Why is the prostate cancer death rate declining in the united states?
Article first published online: 31 OCT 2000
Copyright © 1998 American Cancer Society
Volume 82, Issue 2, pages 249–251, 15 January 1998
How to Cite
Mettlin, C. J. and Murphy, G. P. (1998), Why is the prostate cancer death rate declining in the united states?. Cancer, 82: 249–251. doi: 10.1002/(SICI)1097-0142(19980115)82:2<249::AID-CNCR1>3.0.CO;2-M
- Issue published online: 31 OCT 2000
- Article first published online: 31 OCT 2000
- Manuscript Accepted: 2 SEP 1997
- Manuscript Received: 25 AUG 1997
The National Cancer Institute recently reported data showing that the prostate cancer death rate in the United States declined between 1990 and 1995. The overall decline was from 26.5 to 17.3 deaths per 100,000 men in the population.1 Figure 1 (4K) shows that the percentage of decline was greatest for younger white men, for whom the decline was 11.7%, and smallest for older men and African American men.2 This downward trend represented a sharp break from the prior pattern of increasing mortality. The overall prostate cancer mortality rate in the United States had increased 13.2% during the 5-year interval preceding 1990.
There were no published predictions that this would occur, and no analyses of other data that might explain the reasons for this trend have been reported. Knowing that to "look a gift horse in the mouth" can tarnish the event, it is tempting not to raise questions about the meaning of this trend. On the other hand, shifts in patterns of disease in populations across time are the results of experiments of nature from which disease control professionals can measure the effectiveness of preceding health care and disease control interventions. Once learned, the lessons of past experience can guide efforts that extend and accelerate trends that previously had occurred only fortuitously.
It is possible that the recent declines in prostate cancer mortality are related to shifts in detection and treatment dating back to the 1970s. If true, this raises important questions about the impact of the even greater shifts in detection and treatment that occurred after the widespread introduction of PSA screening in the late 1980s. Skeptical observers have cautioned that an increase in early detection without an ensuing decline in mortality should cast doubt on the effectiveness of early detection as a prostate cancer control strategy.3 Contrariwise, does the recent favorable trend in mortality now make the case for proponents of PSA screening? In other words, does declining prostate cancer mortality argue for increasing efforts to identify men in the population who have treatable prostate cancer and offer them the appropriate treatment choices?
Unfortunately, it is easier to speculate about the meaning of prostate cancer death rate trends than to determine their cause. Historical trends cannot be studied by experimental or even quasi-experimental methods, and the death rate trend for any disease is usually the product of many different factors occurring simultaneously. Prominent among the variables that can influence mortality trends are changes in patterns of disease occurrence, developments in disease specific methods of detection and diagnosis, and shifts in treatment practice. All of these changes have been documented in recent years as pertaining to prostate cancer. In addition to these obvious factors are more perplexing ones that can give the appearance of significant progress when little has actually occurred. Such factors as shifts in competing causes of death or changes in standards of recording cause of death are examples of artifacts that can confound understanding of the underlying trends.
Although the number of uncontrolled variables and the historical nature of the event make it impossible to prove any particular theory of causation, it is possible to examine the sequence and magnitude of events that may be used to evaluate what explanations are plausible and reasonable. For example, prevention of prostate cancer can fairly readily be ruled out as an explanation for declining prostate cancer mortality. Although significant research progress has been made concerning the environmental and genetic etiology of prostate cancer, the cause(s) of the disease remain largely unknown. Unlike lung cancer trends, in which reductions in cigarette smoking can be correlated to declines in deaths from the disease, no preventive intervention can be pointed to as a factor in prostate cancer trends. In fact, the decline in prostate cancer mortality has occurred in the face of rising incidence rates.4
It also seems implausible that the observed trends can be attributed to general change in the accuracy or procedures for recording the cause of death. First, the change in direction in mortality does not date back to some prior era when determination of cause of death for a man having prostate cancer might have been ambiguous or death certification standards were not highly developed. Prostate cancer mortality was rising immediately preceding the recent decline, and modern standards for recording cause of death are likely to have been applied equally to both intervals. Secondly, as the data from the National Cancer Institute show, the decline in the death rate is more pronounced for younger men and white men. Any explanation based on the premise that the decline is the result of an artifact of changes in death certification would require a subsidiary explanation for this to occur to a greater extent in some age and race groups than in others. Finally, and possibly most compelling, is the finding that no specific revision of coding procedures pertaining to prostate cancer as a cause of death was introduced at the time the death rate began to change.
Another possibility to consider is that there has been some breakthrough in treatment that has rendered prostate cancer more curable. There is ample precedent for this in oncology. Childhood leukemia, testicular cancer, and Hodgkin's disease are well-known examples of diseases for which new treatments or new combinations of treatments have resulted in longterm, populationwide declines in the death rates. Unfortunately, there have been no comparable prostate cancer treatment advances that would similarly impact on mortality. Although the major treatment options of radical prostatectomy and radiation therapy for localized disease are continually being refined, they have been in use for many years, and there are no curative treatments for advanced prostate cancer.
Although the basic treatment options may not have changed much, there is considerable evidence that the pattern of use of these treatments has changed. The repeated studies of the American College of Surgeons document that the proportion of prostate cancer patients receiving radical prostatectomy or radiation has increased.5 In 1974, 9.2% of all prostate cancer patients were treated by radical prostatectomy, and by 1993 this proportion had increased to 29.2%. The comparable increase for radiation therapy for the period 1974-1993 was from 5.5% to 30.1%. This increased use of potentially curative treatment (as opposed to observation or palliation) happened in concert with a longterm trend toward earlier prostate cancer detection. Between 1973 and 1993, the proportion of prostate cancer diagnosed at localized stages increased from 56.7% to 74.0%. The public health consequences of increasing treatment rates for prostate cancer may be reflected in the results of a recent analysis of national SEER data. Those results showed that the disease specific mortality rate for men treated for prostate cancer (i.e., by surgery or radiation) declined significantly between 1973 and 1990.6
These observations may support the hypothesis that prostate cancer death rates are declining because men are increasingly diagnosed when the prostate cancer is localized and increasingly receive treatments that reduce the likelihood of death from the disease. Although not rising to the level of proof, this hypothesis is consistent with a sequence of changes in patterns of detection and treatment that is well documented in the literature. Further research is needed to quantify the association between early detection and declining mortality. Further research also is needed on competing hypotheses. Finally, continuing close monitoring of the mortality trend for prostate cancer in the United States is needed to establish that the downward trend is not transient and is responsive to continuing changes in the cancer control interventions applied to this disease.
As this needed research goes forward, it seems reasonable to accept the recent decline in prostate cancer mortality at face value. It is a favorable trend that probably reflects the impact of increasing emphasis on early detection and appropriate treatment. The American Cancer Society recently reissued recommendations that men at risk for prostate cancer be offered prostate specific antigen testing and digital rectal examination as part of their routine healthcare.7 Based on the recent evidence, this appears to be a good recommendation for promoting the public's health.
- 1Surveillance, Epidemiology, and End Results (SEER) Program. Age-adjusted U.S. cancer death rates. J Natl Cancer Inst 1997; 89: 12.
- 2Office of Cancer Communications. Questions and answers on trends in cancer mortality [press release, November 15, 1996]. Bethesda, MD: National Cancer Institute.