The Effectiveness of Acupuncture in the Treatment of Frozen Shoulder: A Systematic Review and Meta-Analysis

Table 1

Main findings and characteristics of studies included in the systematic review.

Study	Cases	Intervention	FS duration (mean)	Sessions/weeks	Outcome measurement	Results
[18]	21	(1) EA + PT (n = 11): LI15, TE14, SI10, GB34, ST38; (2) Sham EA + PT (n = 10): sham acupoints 1-2 cm around the actual points	Did not describe. Inclusion criteria less than 3 months	18 sessions 2-3 sessions per week. Follow-up post 18 sessions and 1, 3, and 6 months post 18 sessions.	VAS, ROM, SPADI	A significant decrease in VAS and an increase in SPADI, ROM between the baseline and the follow-up periods, no significant change between the two groups in VAS, SPADI. ROM in the follow-up periods.
[19]	57	(1) MA + PT exercise + ESWT (n = 28): acupoint in the middle of the forehead (between GV24 and Yintang), BL1, LU10, LI4, LI15, TB3, SI3, SI10; (2) PT exercise + ESWT (n = 29)	4.13 ± 1.96 months in the MA group and 4.70 ± 2.26 months in the control group	PT exercise, MA, ESWT, 6 times a week for 4 weeks; Follow-up: after the treatment course, telephone calls 30 days after the treatment course.	VAS, CMS, HAMA, cured, effective, or ineffective	Significant improvement in both groups in all measurements from baseline. MA group had a significant reduction in VAS and in HAMA compared to the control group at both follow-up periods. CMS was significantly increased in the MA group compared to the control in both follow-up periods.
[20]	64	(1) MA-Ashi points (n = 32): LI4, LI20, TE3, TE23, SI3, SI19; (2) Ashi points (n = 32)	4.56 ± 1.25 months in the treatment group and 4.84 ± 2.62 in the control group	Both groups received the interventions once a day, 6 times per week. The total treatment time was 4 weeks. Follow-up at 1 and 2 months.	VAS, CMS, HAMA, cured, effective, or ineffective	Significant improvement of both groups posttreatment and in the follow-up compared to baseline in all measurements. The treatment group has significantly improved in all measurements and in both follow-up times compared to the control.
[21]	60	(1) A double-blind; (2) ACU with press tack needles (n = 30): using reflex areas in the distant extremities; (3) Press tack placebos (n = 30): same method with placebo needles	ACU 16.0 ± 23.6 months; Placebo 15.6 ± 18.8 months	Participants received just one treatment; Effect measured immediately after treatment	CMS pain subscore	Improvement of 3.3 points in ACU vs. 1.6 points in the placebo group ().
Follow-up study [21]	47	(1) MA + conservative therapy (NSAID) (n = 34); (2) Conservative therapy only, (n = 13) (treatment includes NSAIDs, steroid injections, oral opioid, TENS)	16 months (47 of the 60 proceed to the follow-up study)	MA 10 treatments in 10 weeks; Conservative therapy for up to 1 year; 1 year follow-up	CMS pain subscore	After 1 year increase of CMS pain subscore of 5.7 ± 3.8 points in the MA group and 5.1 ± 4.5 points in the conservative therapy group. MA group reaches the final goal significantly faster ().
[22]	40	(1) MA + PT (n = 20) acupoints did not describe; (2) PT (n = 20)	MA + PT group 4.05 ± 2.06 months and control group 4.10 ± 2.17 months	MA: twice a week for 20 treatments (1.5 months); PT: each the other day until the last session of MA for all groups (1.5 months)	SPADI, VAS, ROM; Follow-up: 1.5 months 3 months	In ROM, MA + PT was higher than the control (), after 3 months, the VAS reduced significantly in the MA + PT comparing to the control group (). No difference in SPADI between the 2 groups.
[23]	20	(1) Home exercises (n = 5); (2) Corticosteroid injection with home exercises (n = 8); (3) MA with home exercises (n = 7): LI15, SJ14, LI14, SI11, SI12, SI14, LI4, ST38	7 months	Exercise twice a day for 6 weeks (all groups). MA once a week for 6 weeks Measurements after treatments.	OSS, ROM, VAS scores	At 6 weeks, improvements in pain, function, and ROM in all groups. No significant difference between the groups.
[24]	64	(1) MA + AQF (n = 32); (2) MA (n = 32) acupoints: LI15, TE14, SI9, LI 14, LI11, TE5 in both groups	MA + AQF 3.46 ± 1.57 months; MA 3.42 ± 1.48 months	Treatment once daily for 12 days Effect measured after treatment	Cured, improved, or failed ADL	Statistically significant improvement in both groups (). The MA + AQF group was better than the MA group in all aspects.
[25]	174	(1) MA group (n = 56); (2) EA group (n = 57); (3) MOXA group (n = 61) acupoints: Jianqian, LI15, TE14, SI10, TE5 (EA), LI4 (EA)	2.97 ± 0.33 months	5 treatments in total 10 days After 10 days, the effect was evaluated.	VAS, ROM, cured, improved, or failed	Significant improvement in all 3 groups in ROM and analgesic effect. EA made more analgesic effects and MOXA made more improvement in the shoulder ROM.
[26]	70	(1) EA + exercise (n = 24): LI15, 1 trigger point (EA), ST38 (MA); (2) IFE + exercise (n = 23); (3) Control group (n = 23): no treatment	EA: 6.71 ± 6.5 months; IFE: 6.70 ± 6.05 months; control: 8.26 ± 7.94 months	10 sessions over a 4-week period (2-3 times a week); Effect measured after treatments and at 6 months	VAS, CMS	In EA and IFE groups, the increased CMS and VAS decreased significantly (both ). No significant in the control group and no significant difference between the 2 intervention groups.
[27]	75	(1) MA only (n = 30): TB14, LI15, GB20, LI4, GB34; (2) PT: control (n = 30); (3) MA + PT (n = 15)	Mean 25.8 weeks in all 3 groups	MA: 4 weeks, 8 treatments; PT: 4 weeks, 20 treatments. The effect was measured at 2 and 4 weeks.	ROM, VAS, SF-36 Health Survey (ADL)	All groups improved in quality of life. The pain was reduced more by MA while ROM improved more in PT. MA + PT had the best outcome.
[28]	14	(1) EA: ST38 (n = 14); (2) Sham: 2 sham acupoints + EA(n = 14) (all of the patients received both true and sham ACU)	Did not describe	3 times for treatment; Measurement before and after the interventions	Video-based humeral elevation	Humeral elevation ROM increased in 8.34 degrees in the true ACU group, significantly better than in the sham group.
[29]	272	(1) MA + exercise (n = 210): LI3 and Lingxia; (2) Control (n = 62): EA—LI15, LI14, TB14, SI9, SI11, GB21	Mean 4.6 months (in both groups)	20 treatments in 22 days. The effect was measured after treatment.	Cured, improved, or failed	MA + exercise: 158 cases were cured, 40 cases improved, and 12 cases failed. Control: 26 cases were cured, 21 improved, and 15 failed. MA + exercise compared to control ().
[30]	35	(1) PT exercise (n = 22); (2) MA + PT exercise (n = 13): Zhongping extra point Contralateral deep needling	7.1 ± 3.9 months in the PT group and 5.5 ± 1.6 months in the PT + MA group	6 weeks twice a week. The effect was measured after treatment and after 20 weeks.	CMS	Significantly higher CMS in the PT exercise + MA group compared with the PT exercise group at 20 weeks ().

ACU: acupuncture; ADL: ability in daily life; AQF: accelerating qi-flow along meridians; CMS: Constant-Murley Score; EA: electroacupuncture; IFE: interferential electrotherapy; MA: manual acupuncture; MOXA: moxibustion; OSS: Oxford Shoulder Score (scoring system of 12–60, with 12 being the best outcome); PT: physical therapy; ROM: range of motion; TCM: traditional Chinese medicine; TENS: transcutaneous electrical nerve stimulation; ESWT: extracorporeal shockwave therapy; VAS: visual analogue scale; SPADI: Shoulder Pain and Disability Index; HAMA: Hamilton Anxiety Scale. Chinese name of acupoints: Ju Gu (LI16), Jian Yu (LI15), Bi Nao (LI14), Qu Chi (LI11), He Gu (LI4), San Jian (LI3), Jian Liao (TB14), Wai Guan (TB5), Jian Wai Shu (SI14), Bing Feng (SI12), Tian Zong (SI11), Nao Shu (SI10), Jian Zhen (SI9), Feng Chi (GB20), Jian Jing (GB21), Yang Ling Quan (GB34), Tiao Kou (ST38), Cheng Shan (BL57), Shen Ting (GV24), Jing Ming (BL1), Yu Ji (LU10), Zhong Zhu (TB3), Hou Xi (SI3), Ying Xiang (LI20), Sizhukong (TH23), and Ting Gong (SI19).