THE "NEW" PROSTATE CANCER INFOLINK

A new paper in the International Journal of Epidemiology makes very little sense to us … but then we never claimed to be the smartest chickens in the coop!

Eisenberg et al. have used data from the National Institutes of Health – American Association of Retired Persons Diet and Health Study to investigate whether there is any appearance of an association between fatherhood status and risk for prostate cancer. They apparently aren’t the first to look at this question, and they themselves state clearly that the prior evidence “is limited and contradictory.”

In 1995 the Diet and Health Study enrolled a total of 161,823 men aged between 50 and 71 years. At baseline, none of these men had any evidence of any form of cancer.

The researchers were therefore able, prospectively, to track the occurrence of prostate cancer in this cohort of men and to assemble data on the occurrence of a diagnosis of prostate cancer and the number and sex of the children of the men who did and didn’t have a diagnosis.

We hasten to point out that we have not seen the full research report — only the abstract. The abstract reports the following findings:

• Data was analyzed from 8,134 men with a diagnosis of prostate cancer (but it is not clear that these 8,134 men included all of the cases of prostate cancer within the original study cohort of 160,000+ men).
• There was no identifiable relationship between fatherhood status and the overall incidence of prostate cancer (hazard ratio [HR] = 0.94).
• After stratifying the men into groups based on whether they did or did not get PSA tests … 
  • Untested and childless men had a significantly lower risk for prostate cancer (HR = 0.73) than fathers (“due to the interaction between PSA screening and fatherhood”).
  • Untested fathers had a lower risk  for prostate cancer than childless men (for low-grade prostate cancer, HR = 0.78; for high-grade prostate cancer, HR = 0.62; and for fatal prostate cancer, HR = 0.28), but this was not a statistically significant result — only a trend.
• There was no evidence of a relationship between the number of children fathered and risk for prostate cancer.
• There was a weak association between having no female children and increased risk for prostate cancer among men who had not had PSA tests for risk of prostate cancer.

The authors manage to conclude from this study that, “Our findings suggest fatherhood status and offspring gender is associated with a man’s prostate cancer risk.”

The “New” Prostate Cancer InfoLink would respectfully suggest that these data are rather less than compelling. The study appears to be rife with assumptions needed to conduct the statistical analyses, and the validity of some of those assumptions may well be open to question. There is the implication of evidence that fatherhood was associated with a greater likelihood of PSA testing, but the significance of that finding to actual risk for prostate cancer seems to be unexplored. There is also an implication in the abstract that “socio-demographic and lifestyle characteristics” were “accounted for” in some manner, but it is not clear what the authors mean by this.

Is is possible that being a father or not having female children significantly increases or decreases one’s risk for prostate cancer? Sure … it’s possible! Is it likely to be one of the important factors driving that risk? Frankly, we doubt it — by comparison with the other known risk factors. Our real question here would be whether an analysis of this type represents good use of money in the search for strategies to prevent prostate cancer. Something tells us that men are unlikely to stop fathering children because of a very small increase in their risk for prostate cancer 30 to 50 years later … always assuming that such a risk could really be documented with some really compelling data. In our humble opinion, the evidence for any association between fatherhood and prostate cancer risk remains “limited and contradictory” at best.

Filed under: Diagnosis, Prevention, Risk | Tagged: fatherhood, risk |