Article Text


Laser acupuncture for chronic non-specific low back pain: a controlled clinical trial
  1. Gregory Glazov1,
  2. Peter Schattner2,
  3. Derrick Lopez3,
  4. Kerrie Shandley4
  1. 1
    Department of General Practice, Faculty of Medicine, University of Western Australia, Perth, Australia
  2. 2
    Department of General Practice, Faculty of Medicine, Monash University, Victoria, Australia
  3. 3
    University of Western Australia, Perth, Australia
  4. 4
    Monash University, Victoria, Australia
  1. Correspondence to Dr Gregory Glazov, School of Primary, Aboriginal and Rural Health Care, University of Western Australia, 328 Stirling Hwy Claremont WA 6010; glazog01{at}


Objective: The primary aim was to determine if laser acupuncture (LA) is more effective than sham laser in reducing pain and disability in adults with chronic non-specific low back pain.

Methods: The design was a double blind, two-group parallel randomised controlled trial. The active intervention was an 830 nm (infrared), 10 mW, Ga-Al-As laser diode laser for acupuncture and a sham control. The primary outcome measures were changes in pain (visual analogue scale) and disability (Oswestry Disability Index) at the end of 5–10 treatment sessions. Secondary outcomes were patient global assessment, psychological distress (Depression Anxiety Stress Scale) and subjective wellbeing (Personal Wellbeing Index). Follow up was performed at 6 weeks and 6 months after completion of treatment.

Results: 100 participants were enrolled and treated in a general practice setting. Per protocol analysis of the primary outcome measures using ANOVA suggested that although there was a significant overall improvement in pain and disability after the course of treatments (p<0.01), there was no significant difference between the intervention and control group in both the primary and most secondary outcome measures.

Conclusion: This study did not show a specific effect for LA using infrared laser at 0.2 Joules per point for chronic low back pain. The overall intervention appeared effective because of placebo and other factors. As there was some concern about baseline inequality between the groups further research using tighter inclusion criteria should attempt to replicate the result and examine if a dose response may exist.

Statistics from

In the past, there has been criticism that interventions such as acupuncture used in the treatment of back pain are no more effective than sham treatment or placebo.1 However, there is now evidence emerging from larger more rigorous trials for the efficacy of needle acupuncture in this condition.2 3

Low level laser stimulation of acupuncture points, otherwise known as laser acupuncture (LA), using laser pointer devices as an alternative to needles, has been used in the last 35 years. Although there is lack of experimental evidence from controlled trials, a dose of 0.1–0.5 Joules (J) per point is commonly used for acupuncture point stimulation.4 G Greenbaum, an Australian acupuncture teacher, stated in a postgraduate medical acupuncture course offered by Medicine Monash University in 1998 that even lower doses of 0.01–0.05 J per point are sufficient to produce a therapeutic effect.

The effect of LA on back pain is still unclear but recently a small systematic review examining low level laser therapy for non-specific low back pain concluded that there was “a small effect on pain intensity”.5 This review, however, excluded trials with reference to the use of LA.

A search of the literature in Medline using search terms (“controlled trials”, “acupuncture”, “laser therapy”, “back pain”, “neck pain”), references lists from articles,6 7 and access to an original thesis from Monash University,8 located only six controlled trials since 1981 involving LA in treatment of back or neck pain.8 9 10 11 12 13 Four out of six trials reported a “positive” effect in favour of LA, although the higher proportion of positive studies may have reflected a publication bias. The trials were of variable methodological quality with small sample sizes, single blinding and cross-over designs being obvious problems. There was only one double blind trial,8 which examined the effectiveness of LA against a sham laser control for chronic low back pain (CLBP). This was a small study (41 participants), in which an infrared laser was used (1.1 J were applied per point). Individualised treatment was given for 5 weeks, there was no co-intervention and no follow up. A small significant positive benefit in only one pain outcome of many measured, and a 6% reduction disability (Oswestry Disability Index) was detected in the active laser intervention group. The primary outcome was not pre specified in this study with problems resulting from multiple analyses.

As LA is currently widely used to treat chronic back pain, it was considered important to conduct further research with a larger participant group, longer treatment and follow up. A 10 mW research laser machine was used as this was available to the researcher and could deliver an energy dose per point treated (0.2 J) commonly used in LA practice.


The study design was a double blind, prospective, two-group parallel randomised controlled trial using sham laser in the control group. This study was approved by a Monash University ethics committee in October 2005 and all participants signed a consent form before entry into the trial.


Participants were recruited by way of advertisements placed in local community newspapers (68%), GP waiting room notices (15%), a notice in an arthritis foundation circular (7%) and by word of mouth (9%). Respondents to recruitment notices initially underwent telephone screening by the principal investigator. They were informed that they would be randomised to receive an active or sham laser treatment but would be blinded to exact mode of this. If they satisfied criteria and were agreeable, they were mailed an explanatory statement and consent form, and invited to attend an assessment at the premises of one of three Perth general practices.

Subjects included in the trial had chronic non-specific low back pain (pain was present “most of the time”) with a duration of at least 3 months, with or without pain radiation. Low back pain was “non-specific” if no significant neurological entrapment was evident on available imaging reports and there was no suggestion of this on history/examination (straight leg raising and other neurological testing). There was also no underlying systemic or visceral disease, or recent vertebral fracture present. They were in age range of 19–70 years, non pregnant and their pain before initial assessment was greater or equal to 3.0 cm on the visual analogue scale.

Participants were excluded if they were on strong opioid analgesia (use of codeine/paracetamol compounds or tramadol were allowed), on regular benzodiazepines, had current involvement in workers’ compensation or other insurance claims or had had any form of acupuncture, dry needling or procedures such as facet block injections in the previous 3 months. They were also excluded if any back surgery had been performed in the previous 2 years.

Participants with “widespread body pains” were also excluded, however the full criteria for fibromyalgia14 were not used in this assessment.

The investigator

The principal researcher and therapist/assessor (GG) is a general practitioner who has practised medical acupuncture since 1990, and is a Fellow of the Australian Medical Acupuncture College. He uses both needle and laser acupuncture techniques routinely in his private clinical practice.

Outcome measures

Parameters assessed included:

1. Visual analogue scale (VAS) for pain.15 Participants were asked to mark their average pain level over the previous week on a 10 cm horizontal line.

2. Oswestry Disability Index Version 2.0 (ODI).16

3. Global assessment (GA) of effectiveness of treatment.17

4. The short version of Depression Anxiety Stress Scale (DASS-21).18

5. Subjective wellbeing (Personal Wellbeing Index-Adult (PWI-A)).19

6. Levels of exercise and analgesic use compared to baseline. At the start of each session the assessor directed enquiry about each participant’s exercise level and analgesic use in previous week and obtained a score on this relative to baseline (less (−1), the same (0) or increased (+1)). At 6 week follow up a corresponding score was obtained from (i) questionnaire item on exercise level “since completion of treatment”, and (ii) from a diary of analgesic use “during previous week”. Scores were averaged and compared between groups.

7. Adverse effects and exacerbations of pain during the course of treatment.

The timing for assessing each parameter is shown in table 1.

Table 1

Timing for assessed parameters (primary outcome measures indicated in bold lettering)


Subjects identified by screening attended an initial assessment, which involved a structured interview and brief examination. Previous imaging investigations were reviewed and the nature of the chronic pain problem and general aspects of the trial/acupuncture were discussed. Subjects who satisfied entry criteria and gave informed consent were enrolled into the study. Baseline self-administered questionnaires were completed before the first treatment. All assessment and treatment was provided by the principal investigator (GG) who remained blinded throughout.

The completion of between five and 10 weekly sessions satisfied the protocol. An interval of some weeks between sessions was permitted (eg, for holiday or illness). At the start of each session participants were questioned about symptoms, analgesic use, exercise levels and pain level during the previous week and their responses were recorded. Tender points were located, followed by application of the laser treatment. During the course of the sessions further education was provided and participants were encouraged to gradually increase stretching and general exercise.

This was facilitated using written material and hand-outs based on material from a popular self-help book.20 Participants were requested not to start any new physical treatments or medications during the treatment period or for 6 weeks after. They could continue any of their current analgesics according to their pain. Follow up survey forms were mailed, enclosing stamp addressed return envelopes.

The device was an 830 nm (infrared), 10 mW, Ga-Al-As laser diode (Acupak, Melbourne, Australia). Power density at the probe/skin interface was 0.05 W/cm2. The basis of the treatment protocol reflected the usual clinical practice of the author. Acupuncture point selection was individualised to each patient and was chosen according to the distribution of pain and palpation of local tender points. Local points on the three main acupuncture meridians21 (Bladder, Gallbladder and Governor vessel) and ah shi points were most often used. Regional palpation was first conducted by firm finger pressure and those points which were tender were marked with a ball point pen. These were unilateral or bilateral and usually at site of maximum pain. Distal acupuncture points on limbs or head and neck, often at the site of radiation of pain or according to common rules of acupuncture point selection, were also selected by discretion of the therapist. No auricular or microsystem acupuncture was used. Acupuncture points for psychological distress and other symptoms, such as headache, neck pain and so on, were also treated if present. Points were stimulated consecutively after they had been all marked. The laser probe rested lightly under its own weight perpendicular to the skin with no change in pressure during stimulation. Twenty seconds treatment (0.2 J) was usually applied consecutively to each point selected. On some occasions patients were treated with 10 seconds per point, contra-lateral to their maximum pain or with fewer points if an exacerbation of pain had occurred with a previous session. A verbal description was provided to participants in simple terms on why certain points were used during their treatment.


Apparatus modified for use in double blind research was used.7 The mode (laser or sham) was set by turning a four digit cogwheel dial. One-hundred consecutive settings of the dial were programmed into the circuitry of the device in a random pattern at time of manufacture with only 50 of these settings resulting in production of a laser beam when the machine was operated. Before commencement of the study, an independent person further randomised this sequence of 100 numbers using permuted block randomisation technique and a copy of this allocation sequence was sent to the therapist/assessor.

Upon enrolment, each participant was allocated the next ID/switch code in the sequence provided. This same code was dialled each time the participant presented for treatment, resulting in the same treatment mode each time the patient was treated. The code could not be re-used if a participant dropped out.

The participant and the assessor/therapist were both blind to the mode of laser intervention. The laser probe had a translucent Perspex conical tip and a red (LED) decoy light issued from the tip each time the unit was operated, regardless of mode of laser production. The success of blinding for this machine (which prevented detection of a diode glow) has been described in a different paper.7

Statistical analysis

In planning the study the statistician calculated that a sample size of 60 would show a difference between groups of 1.5 cm in VAS for pain with a baseline standard deviation (SD) of 2.1 which was present in a previous study,8 at power 0.8 and significance level 0.05. We therefore planned to recruit 100 patients to account for drop outs and other variations in this trial.

During subsequent analysis of the trial data, retrospective power calculation based on the observed SD of 2.4 after treatment, indicated that the analysed sample of 90 would be sufficient to detect a difference of 1.4 cm between treatment groups at the same power and significance levels.

This is below the minimal clinically significant change for low back pain of 15 for VAS (0–100) obtained in an international consensus during 2006.22

Statistical analyses were performed using SPSS Version 15.0 for Windows. Categorical data were compared using Fisher’s Exact test, while numerical data were assessed using the Mann-Whitney U test. To assess the effect of time and treatment (ie, control versus intervention) on outcomes, repeated measures ANOVA was used.


Progress in study

Recruitment and treatments were conducted from February 2006 to May 2007. There were a total of at least 220 enquiries from subjects about participation in the study. 100 participants were enrolled into the trial and 107 were considered either unsuitable after initial telephone screening, not contactable after initial enquiry or cancelled/did not attend after initial appointment was made. Another 13 subjects were reviewed at the assessment interview but were excluded for the following reasons—recent acupuncture, nerve root compression on scan causing symptoms, spinal stenosis (n = 3), recent osteoporotic vertebral fracture, VAS <3 cm (n = 2), age >70 years, on oxycodone, on diazepam 15 mg/day and one patient declined to enter trial after assessment interview.

An average of 9.1 treatment sessions were completed per participant. Ninety enrolled participants completed the required 5–10 treatments, but four others did not start any treatment and another six dropped out later, usually due to other commitments or non-compliance. No subjects discontinued due to any adverse effect of treatment (see fig 1 flow chart).

Figure 1 Flowchart for analysis of primary outcome measure: “pain immediately at the completion of treatment”.

Details of acupuncture treatment given

Participants received an average of about eight points per treatment session (for further details see table 2).

Table 2

Meridians and individual acupuncture points frequently used in this study22

Baseline characteristics of whole study group (n = 100)

Most of the participants (97%) were white Caucasian, females (56%), and their average age was 51 years (SD = 12.6). The mean body mass index was 28. Almost a fifth (17%) were smokers and 92% consumed “safe” quantities of alcohol (less than 4 and 2 standard drinks/day in men and women respectively). Nearly half of the sample (49%) were currently employed in the workforce. 47% were receiving a pension (19% aged and 15% disability support for a health related condition).

Most participants (90%) had the pain for more than two years with the range in duration of 6 months to 40 years. The majority (61%) had a moderate disability at the baseline measurement. Baseline psychological distress level was low, with 70%, 79% and 74% assessed as rating “normal” or “mild” for depression, anxiety and stress respectively.

Comparability of baseline factors and treatment protocol in study groups

Characteristics of the control and intervention groups are shown in table 3 for the 90 participants (drop-outs excluded). Despite the randomisation procedures there was some variability of baseline characteristics between groups, this reaching statistical significance for distributions involving analgesic usage, previous use of acupuncture and current physical therapies.

Table 3

Distribution of baseline characteristics and comparison treatment protocol variables between study groups

Despite the intention to exclude some diagnoses from the whole sample it became apparent on record review after the study that four patients with spinal stenosis and seven patients with possible fibromyalgia were included in the population treated. The few participants with spinal stenosis either had mild changes on available imaging with no other clinical features present, or imaging was ordered by their GP after enrolment into the trial at which the condition was demonstrated. It also became apparent that there were some patients with possible fibromyalgia that was not detected at time of initial assessment. Their final distribution between treatment groups is shown.

There was no significant difference between subject groups in implementation of treatment protocol details (see table 3).

Comparison of outcome measures in study groups

Overall there was almost 40% reduction of pain at the end of treatment and a 30% reduction which was maintained at short term (6 weeks) and intermediate term (6 months) follow up. A 20% reduction in disability at the end of treatment was maintained in the short term.

A summary of the means for baseline and post intervention data for VAS (pain), Disability, Depression/Anxiety/Stress and Subjective Wellbeing in both treatment groups are presented in Table 4 below. Baseline values for the primary outcome measures of pain and disability were similar in the two groups.

Table 4

Means and assessments for statistical comparability of outcome measures between study groups

This study demonstrated a significant reduction in pain and disability scores across time, however there was no difference in primary outcome of reduction in pain or disability at completion of treatment between the groups (ANOVA). This relation also held during follow up.

The primary outcome findings involved a “per protocol” analysis of participants who completed the treatment protocol (n = 90). It was considered justifiable to exclude the four dropouts who did not commence any treatment. As the resulting drop-out rate was less than 10% (6%), an intention to treat analysis was not performed. A sensitivity analysis (n = 96) however was done including the six subjects who participated in a certain number of treatment sessions before dropping out. The drop-out cases who were in the intervention group were imputed to obtained “best” pain relief and disability reduction after treatment while the cases in the control group were imputed to show no improvement in pain and disability after treatment. This did not alter the findings previously described with no significant difference between groups (ANOVA).

Per protocol analysis was also used when comparing the secondary outcomes. There was no significant difference in these outcome measures between the intervention and control group, apart from a slight reduction (see below) in analgesic use in the laser group during treatment phase.

For domains of psychological distress (depression, anxiety and stress), when measured at completion of treatment and at 6 weeks follow up, there was no significant differences between the groups in response to treatment. Overall there was a statistically significant improvement in the domains of depression and stress (p<0.01, ANOVA) between baseline and completion of treatment, and 6 weeks follow up. There were no changes demonstrated in the anxiety domain.

Overall and between groups, there was no statistically significant improvement in subjective well-being (PWI-A), when this was measured at the 6 week follow up. There was a larger improvement in the mean score of the individual domain “satisfaction with health”, but this still failed to reach statistical significance.

In the overall global assessment by the participants of the effectiveness of treatment 59%, 48% and 38% of participants reported that treatment had been moderately or extremely effective at completion of treatment, at 6 weeks and 6 months follow up respectively with no significant difference (p>0.05, Fisher’s Exact) between the treatment groups at any of these follow up times.

Overall, the exercise level of subjects increased relative to baseline, both during the course of treatment and at 6 weeks follow up but again there was no statistical difference between the groups. The use of analgesics and anti-inflammatory agents was also assessed, and there was a small reduction in use of these medications in both treatment groups only during treatment. The difference between groups was significant with a greater reduction in the laser group (p<0.05, ANOVA). At 6 weeks follow up there was no change in analgesic use from baseline in both groups, and no significant difference between them.

Assessment of side effects involved the therapist/assessor noting comments of adverse effects experienced during the course of treatment and on follow up surveys. Participants were simply asked “how they were over the previous week” and there were no side effects documented apart from variation in pain levels reported. It is likely that a greater response for minor complaints would have been obtained had a more formal method of recording been used. It was also anticipated that exacerbations of pain could be a possible result of undergoing treatments. This was expected to occur in patients with chronic pain, where attacks of pain can follow minor events. Overall, a minor flare up of pain (defined as worsening of pain after a treatment which settled within a week) occurred in 59%, and a major flare up (worsening of pain after a treatment described as “worst ever” or not settling over a number of sessions) in 6% of the participants, but there was no statistically significant difference between the groups (p>0.05, Fisher’s Exact).

As an incidental finding, one patient in the laser group attended hospital after suffering a small myocardial infarction during the treatment phase but then elected to continue in the trial, another patient in the sham group developed a kidney infection and had to discontinue her treatments early and one patient in each intervention group developed acute gout in their lower limbs during their course of treatments.


This randomised placebo controlled trial attempted to determine if laser acupuncture (Ga-Al-As laser diode, 830 nm (infrared), 10 mW, applied for 20 seconds (0.2 J) at individualised acupuncture points by one experienced therapist for an average of 9.1 weekly sessions) is more effective than sham laser to reduce pain and improve function in patients with chronic non specific low back pain. The results however suggested that both sham and laser have acupuncture-like effects and are equally effective in treatment of this condition.

There could be a number of possible reasons for improvement to the overall intervention regardless of the laser being switched on. Acupuncture is known to be associated with a significant placebo effect. It is important to note that the acupuncture treatment provided in this trial was a complex non-pharmaceutical intervention,23 where the characteristic aspects of acupuncture in addition to palpation and application of the probe, included the diagnostic process and aspects of listening and talking. Suggestion was provided to all participants during the course of treatment including an explanation about the effect of acupuncture and why certain points were used. In this study all participants were informed at the onset that they could be getting either a “real laser” or an “inactive” red light beam which may have reduced the placebo effect.

Chronic pain patients experience a natural history of fluctuating pain levels. Patients were more likely to enrol into the trial when their pain was worse. Both groups in the trial had a comparable proportion of participants who were experiencing an exacerbation at the initial assessment. Also as the majority of participants had experienced their pain for over two years, spontaneous improvement over the study period was unlikely. The phenomenon of regression to the mean would have resulted in improvement in both groups.

A previous trial indicated that sub-CLBP could be managed successfully with an approach that included clinical examination combined with information and encouragement to exercise,24 which was a co-intervention in this study.

Another acupuncture-like effect due to skin stimulation, separate from the effect of laser could have been of importance. Acupressure from palpation of points, marking the skin with a ball-point pen and pressure of probe on skin could have played a role. It has been previously shown,25 that disability and pain in CLBP were significantly reduced after a course of acupressure. Other research related factors such as “hello-goodbye” and Hawthorne effects may have also played a role.17 26

In summary, there are many more factors than the placebo effect which may have contributed to the positive therapeutic response in both groups. From this study it is not possible to determine their relative contribution. It would also be incorrect to state that the LA intervention is only a placebo.

Strengths and limitations of this study

The importance of this study is in focusing on a widely used non invasive approach with little evidence base to treat non specific CLBP. This is the first trial to employ a novel laser device7 which alleviates previous problems in blinding the patient and therapist, and is a real advance in this field of research. The project examined an adequate sample size which in principle would detect a clinically significant effect of LA if it existed.

Although a negative result was obtained, some doubt remains despite a rigorous randomisation and blinding process being undertaken. Certain baseline inequalities arose between the treatment groups which may have confounded the results (see table 3). If factors with a negative effect on pain outcome aggregated by chance in the laser group, this could have produced a false negative conclusion. This will be analysed in detail in a subsequent paper by the author.

The results obtained only apply to infrared lasers stimulating about 0.2 J per point and may not exclude efficacy for LA using a machine of different wavelength or power, or for treatment of other conditions. Treatment was conducted by a single therapist with a regimen that reflected his own practice which could affect the external validity of the results. It was also possible that with the sham laser intervention, palpation of the skin during the course of location of acupuncture points may also have produced a physiological acupressure effect.

Recommendations for further research

The question of efficacy of LA for CLBP has not been resolved by the current trial and further research is recommended. It is important to assess dose effect across an energy range such as 0.05–1.0 J per point which has been used in LA practice. A much higher laser dose (8 J per point) for laser treatment of lumbar spine is recommended by World Association for Laser Therapy (WALT),27 and is used by laser therapists who believe that a local effect on tissues produces a therapeutic effect without regard to acupuncture principles. It is not clear if this approach produces better results and further research still needs to be performed comparing treatment according to these different philosophies.

Tighter inclusion criteria also need to be implemented excluding poor responders to acupuncture intervention, as these make an acupuncture effect difficult to detect and confound results if present unequally between intervention groups. This issue is extremely important in planning acupuncture trials in general and may explain many previous equivocal results in this type of research. The analysis of responders in this study is the subject of another report submitted for publication by the author.

Consideration should also be given to avoiding co-interventions, and reducing other “acupuncture- like” stimulation such as tactile pressure in LA trials in location of points and during skin application of the laser pointer.

Summary points

  • Laser acupuncture is used to treat back pain but it is still not clear whether it is effective.

  • This double-blind RCT compared infra-red laser (0.2 J per point) with sham laser in 90 patients with chronic low back pain.

  • Laser acupuncture had no effect on pain, disability or most other measures.


I would like to thank the participants of the trial for contributing to this research.


View Abstract


  • Funding Australian Medical Acupuncture College purchased the Acupak research laser and also provided some funding for the project and to enable presentation of its findings at conferences.

  • Competing interests None declared.

  • Ethics approval This study was approved by Monash University ethics committee.

  • Provenance and peer review Not commissioned; externally peer reviewed

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.