We read with great interest the recent paper by Hawk et al. (1), who reported that male pattern baldness was significantly associated with increased risk for prostate cancer. Given that the study relied upon a prospective design, a large sample, and data that were appropriately analyzed, the report makes a substantive contribution to the literature related to potential risk factors for prostatic carcinoma. However, there are some weaknesses in the study. First, as the authors acknowledge, male pattern baldness was defined as any sort of baldness that could be rated as “mild,” “moderate,” or “severe.” Thus, specific data were not collected on the two major types of baldness patterns, i.e., vertex and frontal. Given evidence from previous studies that both androgen levels and androgen receptor status may differ between men displaying these two distinct types of baldness patterns (2, 3, 4), as well as direct evidence that vertex baldness may be associated with increased risk of prostate cancer, whereas frontal baldness may not (5), we suggest that the odds ratios reported by Hawk et al. may be attenuated. Second, the reported odds ratios may be attenuated further given evidence that there may be an optimal period to assess baldness; i.e., by age 30, if indeed it is used to portend risk (5). Finally, the age distribution of the men who were bald is also older than men who were not bald, which also may have lowered the magnitude of the reported odds ratios. In concluding this letter, we applaud the efforts of Hawk et al. and put forth the premise that the association that they found between baldness and prostate cancer may be even stronger, if indeed they had the ability to discriminate between vertex and frontal baldness and if baldness were assessed by age 30 in a higher proportion of their sample. Future studies that assess baldness at younger ages as well as ascertain data on specific baldness patterns are needed to determine whether or not baldness truly is a strong risk factor for prostate cancer.
To whom requests for reprints should be addressed, at National Cancer Institute, Gastrointestinal and Other Cancers Research Group, Executive Plaza North, Suite 201, MSC-7322, 6130 Executive Boulevard, Bethesda, MD 20892.
Reply
We thank Drs. Demark-Wahnefried and Schildkraut for their thoughtful comments about the merits and potential limitations of our study (1). We agree that our inability to specify the pattern (i.e., frontal versus vertex) or rate of male pattern baldness may have attenuated our risk estimates for prostate cancer. As we discussed in the article, the source of data for our study, the Epidemiologic Follow-up Study of the first National Health and Nutrition Survey (NHANES I), did not include these characteristics. Whether our risk estimates were attenuated further by failing to assess baldness at some “optimal” point in time is an interesting question. Like Drs. Demark-Wahnefried and Schildkraut, we were interested in whether men with an earlier onset of baldness might be at greater risk for prostate cancer, but as we reported, we did not find this to be the case. To fully address this question however, additional follow-up of prostate cancer outcomes in the NHANES I cohort will be useful. Indeed, our study included data on the prostate cancer status (as of 1992) for 793 men <35 years of age at the time of entry into the cohort in the early 1970’s. These men are just now reaching the age at which prostate cancer becomes most common. We agree that in addition to further follow-up of the NHANES I cohort, other longitudinal studies with repeated, detailed assessments of baldness starting at younger ages that provide information on the specific patterns and rates of baldness throughout adult life would be helpful to determine whether baldness is, in fact, a strong risk factor for prostate cancer.
