| Relative values of different outcomes | The outcomes considered for this question were:
PASI75 PASI50 Change in PASI Clear or nearly clear Improved (for palmoplantar pustulosis population only) Time to relapse (loss of PASI50) Time to remission/maximum response Change in DLQI Burn Cataracts Severe adverse events Withdrawal due to toxicity
The GDG considered which outcomes were most important when formulating recommendations for this review question. It was noted that it would be helpful to have consistency with the outcomes prioritised for the question on systemic non-biological therapies.
Trials for phototherapy tend to report time to clearance, whereas trials for systemic non-biological therapies tend to report PASI75 or PASI50.
Clear or nearly clear is a key outcome from the patient perspective, and there was most evidence for this outcome. Time to relapse and time to remission were felt to be important, as phototherapy is given intermittently, and so a longer duration of action is beneficial.
There was no evidence for change in DLQI, cataracts or severe adverse events.
The GDG discussed measures of toxicity. Toxicity from cumulative UV exposure was felt to be an inappropriate measure of toxicity, due to known inconsistencies in the metering of UV dose between centres. Number of treatments could be used instead of cumulative dose, but this was not an outcome for this question (although it is an outcome for the phototherapy plus acitretin or plus topical therapy reviews). Very few trials followed up participants at six or twelve months. There was no data on serious adverse events, so the GDG agreed on withdrawal due to toxicity as a measure of toxicity.
There was limited evidence for the rest of the outcomes.
Therefore the outcomes prioritised by the GDG were:
|
| Trade off between clinical benefits and harms | The phototherapy efficacy data were considered in the context of adverse effects in the short term (in this evidence review) and also for longer term skin cancer risk (see section 9.4). UVB (either NBUVB or BBUVB) was effective for inducing remission for plaque and guttate psoriasis, and well tolerated in the short term. Only very limited data were available for skin cancer risk. There was no statistically significant benefit of NBUVB over BBUVB in terms of efficacy but there was a trend favouring NBUVB over BBUVB for clearance at the end of treatment.
NBUVB three times a week is as effective as NBUVB twice a week, although time to clearance is shorter with three times weekly. The GDG agreed that either dosing schedule could be used depending on patient preference.
Following treatment with UVB, most patients relapse. Time to relapse is variable. In patients who relapse rapidly, the time, inconvenience, cost incurred when multiple courses of UVB are required to maintain disease control, together with the potential aging and any (unknown) risk of skin cancer, mean that further courses of UVB may not be appropriate and other treatments considered.
PUVA is more effective than NBUVB for achieving clearance of plaque psoriasis when both are used twice a week, but the two interventions are comparable when NBUVB is given three times a week. For people with palmoplantar pustulosis, oral PUVA was effective in terms of clearance compared to no treatment. There is a trend towards topical PUVA being more effective than NBUVB, but this was not statistically significant. From the evidence it is not known whether topical PUVA is as good as oral PUVA, as this comparison was not made.
Taking all the evidence into account, the risks of skin cancer with PUVA for psoriasis are significant, so UVB should be used in preference to PUVA as a first line phototherapy intervention. In patients who fail UVB, PUVA could be considered but only subject to the caveats and considerations discussed in section 9.4. |
| Economic considerations | There was limited health economic evidence to inform the GDG on the cost-effectiveness of BBUVB, NBUVB and PUVA. The GDG considered the partially applicable evidence whilst being mindful of its various methodological limitations. Two studies showed that PUVA was more costly and slightly more effective than broadband and narrowband UVB, but because neither study measured outcomes in terms of QALYs, the relative cost-effectiveness of PUVA remains indeterminable. When the result of one study was roughly translated from additional days in remission to QALYs, the incremental cost-effectiveness of PUVA was nearly £ 40,000 per QALY gained compared to broadband UVB.
The GDG considered whether de novo economic modelling would help to reduce uncertainty in the cost-effectiveness of phototherapy and PUVA, but concluded that it was unlikely to provide any additional information other than that which was already available. This was largely due to a lack of long-term trial data, and the fact that it would be difficult to robustly incorporate the risk of skin cancer into a model. In the absence of high quality, UK specific evidence, the GDG considered the unit cost of delivering phototherapy, for which NHS reference costs from 2010–11 indicate that PUVA is £ 59 more costly per session compared to UVB.
The clinical evidence suggests that there is very little difference in terms of effect (i.e. proportion achieving clearance of their psoriasis) between narrowband UVB administered at different frequencies (2×, 3× or 5× weekly). The main differences in effect appear to be related to the time and number of exposures by which clearance is achieved. The evidence suggests that an increased frequency of exposures per week may result in a slightly greater number of total exposures by the end of the treatment period (non-significant trend) and quicker clearance. This would translate to potentially higher costs, but also more QALYs. The combination of a vitamin D or vitamin D analogue to narrowband UVB may reduce the total number of exposures required to induce clearance, but the results did not reach statistical or clinical significance.
The clinical evidence suggested that PUVA, if offered at the same frequency, may be slightly better than narrowband UVB in terms of the proportion achieving clearance, time to clearance and total exposures to clear. In deciding to recommend narrowband UVB over PUVA, the GDG considered that the cost of delivering PUVA is £ 59 more per session than narrowband UVB. If 24 sessions (2× weekly for 12 weeks or 3× weekly for 8 weeks) were required to induce response, treatment costs would amount to an extra £ 1,416 for PUVA compared to UVB; to be considered cost saving compared to narrowband UVB, PUVA would need to generate the same response in 14 sessions or less. Combined with the evidence that the longer term risks of skin cancer associated with PUVA appear to be high and potentially higher than with narrowband UVB, they concluded that PUVA was unlikely to represent better value for NHS resource than narrowband UVB.
The GDG considered whether they should make a recommendation for phototherapy delivered in the home, given that clinical and cost-effectiveness evidence from the Netherlands suggested that it might be cost-effective. There were some concerns about the study and its application to decision-making for the NHS, including the inclusion of direct and indirect costs (productivity losses and travelling expenses) and the method by which QALYs were estimated during follow-up. The GDG was aware of home phototherapy being delivered in certain regions of the country, but did not consider the evidence robust enough to support its implementation across the entire NHS. In summary, the GDG recommended that it should only be considered in a select group of patients who may be unable to access hospital based services. |
| Quality of evidence | There were a number of important variables between the study designs that the GDG considered in reaching their recommendations:
Treatment frequency: not all trials used the standard number of treatments per week and the treatment frequency varied between treatment arms Within- and between-patient randomisation (and few studies provided sufficient information to correct for paired data in the analysis Treatment period and how accurately this was reported, which would influence the numbers experiencing improvement or toxicity
The GDG had reservations about the validity of the evidence comparing NBUVB and BBUVB, because some of the studies used BBUVB UV6, which is not true BBUVB as its wavelength lies somewhere between BBUVB and NBUVB.
The Cameron study found that NBUVB three times a week is better than NBUVB two times a week, but the data could not be included in the meta analysis (because the standard deviation was not available and mean time-to-event data cannot be used).
The GDG noted that NBUVB treatment regimes were likely to be suboptimal in some studies owing to a low treatment frequency. |
| Other considerations | It was noted that in many departments, NBUVB had become the main form of UVB phototherapy. The GDG considered the evidence (for superior efficacy or safety of NBUVB over BBUVB) not strong enough to recommend disinvesting in BBUVB, and also noted that BBUVB was used for other dermatoses.
The GDG considered home UVB treatment. The consensus view was that home UV treatment should be made available to people who are unable to access hospital treatment due to physical impairment or geographical reasons, and when other treatment options have failed or could not be used. However, given the unknown costs and lack of HE evidence, the GDG were unable to make a national recommendation.
From the GDG clinical knowledge PUVA itch and or pain is associated with PUVA use and can continue two years after stopping therapy. It affects up to 20% of patients.
The GDG noted that phototherapy is absolutely contraindicated in certain groups of people (for example xeroderma pigmentosum and other skin tumour prone photogenodermatoses), and those with photosensitive dermatoses (for example lupus erythrematosus, particularly systemic type). There are also a number of relative contraindications (for example epilepsy). The GDG agreed that provision of an exhaustive list was beyond the scope of the guideline, and that a recommendation that encompassed the fact that HCPs should be aware of the indications and contraindications to phototherapy, and the optimal administration of phototherapy, would be more appropriate.
The GDG noted that the response rates for PPP in the PUVA versus NBUVB study were potentially clinically relevant when considering response rates documented in the placebo controlled PUVA studies; this condition is difficult to treat, often functionally disabling, and NBUVB is a well tolerated intervention. The GDG considered the use of NBUVB an area for future research. |
