BB&T Washington Editor
The typical layperson finds medical opinion confusing partly because medical science is far from perfectly exact. And it seems that medicine can’t make up its collective mind about what therapies and diagnostics are best, since research opinion on these issues often shifts, further feeding public confusion.
This state of affairs won’t change much, thanks to an article appearing in the Dec. 13 issue of the Journal of the American Medical Association concerning prostate cancer and its management via “watchful waiting,” frequently considered superior to interventional strategies, given the slow growth of this type of cancer.
A research team led by Yu-Ning Wong, MD, of the Fox Chase Cancer Center (Philadelphia), tracked nearly 45,000 men who had been diagnosed with non-metastatic prostate cancer between 1991 and 1999 and were between the ages of 65 and 80 at diagnosis. Those who were included had survived for at least a year, and the follow-up ran until the end of 2002. Of that group, slightly more than 32,000 opted for treatment with radical prostatectomy or radiosurgery, and the balance, a bit more than 12,000, decided against use of either intervention. The protocol excluded those who had decided on hormonal therapy.
More than 4,600 (37%) of those who chose observation passed away during the follow-up period. This compared to 23.8% of those choosing the interventional treatments who passed away during the course of follow-up.
The article states that the authors used “propensity scores to adjust for potential confounders (tumor characteristics, demographics and co-morbidities),” but still found “a statistically significant survival advantage associated with treatment.” The authors acknowledge that this observational study has its flaws, including the possibility of selection bias, and they indicate that “these results must be validated in randomized controlled trials of alternative management strategies.”
On the subject of selection bias, Gary Hudes, MD, the director of genitor-urinary malignancies program at Fox Chase and a co-author of the study, told Biomedical Business & Technology that “when you’re dealing with a non-randomized prospective study, physicians could be exercising bias in who they assign to treatment.”
The researchers tried to flatten out this influence by dividing the subjects into subcategory for co-morbidity, tumor characteristics and age based on the premise that these factors are most likely to influence a physician’s recommendation. All the same, a statistical analysis of such data is limited, Hudes noted. “At the end of the day, you can only do this for things you know about.”
Still, the sample size in question gives a statistician plenty of data to root out known sources of bias. “You would have had to have had a skewed distribution of characteristics . . . that was ridiculously imbalanced in order to affect the result” for this source of bias.
Hudes expressed confidence in the validity of the analysis. “It really shifts the view that instead of having to have a reason to recommend intervention, you now have to have a reason not to recommend an intervention.” Serious co-morbidities, and anything else suggestive of limited life expectancy, might render an intervention questionable, but “what most urologists and physicians would use is 10 years or more expectancy” as a justification, he said.
One point of interest in any study involving serious conditions is that of the cause of death.
Cause of death data “is the least reliable of all,” Hudes said, and most difficult to obtain.
Actual cause unclear
The actual cause of death is not always as stated on the death certificate, partly because it is often difficult to determine which of several conditions killed the patient with more than one disease. “To know exactly how many died of prostate cancer vs. other conditions is not something we can vouch for,” he said.
Hudes said that not much has changed in the past year or so as to which biomarkers are the strongest determinants of a patient’s risk of death. “Other than the conventional histological things like Gleason’s scores, we know of none that have been validated” to indicate greater risk, he said.
As for exactly which patients should opt for watchful waiting, he said it should depend “on the individual case in terms of other medical problems.”
He also noted that nowadays, a man who is 65 has 12 years or more of life expectancy, but a substantially older and/or sicker man presents a different picture.
“Someone with poor heart function, for example,” might not be a great candidate, and the side effects of surgery are a huge deterrent, including incontinence and erectile dysfunction. “We will always be faced with the choice of length of life vs. quality of life,” Hudes said.