Background: Systemic methoxsalen PUVA increases nonmelanoma skin cancer risk in a dose-dependent manner, whereas trioxsalen bath PUVA treatment has been suggested to be less carcinogenic.

Objective: Our purpose was to study the carcinogenicity of topical trioxsalen PUVA.

Methods: We performed a record linkage study of 337 male and 190 female patients with psoriasis treated with trioxsalen bath PUVA during the period 1977 to 1988 and the Finnish Cancer Registry (cancer incidence in the period 1977 to 1993). The mean follow-up period per person was approximately 11 years. Data on the total cumulative UVA dose and other potentially carcinogenic treatments were collected from the patients' files. The standardized incidence ratio (SIR) was calculated, in which the expected number of cases was based on the national cancer incidence rates.

Results: During the follow-up, 26 cancer cases were observed in the cohort versus 30 expected (SIR, 0.88; 95% confidence interval [CI], 0.57–1.28). The only primary sites showing high SIRs were cancer of the kidney (SIR, 3.56; 95% CI, 0.97-9.10) and non-Hodgkin's lymphoma (SIR, 2.94; 95% CI, 0.36–10.6). There was only one case of nonmelanoma skin cancer; the expected number was 0.8 (SIR, 1.26; 95% CI, 0.03–7.04). The average cumulative UVA dose was 65 J/cm²; 40 patients had received more than 200 J/cm². The average number of treatments was 112; 65 patients had received more than 200 treatments.

Conclusion: No excess of squamous cell carcinoma of the skin was found in patients treated with trioxsalen bath PUVA. However, because of the small size of the cohort, only a sevenfold excess risk can be excluded. The possible associations between psoriasis or its treatment and kidney cancer and lymphoma need to be studied further in larger series.