Within-subject comparisons in phototherapy patients (n = 17). Change in scaling, erythema, and induration (SEI) score (from beginning to end of UV-B course) in plaques covered during UV-B, plaques exposed once weekly only, and plaques exposed normally 3 times weekly. The hatched line indicates mean change in SEI score.
This photograph, taken at end of treatment course, shows the clear cutoff in clearing response to UV-B therapy between part of plaque directly exposed (E) and covered plaque (C).
Mean scaling, erythema, and induration (SEI) score at each weekly assessment of plaques in 17 patients receiving UV-B therapy. AUC indicates area under the psoriasis severity time curves, which were calculated separately for each individual's selected plaques.
Change in scaling, erythema, and induration (SEI) score over 3 weeks (with correction of scores to 3-week values for control patients who were assessed later than 3 weeks) in plaques covered during UV-B treatment in phototherapy patients compared with that in plaques in control patients awaiting UV-B treatment. The hatched line indicates mean change in SEI score.
Dawe RS, Cameron H, Yule S, Man I, Ibbotson SH, Ferguson J. UV-B Phototherapy Clears Psoriasis Through Local Effects. Arch Dermatol. 2002;138(8):1071-1076. doi:10.1001/archderm.138.8.1071
To determine if UV-B phototherapy clears psoriasis through systemic
Randomized, within-subject comparison of change in psoriasis in 3 plaques
in patients attending for whole-body UV-B therapy. Change in patients' psoriasis
plaques covered during UV-B treatment was compared with plaques in an untreated
University hospital phototherapy unit.
The study population comprised 17 patients with chronic plaque psoriasis
treated with UV-B and 24 psoriasis control patients awaiting UV-B phototherapy.
Treatment with a standard 3-times weekly narrowband TL-01 UV-B regimen.
Three similar plaques were randomly allocated to be covered every treatment,
covered for 2 of 3 weekly treatments, and exposed to local UV-B every treatment.
Similar plaques were selected in control patients (awaiting but not yet started
UV-B therapy). Severity of psoriasis plaques was assessed using a scaling,
erythema, and induration (SEI) scoring system.
Main Outcome Measures
Change in SEI score of the selected plaques over the complete treatment
course for UV-B–treated patients and change over 3 weeks in SEI score
of plaques covered during UV-B treatment compared with that of plaques in
There was a significant (P<.001) difference
in how much the SEI score changed in the 3 plaques in UV-B–treated patients.
It fell by a mean of 7.6 for uncovered plaques compared with 3.2 for plaques
covered during each UV-B exposure (95% confidence interval for difference,
3.0 to 5.8). In patients awaiting UV-B, SEI score of plaques fell by a mean
of 0.4 over 3 weeks, compared with a mean fall of 1.4 for covered plaques
in UV-B-treated patients (95% confidence interval for difference in means,
0.1 to 2.0).
If UV-B therapy has any systemic effect capable of improving psoriasis,
this effect is small and unlikely to be of clinical importance. It is insufficient
to alter interpretation of findings of within-subject comparative phototherapy
studies. UV-B phototherapy works for chronic plaque psoriasis through local
DOES UV-B phototherapy clear psoriasis through direct local effects,
systemic effects, or a combination of both? The answer to this has implications
for the interpretation of within-individual comparisons of different phototherapy
regimens. It should also contribute to our attempts to understand how UV-B
Conceivably, UV-B may work to clear psoriasis on sites not directly
irradiated. Ultraviolet B has various effects on parameters of immune function
detectable in the circulating blood.1 Whole-body
broadband UV-B treatment of humans has effects on the function of blood neutrophils,2 and narrowband TL-01 UV-B reduces circulating natural
killer cell activity.3 In mice, systemic inhibition
of induction of delayed type hypersensitivity appears to be mediated by UV
radiation–induced interleukin 10 release.4 According to findings from laboratory studies, UV-B may have anti-inflammatory
effects resulting from the release of neuropeptides from the skin into the
circulation.5,6 In healthy human
volunteers, systemic effects of UV-B on nonimmunological responses to benzoic
acid and methyl nicotinate have been reported.7 Also, increased activity of interleukin 6 and tumor necrosis factor α
may be important in systemic symptoms associated with sunburn,8,9 and therapeutic broadband UV-B leads to increased interleukin 1 activity.10 Such demonstrated systemic effects could possibly
be relevant in psoriasis clearing.
Other treatments have been reported to have a systemic effect on psoriasis
clearing: untreated plaques in patients treated elsewhere with dithranol improved
significantly in one study.11 It was suggested
that this was due to circulating factors (whether cellular or humoral) released
from the treated plaques, rather than to a direct systemic effect of absorbed
In clinical practice, we recognize that sites not locally irradiated
tend not to clear as a result of UV-B therapy alone, and routinely adjunctive
topical therapies for shadow sites have to be prescribed. This does not, however,
rule out a systemic effect that might be important, even if insufficient to
consistently clear psoriasis without other treatment.
The within-patient study design (ie, comparing one phototherapy regimen
to one area [usually half] of each patient's body with another regimen to
another area) has been widely used to compare different TL-01 UV-B regimens12- 14 and to compare TL-01
UV-B with other phototherapies.15- 22 This design has the major advantage of ensuring that all variables that might
influence clearing of psoriasis, apart from the one being studied, are the
same. However, if TL-01 UV-B exerts an important systemic effect on psoriasis
clearance, even if this effect might only be apparent in psoriasis that also
receives some local UV radiation, it is possible that the results of some
of these studies were misleading. For example, following a comparison of 3-times
weekly with 5-times weekly TL-01 UV-B treatment for psoriasis, we favored
3-times weekly treatment, but could a systemic effect of the 5-times weekly
treatment to 1 body half have aided clearing on the 3-times weekly side?
This study was a prospective, randomized, controlled, interventional
study approved by the Tayside Committee on Medical Research Ethics. Consecutive
patients with chronic plaque psoriasis referred to the phototherapy unit at
Ninewells Hospital and Medical School (a university hospital serving Tayside,
Scotland) between April 1998 and June 2000 were invited to participate. Exclusion
criteria were age younger than 18 years; a history of skin cancer or solar
keratoses; or phototherapy, psoralen–UV-A therapy, or systemic therapy
for psoriasis within the preceding 3 months. The enrollment criteria for patients
treated with phototherapy during the study and the concurrently recruited
control group (awaiting but not yet started phototherapy) were identical.
Ultraviolet B phototherapy was administered according to our standard
regimen. The starting dose was 70% of each individual's minimal erythema dose,
with 20% increments, reduced to 10% according to erythemal response. Either
a Waldmann UV5000 cabinet (Herbert Waldmann GmbH & Co KG, Villingen-Schwenningen,
Germany) fitted with 24 Philips 100W TL-01 lamps (Philips, Eindhoven, the
Netherlands) or a cabinet constructed at the Ninewells Medical Physics Department
with 50 Philips 100W TL-01 lamps was used.
Adjunctive therapy was limited to emollients known not to significantly
impede UV transmission23,24 and
standard topical treatments for scalp, face, and flexures. End points for
stopping therapy were, as in our routine practice, clearance (no palpable
psoriasis remaining) or "minimal residual activity," defined as trace disease,
below knees or on sacrum only. To give all patients a chance of achieving
complete clearance, treatment was stopped either at clearance or after the
fourth treatment at which minimal residual activity was documented, whichever
came first. The control patients awaiting phototherapy were permitted the
same topical therapy (emollients and scalp and flexure treatments) as those
treated with UV-B.
For the group attending for phototherapy, 3 similar plaques on the trunk
or limbs were selected before treatment was started. Only plaques above the
knees were selected on lower limbs. These were assessed for scaling, erythema,
and induration scores (see "Outcome Measures") and randomly allocated (as
described under "Assignment") to the following interventions: 1 was covered
with Tegaderm dressing (3M Health Care, St Paul, Minn) with backing kept on
(which allowed no UV radiation transmission when assessed with Hitachi U-3210
[Tokyo, Japan] double-beam reflectance spectrophotometer) during each treatment,
1 was kept uncovered (ie, exposed normally during treatment), and 1 was covered
for 2 of the 3 treatments delivered each week (and treated independently following
the same standard regimen as for whole-body exposure). This once-weekly exposed
plaque was included because we hypothesized that systemic effects might only
significantly influence psoriasis clearing in conjunction with some "priming"
For the control group, a single plaque, similar to the selected plaques
in the patient group receiving phototherapy, was chosen in each of 24 patients
referred for, but not yet receiving, phototherapy. Topical therapies for these
patients were limited to the same preparations as were allowed for the phototherapy
patient group. Scaling, erythema, and induration score for this plaque was
recorded at baseline and 3 weeks later (or as soon after as feasible).
For pragmatic reasons, to aid control patient recruitment, this assessment
was done when the patient attended to commence phototherapy (hence, some variation
in exact interval from baseline assessment).
We used the sum of scaling, erythema, and induration (SEI) scores (each
on a 0 to 4 scale) as a measure of psoriasis severity for each selected study
plaque. This scoring scheme was based on the standard psoriasis area and severity
index, and we had experience of its use in earlier studies.12,14
The main end points were based on change in SEI scores for the selected
plaques. We planned to compare the amount of change from baseline to end of
treatment course between the 3 within-patient selected plaques and compare
the change over 3 weeks between the selected control patient plaques and the
plaques covered during UV-B treatment. A secondary end point was comparison
of area under the psoriasis severity score time curves for plaques covered
every treatment, covered for 2 treatments per week (ie, exposed to local irradiation
only once weekly), and plaques exposed normally during treatments 3-times
We estimated that we would require at least 16 patients to have 90%
power to detect (as significant at the 5% level) a difference of 1.5 in mean
change in SEI score for plaques covered vs plaques not covered during UV-B.
Variation (SD, 1.3) of within-patient difference in amount of change in SEI
score was derived from an 8-patient pilot study. For the within-patient comparisons,
repeated measures analysis of variance was used to look for any difference
across the 3-plaque treatment allocation groups. This was followed by the
Tukey wholly significant difference pairwise comparison procedure, using the
variance measure derived from the analysis of variance model. For the comparisons
between patients (covered phototherapy patient plaques and untreated patient
plaques), the unpaired t test was used. Not all control
patients were assessed at exactly 3 weeks, and therefore we corrected SEI
scores to those that would have been expected at 3 weeks after baseline, assuming
that the rate of change for each plaque would follow the slope of a simple
linear regression model (based on all the control patient plaque scores at
the second visit plotted against days since baseline). When comparing demographic
and treatment course data for the phototherapy patients and control (awaiting
phototherapy) patients, the χ2 test and the unpaired t test (with log transformation when appropriate to ensure
approximation to a Gaussian distribution and using the Satterthwaite method
when the requirement for equality of variances was not met) were used. Stata
(Intercooled Stata for Windows, release 7; Stata Corp, College Station, Tex,
2001) statistical software and a user-written macro for pairwise comparisons25 were used.
Random allocation determined what intervention each of the 3 selected
within-patient plaques received. The unit of randomization was plaque within
patient. After selection, plaques were numbered (1, 2, and 3), and a randomization
list generated from random number tables26 held by a member of department not directly involved in the study was used
to determine which intervention each patient was to receive.
Patients and those administering therapy could not be kept unaware of
allocation. Assessors were not told each plaque's treatment allocation. However,
this observer masking was unavoidably incomplete because clearly demarcated
pigmentation developed around the covered plaques as treatment courses progressed,
making it possible to guess which these were.
Seventeen patients attending for UV-B phototherapy and 24 control patients
awaiting phototherapy participated in the present study (Table 1). All those attending for UV-B phototherapy completed the
study. Of the control patients, 5 did not complete the study: 3 withdrew (1
to start UV-B immediately, 1 used a coal tar solution cream, 1 moved and started
UV-B therapy elsewhere), and 2 did not have their selected plaques assessed
at their second (pre–UV-B therapy) visit.
Phototherapy patients and controls were similar in age, sex ratio, and skin phototypes (Table 1). Within these groups, the psoriasis responded to a similar degree, as indicated by the almost identical treatment numbers and UV-B dose required for treatment.
Most plaques improved over the period from the beginning of the UV-B treatment course to the end of the course (Figure 1). There was a significant (P<.001) difference in how much the SEI score changed between the 3 plaques (covered during treatment, covered for 2 of 3 treatments per week, or exposed during treatment). The mean reduction in SEI score was 7.6 for plaques directly exposed to UV-B compared with 3.2 for plaques covered during each UV-B exposure, that is, a greater reduction in mean SEI score of 4.4 (95% confidence interval [CI], 3.0-5.8) in the normally exposed plaques). The difference in clinical appearance of psoriasis covered during each treatment and neighboring exposed psoriasis was clear by the end of most patients' treatment courses (Figure 2). There was no significant difference between the mean fall in SEI score for covered vs once-weekly exposed plaques (95% CI, –0.8 to 2.0). The markedly greater decline in mean SEI score throughout the treatment courses in the 3–times weekly exposed plaques compared with the covered and once-weekly–exposed plaques is shown in Figure 3.
Initial comparison of change in SEI score over 3 weeks for covered plaques in phototherapy patients with change from baseline to second assessment (≥3 weeks [mode, 3½ weeks; median, 4 weeks]) for control patient plaques showed no significant difference. The mean score for covered plaques in phototherapy patients fell by 0.4 (95% CI, –0.4 to 1.3) more than the mean control patient plaque score. When the change in score for control plaques was corrected (to allow a fairer comparison, since control plaques were assessed at intervals up to a maximum of 7 weeks after initial assessment), the difference remained small and of no likely clinical importance, although it was unlikely to have occurred by chance. The mean SEI score for covered plaques in phototherapy patients fell by 1.0 (95% CI, 0.1-2.0) more than it fell over 3 weeks in control patient plaques (Figure 4).
There were no differences in variables that might influence response of the 3 plaques in patients receiving phototherapy. Prognostic features attributable to the patient as a whole were of course identical, and by random allocation of intervention to the individual plaques, we minimized the possibility that the plaques (allocated to any 1 of the 3 interventions) would be expected to be more or less likely to improve. We attempted to ensure that the plaques in control patients would have similar prognostic variables to the covered plaques in the phototherapy patients by selecting a control group of patients referred for phototherapy for chronic plaque psoriasis, that is, a group expected to have similarly severe psoriasis. In practice, this group showed similar responses to treatment when they received phototherapy after the study (Table 1).
The present study has shown that narrowband TL-01 UV-B phototherapy does not have a clinically important systemic effect contributing to its ability to clear psoriasis. It also showed that once-weekly treatment for up to 10 weeks has little effect on psoriasis. If, as we had hypothesized, UV-B therapy cleared psoriasis through a combination of systemic effects and minor local exposure, then we would have expected a much greater reduction of psoriasis severity in the once-weekly–treated plaque. Incidentally, this study makes it clear that once-weekly treatment of psoriasis would be inappropriate; even if once-weekly treatment might eventually clear psoriasis, it would be unacceptably slow.
To our knowledge, there has been no previous study attempting to determine whether systemic effects of UV-B, or any other form of phototherapy, contribute to psoriasis clearance. One earlier study assessed the possible systemic effect of dithranol treatment.11 That study did not include a control group, although the magnitude of improvements in plaques not directly treated suggested a systemic effect, possibly related to effects resulting from clearing of psoriasis elsewhere rather than directly due to dithranol application. We included a control group in our study because we expected some improvement in the study plaques, regardless of the presence or absence of systemic UV-B effects. Such improvement, as observed in the covered plaques of the phototherapy patients in our study (Figure 3), was expected as a result of the following:
The regression to the mean phenomenon. These patients were all seen when their psoriasis was likely to be worse than usual (to enter the study they had to have sufficiently severe psoriasis to require referral for UV-B). We could anticipate, regardless of any medical intervention, a fall in psoriasis severity from baseline as a result of simple random fluctuation in severity with a downward tendency toward the mean population severity.
The effects of allowed treatment other than UV-B, that is, the emollients and the encouragement to apply these and possibly the psychological support associated with 3-times weekly attendance.
We aimed to recruit a control group with similar psoriasis to those attending for phototherapy. This meant asking patients with chronic plaque psoriasis referred for UV-B phototherapy to participate. We considered it inappropriate to withhold treatment from such patients for 10 weeks (the maximum duration of a normal treatment course). So, we did not recruit the ideal comparison group of patients requiring, but not yet receiving, UV-B therapy and willing to be treated with emollients alone, while attending for review 3-times weekly for 10 weeks. Although it would have been possible to recruit patients with psoriasis not requiring UV-B as such a control group, it would not have been appropriate. Such patients might be expected to have milder psoriasis and more likely to respond well to emollients and encouragement of regular review alone, so use of such a control group might have hidden a systemic effect of UV-B on the plaques of patients attending for phototherapy. Our compromise was to ensure our control group was similar by recruiting those referred for UV-B therapy, but accepting that we could only ask them to do without treatment for 3 weeks (not unreasonable, since we then had such a waiting list to start treatment) and not insisting on them making extra visits 3 times weekly while not being treated.
As not all control patients were assessed at 3 weeks, we corrected their second SEI score to that expected at 3 weeks, making the assumption that control patient plaques showed an equal rate of decline in SEI score. If we did not do this and compared the change over 3 weeks in phototherapy patients, there would be no significant difference, but with correction for the fact that several control patients had longer than 3 weeks to improve, the SEI score fell significantly more in the covered plaques of UV-B treated patients. However, it should be noted that this difference in fall in SEI scores was small, and although unlikely to be a chance finding, was not large enough to be clinically important. It is probable that such a small difference in decline in SEI score can be explained by the fact that the patients receiving phototherapy were encouraged to apply emollients during their 3-times weekly visits, while the controls were not. It remains possible that a systemic UV-B effect contributed to this slightly greater improvement in the covered UV-B–treated patient plaques, but such an effect would have to be minor.
We conclude that UV-B has no clinically important systemic effect on psoriasis clearance and that the possibility of what seems to be at most a minor systemic effect should not influence our interpretation of within-subject study comparisons of UV-B phototherapy regimens for psoriasis. Ultraviolet B phototherapy works for psoriasis through local effects.
Accepted for publication October 10, 2001.
This study was presented in part at a meeting of the Scottish Dermatological Society, Edinburgh, Scotland, February 8, 2001, and at the annual meeting of the European Society for Photobiology, Lillehammer, Norway, September 3-8, 2001.
Rhoda Hodgson and all the phototherapy unit nurses treated the participants and helped ensure the smooth running of the study. Lynn Fullerton, Chesarea McGeoghie, and Dee Watson provided medical physics support and ensured accurate dosimetry.
Corresponding author and reprints: Robert S. Dawe, MRCP, Photobiology Unit, Department of Dermatology, Dundee University, Ninewells Hospital and Medical School, Dundee DD1 9SY, Scotland (e-mail: firstname.lastname@example.org).