Bisphosphonates are a class of medicines that can reduce the risk of fractures in persons with osteoporosis. They are valuable medicines in the right patients. But, over the past week, a number of reports in the media have attributed miraculous life-extending properties to these medicines.
For example, MedConnect (a site that will help you “stay connected to the most important developments in medicine”) proclaimed,
“Study Reveals Possible Secret to Long Life – Biophosphonates.”
One of America’s leading newspaper introduced its storywith the headline,
“Osteoporosis medication extends life by five years.”
Wow! This must be a really great medicine! Mortality benefits of this magnitude are pretty much unheard of in clinical medicine. The headlines suggest that just increasing use of these common medicines could yield an unprecedented increase in life expectancy marvelling some of the greatest public health interventions of this century.
But the stories are not true. These headlines are inaccurate and very misleading. They do not properly represent the study they are reporting on, as this study did not show bisphosphonates extend life.
The headlines are based on a recent study in the Journal of Clinical Endocrinology and Metabolism that followed women in Australia and compared rates of death while women were or were not taking bisphosphonates. The study showed that women had lower rates of death while taking bisphosphonates than while not taking bisphosphonates.
So, what’s the problem? This press report, and the study, makes the serious error of assuming that an association between bisphosphonate use and mortality means that bisphosphonates are the cause of lower mortality.
This distinction may seem subtle, but it is absolutely critical. Elders taking bisphosphonates are almost certainly different than elders not taking them, and it is quite likely these patient differences explain the mortality effect. Most, if not all of the mortality difference may have nothing to do with the bisphosphontes. This study had very limited capability of accounting for these patient differences.
In addition, this study examined bisphosphonate use as a time dependent variable. This type of analysis is well known to be prone to serious biases when examining the effects of drugs. For example, if bisphosphonates are discontinued when patients are seriously ill, it would inaccurately make bisphosphonates look very good.
So, where did all these misleading media reports come from? It appears most of the media reports essentially paraphrased a press releasefrom the Garvan Institute, the institute that led this study. The press release has the absurd headline,
“An extra 5 years of life an unexpected benefit of osteoporosis treatment.”
The press release goes on to suggest that these medicines can reduce the risk of death by 80%:
“In a group of women with osteoporotic fractures over the age of 75, you would expect 50% to die over a period of five years. Among women in that age group who took bisphosphonates, the death rate dropped to 10%.”
It is possible that bisphosphonates do reduce mortality. In fact, a randomized trialshowed one agent (zoledronic acid) may lead to a mortality reduction in persones with hip fracture. But the study that is the subject of these media reports does not provide substantial evidence a mortality benefit. And if a benefit exists, it is definitely much more modest—not the fountain of youth like benefits claimed in the press surrounding this study.
The study and media coverage provides a cautionary lesson for researchers who work with the media to publicize their work. We all have a tendency to believe our studies are more important than they actually are. And there is a natural tendency to make our results sound more spectacular than they are in order to better disseminate our work. But we have a responsibility to the public to guard against these impulses.
This story should also serve as a cautionary tale for the media. Media really needs to proceed with caution when writing a story based entirely on a press release without any independent reporting. This is a practice that is best avoided, but at a minimum, it should be disclosed to readers when the true souce of a story is a press release than objective reporting.
by: [ken covinsky]