Noora Neittaanmäki-Perttu1,2, Mari Grönroos2, Toni T. Karppinen2, Erna Snellman3 and Pekka Rissanen4

Departments of Dermatology and Allergology, 1Helsinki University and Helsinki University Hospital, Helsinki, 2Päijät-Häme Central Hospital, Päijät-Häme, 3Department of Dermatology, Tampere University and Tampere University Hospital, and 4School of Health Sciences (Health Economics), Tampere University, Tampere, Finland

Daylight-mediated photodynamic therapy (DL-PDT) is considered as effective as conventional PDT using artificial light (light-emitting diode (LED)-PDT) for treatment of actinic keratoses (AK). This randomized prospective non-sponsored study assessed the cost-effectiveness of DL-PDT compared with LED-PDT. Seventy patients with 210 AKs were randomized to DL-PDT or LED-PDT groups. Effectiveness was assessed at 6 months. The costs included societal costs and private costs, including the time patients spent in treatment. Results are presented as incremental cost-effectiveness ratio (ICER). The total costs per patient were significantly lower for DL-PDT (€132) compared with LED-PDT (€170), giving a cost saving of €38 (p = 0.022). The estimated probabilities for patients’ complete response were 0.429 for DL-PDT and 0.686 for LED-PDT; a difference in probability of being healed of 0.257. ICER showed a monetary gain of €147 per unit of effectiveness lost. DL-PDT is less costly and less effective than LED-PDT. In terms of cost-effectiveness analysis, DL-PDT provides lower value for money compared with LED-PDT. Key words: actinic keratoses; cost-effectiveness; daylight-photodynamic therapy.

Accepted Aug 4, 2015; Epub ahead of print Aug 10, 2015

Acta Derm Venereol 2015; XX: XX–XX.

Noora Neittaanmäki-Perttu, Department of Dermatology and Allergology, Helsinki University Hospital, Box 160 Meilahdentie 2, FIN-00029 Helsinki, Finland. E-mail: noora.neittaanmaki@fimnet.fi

Actinic keratoses (AK) are precursors with potential to develop into metastatic squamous cell carcinoma (SCC) (1). High incidence rates of AKs raise a significant economic issue. The prevalence of AK is estimated to be approximately 50% in Australia and 11–34% in the Northern hemisphere, with higher prevalence among elderly subjects (2–4). The total annual direct costs for AKs were 18 million Euros (€) in Sweden in 2011 and 1.2 billion USD in the US in 2004 (5, 6).

There are several topical treatment options for AKs (7, 8). Photodynamic therapy (PDT) is a recommended treatment option (9–11). A novel approach, using daylight (DL-PDT) in the treatment of AKs is considered to be as effective as treatment with conventional PDT using artificial light (light-emitting diode (LED)-PDT) (12, 13). The use of DL-PDT is thought to be less costly due to shorter visiting times at the clinics (13, 14). To our knowledge, the cost-effectiveness of DL-PDT for the treatment of AK has not been studied previously.

This study was a prospective randomized non-sponsored trial to assess the cost-effectiveness of DL-PDT compared with LED-PDT in the treatment of AKs.

MATERIALS AND METHODS

The study protocol followed the Declaration of Helsinki and was approved by the local ethics committee. All participants provided informed consent. For full details, see Appendix S11.

All patients who fulfilled the inclusion criteria were sequentially recruited from the Department of Dermatology, Päijät-Häme Central Hospital between 2011 and 2013. Inclusion criteria included a minimum of 3 clinically clearly detectable AKs on facial or scalp areas. Exclusion criteria are detailed in Appendix S11.

Patients were randomized to receive either DL-PDT or LED-PDT. Before treatments AK lesions were photographed, counted and classified into grades I–III (16).Three clearly detectable target lesions per patient were chosen for the study follow-up. If patients had more than 3 lesions, all lesions were treated according to the randomization. Grade I lesions were treated once and grades II–III lesions twice. The treatment procedure is detailed in Appendix S11.

Response was evaluated clinically (patient complete response, 3 target lesions cleared) at 6 months. Effectiveness was defined by the level as the probability of patient’s complete response at 6 months.

Costs included societal costs (including the working time of the nurse and doctor, treatment room rent, medication and equipment) and patients’ costs (including treatment time and travel costs). The analysis adopted a societal perspective, including both healthcare and patients’ costs (Table I). The costs of control visits or further treatments were not included.

Cost-effectiveness analysis (CEA) was performed using DL-PDT and LED-PDT as the intervention and control treatments, respectively, generating the incremental cost-effectiveness ratio (ICER). For a detailed description of the method, see Appendix S11.

RESULTS

Patients

Of the 73 randomized patients 3 were excluded; one due to an unrelated death, one for diagnosed SCC on the studied area before treatment, and one for severe diffuse photo-damage that hindered the count of AKs (Fig S11).

A total of 70 patients completed the study, 39 men and 31 women, age range 59–93 years (mean 76 years). Eight patients had anamnestic skin photo-type I, 25 photo-type II, 34 photo-type III, and 3 photo-type IV. Forty-six patients had earlier received treatment for their AKs and 3 for carcinoma in situ. Six had had operations for SCC, 18 for basal cell carcinoma (BCC), 1 for melanoma and one for verrucous carcinoma, of which 2 BCCs, one SCC and verrucous carcinoma were in the studied skin areas. Twenty-four patients had previously received cryosurgery, 8 DL-PDT and 11 LED-PDT for AKs on the studied areas. In addition, 5 patients had received LED-PDT, 23 cryosurgery, 1 imiquimod, 1 diclofenac and 1 topical retinoid treatment for AKs in other skin areas. None of the patients had received both DL-PDT and LED-PDT. There was a wash-out period of at least 6 months from previous treatments to the studied areas.

Thirty-five patients received DL-PDT and 35 received LED-PDT. Seven patients in both groups received 2 treatments for thicker lesions 7–23 days apart (mean 12 days) and the rest received one treatment session. The mean time for daylight-exposure was 161 min (range 120–480 min). Patients did not visit or contact the hospital between treatments and control visits.

Three target lesions per patient were included in the study; thus, the number of lesions studied was 210, of whom 105 (92 grade I, 13 grade II–III) were treated with DL-PDT and 105 (93 grade I and 12 grade II–III) with LED-PDT (p = 0.46). Patient complete response rates (3/3 lesions cleared) were 15 of 35 (42.9%) with DL-PDT and 24 of 35 (68.6%) with LED-PDT (p = 0.030) (Fig S11).

To better compare our results with previous DL-PDT studies, we also conducted per lesion clearance analysis. At 6 months LED-PDT cleared 94 of 105 (89.2%) and DL-PDT 76 of 105 (72.4%) lesions (p = 0.0025). In the LED-PDT group 93 grade I AKs receiving 1 and 12 grade II–III AKs 2 treatments showed 88.2% and 100% complete clearance, respectively (p = 0.36). In the DL-PDT group 92 grade I AKs received 1 treatment and 13 grade II–III AKs 2 treatments with 73.9% and 61.5% clearance rates (p = 0.34).

Costs

The imputed mean total costs per patient (1–2 treatments) were €132 (95% confidence interval (CI) 111.3–152.6) for DL-PDT and €170 (95% CI 126.0–213.5) for LED-PDT, resulting in incremental costs savings of –€38 (p = 0.022) (Table I).

Cost-effectiveness

The probabilities for patients’ complete response were 0.429 (95% CI 0.414–0.443) for DL-PDT and 0.686 (95% CI 0.674–0.698) for LED-PDT, thus yielding a loss in the probability of being healed of 0.257. The ICER revealed a monetary gain of €147 per unit of effectiveness lost (Table II). The cost-effectiveness (CE)-plane showed that approximately 96% of the bootstrapped replica data yielded a result that lies in the south-western part of the plane, indicating that DL-PDT provided a lower value for money compared with LED-PDT (Fig S21).

Additional data analysis

When dividing the mean cost per patient with the estimated probability for patient’s complete response, the costs per complete responders were calculated as €308 for DL-PDT and €248 for LED-PDT (p = 0.004).

DL-PDT required significantly less of the nurses’ time (median 51 vs. 78 min per treatment, p = 0.003) and less of the patients’ treatment time (median 194.5 vs. 271 min, p < 0.0001) compared with LED-PDT.

As measured with the visual analogue scale (VAS, range 0–10) during and after the treatments until the pain had vanished, DL-PDT was significantly less painful, with a mean maximal pain value of 1.53 (range 0.1–6.0) compared with LED-PDT 4.36 (range 0.3–8.4), p < 0.001).

DISCUSSION

The results of this study show that DL-PDT is less costly, but also less effective, than LED-PDT. DL-PDT was associated with an incremental cost saving of €147 and a decremental probability of being healed of –0.257. Thus, DL-PDT provides a lower value for money compared with LED-PDT.

The cost-effectiveness of LED-PDT compared with other treatments for AK has been evaluated in several studies. A limitation of many of these studies compared with our prospective study using accurate costs is that they use estimated costs; some of the studies included only the cost of the topical drug. To our knowledge there have been no cost-effectiveness evaluations of DL-PDT. The simulated costs per complete responder were €379 for MAL-PDT and €363 for cryotherapy, including the cost of yearly re-treatments, and valuing the cosmetic outcome. The incremental cost per extra complete responder was €401, with MAL-PDT being more expensive (18). When LED-PDT was compared with imiquimod using a decision-tree model estimating quality-adjusted life years (QALYs) and treatment costs, it was implied that imiquimod might be more cost-effective (19). A recent prospective head-to-head study found that MAL-PDT was more cost-effective compared with diclofenac + hyaluronic acid gel (DHA), with the costs per complete responder being €566.7 and €1,026.2 for MAL-PDT and €595.2 and €2,295.2 for DHA at 3 and 12 months, respectively (20). Our study showed that the costs per complete responder were €308 for DL-PDT and €248 for LED-PDT at 6 months. A limitation of our study is that we did not include the costs of further treatments or the quality of life analysis (QALY), and thus the numbers are not directly comparable with previous data.

PDT studies rarely report response rates at the patient level. In the Italian study, the patient complete response rate for LED-PDT was 68.4% at 3 months and 55.2% at 1 year (19). In another study, MAL-PDT was superior to a placebo, with a patient complete response of 59.2% vs. 14.9% at 3 months (21). Our results are in concordance with these findings, showing a 68.6% patient complete response rate for LED-PDT at 6 months. Our patient complete response rate for DL-PDT was 42.9%. To our knowledge, the patient complete response for DL-PDT has not been reported earlier. As patient complete response is a major indicator of the need of further treatments and further costs, future research should focus more attention on this subject.

The majority of PDT studies report lesion clearance rates ranging from 71% to 92% for LED-PDT, and from 75.9% to 79.5% for DL-PDT (10, 13, 14, 22, 23). Our results show equal clearance rates for LED-PDT (89.2%), but slightly lower clearance rates for DL-PDT (72.4%). The slightly lower efficacy rates for DL-PDT could be explained by a longer follow-up period (6 months) than in the previous studies reporting 3-month clearance rates.

A few facts might have affected our DL-PDT effectiveness results2.

2Meteorological studies have shown that DL-PDT can be implemented in Reykjavik (located in the same latitude (64°N) with Finland) until the middle of September (24). In the first 2 years we continued treatments until early October, which influenced the response rates. Of the 14 patients treated in late September to early October only 16.7% (1/6) in the DL-PDT group and 87.5% (7/8) in the LED-PDT group were completely cleared. Thus, the treatment should be limited to the summer months in northern countries. However, the most effective light-dose needed for DL-PDT remains unclear. We did not use light-dosimeters for DL-PDT and thus accurate light doses are not available. We used a different sunscreen (ACO Sun Kids High Protection Sun Spray® SPF 30, ACO) from that used in previous studies (14). However, the absorption spectrum only minimally overlapped the blue light region and thus was assumed not to affect treatment outcome. A 0.25-mm thick layer of MAL cream has been approved as sufficient for DL-PDT, resulting in 74% lesion complete clearance (25). To our knowledge the efficacy of <1-mm layer photosensitizer has not been studied for LED-PDT. Thus, in our study, the DL-PDT group received a thinner layer of MAL cream (0.25 mm) than the LED-PDT group (1 mm), which reduced the costs for DL-PDT, but also may have affected the response rate. In LED-PDT we used the standardized protocol of a 1-mm thick layer of the photosensitizer under 3 h occlusion, while no occlusion was used in DL-PDT. This results in higher amounts of protoporphyrin IX (PpIX) in the tissues of the LED-group than in the DL-group; in the latter PpIX is activated while developing (13). A larger trial evaluating different thicknesses of MAL cream in LED-PDT is warranted.

The present study found DL-PDT to be significantly less painful compared with LED-PDT. A limitation is that this finding and patient preference were not included in assessment of QALYs and the CEA analysis. LED-PDT might not be an attractive option for patients, as the use of DL-PDT significantly lowers pain and time spent at the clinic. An earlier split-face study comparing LED-PDT and DL-PDT reported higher patient preference for DL-PDT (13). Further limitations to our study include the lack of investigator-blinded outcome evaluation and assessment of adverse reactions. We did not perform skin biopsies to verify AK diagnoses, and despite the high accuracy of clinical diagnoses this should be considered a limitation (26). To simplify the treatment process for cost evaluation we used a treatment method targeted to lesions. Had we targeted the whole field, the cost of the treatment would have been higher and this may have affected the CEA analysis. The results may not be easily generalized to the working age population as it focused on pensioners.

Despite its higher efficacy, LED-PDT might not be an attractive option for patients, as the use of DL-PDT results in significantly less pain and time spent at the clinic. Thus, we still prefer the use of DL-PDT over the conventional treatment during the summer months. With its short visit time, DL-PDT can be implemented in private practices, which could reduce the burden on public clinics. A further CEA analysis of DL-PDT should be conducted including QALYs and patient preference. Furthermore, the cost-effectiveness of newer low-concentration photosensitizers in DL-PDT needs to be studied (25, 27).

In conclusion, this study assessing the detailed costs of DL-PDT and LED-PDT for treatment of AKs found that, in terms of CEA, DL-PDT provides a lower value for money compared with LED-PDT.

ACKNOWLEDGEMENTS

We would like to thank research nurse Ulla Oesch-Lääveri from Päijät-Häme Central Hospital for her dedication to this study. The corresponding author received a research grant from Orion Pharmos Foundation (not involved in any of the products used in the study) and from Foundation for Clinical Chemistry Research.

The authors declare no conflicts of interest.

1http://www.medicaljournals.se/acta/content/?doi=10.2340/00015555-2205

REFERENCES