Five double-blind RCTs were included in the analysis. Quality scores ranged from one to five, the median score was three. All compared laser therapy with placebo. Trials varied in the number of laser treatments given, type of laser and wavelength used and outcome measures. Treatment sessions were two to three per week for three to four weeks for all trials except one, which treated patients for 10 weeks, three times a week. Pain was reduced from baseline in LLLT group but not in placebo group by 14 to 36% in three trials. Standardised mean differences (SMD) of pooled results found a statistically significant improvement in pain with LLLT, SMD -0.53 (95% CI: -0.85 to -0.22). Statistically significant improvements were also found for morning stiffness and tip to palm flexibility. All other outcomes were negative.
Adverse effects
No adverse effects were reported and no evidence of harm was found.


The Effect of Low-Level Laser in Knee Osteoarthritis: A Double-Blind, Randomized, Placebo-Controlled Trial
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2957068/pdf/pho.2008.2297.pdf
PDF has before and after pictures

Discussion

Our measurement results provide evidence that treatment with the active LLLT probe resulted in significant improvement for all evaluated parameters. In the placebo LLLT group, we found non-significant changes in joint flexion and pain. In the active LLLT group, we found significant improvement with regard to joint flexion, pain, and pressure sensitivity in the active group in comparison with the placebo group at the times examined. The positive effects obtained from active LLLT still persisted 2 mo after treatment. The lack of effect on knee circumference was expected and has not been demonstrated with other therapies. In the pla-
cebo LLLT group, three patients gave an account of an explicit reduction in their complaints, which is in line with placebo improvement in studies of other KOA therapies. It is a weakness of the study that we did not use other validated tools for measurement of KOA pain and disability such as the WOMAC questionnaire or the Lequesne index. However, there is a high correlation between pain scores and these tools, and there is little reason to believe that incorporation of these tools would have altered our results.

Over the years more than 100 double-blind, placebo-controlled studies have been published on the effects of LLLT. These articles also showed the favorable anti-inflammatory effect of LLLT.  11–13 Based on the objective, semi-objective, and subjective measurements after laser and placebo treatments in patients with seropositive rheumatoid arthritis, Barabas came to the conclusion that laser treatment exerts a positive influence on the clinical signs and laboratory parameters of this disease. 14 Ohshiro also demonstrated a positive effect on microcirculation and verified changes by thermography in parallel with the reduction of pain. 15 

Low-Level Laser Therapy for Zymosan-Induced Arthritis in Rats: Importance of Illumination Time
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2935792/pdf/nihms232379.pdf
Abstract

Background

It has been proposed for many years that low-level laser (or light) therapy (LLLT) can ameliorate the pain, swelling, and inflammation associated with various forms of arthritis. Light is thought to be absorbed by mitochondrial chromophores leading to an increase in adenosine triphosphate (ATP), reactive oxygen species and/or cyclic AMP production and consequent gene transcription via activation of transcription factors. However, despite many reports about the positive effects of LLLT in arthritis and in medicine in general, its use remains controversial.