Correspondence
Melanoma after PUVA Therapy for Psoriasis
N Engl J Med 1997; 337:502-503August 14, 1997
- Article
To the Editor:
The report by Stern et al. (April 10 issue)1 of a cohort of 1380 patients with psoriasis who were treated with oral methoxsalen (psoralen) and ultraviolet A radiation (PUVA) suggests that PUVA therapy is associated with a higher incidence of melanoma than expected, but this finding may be due to inaccurate statistics on melanoma, confounding variables, and surveillance bias.
Their estimate of relative risk compared the observed number of melanomas in the PUVA cohort over a 20-year period with the expected number, based on cancer statistics from the Surveillance, Epidemiology, and End Results (SEER) data. These data estimate incidence rates of cancer from tumor registries of hospitals and “searches of private laboratory records” in designated regions of the United States, thereby providing a sample of approximately 9.6 percent of the population.2 These data may grossly underestimate the true incidence of melanoma, because melanoma is frequently excised in doctors' offices and specimens may be sent to distant laboratories. A survey of the membership of the American Academy of Dermatology estimated that in 1992 the incidence of melanoma and melanoma in situ was double the incidence of invasive melanoma estimated by the SEER program. Even that survey may have underestimated the incidence of melanoma, because it excluded nondermatologists.3
Another concern about this cohort study is that it lacked a control group of patients with psoriasis who never received PUVA. It cannot be determined whether one or more confounding variables, including exposure to sunlight, a history of sunburn (particularly before adulthood),4 previous phototherapy with ultraviolet B, coal-tar therapy, systemic methotrexate therapy, and other, unknown endogenous or exogenous factors, may partly or completely account for the association found.
Surveillance bias may have also affected the results. The study cohort was made aware that they were being followed specifically for possible adverse effects of PUVA therapy, among them the development of skin cancer. Unlike this cohort, the general population does not have the same heightened concern about possible growths and skin changes that may represent skin cancer and is not followed by physical examination, telephone calls, or correspondence about skin cancer. The cohort surveillance may have led to the diagnosis of melanomas that might have been missed in the general population — for example, melanomas with periods of latency before metastasis that surpass the patient's lifetime,5 melanomas that may regress spontaneously,5 and misdiagnosed “atypical pigmented skin tumors of limited or nonexistent potential for malignant behavior.”6
6 ReferencesS. Elizabeth Whitmore, M.D.
Warwick L. Morison, M.D.
Johns Hopkins University School of Medicine, Baltimore, MD 212051
Stern RS ,Nichols KT ,Vakeva LH . Malignant melanoma in patients treated for psoriasis with methoxsalen (psoralen) and ultraviolet A radiation (PUVA). N Engl J Med 1997;336:1041-1045
Full Text | Web of Science | Medline2
Miller BA, Ries LAG, Hankey BF, eds. SEER cancer statistics review, 1973–1990. Bethesda, Md.: National Cancer Institute, 1993:I.1-I.2. (NIH publication no. 93-2789.)
3
Rigel DS ,Friedman RJ ,Kopf AW . The incidence of malignant melanoma in the United States: issues as we approach the 21st century. J Am Acad Dermatol 1996;34:839-847
CrossRef | Web of Science | Medline4
Koh HK . Cutaneous melanoma. N Engl J Med 1991;325:171-182
Full Text | Web of Science | Medline5
Truchetet F ,Heid E ,Friedel J ,Chartier C ,Grosshans E . D.N.C.B. for malignant melanoma: significance in the treatment strategy. Anticancer Res 1989;9:1531-1536
Web of Science | Medline6
Swerlick RA ,Chen S . The melanoma epidemic: is increased surveillance the solution or the problem? Arch Dermatol 1996;132:881-884
CrossRef | Web of Science | Medline
Author/Editor Response
Dr. Stern replies:
To the Editor: Whitmore and Morison claim that our finding of a dose-related increase in the risk of melanoma that begins 15 years after the start of PUVA therapy is attributable to the SEER data we used, to a higher risk of melanoma in patients with psoriasis, and to surveillance bias. They refer to incidence estimates from a survey with a 35 percent response rate that relied on practitioners' recollections of the number of cases of melanoma seen in the previous decade.1 That survey included melanoma in situ (we excluded from our analyses the five cases of this that were found in our cohort). The SEER data on the incidence of cancer are the national standard, cited more than 250 times since 1988.
As we noted, from 1975 to 1990 the observed number of melanomas in our cohort and the number expected on the basis of the SEER data were nearly identical. Except for exposure to PUVA and the passage of time, there is no other apparent explanation for the dose-dependent increase in risk after 1990. A large case–control study found no association between psoriasis and the risk of melanoma.2 Because the surveillance was more intense during the first 15 years of our study (when there were periodic sponsored physical examinations) than after 1990 (when there were no such examinations), any surveillance bias would almost certainly have biased our results toward a lower risk estimate in the later period of surveillance than in the earlier period — the opposite of what we found.
In February 1996, the PUVA investigators discussed the melanoma experience of the cohort as of that date (10 melanomas in 8 patients in 20 years). The group's consensus was that if there was an additional case in two years, especially in a patient who received a high dose of PUVA, the association between a high level of exposure to PUVA and an increased risk of melanoma would be strongly confirmed. In the next 14 months, we detected malignant melanoma in three additional patients in the cohort — Patient 9, a patient in whom melanoma was detected after our manuscript was accepted, and another in whom melanoma was detected after the publication of the paper. This incidence equaled that observed during the first 14 years of the study. Two of the three patients had received more than 450 PUVA treatments. When the two additional patients not counted in our published risk estimates were included in the analysis, the risk of melanoma associated with long-term PUVA therapy was greater and more significant.
Many members of our cohort have had high levels of exposure to PUVA and many other treatments for psoriasis. A variety of other exposures in concert with PUVA may increase the risk of melanoma. However, there was no substantial difference between the patients with melanoma and the other members of the cohort in the distribution of phenotypic characteristics or high-dose exposure to suspected carcinogens other than PUVA.
2 ReferencesRobert S. Stern, M.D.
Beth Israel Deaconess Medical Center, Boston, MA 022151
Salopek TG ,Marghoob AA ,Slade JM , et al. An estimate of the incidence of malignant melanoma in the United States: based on a survey of members of the American Academy of Dermatology. Dermatol Surg 1995;21:301-305
CrossRef | Web of Science | Medline2
Elwood JM ,Gallagher RP ,Stapleton PJ . No association between malignant melanoma and acne or psoriasis: results from the Western Canada Melanoma Study. Br J Dermatol 1986;115:573-576
CrossRef | Web of Science | Medline
- Citing Articles (3)
Citing Articles
1
Stacy L. McClure, Jayme Valentine, Kenneth B. Gordon. (2002) Comparative Tolerability of Systemic Treatments for Plaque-Type Psoriasis. Drug Safety 25:13, 913-927
CrossRef2
LindelÖF, Sigurgeirsson, Tegner, LarkÖ, Johannesson, Berne, Ljunggren, Andersson, Molin, Nylander-Lundqvist, Emtestam. (1999) PUVA and cancer risk: the Swedish follow-up study. British Journal of Dermatology 141:1, 108-112
CrossRef3
Wolf, SchOllnast, Hofer, Smolle, Kerl. (1998) Malignant melanoma after psoralen and ultraviolet A (PUVA) therapy. British Journal of Dermatology 138:6, 1100-1101
CrossRef
