691 to 1000)
Xiong et al. 2011 [68]; Subject: acupuncture for primary dysmenorrhea; Setting: outpatients from Tongji Hospital, Wuhan, China	Effective rate for pain relief after treatment assessed with the efficacy assessment guideline for TCM for primary dysmenorrhea	Study population		RR 2.24  (1.51 to 3.32)	90 (1 study)	⊕⊕⊕⊝; moderate^2,4,7,8
		378 per 1000	847 per 1000    (571 to 1000)
	Pain intensity after treatment assessed with VAS, scale from 0 to 10	The mean pain intensity score in the control group was 4.48	The mean pain intensity score in the intervention group was 2.78 lower (3.61 to 1.95 lower)		90 (1 study)	⊕⊕⊕⊝; moderate^2,4,7,8
Chen et al. 2011 [22]; Subject: acupuncture plus intradermal needle placement for cervical spondylosis; Setting: volunteers recruited in two Guangdong hospitals in China	Neck pain after treatment assessed with the Northwick Park questionnaire, scale from 0 to 100	The mean neck pain score in the control group was 23.56	The mean neck pain score in the intervention group was 17.86 lower (23.65 to 12.07 lower)		70 (1 study)	⊕⊕⊝⊝; low^2,4,7,9
	Pain after treatment   assessed with the McGill pain questionnaire, scale from 0 to 60	The mean pain score in the control group was 20.35	The mean pain score in the intervention group was 7.80 lower (10.3 to 5.3 lower)		70 (1 study)	⊕⊕⊝⊝; low^2,4,7,9
Zheng 2012 [70]; Subject: acupuncture for migraine Setting: outpatients from two acupuncture clinics in Beijing, China; Note: this study is ongoing by the time of publication, so incomplete data was reported.	Total migraine hours per 4 weeks  after treatment	The mean total migraine hours per 4 weeks in the control group were 21.95 hours	The mean total migraine hours per 4 weeks in the intervention group were 19.33 hours longer   (9.19 to 29.47 hours longer)		19 (1 study)	⊕⊝⊝⊝; very low^2,7,10,11
	Migraine pain intensity (total VAS score per 4 weeks)  after treatment assessed with VAS, scale from 0 to 10	The mean migraine pain intensity score in the control group was 11.70	The mean migraine pain intensity score in the intervention group was 7.01 higher   (2.8 to 11.22 higher)		19 (1 study)	⊕⊝⊝⊝; very low^2,7,10,11

1Randomization method and blinding of the patient not mentioned. Randomized assignment, binding of outcome assessor, and allocation concealment mentioned. For acupuncture trials, blinding of the practitioner is impossible; 3/70 patients were lost to followup; reasons explained; no ITT analysis. No selective outcome reporting.
²This item was omitted here because we assessed one single study.
³For this single study, findings presented evident individual variations in both groups. The calculated CI equals 1/3–1/2 effect size.
⁴A single study is very likely to be biased. However, it was omitted here to avoid all evidence being “very low” in quality and therefore indistinguishable.
⁵Patient-important outcome.
⁶Randomization methods, allocation concealment, and blinding of the patient not mentioned. Randomized assignment and binding of outcome assessor mentioned. For acupuncture trials, blinding of the practitioner is impossible; 4/86 patients were lost to followup; reasons explained; no ITT analysis. No selective outcome reporting.
⁷Subjective assessments. The calculated CI equals 1/9–2/3 effect size.
⁸Random number table, randomized assignment, allocation concealment, and blinding of the patient and assessor described. For acupuncture trials, blinding of the practitioner is impossible. No dropouts. No selective outcome reporting.
⁹Central randomization, randomized assignment, and the use of sealed envelope described. Placebo acupuncture was used, and the patient was blinded. However, blinding of the outcome assessor was not mentioned. For acupuncture trials, blinding of the practitioner is impossible. It is highly suspected that the physicians act as assessors; hence, the risk for measurement bias is high. No dropouts. No selective outcome reporting. Imbalanced baseline was reported.
¹⁰Central and block randomization and allocation concealment described. The outcome assessor was blinded, but both the patient and the acupuncturist were aware of the allocation. For acupuncture trials, blinding of the practitioner is impossible. High dropout rate (22/59); reasons explained. No selective outcome reporting.
¹¹The trial is ongoing by the time of publication. Preliminary results were published (19/48 cases planned), with high risk of biases.