Dibenzoyl peroxide

Epidemiological data

Administrative data

Endpoint:

epidemiological data

Adequacy of study:

supporting study

Reliability:

2 (reliable with restrictions)

Data source

Reference

Reference Type:

review article or handbook

Title:

BENZOYL PEROXIDE: REVIEW OF EXPERIMENTAL CARCINOGENESIS AND HUMAN SAFETY DATA

Author:

Binder RL, Aardema MJ and Thompson ED

Year:

1995

Bibliographic source:

Growth factors and tumor promotion: Implications for Risk Assessment, pages 245-294, Wiley-Liss, Inc

Materials and methods

Study type:

other: Review

Endpoint addressed:

carcinogenicity

Test material

Test material information

Constituent 1

Reference substance name:

Dibenzoyl peroxide

EC Number:

202-327-6

EC Name:

Dibenzoyl peroxide

Cas Number:

94-36-0

Molecular formula:

C14H10O4

IUPAC Name:

diphenylperoxyanhydride

Method

Type of population:

general

Results and discussion

Results:

The epidemiology data, including a large case-control study, provide evidence that the use of BPO to treat acne does not increase the risk of skin cancer.

Applicant's summary and conclusion

Conclusions:

The epidemiology data, including a large case-control study, provide evidence that the use of BPO to treat acne does not increase the risk of skin cancer.

Executive summary:

The reports that BPO bas tumor promoting activity in animal models have lead to epidemiology studies in exposed populations. Wright, et al. (1983) conducted a small hypothesis generating study using data from the Los Angeles County Cancer Surveillance program. They reported that chemists with malignant melanoma were younger, better educated and exposed to more organic chemicals, pesticides, ionizing radiation and BPO than chemists with other cancers. While this study was too small to be subjected to formal statistical analysis, the authors concluded that the hypothesis that chemical exposures increase the risk of malignant melanoma is plausible. Because of the very small size of this study, and the fact that cases were exposed to many different agents, no conclusions regarding BPO can reached from the reported results.

In Sweden BPO-containing acne medications were introduced in 1977. Based on data in the Swedish Cancer Registry, Lidén, et al. (1990) noted that there was no increase in either melanoma or nonmelanoma skin cancers of the face (excluding basal cell carcinomas), in the first 8 - 9 years since the introduction of BPO. The data evaluation was limited on a yearly basis to the age group 10-29, who are the primary users of BPO. However, very few skin cancers occur in this group.

Cartwright, et al. (1988) examined the potential association between the use of benzoyl peroxide to treat acne and malignant melanoma. They conducted a pilot case­ control study with 159 cases and 213 age- and sex-matched controls. The cases were identified from a regional cancer registry in the United Kingdom, and controls were identified by the physicians who treated the melanoma cases. Cases and controls clinically diagnosed with acne and prescribed BPO were identified from physician records. No significant associations were found between the prescribed use of BPO to treat acne and malignant melanoma, or between the occurrence of acne and malignant melanoma. Limitations of this study were the small size, and the fact that patients were not contacted to determine if they used nonprescription acne medications containing BPO.

A much larger case-control study to evaluate acne treatments as risk factors for skin cancer of the head and neck was conducted in Saskatchewan (Hogan, et al., 1990, 1991; To, et al., 1991). The cases were identified from the Saskatchewan Cancer Registry, which records essentially 100% of all melanoma and nonmelanoma skin cancers that occur in the Province. In the U.S. most nonmelanoma skin cancers are not registered because they often are treated in physician offices. Wbereas, in Saskatchewan there are legal and monetary incentives for physicians to register nonmelanoma skin cancers. Age­ and sex-matched controls were identified from the Saskatchewan Medicare Plan.

There were two phases to the Saskatchewan study. Phase 1 linked computer records from the Saskatchewan Cancer Registry and Medicare and Drug Prescription Plans to evaluate whether the prescribed use of BPO to treat acne was associated with increased risk of skin cancer. Phase 2 was a population-based case-control study in which a questionnaire was mailed to cases and controls. No statistically significant association between the prescribed use of BPO and skin cancer was found in Phase 1, which involved 931 cases and 7630 controls (Hogan, et al., 1990; To, et al., 1991). The overall odds ratio was 1.27 (with a 95% confidence interval of 0.69-2.36), which was not significantly different from 1.0, a value indicating no increased risk of cancer. Limitations on of Phase 1 were that the nonprescription use of BPO by cases and controls was not considered, and the Drug Prescription Plan records did not cover the full period during which BPO­ containing medications were available. However, Phase 2 covered all prescription and nonprescription use of BPO-containing acne medications.

The Phase 2 analysis involved 964 cases and 3856 controls, which were initially eligible (Hogan, et al., 1991). The population studied was restricted to an age group that may have used BPO from the time it came on the market in Canada in 1966 until 1988. Therefore, the maximum possible period from the start of use to the diagnosis of skin cancer was 22 years. The minimum period was less than one year. A detailed questionnaire was used to survey for potential risk factors for skin cancer, and to aid recall the survey listed 33 widely used acne medications. The response rate to the questionnaire was high and about 90% of all patients, in the appropriate age group, with skin cancer of the head and neck were studied. Of the cases studied, 92.3% were basal cell carcinomas, 4.8% were squamous cell carcinomas and the remainder were melanomas.

In Phase 2 no statistically significant association was found between the use of acne medications containing BPO and skin cancer of the head and neck. Acne also was not associated with increased risk of skin cancer. The odds ratio for use of any BPO­ containing acne medication by multiple regression analysis was 0.8 (95% confidence interval, 0.48-1.33). The study was designed to have the power to detect an odds ratio of 1.5 or greater, but because of the excellent questionnaire response rate, an odds ratio >1.33 would have been statistically significant. Importantly, this epidemiology study was sufficiently long and adequately sensitive to detect major known risk factors for skin cancer including family history of cancer, light skin color, sun burning easily and Celtic descent, which ail had highly significant odds ratios. If acne medications containing BPO did have human skin tumor promoting activity, or enhanced the progression of preneoplastic lesions, they would be expected to decrease the latent period for tumorigenesis, or induce tumors in individuals exposed to subtumorigenic levels of UVR or other carcinogens. Therefore, a study sufficiently long to detect other risk factors would be expected to detect an agent that had tumor promoting activity. In support of this conclusion, the univariate odds ratio for use of any acne medication containing BPO was 0.84 (Hogan, et al., 1990), compared to 0.8 for the multiple regression odds ratio, noted above. Since the multiple regression analysis controls for other risk factors, the small difference between these values provides evidence that BPO did not exacerbate the risk associated with the presence of identified skin cancer risk factors, such as fair skin or tendency to sun bum easily. In summary, this study provides evidence that the use of BPO to treat acne does not increase the risk of skin cancer of the head and neck through any mechanism, including complete carcinogenesis, tumor promotion or enhanced progression.