Dokumentasjon
The Efficacy of Laser Therapy in Wound Repair:
A Meta-Analysis of the Literature
LYNDA D. WOODRUFF, P.T., Ph.D.,1 JULIE M. BOUNKEO, M.S.,1 WINDY M. BRANNON, M.S.,1
KENNETH S. DAWES, Jr., M.S.,1 CAMERON D. BARHAM, M.S.,1
DONNA L. WADDELL, Ed.D., R.N., C.S.,2 and CHUKUKA S. ENWEMEKA, Ph.D., FACSM3,4
ABSTRACT
Objective: We determined the overall effects of laser therapy on tissue healing by aggregating the literature
and subjecting studies meeting the inclusion and exclusion criteria to statistical meta-analysis. Background
Data: Low-level laser therapy (LLLT) devices have been in use since the mid sixties, but their therapeutic
value remains doubtful, as the literature seems replete with conflicting findings. Materials and Methods: Pertinent
original research papers were gathered from library sources, online databases and secondary sources.
The papers were screened and coded; those meeting every inclusion and exclusion criterion were subjected to
meta-analysis, using Cohen´s d. statistic to determine the treatment effect size of each study. Results: Twentyfour
studies with 31 effect sizes met the stringent inclusion and exclusion criteria. The overall mean effect of
laser therapy on wound healing was highly significant (d = +2.22). Sub-analyses of the data revealed significant
positive effects on wound healing in animal experiments (d = +1.97) as well as human clinical studies
(d = +0.54). The analysis further revealed significant positive effects on specific indices of healing, for example,
acceleration of inflammation (d = +4.45); augmentation of collagen synthesis (d = +1.80); increased tensile
strength (d = +2.37), reduced healing time (d = +3.24); and diminution of wound size (d = +0.55). The Fail-Safe
number associated with the overall effect of laser therapy was 509; a high number representing the number of
additional studies-in which laser therapy has negative or no effect on wound healing-required to negate the
overall large effect size of +2.22. The corresponding Fail-Safe number for clinical studies was 22. Conclusion:
We conclude that laser therapy is an effective tool for promoting wound repair.
INTRODUCTION
TO MANY CLINICIANS and scientists, the idea that low-power
laser light (so low in intensity that some have compared its
power to dull sunlight) can be therapeutic enough to relieve
pain and promote tissue repair in collagenous tissues seems
preposterous. Yet, reports abound which indicate that these
lasers, that is, lasers with 500 mW average power, promote
the repair processes of skin, ligaments, tendons, bone, and cartilage
in experimental animals,1-28 as well as wounds and ulcers
of a wide range of etiologies in humans.7,29-34 The
availability of other studies35-42 that suggest the contrary, that
is, that low-intensity lasers and other monochromatic light
sources are not effective in promoting tissue repair, further
complicates the matter, creating the present scenario in which
low-intensity lasers are viewed with doubt and cynicism.
There is little disagreement that a majority of animal experiments
suggest that low-intensity lasers enhance wound healing
by promoting cell proliferation,1,7,43-53 accelerating collagen
synthesis and promoting the formation of granulation tissue,
2-6,55-60 fostering the formation of type I and type III procollagen
specific pools of mRNA,58 increasing ATP synthesis
within the mitochondria, activating lymphocytes, and increasing
their ability to bind pathogens.7,61-66 In contrast, clinical reports
concerning the effects of low-intensity lasers remain, at
least prima facae, contradictory, with some studies reporting
beneficial effects on tissue repair and others showing no effect
whatsoever.7,29-42,67-76 Given the multitude of variables involved
in treatments with laser therapy devices, that is, wavelength,
power, power density, energy, energy density, treatment
duration, treatment intervention time post-injury, and method
of application (contact mode versus non-contact mode), a traditional
review of the literature does not leave one with a clear
impression concerning the true effects of laser therapy on tissue
repair. Consequently, the purpose of this study was to aggregate
the literature, and subject every study meeting every
inclusion and exclusion criteria to statistical meta-analysis, in
order to determine objectively the overall effects of laser therapy
on tissue repair processes. A secondary goal was to elucidate
information that might be helpful to clinicians and
researchers in developing effective treatment guidelines.
In this study laser therapy or laser is used operationally to
treat with monochromatic light devices, including low-power
lasers, light-emitting diodes, and superluminous diodes.
MATERIALS AND METHODS
Subjects and design
Original research papers, investigating the effects of laser
therapy on tissue healing, were gathered and used for this
study. The papers were sought and obtained from library
sources and online data bases, including Medline, Index
Medicus, Excerpta Medica, Citation Index for Nursing and
Allied Health Literature (CINAHL), and Psychology Information
(PsycInfo). Search terms used include "laser therapy,"
laser biostimulation," "soft laser," laser photostimulation,"
"biostimulation," "photostimulation," "light therapy," "laser
therapy and wound healing," and "biostimulation and wound
healing." Additional secondary sources of information include
papers cited by authors whose articles were obtained from the
aforementioned sources, Internet Web pages, and pertinent papers
published in journals that were not found from any of the
above data bases.
Inclusion Criteria: We included studies that met the following
criteria:
• The type of laser and precise wavelength were defined.
• Laser or other light source is clearly identified as the independent
variable.
• At least one index of wound healing, that is, collagen content,
healing time-defined as the total amount of time that
treatment was performed to achieve full healing of tissues-
reduction in wound area, tensile strength, acceleration of
inflammation, and prevention of dermal necrosis was identified
as the dependent variable.
• The authors either stated or we were able to determine the
following treatment parameters: power, power density, energy,
energy density, number of treatments given, duration of
each treatment, frequency of treatment, beam and spot size,
dose (expressed in J/cm2), size of the area treated, and mode
of treatment (contact or non-contact mode).
• The condition treated, for example, bed sores, venous ulcers,
diabetic ulcers, or surgical wounds, was clearly stated.
Exclusion criteria: Studies excluded from this analysis
include:
• In vitro studies involving cells and tissues, not whole animals.
• Case reports and single case studies regardless of etiology
• Studies with data from which Cohen´s d statistic,77 that is,
treatment effect, could not be calculated using one of the statistical
formulas detailed below.
• Studies reported in languages that we could not interpret.
Reliability study and data coding:
To determine the presence
or absence of our inclusion or exclusion criteria in each
study reliably, a coding form was developed and used to list the
essential parameters and other pertinent information obtained
from each primary study as detailed in Table 1. Then, a pilot
study was conducted to ascertain the level of agreement among
raters as they calculated the treatment effect sizes, that is,
Cohens´s d, from each study. The analysis was continued when
90% agreement was achieved among raters. Thereafter, studies
included in the meta-analysis were coded by one of the investigators
using the nineteen parameters on the coding form. Thus,
the overall result of each study was transformed into a standardized
effect size statistic using Cohen´s d formula.77
Data analysis
Calculation of Cohen´s d:
Treatment effect sizes were
calculated using the formulae for determining Cohen´s d.77
According to Wolf,77 Cohen´s d may be defined as the standardized
difference between the means of the experimental
group and the comparison group divided by a standard deviation
of the comparison group. This definition and Cohen´s
classification of effect sizes were applied to each study included
in our meta-analysis. According to Cohen, the values
of 0.2, 0.5, and 0.8 indicate a small, medium, and large average
effect size, respectively.
Conceptually Cohen´s statistic may be expressed as follows:
d = |x1-x2| / SDcomparison
where d is the effect size, x1 is the mean of the laser treated
group, x2 is the mean of the comparison group, and SDcomparison
is the standard deviation of the comparison group.
If means or standard deviations were not reported and percentages
were reported in a study, then, the following t formula
was used:
where P2 is the population of the laser group, P1 is the population
of the comparison group, N2 is the number of subjects in
the laser group, and N1 is the number of subjects in the comparison
group.
Once a t-value was calculated, then it was converted to d as
follows:
d=2t / √df
where d = effect size; t = t-value, and df = the degree of freedom.
The effect size (d) was assigned a positive or negative value
depending on the outcome of the study. For example, positive
values were assigned to experiments whose results were positive,
that is, indicated that laser therapy promotes wound healing.
Negative values were assigned to experiments that showed
negative effects of laser therapy, that is, that laser therapy had no
effect or retarded wound healing. After calculating the treatment
effect size of each study independently, the mean overall effect
size was calculated by summing all the effect sizes and dividing
by the total number of effect sizes using the following formula77:
daverage = ∑d / N
where, in this formula, daverage = mean effect size, ∑d = the sum
of the effect sizes, and N = the total number of effect sizes calculated
and used.
Multiple effect sizes were calculated in several studies. In
order to avoid violating the assumption of independence, no
more than two effect sizes were taken from each study.
Calculation of fail-safe number:
Given the likelihood that
we did not obtain every study that ever examined the effects of
lasers on wound healing, a fail-safe number (Nfs) was calculated.
The Nfs reveals the number of additional studies with effect
sizes below a set criterion value that would have to be
included in the meta-analysis in order to change the outcome
of the study. We used 0.10 as the criterion, a number that is remarkably
lower than the small effect size of 0.2 suggested by
Cohen. The following formula was used to calculate the failsafe
number:
Nfs = N(δ - dc) / dc
where Nfs is the fail-safe number (N) value; N = the number of
studies in the meta-analysis; and δ = the average effect size of
all studies, and dc = the criterion value.
Sub-analysis of effect sizes:
The data obtained were grouped
into categories to permit further sub-analysis of the overall treatment
effects of laser therapy on each outcome parameter. The
outcome categories used were: collagen content, healing time
(i.e., the total amount of time that treatment was performed to
achieve full healing of tissue), reduction in wound area, tensile
strength, acceleration of inflammation, and prevention of dermal
necrosis. Furthermore, the data were analyzed to compare the
overall effects of the different types of lasers, as well as the relative
effects of treatment in animal and human experiments.
RESULTS
The overall effect of laser therapy on tissue repair
Our literature search revealed hundreds of studies that examined
the effects of laser therapy on tissue repair, but the final
set of papers from which treatment effect sizes could be calculated
was just 24. These 24 publications had 31 computable effect
sizes. Insufficient data with which to calculate treatment
effect size and/or inadequate reporting of treatment parameters
were the major reasons that so many studies were not included
in the meta-analysis. In addition, there were several studies in
which data were only summarized in the form of graphs and illustrations
from which it was not possible to extrapolate the
data needed to compute Cohen´s d.
The overall mean treatment effect size (Cohen´s d), determined
from the 24 studies was +2.22, indicating that laser therapy
is highly effective in promoting wound healing. The fail
safe number (Nfs) associated with this finding was 509, meaning
that 509 additional studies in which laser therapy had little
or no effect on tissue repair, would be needed to negate the
large treatment effect of this meta-analysis. When animal and
human clinical studies were analyzed separately, our subanalysis
revealed an overall treatment effect size of +1.97 and
+0.54, respectively, indicating that laser therapy strongly promotes
healing in experimental animal models and moderately
so in humans. Given the overwhelming interest in the clinical
effects of laser therapy on wound healing in humans, we subanalyzed
the fail-safe number for human studies, using the same
set criterion value of 0.10. The fail-safe number (Nfs) associated
with this sub-analysis was 22 additional human studies in
which laser therapy had little or no effect on tissue repair. This
number corresponds with the moderate healing effect of +0.54
previously determined from the diverse clinical studies analyzed.
Consistent with Figure 1, the distribution of the calculated effect
sizes was trimodal, with all three peaks occurring above
zero, and with the largest peak between zero and 1.5; again showing
an overall positive effect of laser therapy on wound healing.
Treatment effect sizes ranged from 0.33 to +9.10; mean ± SD =
2.22 ± 2.86. The associated 95% confidence interval was calculated
as 3.41 to +7.81, reflecting the wide variability in the parameters
of treatment used in these studies (Table 2).
In order to verify the accuracy and dependability of the
overall effect size and its implications, we examined the data
points that were used to calculate the effect size and determined
whether extremely large effect sizes skewed the data.
An analysis of the trimodal frequency distribution curve revealed
no obvious differences between the studies in the positively
skewed cluster and the rest of the data. Moreover, the
median was found to be +0.95, signifying that more than half
of the effect sizes were greater than or equal to +1.00.
Furthermore, 15 studies, which examined the effects of laser
therapy on wound healing in experimental animal models and
humans but were excluded from this meta-analysis because the
data needed to calculate effect size narrowly missed our stringent
inclusion and exclusion criteria, i.e., data were either insufficient
or presented in graphs from which effect sizes could
not be derived, were closely examined after the meta analysis.
Of these studies, 12 reported positive outcomes and three reported
negative outcomes. Moreover, four were human studies
and 11 were animal studies. All the human studies showed positive
outcomes favoring the use of therapeutic lasers on
wounds. Thus, it appears that if these studies were included in
the meta-analysis, the overall treatment effect size reported
here would have been higher than +2.22.
Sub-analyses of effect sizes
Further sub-analyses of our data showed a positive influence
of laser therapy on several indices of tissue repair, including
reduction of skin necrosis, acceleration of inflammation,
wound tensile strength, reduction of wound sizes, healing time,
and total collagen content (Table 3). Similarly, as shown in
Table 4, there were notable differences in the treatment effects
of the different lasers used in the literature. In this analysis,
Krypton laser had the highest treatment effect size (+4.29), and
Gallium Arsenide laser, the least (+0.63).
Since treatment parameters varied widely, the calculated effect
sizes did not correlate with any treatment parameter.
Nonetheless, a trend was discernible with respect to energy
density. Energy densities from 19 to 24 J/cm2 were more effective
than energy densities at or below 8.25 J/cm2 and at or
above 130 J/cm2.
DISCUSSION
Overall, our findings show that low power lasers promote
wound healing in both experimental models of tissue repair
and human cases of wounds and ulcers. This finding is consistent
with several reports that indicate that laser therapy accelerates
tissue repair processes. To many clinicians and scientists,
the notion that imperceptible doses of low-energy laser light
can promote healing seems impossible. Yet, there are abundant
reports that indicate the low-power lasers do not just advance
several of the metabolic processes involved in wound healing,
they accelerate healing of recalcitrant and non-healing wounds
and ulcers in experimental animal models and humans.
This state of affair seems understandable given the availability
of reports suggesting that no conclusions can be drawn
concerning the effects of therapeutic lasers on wound healing.
35,36,38,39 One meta-analysis on this subject concluded that
laser therapy studies are remarkably flawed and poorly reported.
70 In supporting this assertion, others have attributed
this situation to inconsistencies in reporting the treatment
parameters and the methods used in several published research
studies.44 These observations are consistent with our experience
and our finding that only 24 research publications met our
inclusion and exclusion criteria. The major flaws in most studies
were inadequate reporting of treatment parameters, inexplicable
reporting of actual doses used, and lack of hardcore
numerical data.
The large standard deviation associated with our overall
finding and the 95% confidence interval suggest a high degree
of variability within the computed effect sizes, and provide further
support for this assertion, that is, that treatment methods,
outcome measures, and subjects differ markedly from one
study to the next,78 a situation which has been declared a major
problem with the literature.70 As shown in Table 4, inconsistent
reporting of treatment parameters remains a problem with the
current literature, as a consensus on effective treatment parameters
remains elusive. Notwithstanding the large standard deviation
and the 95% confidence interval, the large fail-safe
number and the frequency distribution of effect sizes strongly
support the large treatment effect size found in this study.
Moreover, the large treatment effect obtained would have been
larger, had we not eliminated fifteen studies that almost met
our criteria, but lacked the data needed to calculate effect size.
Of these, 75% reported positive outcomes, and all the human
studies had positive results. Overall, the data indicate that laser
therapy promotes wound healing in experimental animal models
and human cases, but the outcome of treatment varies with
treatment parameter.
In this regard, our results reveal that energy density is the
only treatment parameter with predictable dose dependent
treatment effects. Five studies with energy densities ranging
from 19 to 24 J/cm2 had the largest average effect size.16,47-50
This suggests that this range of energy densities has more positive
effect than other dose levels, but the variability in the experimental
models used in these and other studies suggests the
need for a considerable amount of caution in drawing any conclusions
from this aspect of our findings. Nonetheless, the
range of energy density may serve as a reference starting point
for research programs aimed at comparing the effects of various
energy densities on tissue repair.
Several indices of tissue repair are positively affected by laser
treatment. This finding supports experimental animal and clinical
reports that indicate that laser therapy promotes wound healing
by accelerating collagen synthesis,2-6,55,60 inflammation,7,43,44
healing time and strength acquisition.2-7,29-34,79-85 This is consistent
with previous reports that have demonstrated elevation of
several metabolic indices of ATP synthesis,61-66 fibroblast proliferation,
7,8,55,60 and collagen synthesis,2,8,10,11,21,59 as well as increases
in the biomechanical indices of tissue healing.2-6,86
The highest mean treatment effect size was calculated from
studies that used cows, followed by rats, pigs, and humans in
that order. This finding supports previous studies15,35 that indicate
that the effects of low energy lasers appear to be more
prominent in loose skinned animals than pigs and humans.
15,35,51 However, these reports must await further confirmation
by well-designed studies comparing the effects of laser
therapy on loose-skinned animals, pigs, and humans, under the
same experimental conditions.
CONCLUSION
The results of this statistical meta-analysis mandate the following
conclusions:
• Laser therapy, also referred to as low-level laser therapy
(LLLT), is an effective modality for treating wounds.
• The outcome of treatment varies with treatment parameters
(i.e., power, power density, wavelength, beam profile, energy,
energy density, number and frequency of treatment, duration
of treatment).
• Energy density appears to be the only treatment parameter
with predictable dose-dependent treatment effects.
Mer dokumentasjon her!
© Bergen Kvantemedisin
Produsert av Datakilden AS
|
