“Doc, I want to live forever”. This is not an uncommon thought that we hear frequently from older men. It is not easy to accept the concept of our own mortality. Mankind has perpetually been looking for the fountain of youth. For many, they believe we have found it in modern medicine—antibiotics, vaccines, aspirin, statins, etc. Collectively “modern medicine” has successfully delayed death such that a male child born in the United States today can expect to live into his mid-70s. Indeed, quips like “80 is the new 60” are all too common in the media. So how does this quest to live forever play out when dealing with older men diagnosed with prostate cancer?
Chamie and colleagues in this issue of the journal reviewed the records of 1,031 veterans diagnosed with non-metastatic prostate cancer between 1997 and 2004 at two academically-affiliated VA hospitals in Los Angeles County, California.1 They found that men with comorbidities, with the exception of moderate–severe chronic obstructive pulmonary disease (COPD), were equally likely to receive prostate cancer treatment as men without any comorbidities. Despite similar treatment rates, non-cancer survival differed dramatically with, as expected, men with comorbidities having worse overall survival. In other words, men with a shorter non-prostate cancer survival because of their other medical illnesses were receiving cancer treatment at equal rates as “healthy” men.
It has been relatively well documented that prostate cancer screening is poorly targeted according to the health status of patients.2 For example, Walter et al.2 using VA data from 2003 (which would have overlapped with the data in the Chamie et al. study1), found that among men aged 85 years or older 34% in the best health had a PSA test vs. 36% in the worst health. In a follow-up study published in this journal, So et al.3 found that within each VA medical center, the correlation coefficient between PSA screening among healthy and less healthy men was extremely high (0.90). In essence, centers as a whole either screened many men or few men, but “healthiness” was not a major factor in deciding who to screen. While some might argue that screening itself poses little harm and therefore “healthiness” should not factor in, it is clear that prostate cancer treatments can have major side effects. Thus, it often assumed that while screening may be somewhat indiscriminate, decisions regarding treatment should be highly tailored to the patient’s health.
Indeed, as noted by Chamie et al.,1 several studies have shown that health status of the patient does influence treatment decisions (references 11–15 in Chamie et al.1). However, others have shown that worsening comorbidity (up to a Charlson score of 2) did not predict treatment patterns (reference 16–18 in Chamie et al.1). Given this mixed literature, what do the current findings tell us? While the numbers in certain comorbidity subsets are modest, they suggest that on the whole, the presence of a single Charlson comorbidity does not influence treatment decisions within two VA hospitals in Southern California. While the true severity of each comorbidity is not captured by the Charlson Comorbidity Index, all comorbidities in this index were originally selected because they increase mortality. Yet, these data suggest physicians may not consider many of these comorbidities to be significant since they had little influence over treatment decisions. However, treatment decisions were very heavily influenced by age.
The first question these findings raise is, “are these results generalizable”? There are several unique factors about these VA hospitals that would tend to argue both sides of this question. First, both VA hospitals are academically affiliated and have training programs for residents and fellows. Trainees may be more motivated to operate on men with comorbidities to gain additional experience. This would support the findings of Chamie et al.1 Alternatively, VA hospitals are not fee-for-service and thus there are no financial incentives to treat men and best practices should prevail. Viewed in this light, it remains unexpected that for most of the comorbidities constituting the Charlson Index, the presence of that comorbidity had almost no influence on treatment decisions.
Assuming these results may be generalizable outside the VA system, the question is why does the presence of a comorbidity have limited effect on our treatment decisions for prostate cancer? The most likely explanation is that we as physicians simply are poor judges of overall survival. It has already been shown that urologists—who in theory live and breathe prostate cancer—cannot predict recurrence rates much better than a coin toss.4 Thus, what makes us think we can predict overall survival any better? Moreover, this is likely not a urology specific issue, but an issue for all health care providers. Again, drawing upon the data for PSA screening, which is generally performed by primary care providers, there appears to be little association between comorbidities and PSA screening rates.2
Are physicians truly unable to estimate mortality, or are we simply unwilling to talk about it and go along with our patients when they tell us they are going to live forever? We would submit that it is likely a combination of both. We have all been pushed to perform more aggressive treatments with some less healthy patients because they are insistent that their goal is to live as long as possible. While aggressive treatment is unlikely to fulfill this goal for patients who are not “healthy enough”, it is not easy to refuse to treat a patient’s cancer. It is unclear whether this is driven by medical-legal issues, not wanting to hurt people’s feelings or diminish hope, not wanting to “lose” the patient to another surgeon who is willing to treat despite comorbidities, or truly not appreciating what “healthy enough” means. However, it is much easier to acquiesce to treat a person with comorbidity than spend the time necessary to explain why treatment is very unlikely to be helpful and very likely to be harmful.
Given health care providers and patients alike struggle with perceptions of mortality risk, what can we do about this? The first thing is to spend time talking with patients. While many men do say they want to live forever, there are large percentages who do accept their own mortality and put a greater value on quality of life. Though it may be uncomfortable for us to have that discussion, it is a discussion that is vital to have. Second, we need to better understand and appreciate the implications of comorbidity. The study by Chamie et al.1 nicely illustrated that even a single Charlson comorbidity can increase mortality risk by over twofold. Third, in terms of prostate cancer, we need to accept that not “treating” (i.e. a watchful waiting/active surveillance approach) is not only acceptable but preferred for many. As active surveillance studies are maturing, we are learning that the 10-year risk of prostate cancer death is very small among men on active surveillance.5 This compares to a reasonably high risk of non-prostate cancer death over this same time period. Moreover, a randomized trial of surgery vs. watchful waiting within the VA showed that aggressive treatment had no survival benefit over the next 10 years, except in men with higher risk disease (PSA>10).6
In summary, the study by Chamie et al.1 should be a wake-up call to think differently about prostate cancer treatment decisions. While we currently heavily rely on age to determine longevity and form the basis of treatment decisions, we need to better focus on the whole patient—not just age. This will not be easy as this will essentially be a sea change in how we think about our patients. However, the data are clear—comorbidities matter for survival and they should matter in our treatment decisions.