LLLT Systematic reviews with meta-analysis are hard to do for all sorts of reasons:
1) LLLT has had a lot of different names (cold laser, laser biostimulation, photobiomodulation, low intensity laser laser therapy, LEDT, etc) so the literature search is hard (see my post on this from 2008, recently updated)
2) The authors need to have an advanced knowledge of the pathology in question.
3) The authors need to have an advanced knowledge of LLLT too, particularly the matter of irradiation parameters and dose*
4) Then the meta-analysis requires a high degree of competence with medical statistics
* So we should not be surprised when someone who looks at a broad range of interventions for a pathology with limited knowledge of LLLT struggles (and often fails) to stratify the data by effective irradiation parameters (wavelength, power, beam area, irradiance, pulses), dose (time, energy, fluence), treatment location, number of treatments and interval between treatments.
It may be that LLLT is not a suitable intervention for the pathology in question but if LLLT appears to be effective sometimes and not others then you have to stratify the data in the way that others have done to discover that there may dose, dose rate (irradiance/power density) or other irradiation dependent effects.
The WALT guidelines go some way towards helping the reviewers in this regard though there are gaps due to lack of data, and even then there are mistakes in the original papers which makes writing those guidelines hard too.
I will get around to preparing a blog post on my extended version of the WALT guidelines which attempt to fill in the gaps based on extrapolation and first principles.