4 Evidence and interpretation

The Appraisal Committee (Appendix A) considered evidence from a number of sources (Appendix B).

4.1 Clinical effectiveness

Ten randomised controlled trials (RCTs) of tacrolimus and eight RCTs of pimecrolimus for the treatment of atopic eczema were considered. In most of the studies, comparators were topical corticosteroids or vehicle (a placebo consisting of the base cream or ointment without the active ingredient).

Effectiveness was assessed in different ways across the trials, but common measures included the following:

  • changes in severity using the Eczema Area Severity Index (EASI) or modified Eczema Area Severity Index (mEASI), which are well established assessment tools that account for individual symptoms and disease severity in four predefined regions of the body

  • the Investigator's Global Assessment (IGA), a physician's rating scale based on interpretation of signs of eczema

  • the Physician's Global Evaluation (PGE) of treatment success, which estimates the percentage change in the condition since the person was last seen

  • the physician's evaluation of reduction in affected body surface area (BSA)

  • patient-based measures such as number of flare-ups, pruritus and use of alternative treatments, such as topical corticosteroids.

    The main quality-of-life measure used in the trials was the Dermatology Life Quality Index (DLQI), which is a 10-item, self-administered questionnaire designed to measure the impact of skin diseases on patients' lives over the previous week.

Tacrolimus

4.1.1.1 Of the ten RCTs of tacrolimus, four were conducted in children and six in adults, all in populations with moderate to severe disease, reflecting the licensed indications.

4.1.1.2 Two of the studies in children compared tacrolimus with mild topical corticosteroids (n = 374 and n = 417 ['n' refers to overall study size]) and two compared it with vehicle (n = 180 and n = 352). Results were considered for 0.03% tacrolimus only, because 0.1% tacrolimus is not licensed for use in children. No trials in children have compared tacrolimus with potent or moderately potent topical corticosteroids.

4.1.1.3 Three of the RCTs in adults used potent topical corticosteroid regimens as a comparator (n = 181, n = 570 and n = 975) and three used vehicle (n = 14, n = 215 and n = 632).

4.1.1.4 For children, a pooled analysis of two studies that compared 0.03% tacrolimus with mild topical corticosteroids showed 0.03% tacrolimus to be more effective at 3 weeks, as measured by an improvement of at least 90% ('cleared' to 'excellent improvement') in the PGE (relative risk [RR] 2.56, 95% confidence interval [CI] 1.95 to 3.36). Based on the median improvement in mEASI score, both studies also showed 0.03% tacrolimus to be statistically significantly more effective than mild topical corticosteroids.

4.1.1.5 Both studies in children comparing 0.03% tacrolimus and vehicle found tacrolimus to be more effective using the PGE categories of 'cleared' to 'marked improvement' (56.5% vs 15.7% in the larger study [p < 0.001] and 69.0% vs 38.0% in the smaller study [p < 0.01] for 0.03% tacrolimus vs vehicle). Statistically significant differences were also demonstrated for a number of other measures, including changes in EASI score, pruritus score and patient reports of 'feeling better'. Loss to follow-up was higher in the vehicle arms of the trials (56.0% vs 17.0% in the larger study and 15.9% vs 8.1% in the smaller study).

4.1.1.6 For adults, a pooled analysis of two studies (n = 181 and n = 570) comparing 0.1% tacrolimus and potent topical corticosteroids found no statistically significant difference in the proportions of people with an improvement of at least 75% on the PGE ('cleared' to 'marked improvement') at 3 weeks (RR 1.08, 95% CI 0.97 to 1.21).

4.1.1.7 The pooled analysis did not include the largest study (n = 975), which compared 0.1% tacrolimus with mild topical corticosteroids on the face and with potent topical corticosteroids on the body. Overall, this study found a greater improvement on the PGE with tacrolimus (62.9% vs 40.7% at 3 months, difference = 22.2%, 95% CI 16.1 to 28.4). This study reported other statistically significant differences favouring tacrolimus compared with a topical corticosteroid regimen. This included a higher proportion of people achieving an improvement in mEASI by at least 60% at 3 months (the primary outcome), a greater median improvement in mEASI (83.3% vs 76.9% at 3 months, p < 0.001; also statistically significant at 4 and 6 months, 95% CIs not stated), a greater reduction in the affected BSA, and a greater proportion of people 'feeling better'. However, there was a high loss to follow-up (25.5% with tacrolimus and 42.1% with topical corticosteroids).

4.1.1.8 The other two studies that compared 0.1% tacrolimus and potent topical corticosteroids in adults did not report any statistically significant differences, although the larger of the two (n = 570) found smaller improvements in median mEASI score with 0.03% tacrolimus compared with potent topical corticosteroids (71% vs 83%, p < 0.05). Loss to follow-up in this study was 11.5% with tacrolimus and 9.1% with topical corticosteroids.

4.1.1.9 Compared with vehicle, treatment of adults with tacrolimus was shown to be statistically significantly more effective using a number of outcome measures, including the PGE categories of 'cleared' to 'marked improvement', reduction in BSA affected, and improvement in EASI score. Loss to follow-up was noticeably high in the vehicle arms of the trials (68.4% vs 26.7% and 38.9% vs 13.2% in the two largest trials).

4.1.1.10 Quality of life measures were reported by one study with an active comparator (n = 975) and one vehicle-controlled study in which participants were drawn from subsets of the largest studies of adults and children (n = 985). Both studies measured quality of life using the DLQI. Compared with a topical corticosteroid regimen (potent topical corticosteroid on the body and mild topical corticosteroid on the face), greater percentage improvements were found with 0.1% tacrolimus (66.7% vs 58.5% at 3 months and 74.3% vs 69.2% at 6 months, statistical significance was not reported), although the results for the different dimensions of 'quality of life' were not reported separately. Compared with vehicle, both strengths of tacrolimus were associated with greater improvements across all measurement dimensions in adults (symptoms and feelings, daily activities, leisure, work/school, personal relations, treatment [p = 0.0001]). In children (using the Children's DLQI [CDLQI] completed by children with help from parents/guardians), greater improvements were found with 0.03% tacrolimus than with vehicle across all measurement dimensions except one (the personal relationships dimension [p < 0.0001 to p = 0.02]) and across all dimensions for toddlers (using the CDLQI for toddlers completed by parents/guardians [p < 0.0001 to p = 0.001]).

4.1.1.11 For adverse effects, pooled analyses showed no statistically significant differences in skin infection rates between tacrolimus and topical corticosteroids. However, there was more skin burning in children with 0.03% tacrolimus compared with mild topical corticosteroids (RR 1.97, 95% CI 1.25 to 3.11) and in adults with 0.1% tacrolimus compared with potent topical corticosteroid regimens (RR 4.17, 95% CI 3.36 to 5.18). Withdrawal because of adverse effects was similar for tacrolimus and topical corticosteroids (1.6% to 3.9% of the tacrolimus group vs 1.6% to 3.3% of those using topical corticosteroids) but was consistently higher in the vehicle arms of the trials (4.5% to 12.3% with vehicle vs 2.9% to 5.7% with tacrolimus).

4.1.1.12 Overall, tacrolimus is more effective in treating moderate to severe disease in adults and children than vehicle alone. In children, a number of measures of treatment effect suggest that tacrolimus is also more effective than mild topical corticosteroids but no trials have compared it with more potent topical corticosteroids. In adults, compared with topical corticosteroids, 0.1% tacrolimus was statistically significantly more effective in one study (n = 975) but not statistically significantly different in the other two studies (n = 570 and n = 181).

Pimecrolimus

4.1.1.13 Of the eight RCTs of pimecrolimus, three were conducted in children and five in adults. Three of the studies were provided by the manufacturer as commercial in confidence (two studies in adults and one study in children).

4.1.1.14 None of the RCTs in children compared pimecrolimus with topical corticosteroids. Two studies used vehicle as a comparator (n = 403 and n = 713 ['n' refers to overall study size]) and one compared pimecrolimus with 0.03% tacrolimus (provided in confidence by the manufacturer [n = 141]).

4.1.1.15 Two of the RCTs in adults used potent topical corticosteroids as a comparator (n = 260 and n = 658); the larger of these two RCTs was provided in confidence by the manufacturer. Three studies used vehicle as a comparator (n = 34, n = 192 and n = 260). A further vehicle-controlled study was provided in confidence by the manufacturer (n = 200).

4.1.1.16 One of the studies in children (n = 713) and one in adults (n = 192) permitted the use of moderately potent topical corticosteroids to treat flare-ups in the treatment and control groups.

4.1.1.17 Three of the studies in adults were conducted in people with moderate to severe disease, which does not match the licensed indication (see Section 3.2.1). They were included in the review for pragmatic reasons – firstly, on the basis of expert advice that there is considerable overlap between categories of eczema severity and, secondly, because of the limited evidence available from studies comparing pimecrolimus with an active treatment.

4.1.1.18 Compared with potent topical corticosteroids, evidence from the smaller study suggests that pimecrolimus is less effective in treating moderate to severe atopic eczema. This study included four strengths of pimecrolimus (0.05%, 0.2%, 0.6% and 1.0%) but only the results for 1.0% pimecrolimus are presented, reflecting the product licence (see Section 3.2.1). Fewer people treated with pimecrolimus were judged to have 'clear' or 'almost clear' eczema at 3 weeks using the IGA score (11.1% vs 50.0%, difference = 38.9%, 95% CI 21 to 56). The study also reported a smaller percentage reduction in EASI for those using pimecrolimus (47.0% vs 78.0%, statistical significance was not reported) and that fewer people treated with pimecrolimus had mild or absent pruritus (46.7% vs 81.0%, difference = 34.3%, 95% CI 15.5 to 53.1, p < 0.05). The risk of skin burning was greater with pimecrolimus (RR 5.26, 95% CI 1.97 to 14.0). However, almost a quarter of participants were lost to follow-up (23.5%).

4.1.1.19 Compared with vehicle, pooled analyses of studies in children with mild to moderate disease and in adults with moderate to severe disease showed that treatment with pimecrolimus resulted in better IGA scores at 3 and 6 weeks (RR 4.47, 95% CI 1.40 to 14.27 [3 weeks]; RR 1.87, 95% CI 1.26 to 2.76 [6 weeks]) and fewer flare-ups at 6 months (RR 1.78, 95% CI 1.10 to 2.86). People using pimecrolimus were less likely to use additional treatment with topical corticosteroids (RR 1.82, 95% CI 1.51 to 2.21) and pruritus was more often absent or mild at 3 and 6 weeks among the pimecrolimus group (RR at 3 weeks 1.99, 95% CI 1.53 to 2.58).

4.1.1.20 Results from individual studies showed that pimecrolimus was statistically significantly more effective than vehicle using other measures of treatment effect, such as changes in EASI score and reduction in BSA affected.

4.1.1.21 Quality of life for people treated with pimecrolimus was reported by one comparison with vehicle in adults (n = 192) and for a subset of participants from the smaller study in children (n = 241). The study in adults reported changes in two quality of life measures (the Quality of Life Index – Atopic Dermatitis [QoLIAD] and the DLQI) and found greater improvements among patients using pimecrolimus compared with those using vehicle for both measures (25.6% vs 7.4%, p = 0.002 and 22% vs 6.7%, p = 0.01). The study in children, which used the Parent's Index of QOL in Atopic Dermatitis, also reported a higher quality of life with pimecrolimus (p = 0.023).

4.1.1.22 For adverse effects, a pooled analysis showed no statistically significant differences in rates of bacterial infection and skin burning between pimecrolimus and vehicle, but there was a greater risk of viral skin infection with pimecrolimus (12% vs 6%, RR 1.97, 95% CI 1.21 to 3.19).

4.1.1.23 Overall, studies found pimecrolimus to be more effective at treating atopic eczema in adults and children than was vehicle alone. There is limited evidence on the effectiveness of pimecrolimus compared with topical corticosteroids.

4.2 Cost effectiveness

The Committee considered economic analyses submitted by the manufacturers of both products and a model developed by the Assessment Group. There was only one relevant published economic analysis and this was conducted from the perspective of the US healthcare system and had methodological problems that limited its value. The Assessment Group analysis consisted of eight separate models, each relating to different cohorts of people with atopic eczema.

Tacrolimus

4.2.1.1 For the analysis of cost effectiveness in children, the Assessment Group model was run over 14 years to incorporate the possibility of disease resolution. All patients were assumed to be aged 2 years on entering. The model estimated that the costs per additional quality-adjusted life year (QALY) of first- and second-line treatment of eczema on the body were £9100 and £14,200, respectively, relative to the standard practice of using topical corticosteroids alone. For eczema in sensitive areas (on the face or in areas such as the armpits or groin where the treatment options are more limited because of concerns about skin atrophy), the costs per additional QALY were higher. First-line treatment cost £35,700 per additional QALY and second-line treatment was dominated by topical corticosteroids (that is, tacrolimus was both less effective and more costly than topical corticosteroids).

4.2.1.2 In adults, the cost per additional QALY of first-line treatment of eczema on the body with tacrolimus was £37,400 relative to standard treatment with topical corticosteroids. Second-line treatment of 'body' eczema with tacrolimus cost less than topical corticosteroids alone but also conferred fewer QALYs. For eczema in sensitive areas (see Section 4.2.1.1), the cost per additional QALY of first-line treatment with tacrolimus was £11,900, and second-line treatment with tacrolimus was dominated by topical corticosteroids.

4.2.1.3 Using a lower estimate of the amount of tacrolimus ointment used (2.3 g per day rather than 6.8 g per day) for adults reduced the incremental cost effectiveness ratio (ICER) for the treatment of 'body' eczema with tacrolimus (first-line) to £30,300 per QALY. Second-line treatment of 'body' eczema with tacrolimus still cost less than topical corticosteroids alone whilst also conferring fewer QALYs. The ICER for first-line treatment of eczema in sensitive areas (see Section 4.2.1.1) was also lower (£7000 per QALY), but tacrolimus as a second-line treatment continued to be dominated by topical corticosteroids.

4.2.1.4 Overall, tacrolimus (first- and second-line) was estimated to be more costly than topical corticosteroids alone in all cases except one (second-line treatment of 'body' eczema in adults), whilst absolute differences in QALYs were very small. The Assessment Group also performed probabilistic analyses. The base-case results of all the Assessment Group's models were shown to be associated with considerable uncertainty, due mainly to the very small incremental benefits predicted.

4.2.1.5 The manufacturer provided a cost-effectiveness analysis, in terms of disease-controlled days (DCDs), and compared treatment with tacrolimus (0.1% in adults, and 0.1% and 0.03% in children) after topical corticosteroids (potent topical corticosteroids as comparator for 0.1% tacrolimus, mild topical corticosteroids as comparator for 0.03% tacrolimus) with treatment with tacrolimus before topical corticosteroids in a hospital outpatient setting. The model adopted a semi-Markov approach, organised in four arms (topical corticosteroids in moderate or severe disease and tacrolimus in moderate or severe disease), and each one included four states of progression of atopic eczema (cleared or virtually cleared, moderately controlled, uncontrolled and flare-up). Each arm was run for 27 weeks (adults) or 15 weeks (children), corresponding to the duration of follow-up in the RCTs from which the effectiveness estimates were derived (scenario 1). In a second scenario, the model was run for 51 weeks on the basis of effectiveness estimates obtained from experts, and ciclosporin was included as a comparator for adults with severe eczema. A modified NHS and Personal and Social Services perspective was adopted, including work days lost.

4.2.1.6 The results were presented as average cost-effectiveness ratios rather than ICERs. The Assessment Group calculated ICERs using the data provided by the manufacturer.

4.2.1.7 For children with moderate disease, results from scenario 1 showed that 0.03% tacrolimus was dominated by both topical corticosteroids and 0.1% tacrolimus (that is, it produced fewer DCDs and was more costly than topical corticosteroids and 0.1% tacrolimus) and that 0.1% tacrolimus produced more DCDs than topical corticosteroids but was also more costly (the cost per additional DCD with 0.1% tacrolimus relative to topical corticosteroids was £16.40). In contrast, for children with severe disease, 0.1% tacrolimus was dominated by 0.03% tacrolimus and topical corticosteroids, and 0.03% tacrolimus produced more DCDs than topical corticosteroids but was also more costly (the cost per additional DCD with 0.03% tacrolimus relative to topical corticosteroids was £18.10). In scenario 2, where the model was run over 51 weeks, the cost per additional DCD with tacrolimus relative to topical corticosteroids was £3.30 (moderate eczema) and £16.10 (severe eczema). Overall, differences in the costs and effectiveness of tacrolimus compared with topical corticosteroids were very small and the results were sensitive to a number of cost and effectiveness variables.

4.2.1.8 For adults, results from scenario 1 showed that tacrolimus dominated topical corticosteroids in those with moderate or severe atopic eczema. In scenario 2, where the model was run over 51 weeks, tacrolimus produced more DCDs compared with topical corticosteroids but was also more costly; the cost per additional DCD with tacrolimus was £6.20 (moderate eczema) and £26.80 (severe eczema). However, patients with severe atopic eczema treated with ciclosporin achieved more DCDs than did those treated with tacrolimus, although ciclosporin was more expensive (the cost per additional DCD with ciclosporin was £4.80). These patterns were reflected in further analyses excluding work days lost. The results were sensitive to a number of variables.

Pimecrolimus

4.2.1.9 The Assessment Group's base-case results showed that, compared with topical corticosteroids, first-line use of pimecrolimus cost more and conferred slightly fewer QALYs for children and adults with atopic eczema both on the body and in sensitive areas (see Section 4.2.1.1). Using pimecrolimus as a second-line rather than first-line treatment was less expensive but was still dominated by topical corticosteroids. However, the probabilistic analyses showed that these results were associated with considerable uncertainty.

4.2.1.10 The Assessment Group developed subsidiary models (one for adults and one for children) to evaluate the cost effectiveness of pimecrolimus compared with emollients, with moderately potent topical corticosteroids used as a 'rescue therapy' for all patients with uncontrolled 'problem' eczema. The basic structure of these models was the same as that of the models used to compare active treatments, although no distinction was made between eczema on sensitive areas and eczema on the rest of the body. Compared with emollients alone, the ICERs for treatment with pimecrolimus were £9700 per QALY in children and £16,600 per QALY in adults. Again, the results were associated with a high degree of uncertainty.

4.2.1.11 The manufacturer developed a cost–utility model comparing pimecrolimus with emollients. A Markov approach was used, based on four states of progressive severity (remission, mild, moderate and severe). The use of topical corticosteroids was permitted as a rescue therapy in patients with severe disease. The model was run for 1 year. An NHS perspective was adopted and 2003 was the base year for estimating costs.

4.2.1.12 For all parts of the body, the estimated cost per additional QALY was £24,500 in children and £27,400 in adults and, for head and neck only, £4668 in children and £21,766 in adults. The results were sensitive to a number of variables. Probabilistic analyses of the model for children showed that, when £30,000 is considered to be an acceptable amount to pay for an additional QALY, the probability of pimecrolimus being most likely to be cost effective is 60%, the probability that it will dominate vehicle is 20% and the probability that the cost for an additional QALY will be greater than £100,000 is 17%. In patients with mainly head and neck atopic eczema, the probabilities were 75%, 35% and 20%, respectively. Probabilistic analyses were not presented for adults.

4.3 Consideration of the evidence

4.3.1 The Committee reviewed the evidence available on the clinical and cost effectiveness of tacrolimus and pimecrolimus, having considered evidence on the nature of the condition and the value placed on the benefits of tacrolimus and pimecrolimus by people with atopic eczema, those who represent them, and clinical experts. It was also mindful of the need to take account of the effective use of NHS resources.

4.3.2 The Committee was persuaded that tacrolimus and pimecrolimus have both been shown to be effective in treating atopic eczema when compared with vehicle. They noted, however, that vehicle is not the most appropriate comparator in the majority of cases.

4.3.3 The Committee observed that no studies had compared 0.03% tacrolimus in children with moderate to severe disease with potent or moderately potent topical corticosteroids, but that there was evidence from two RCTs of greater effectiveness compared with mild topical corticosteroids. It also noted that the two largest RCTs in adults that compared tacrolimus and potent topical corticosteroid regimens had produced inconsistent results: 0.1% tacrolimus was statistically significantly more effective than topical corticosteroids in the larger study, whereas no statistically significant differences were reported in the smaller study, although potent topical corticosteroids were found to be more effective than 0.03% tacrolimus. The Committee concluded that the available evidence suggested that tacrolimus was similar in effectiveness to topical corticosteroids for the treatment of moderate to severe atopic eczema.

4.3.4 The Committee noted that there is limited evidence on the effectiveness of pimecrolimus compared with topical corticosteroids. No studies had compared pimecrolimus with topical corticosteroids in children, and the results of the studies in adults favoured topical corticosteroids.

4.3.5 Consideration was given to the results of the economic models developed by the Assessment Group and the manufacturers of tacrolimus and pimecrolimus. The Committee acknowledged that the base-case results of the Assessment Group's model were associated with very considerable uncertainty, reflecting the closeness of the utility estimates for all the strategies, coupled with the necessity of estimating a number of parameters in the model in the absence of good-quality evidence, and of the need to combine treatment and disease states in order to address the different potential uses of the new immunomodulators. The Committee concluded that, because of the significant parameter uncertainty, the base-case ICERs could not, in themselves, provide the sole basis for a decision. It was agreed, however, that similar benefits accrue to patients with atopic eczema with tacrolimus as with the use of topical corticosteroids but at a greater cost.

4.3.6 The Committee considered that the Fujisawa economic model of tacrolimus did not provide any evidence to override its conclusion in 4.3.5. The analysis was limited in that results were expressed only in terms of DCDs rather than QALYs; and it only compared tacrolimus treatment after the use of topical corticosteroids with tacrolimus treatment before topical corticosteroids, and not with topical corticosteroids alone.

4.3.7 On the basis of the Assessment Group's economic model of pimecrolimus, the Committee agreed that pimecrolimus was not cost effective compared with topical corticosteroids. Although the Assessment Group's model and the manufacturer's model suggested that pimecrolimus was cost effective relative to emollients alone, the Committee was not convinced that this was the most clinically appropriate comparator in the majority of cases.

4.3.8 The Committee was concerned by the possibility that immunomodulators might increase the risk of skin malignancy in the long term – there is some evidence that tacrolimus may reduce the time to development of ultraviolet-induced tumours in experimental animals, although this was associated with much higher systemic levels of tacrolimus than would be seen in clinical use, and it is known that systemic use of tacrolimus and related drugs is associated with the development of skin cancers. For this reason, the Committee believed that topical immunomodulators should not be used without careful discussion of the potential risks and benefits with the patient.

4.3.9 The Committee heard from the clinical experts that topical corticosteroids provide effective first-line management of atopic eczema and that, because of the higher cost of tacrolimus and pimecrolimus and their potential unknown long-term adverse effects, the experts would not recommend either of these products as first-line treatments.

4.3.10 The Committee heard from the clinical experts that topical immunomodulators are useful in cases where continual use of potent topical corticosteroids is required and there is a serious risk of skin atrophy. The experts also agreed that topical immunomodulators are useful alternatives to systemic treatments in severe resistant atopic eczema.

4.3.11 The Committee heard from the clinical experts that concern about long-term adverse effects, particularly the risk of skin atrophy, causes significant anxiety about the use of topical corticosteroids in people with atopic eczema. The clinical experts were in agreement, however, that anxiety around steroid use should not be an indication for treatment with topical immunomodulators, but that this anxiety should be addressed through effective patient information and education.

4.3.12 Taking into account the evidence on the cost and effectiveness of topical immunomodulators, and the views of the clinical experts, the Committee agreed that topical immunomodulators should not be recommended for mild atopic eczema or as first-line treatments for atopic eczema of any severity. The risk of skin atrophy associated with topical corticosteroid use was relevant only to the moderate and severe end of the disease spectrum and to continual and high-dose use. The Committee, however, did recognise the need for second-line treatments in patients with moderate to severe disease that has not been controlled by topical corticosteroids, where there is a serious risk of important adverse effects from further topical corticosteroid use, particularly irreversible skin atrophy.

4.3.13 Only tacrolimus is licensed for severe atopic eczema, and so consideration was given to the relative benefits of pimecrolimus compared with tacrolimus for the second-line treatment of moderate atopic eczema. The Committee noted the considerably weaker evidence base for pimecrolimus compared with tacrolimus in adults and agreed that pimecrolimus should not therefore be recommended for use in adults.

4.3.14 The Committee further considered the effectiveness of pimecrolimus and tacrolimus in children for whom the only licensed tacrolimus strength is 0.03%. The Committee noted the results of the study that compared pimecrolimus and 0.03% tacrolimus in children, and was persuaded by the views of some of the clinical experts that pimecrolimus is a useful treatment option for moderate facial atopic eczema in children because of its different formulation and tolerability to tacrolimus.

4.3.15 The Committee therefore concluded that tacrolimus should be recommended as an option for the second-line treatment of moderate to severe atopic eczema in adults and children aged 2 years and older in the circumstances outlined in Section 4.3.12. The Committee also concluded that pimecrolimus could be recommended as an option for the second-line treatment of moderate atopic eczema on the face and neck in children aged 2 to 16 years, but that it should not be recommended for use on other parts of the body.

4.3.16 It was agreed that, for the purposes of this guidance, atopic eczema that has not been controlled by topical corticosteroids would refer to disease that has not shown a satisfactory clinical response to adequate use of the maximum strength and potency of topical corticosteroid that is appropriate for the patient's age and the area being treated.

4.3.17 Finally, the Committee considered how and by whom tacrolimus and pimecrolimus should be prescribed. Given the uncertainties around the long-term effects of the immunomodulators, the Committee agreed that it would be appropriate to recommend that they be initiated only by physicians (including general practitioners) with a special interest and experience in dermatology, and only after careful discussion with the patient of the potential risks and benefits of all appropriate second-line treatment options.