Background
Phototherapy treatment with UVA or UVB light is used in moderate to severe cases
of eczema. Phototherapy is often used as an alternative to other long-term
treatment regimens such as topical corticosteroids. Patients are usually treated
in hospital two or three times a week for several weeks and sometimes
months.
Existing systematic reviews
Since 2000, three systematic reviews have been published.83,379,380 The AAD94 and
NICE41 guidelines also
cover phototherapy.
Ultraviolet B treatments
Studies
A left/right within-person trial by Selvaag and colleagues,381 reported in 2005, compared
standard UVB fixed-dose increments against UVB using skin reflectance guided
dosing. The trial included 20 adults aged 16–38 years with mild to
moderate eczema. Treatment was given for up to 6 weeks and was stopped early
if a participant’s SCORAD score fell to < 10 on
either side of the body. The whole of the face was given standard UVB
treatment. Emollients and topical corticosteroids were allowed during the
trial as long as they were used symmetrically. It was not reported how many
treatments were given per week.
Assessment of risk of bias
provides the risk-of-bias assessment for the new study.
Ultraviolet B treatments: risk of bias of the included study
Benefits
The severity of eczema was recorded using the SCORAD index. It was reported
that no significant difference was found in the reduction of eczema severity
between the two treatment regimens, but no detailed data were reported. The
times taken for the SCORAD scores to reduce to < 10 were
reported. The standard treatment regimen had a median time (5th and 95th
percentiles) of 3.5 weeks (1.5, 6.0) and the skin reflectance regimen had a
median time (5th and 95th percentiles) of 3.0 weeks (2.0, 5.5). The
cumulative UVB dose was significantly lower for the skin reflectance regimen
(median 39 × 10 mJ/cm2;
p < 0.01) than for the standard
regimen (median 124 × 10 mJ/cm2).
The initial UVB dose was reported as higher for the skin reflectance
regimen, with a median of 3.4 standard erythemal dose (SED)
(×10 mJ/cm2 at 298 nm using the CIE
erythema action spectrum) compared with 2.6 SED
(×10 mJ/cm2 at 298 nm using the CIE
erythema action spectrum) for the standard regimen, but this was not
statistically significant.
Harms
Data on adverse events were not reported for this trial.
Overall implications for research and practice
The potential to administer low-dose UV radiation and achieve the same
clinically beneficial effects as reported with high-dose UV radiation is
encouraging but requires confirmation in larger pragmatic studies. This one
small trial381 indicates that
this could be achievable; however, the results must be treated with caution.
Problems with the study include the lack of blinding, the claim of
equivalence of non-inferiority being based on a very small sample size and
no methodological details about the study design being reported. Such an
important question should be examined in detail in an appropriately designed
and powered trial to give patients and clinicians clear guidelines on the
best treatment regimen for UV phototherapy.
Ultraviolet A treatments
Studies
Broadband UVA treatment is used alone or in combination with broadband UVB
treatment. More recently, UVA treatment has been combined with other agents
such as the photosensitiser psoralen plus ultraviolet A (PUVA)], which is a
naturally occurring chemical found in the common fig and celerys as well as
other plants and seeds. UVA is also being combined with
photo(chemo)therapeutic agents in UVA1 phototherapy treatment, narrowband
UVA therapy and extracorporeal photopheresis. Narrowband UVA and UVA1
phototherapy (high-intensity, long-wavelength UVA
340–400 nm) are currently used for eczema as they have a
high output and narrow emission spectrum and so are expected to be the most
efficacious and safe versions of UVA for eczema treatment.
A trial by Dittmar and colleagues382 in 2001 compared three different doses of UVA1 phototherapy
against each other. The low-dose group was given a maximum single dose of
20 J/cm2 and a maximum cumulative dose of
300 J/cm2; the medium-dose group received a maximum
single dose of 65 J/cm2 and a maximum cumulative dose of
975 J/cm2; and the high-dose group received a maximum
single dose of 130 J/cm2 and a maximum cumulative dose of
1840 J/cm2. The treatment was given five times a week
for 3 weeks. The trial randomised 34 adults with eczema and a SCORAD score
of > 30. No other treatments except for emollient were
permitted during the trial. The trial appeared to compare the treatments
using a parallel-group design, but this is not specifically stated.
A crossover trial by Tzaneva and colleagues383 compared UVA1 phototherapy with PUVA
provided on an outpatient basis. The 40 participants were given UVA1
phototherapy at doses of 20 J/cm2 unless the minimal
erythema dose was below this, in which case this value was used initially
with an incremental increase of 10 J/cm2 for each
subsequent treatment, up to a maximum of 70 J/cm2 as long
as there was no erythema. The PUVA treatment (1.2 mg/kg) was given 2
hours before UVA exposures, which had a starting dose of 70% of the minimal
phototoxic dose. The UVA dose was increased by 20% of the minimal phototoxic
dose if there was no erythema and 10% if there was barely perceptible
erythema. Other than the study treatment, participants were allowed to use
only emollients.
Assessment of risk of bias
provides the risk-of-bias assessment for the new studies.
Ultraviolet A treatments: risk of bias of the included studies
Benefits
In the trial by Dittmar and colleagues382 comparing three different doses of UVA phototherapy, the
low-dose group did not significantly improve in severity from baseline, with
a reduction in SCORAD score from 54 to 46. The medium- and high-dose groups
significantly improved compared with baseline, reducing from 56.29 to 40.16
and from 70.81 to 33.94, respectively. No between-group severity analyses
were reported. No other efficacy outcomes were reported.
The trial by Tzaneva and colleagues383 primarily measured length of remission after each
phototherapy treatment. The PUVA treatment group had a median time to
relapse of 12 weeks (IQR 4–24 weeks) whereas the UVA1 treatment
group had a median time to relapse of 4 weeks (IQR 4–12 weeks),
which was reported as statistically significant
(p = 0.012). Severity was reported as a
secondary outcome. The SCORAD scores in the PUVA group decreased from a mean
of 62.5 at baseline to 36 after 10 exposures and to 28.8 after 15 exposures.
The SCORAD scores in the UVA1 group decreased from a mean of 63.7 at
baseline to 46.9 after 10 exposures and to 40.1 after 15 exposures. The
mean ± SD percentage reductions in SCORAD scores
from baseline were 54.3% ± 25.7% for PUVA and
37.7% ± 22.8% for UVA1. The difference between the
groups was statistically significant.
Harms
The trial report by Dittmar and colleagues382 stated that no adverse events were
observed. In the trial by Tzaneva and colleagues383 only minor adverse events were reported.
Two participants treated with UVA1 and nine treated with PUVA reported mild
palmoplantar erythema. Seven participants treated with UVA1 reported heat
and burning after treatment. Folliculitis was reported by one participant
using UVA1 and by two participants using PUVA. Two participants using PUVA
reported photo-onycholysis (nail degradation).
Overall implications for research and practice
The optimal dosing regimen for treatment with UVA1 is still unclear. The
trial by Dittmar and colleagues382 has large disparities in both baseline eczema severity and
immediate pigmentation dose, and provides no between-group analysis. The
trial does provide a hint of a positive dose–response relationship,
which should be treated with caution. Further methodologically robust
research should attempt to clarify the optimal treatment regimen for
phototherapy with a clinically realistic duration of treatment.
The trial of PUVA treatment383
showed a modest superior effect of PUVA over UVA1 treatment after 15
treatments; however, this is a relatively short treatment period. The
particularly striking result was the significantly longer length of
remission that PUVA induces compared with UVA1. Inducing long periods of
remission in a chronic, long-term condition such as eczema is vitally
important. PUVA therapy is time-consuming, even more so than phototherapy on
its own, and so this must be taken into account.
Short-term adverse events appear to be mild and occur at low levels for both
UVA1 and PUVA treatment, with some indication that UVA1 has a slightly
better safety profile than PUVA.
Ultraviolet A compared with ultraviolet B treatments
Studies
A trial by Reynolds and colleagues384 randomised 73 adults with eczema that was not considered to
be mild to either narrowband UVB, broadband UVA or visible light
phototherapy. All participants had treatment twice a week for a total of 24
treatments. All participants were allowed to use emollients (emulsifying
ointment or aqueous cream were advised as some emollients absorb UV
radiation) and topical corticosteroids (except very potent ones) 2 weeks
before and during the trial.
A left/right within-person trial by Majoie and colleagues385 compared narrowband UVB
(311 nm) against medium-dose UVA1 (350–400 nm) given
three times per week for 8 weeks. Thirteen adults with a symmetrical eczema
distribution were included. For narrowband UVB, the first dose was 70% of
the minimal erythema dose; the dose was increased by 20% if there was no
erythema or by 10% if the previous dose produced slight erythema. For UVA1
phototherapy, the first dose was 30 J/cm2, and this was
increased to 45 J/cm2 in two treatments. The dose was
decreased if the reaction was too strong. No other topical treatments except
for emollients were allowed during the trial. The face was excluded from the
analyses.
A crossover trial by Gambichler and colleagues386 compared UVA1 phototherapy with narrowband
UVB phototherapy each given three times a week for 6 weeks. The trial
randomised 47 participants with eczema diagnosed according to the Hanifin
and Rajka8 criteria, with a
SCORAD score of ≥ 20. For UVA1 treatment the dose was
50 J/cm2. For narrowband UVB treatment, the first
dose was 70% of the minimal erythema dose and this was increased by
10–20% per session up to a maximum of 1.2 J/cm2
for skin type II or by 1.5 J/cm2 for skin type III or IV.
Any prospective participants with an abnormal photosensitivity to UVA1 were
not included in the trial. Participants were allowed to use emollients and
moisturisers during the trial.
Assessment of risk of bias
provides the risk-of-bias assessment for the new studies.
Ultraviolet A compared with UVB treatments: risk of bias of the
included studies
Benefits
The trial by Reynolds and colleagues384 reported that 19 out of 21 participants in the UVB group had
a reduction in itch over the treatment period compared with 12 out of 19 in
the UVA group. The number of participants who improved in the visible light
group was not reported, but it was stated that the UVA and UVB groups had
higher numbers of improvers than the visible light group. In total, 15 out
of 21 in the UVB group and 10 out of 19 in the UVA group had an improvement
in sleep at the end of treatment. Again, the number of improvers in the
visible light group was not provided but the UVA and UVB groups were
reported to have higher proportions of improvers than this group. The total
disease activity score improved in the UVA group by a mean of 4.4 points
(95% Cl –1.0 to 9.8) more than the visible light group and in the
UVB group by a mean of 9.4 points (95% Cl 3.6 to 15.2) more than the visible
light group. All three groups started with a similar mean baseline severity
score (UVA group 32.3, UVB group 29.8, visible light group 30.8). The
maximum severity score obtainable was 90. These data were extrapolated from
graphs.
In the trial by Majoie and colleagues385 the participant-assessed reduction in itch over 12 weeks was
similar for the two treatments. The medium-dose UVA1 group fell from 5.8 to
2.7 and the narrowband UVB group fell from 5.9 to 2.3. In the between-group
analysis it was reported that no significant difference was found, but no
data were provided. Eczema severity was recorded using the Leicester Sign
Score.61 The medium-dose
UVA1 group fell from 19 to 10 and the narrowband UVB group fell from 18 to
9. In the between-group analysis it was reported that no significant
difference was found, but no data were provided.
The trial by Gambichler and colleagues386 found no significant differences in the reduction in
pruritus, eczema severity measured using SASSAD and quality of life measured
using Skindex-29. For pruritus, the UVA1 group had a 16%
SD ± 61.8% reduction at the end of treatment whereas
the narrowband UVB group had a 25.2% SD ± 30.5%
reduction for treatment at the end of treatment
(p = 0.5). Eczema severity was reported as
the mean relative reduction in SASSAD score after 6 weeks. The UVA1 group
had a 43.7% SD ± 31.4% reduction at the end of
treatment whereas the narrowband UVB group had a 39.4%
SD ± 24.1% reduction at the end of treatment
(p = 0.5). Quality of life was also
reported as the mean relative reduction after 6 weeks. The UVA1 group had a
12.7% SD ± 18.8% reduction at the end of treatment
whereas the narrowband UVB group had a 16.5%
SD ± 17.6% reduction at the end of treatment
(p = 0.1).
Harms
In the trial by Reynolds and colleagues384 two participants withdrew because of
‘burning’ (one in the narrowband UVB group, one in the
visible fluorescent light group) and four withdrew because of
‘exacerbation of eczema’ (one in the narrowband UVB group,
two in the broadband UVA group and one in the visible fluorescent light
group). A further three participants withdrew because of ‘dislike of
treatment’ (two in the broadband UVA group and one in the visible
fluorescent light group). The trial by Majoie and colleagues385 did not report any data
about adverse events. In the trial by Gambichler and colleagues386 one participant in the UVA1
group and four in the narrowband UVB group developed mild erythema.
Overall implications for research and practice
Both UVA and UVB phototherapy appear to reduce pruritus and the severity of
eczema after a course of treatment, although broadband UVA therapy did not
appear to fare as well as narrowband UVB or UVA1 therapy. The length of time
for which these benefits are sustained after the cessation of phototherapy
has not yet been addressed; as there does not seem to be much of a
difference in the efficacy of UVA compared with UVB, this is now a key
research gap. Phototherapy is labour intensive for all involved and some of
the adverse effects such as heat loading can be difficult to cope with,
especially for children. There is a large degree of variability observed in
treatment response, which is evident from the wide deviations in severity
score reductions in the trial by Gambichler and colleagues.386 Trials involving two
different phototherapy treatments are easier to blind than trials of
phototherapy compared with other treatments.
Ultraviolet A/B treatments compared with or in combination with other active
treatments
Studies
A trial by Valkova and Velkova387 compared phototherapy with UVA/UVB against phototherapy with
UVA/UVB combined with the topical corticosteroids fluticasone and
hydrocortisone butyrate. Thirty-one adults and children aged between 8 and
45 years with moderate to severe eczema were randomised and underwent UV
treatment five times a week, with one group also applying topical
corticosteroids twice a day, five times a week. The length of time that the
study treatment was given was not reported.
A left/right within-person trial by Tzung and colleagues388 compared narrowband UVB alone against
narrowband UVB in combination with 1% pimecrolimus, twice daily for 6 weeks.
The trial randomised 26 children aged 5–17 years to either half-body
UVB and whole-body pimecrolimus or whole-body UVB and half-body
pimecrolimus. The first dose of UVB treatment was 70% of the minimal
erythema dose and then percentage-based increments up to a maximum of
1.5 J/cm2 were carried out.
A multicentre, two-arm trial by Heinlin and colleagues389 compared synchronous balneotherapy, in
which the participants were immersed in a bath containing dead sea salts at
a concentration of 10% and given UVB (311 nm) phototherapy, with UVB
(311 nm) phototherapy only. In total, 180 adults with
dermatologist-diagnosed eczema were given treatment according to a dose
escalation schedule for their skin type. The bathing time for the
synchronous balneotherapy increased in line with the schedule for the
phototherapy. Participants started with three to five sessions a week and
underwent 35 sessions in total.
The trial by Granlund and colleagues390 is discussed in Chapter 11.
Assessment of risk of bias
provides the risk-of-bias assessment for the new studies.
Ultraviolet A/B compared with, or in combination with, other active
treatments: risk of bias of the included studies
Benefits
In the trial by Valkova and Velkova387 no between-treatment comparisons were reported. There was a
large reduction in the severity of itch in the phototherapy and topical
corticosteroid combination group, from a mean ± SD
of 235.7 ± 16.9 to
78.6 ± 18.7 after treatment. In the
phototherapy-only group the very low score of
3 ± 13.6 increased to
5 ± 12.3 after treatment. Sleep loss decreased in
the phototherapy and topical corticosteroid combination group, from a mean
of 50 ± 20.2 to 21.4 ± 11.4
after treatment. In the phototherapy-only group the score decreased from
76 ± 23.5 to 11 ± 8 after
treatment. Overall eczema clinical severity decreased in the phototherapy
and topical corticosteroid combination group, from
395.4 ± 35 to 36.9 ± 7.3
after treatment. In the phototherapy-only group the score was
360.4 ± 37.6 at baseline and
37.9 ± 6.7 after treatment.
In the trial by Tzung and colleagues388 the primary outcome of change in eczema severity was
measured using EASI scores. For combination treatment compared with
pimecrolimus alone, there was no significant difference in the reduction of
severity from baseline (p = 0.084). This
was also the case for combination treatment compared with narrowband UVB
treatment alone (p = 0.059). The
combination treatment and each of the treatments alone all reduced the
baseline severity of eczema by around 50%. No absolute values were given for
severity in the trial report. All three treatments also reduced the severity
of pruritus by around 3 points. Again, no absolute values for pruritus were
given.
The trial by Heinlin and colleagues389 reported quality of life, measured using the Sickness Impact
Profile, which was evaluated by the patient. There was no significant
difference in mean quality of life between the treatment groups at the end
of treatment [synchronous balneotherapy 4.6 (SD 6.8) vs. phototherapy only
4.0 (SD 5.5); p = 0.98]. Disease-specific
quality of life was measured using the Freiburg Quality of Life Index and it
was reported that there was no significant difference between the groups at
the end of treatment. The participants assessed their global impression of
treatment on a 6-point scale from ‘very good to ‘very
bad’ and the proportion of participants with a score of
‘good or ‘very good’ at the end of treatment was
statistically significantly different between the groups (synchronous
balneotherapy 73.6% vs. phototherapy only 55.4%;
p = 0.002) and was also significantly
different at 1 and 6 months after the end of treatment. There was a
statistically significant difference in the reduction from baseline in the
severity of eczema (primary outcome), measured using the SCORAD index,
between the synchronous balneotherapy group [61.8 (SD 14.1) to 25.6 (SD
22.0)] and the phototherapy-only group [61.5 (SD 12.4) to 34.6 (22.3)] at
the end of treatment (after 35 treatments)
(p = 0.004).
Harms
The adverse events reported were erythema with skin tenderness, burning, skin
xerosis, uncomfortable heat load and intense sweating.387 All of these events except for skin xerosis
were reported with a frequency of ‘five or less’ but it was
unclear whether this was ‘events’ or ‘participants
affected’. Skin xerosis was reported with a frequency of 10 for the
phototherapy-only treatment group and five for the combination treatment
group.
In the trial by Tzung and colleagues,388 two participants in the whole-body narrowband UVB and
half-body pimecrolimus group had intractable generalised pruritus and tender
erythema after the UVB treatment.
The trial by Heinlin and colleagues389 reported that 30 participants in the synchronous
balneotherapy group experienced 46 adverse events compared with 24
participants in the phototherapy group who experienced 31 adverse events.
Eleven out of 46 events in the synchronous balneotherapy group were
‘definitely’ or ‘probably’ related to the
trial treatment whereas 10 out of 31 events in the phototherapy group were
‘definitely’ or ‘probably’ related to the
trial treatment. The most common events related to trial treatment were
erythema and light dermatoses. Eight of the adverse events were serious
(synchronous balneotherapy, n = 2;
phototherapy only, n = 6), but none of
these was classed as related to the trial treatment. Eight participants
withdrew before the end of the trial because of adverse events (synchronous
balneotherapy, n = 2; phototherapy only,
n = 6).
Overall implications for research and practice
Although no formal comparative analyses were reported, it is obvious from the
reduction in scores that there was no difference between a combination of
UVA/UVB and topical corticosteroids and UVA/UVB alone. A huge disparity in
baseline itch scores leads to questions about the method of randomisation
and allocation concealment and makes it impossible to interpret the impact
of the treatments on itch compared with each other. One small trial of
balneotherapy combined with UVB phototherapy compared with UVB phototherapy
only gives some evidence of benefit from the addition of balneotherapy;
however, as there is no mention of a blinded severity of eczema outcome
assessor for this trial, this evidence must be treated with caution until
appropriately blinded trials are carried out to confirm this beneficial
effect.
Full-spectrum light treatments
Studies
One trial by Byun and colleagues391 compared full-spectrum light (320–5000 nm)
for eight irradiations (twice a week for 4 weeks) plus emollient against
emollient only twice a week for 4 consecutive weeks. The 38 children
randomised into the trial were all Korean with a SCORAD score of
> 25 and skin type III or IV. The children were not allowed
to use any other treatments during the trial.
Assessment of risk of bias
provides the risk-of-bias assessment for the new study.
Full-spectrum light treatment: risk of bias of the included study
Benefits
The participants assessed their own clinical improvement in the trial, with
75% of the full-spectrum light group recording a ‘good’ or
‘excellent’ response on a 4-point scale compared with 50% of
the emollient-only control group. The severity of eczema was measured using
the SCORAD index. The score in the full-spectrum light group reduced from a
mean of 47.87 at baseline to 30.76 at week 8. The score in the control group
reduced from a mean of 39.79 at baseline to 33.8 at week 8. Although a
significant reduction from baseline was reported in the full-spectrum light
group, no between-group analyses were reported.
Harms
No serious adverse events were reported. In the full-spectrum light group, 6
out of 20 participants reported erythema, 6 out of 20 reported dryness, 4
out of 20 reported pruritus and 2 out of 20 reported burning. Six out of 20
participants also reported a transient exacerbation of eczema in the first 2
weeks.
Overall implications for research and practice
This trial391 was reported as
open. It is not clear whether the SCORAD assessor was blinded and the
baseline SCORAD scores were noticeably higher at baseline in the
full-spectrum light group than in the control group. This means that the
results should be treated with caution. The trial appears to show a
reasonable improvement in eczema severity matched by the
participants’ assessment of their own response to treatment. With
only a small number of participants and a narrow range of skin types and a
common heritage, much larger studies on mixed populations are needed before
any recommendations about the use of full-light phototherapy can be
made.
Excimer laser (form of ultraviolet laser)
Studies
A within-person trial from the Netherlands by Brenninkmeijer and
colleagues125 compared
once-daily clobetasol propionate (0.05%) against twice-weekly
200 mW/cm excimer laser for 10 weeks. The trial involved 13
participants with atopic eczema (diagnosed according to the millennium
criteria392) and more
than four symmetrical prurigo nodules.
Assessment of risk of bias
provides the risk-of-bias assessment for the new study.
Excimer laser: risk of bias of the included study
Benefits
A blinded physician assessment of individual signs of eczema found a
statistically significant difference in favour of excimer laser treatment at
14, 22 and 34 weeks after starting treatment, which in all cases was after
the end of the 10-week treatment period. There was a mean absolute decrease
in the score over 34 weeks of 6 points on a 15-point scale in the excimer
laser group and of 4.1 points in the clobetasol propionate group. Pruritus,
assessed on a VAS, showed a 63% improvement in the excimer laser group and a
49% improvement in the clobetasol propionate group over the entire 34 weeks
of the trial. The difference between the pruritus scores was reported as
non-significant for weeks 14, 22 or 34, with the absolute values being only
1 point apart on a 10-point scale over this time period.
Harms
Excimer laser treatment resulted in a fairly high level of adverse events,
although there were no reported serious adverse events. Adverse events
included four reports of a burning sensation, five reports of erythema, two
reports of vesicles and one report of blistering. All 10 participants
analysed for the trial experienced hyperpigmentation at the treatment sites.
Two participants withdrew because of an exacerbation of eczema that required
systemic treatment. It is unclear which treatment group the patients
withdrew from, although the exacerbations were described as being unlikely
to be related to the study treatment.
Overall implications for research and practice
This very small trial125 hints
at the potential for using clobetasol propionate laser treatment for a
relatively short-term period to confer long-lasting beneficial effects
compared with a moderate-potency topical corticosteroid. However, the level
of adverse events is a serious cause for concern.
