Witt et al. recruited participants requesting acupuncture for the treatment of low back pain, who were insured by one of the participating social health insurance funds . Eleven thousand, six hundred and thirty participants were randomly allocated into either the acupuncture group that received immediate acupuncture for 3 months or the control group who received delayed acupuncture after 3 months. Patients continued to receive routine care throughout the study. The outcome measures were HFAQ, SF-36 and low back pain rating scale measured at 3 and 6 months via questionnaires.
At 3 months mean HFAQ scores had increased by 12.1 points (15%) in the acupuncture group and by 2.7 points (3.5%) in the control group. The difference was 9.4 points (95% confidence interval) which was statistically significant (p < 0.01) showing acupuncture more effective than routine care. The SF-36 and Low Back pain rating scale were statistically significantly improved at 3 months in the acupuncture group compared to the control (p < 0.01). The non-randomized acupuncture group’s HFAQ increased by 14.6 points, 1.5 points more than the randomized acupuncture group (p < 0.01). After sub-analysis of the data acupuncture was seen to have a greater effect on patients with worse back function (p < 0.01) that were younger (p < 0.01). The changes at 6 months were slightly lower than at 3 months.
Haake et al. compared acupuncture treatment (needle insertion of 5 to 40 mm) with sham acupuncture (needle insertion of 1 to 3 mm) and a control of conventional therapy valign="top" consisting of physical therapy, exercise and drugs . All participants had non-specific low back pain for more than 6 months and were aged over 18 years. The participants received ten, thirty minute sessions usually twice a week and an extra 5 sessions if they showed a 10-50% reduction in pain. One thousand one hundred and sixty two patients were randomized. The outcome measures were Von Korff chronic pain scale HFAQ and SF-12.
There were a variety of interventions given in the conventional therapy group including physiotherapy, massage, electrotherapy, injections and general exercise, of which some are more evidence based than others. There were statistically significant improvements in the acupuncture and sham acupuncture groups in their primary outcome measures (p < 0.01) over conventional therapy but no significant difference at 6 months between acupuncture and sham acupuncture (p = NS). The HFAQ scores and Von Korff pain scores improved in all three groups.
Brinkhaus et al. compared acupuncture treatment with minimal acupuncture (sham) and a control group of patients on a waiting list . Participants were aged between 40 to 75 years with non-specific low back pain for 6 months or more. The acupuncture treatment was semi-standardized with the physicians using local and distal points to needle, with needles of unspecified length and stimulating the needles to achieve De Qi where possible. The minimal acupuncture group had needles inserted 20 to 40 mm, in predefined non acupuncture points, away from the lower back and these were not stimulated. The treatment in both groups consisted of twelve, thirty minute sessions over eight weeks. The outcome measures were Visual Analogue scale (VAS), Pain Disability Index and SF-36. Two hundred and ninety eight participants were randomized in a 2:1:1 ratio in favour of the acupuncture group. All completed questionnaires were sent directly to the study valign="top" centre. The VAS decreased by 28.7 mm (SD +/- 30.3 mm) in the acupuncture group at 8 weeks and by 23.6 mm (SD +/- 31.0 mm) in the minimal acupuncture group. The difference between acupuncture and minimal acupuncture was 5.1 mm (p = NS) and 21.77 mm between the acupuncture group and the waiting list group (p < 0.01). At 8 weeks there were significant differences between the acupuncture group and the waiting list group but not between the acupuncture and minimal acupuncture groups. The differences in outcome measures were reduced at 26 and 52 week follow ups. Comparison was difficult beyond 12 weeks as by then the control group had received acupuncture treatment.
Thomas et al. compared acupuncture treatment with usual care in patients with non specific low back pain of 4 to 52 weeks in duration, aged 18 to 65 years . The acupuncture group received ten individualized sessions over 3 months. The acupuncturists determined the number of sessions and the contents of the sessions. There was no standardization of the acupuncture points used, the number of needles used, the duration of treatment, the depth of needle insertion or if the needles were manually stimulated. The usual care group was given National Health Service (NHS) treatment according to their general practioner’s assessment of their needs.
The primary outcome measure was the bodily pain dimension of the SF-36 at one year. A number of secondary outcome measures measuring pain again were also used, including Oswestry pain Disability Index, the McGill present pain index and the remaining dimensions of the SF-36. Follow up questionnaires were carried out by mail at 3, 12 and 24 months. Two hundred and forty one participants were randomized.
The acupuncture group received an average of 8.1 treatments. Patients in both groups received various other interventions. The results showed an intervention effect of 5.6 points (p = 0.06) in the SF-36 at 12 months and an estimated effect of 8.0 points (p < 0.01) at 24 months in the acupuncture group. No evidence of functional improvement was found and no data at 3 months is reported.
Cherkin et al. randomized 638 adults, with low back pain for more than 3 months, into individualized acupuncture, standardized acupuncture, simulated acupuncture or usual care . Individualized acupuncture was prescribed by the diagnostician for each participant. Standardized acupuncture used set points prescribed by experts. Simulated acupuncture used a toothpick and guide tube. All participants underwent 10 treatment sessions. Outcome measures were assessed at 8, 26 and 52 weeks and included Roland Morris Disability Questionnaire, “Bothersome” Score and SF-36. There was a statistically significant improvement in function (RMDQ) in all groups at 8 weeks (p < 0.01) but was no longer significant at 52 weeks. The real and simulated acupuncture groups did not differ from each other (p = NS). At 8 weeks the proportion of patients improved in RMDQ scores was significantly greater in the real and simulated acupuncture groups compared with usual care (p < 0.01).
Kerr et al. randomly allocated 60 patients, with low back pain for more than 6 months, into 2 groups to receive either acupuncture or placebo TENS. Participants were treated weekly for 6 weeks with outcome measures including McGill Pain Questionnaire . The VAS and SF-36 scores were collected pre and post treatment. Follow up was at 6 months. The acupuncture treatment was standardized using set points and 11 needles only for a duration of 30 minutes. Placebo TENS used 4 electrodes over the lumbar spine for 30 minutes. The machine was switched on but the circuit was broken so no current actually reached the patient. Both groups showed an improvement in their pre and post scores at 6 months. In the acupuncture group there was a significant difference in SF-36 (p < 0.01), MPQ (p < 0.01) and ROM (p < 0.01). There was a statistically significantly improvement in SF-36 for TENS placebo group (p < 0.01) and ROM (p < 0.05) and VAS (p < 0.05). There was no significant difference between the 2 groups for any outcome measure.
Leibing et al. randomly allocated 131 patients aged between 18 and 65 years with non-radiating low back pain of less than 6 months in duration . All patients received 26 sessions of physiotherapy. The control group solely underwent physiotherapy, while the acupuncture group additionally received 20 sessions of traditional standardized acupuncture over 12 weeks. Fixed points were needled to a depth of 10-30 mm and stimulated to achieve De Qi. The sham acupuncture group 20 sessions of minimal acupuncture with needles inserted to a depth of 10-20 mm and not stimulated. Outcome measures were pain intensity, measured using a VAS, pain disability index, psychological distress, using the hospital anxiety and depression scale and lumbar spine flexion.
At 12 weeks acupuncture was superior to the control group in pain intensity (p < 0.01), pain disability (p < 0.01) and psychological distress (p < 0.05). There was no significant difference between acupuncture and sham acupuncture in pain disability or intensity although there was a difference in psychological distress (p < 0.05). At I year acupuncture was still superior to the control in pain disability (p < 0.05) but there were no differences between acupuncture and sham acupuncture. There were no differences in spine flexion throughout.
All 7 studies used outcome scores that assessed pain, disability and function. Witt et al demonstrated a statistically significant improvement in SF-36 (p < 0.01) and HFAQ (p < 0.01) scores at 3 months in the acupuncture group compared to routine care . Cherkin et al similarly found a statistically significant increase in RMDQ scores in both the real and simulated acupuncture group (p < 0.01) over usual care at 8 weeks but there was no difference between real and simulated acupuncture (p = NS) . Kerr et al showed a significant difference in the SF-36, McGill pain score and ROM in the acupuncture group when comparing pre and post scores at 6 months (p < 0.01) . But this study revealed identical results for the placebo TENS patients with no difference in outcome scores between placebo TENS and acupuncture at 6 months. Thomas et al showed no functional improvement in acupuncture over usual care .
Brinkhaus et al revealed a significant reduction in VAS score for pain in the sham and acupuncture groups over patients on the waiting list at 8 weeks (p < 0.01) although there was no significant difference between the sham and acupuncture groups themselves . The study by Haake et al supported these results stating a significant improvement in Von Korff pain scores in both the acupuncture and sham acupuncture over conventional therapy at 6 months (p < 0.01) . As in the previous study there was no significant difference found between the sham and acupuncture groups (p = NS). The study by Leibing et al.  found similar results with a significant improvement in VAS scores at 12 weeks in the acupuncture group over the control (p < 0.01) which was the same at 1 year. Once again, there was no significant difference found between acupuncture and sham acupuncture.