This story about Celebrex’s potential to prevent colorectal cancer satisfied most of our criteria and gives readers the correct bottom line on this study. The story should have taken a bit more care to explain that the investigators were looking at a surrogate outcome, and it could have quantified the harms of treatment more precisely. But even in the areas where it fell short, the story did end up providing enough information to advance the reader’s understanding.
As the story points out, screening via colonoscopy does not always catch polyps before they have a chance to progress to cancer. Drugs that can reduce the number of polyps that develop in the first place (e.g. Celebrex and aspirin) might therefore have a role to play in prevention efforts, especially in individuals at high risk. The challenge is to balance this potential benefit against the well established risk of adverse effects with these medications (i.e. heart problems with Celebrex and gastrointestinal bleeding with aspirin). At present, there isn’t enough evidence that the benefits of treatment outweigh the harms for most people.
The story manages to work costs into the discussion: “At a cost of several dollars per day, depending on dosage, [Celebrex] is much more expensive than older pain relievers such as ibuprofen.”
The story reports the benefits in appropriate statistical terms, noting that “new polyps were found in 58 percent of people on placebo, and only 51 percent on Celebrex.” The story could also have given us the number of advanced polyps that were prevented in each group. And, as noted above under the “Evidence” criterion, the story should have explained that these differences in polyp incidence might not translate directly to the number of actual cancers prevented. Since we’ve already docked the points for this omission, however, we’ll award a satisfactory here to recognize the appropriate use of absolute risk numbers.
If we were tough on the “evidence” criterion grade, we go a little easy on this one.
The story notes that Celebrex users were more likely than the placebo group to develop polyps once they stopped taking the drug, and it quantifies this risk appropriately in absolute terms. However,cardiovascular risk — which is the main concern with this drug — is described only in relative terms. The story states that those treated with the drug had a 66% higher risk of serious cardiac disorders. Although we acknowledge that this is a picky point, the story should have reported the absolute rate of these problems in both groups and the associated difference in risk. This would have helped readers understand how common cardiovascular events were in the study overall and how concerned they should be about the increased risk in the Celebrex group.
Nonetheless, because of the overall handling of harms and repeated mention of safety concerns, we’ll give the story the benefit of the doubt.
We appreciated this story’s generally restrained take on the results, which included mention of “caveats” right in the headline and some discussion of the study’s limitations. There’s also a solid explanation of how one might balance benefits and harms of treatment, and the story ends with appropriate emphasis on the doctor-patient relationship and the importance of shared decision-making regarding treatment.
An important shortcoming, however, is the failure to explain that polyps are a surrogate endpoint. The story should have been clearer about the fact that these were precancerous lesions that were being assessed, and that most such polyps do not progress to cancer. The number of actual cancers prevented would likely be lower than what is suggested by the study results. Although the story does explain that polyps are not in themselves cancerous, but “could lead to colorectal cancer,” there should have been more discussion of how this affects our interpretation of the results.
A tough call, but we think we our observation is important.
The story states that colorectal cancer strikes 1.2 million people each year. It probably should have specified that this figure represents a global tally and not the incidence in the United States. Since the story quotes an American researcher and the study appears in the American Journal of Gastroenterology, readers might assume that the statistics are also U.S. – centric.
With that being said, neither this nor anything else in the story rises to the level of disease-mongering.
The story notes that the study was funded by Pfizer, maker of Celebrex, and includes commentary from an independent expert who didn’t participate in the study.
The story notes that most people should rely on colonoscopies to find colorectal cancer in its early stages rather than take a risky drug to prevent it. It also mentions the fact that aspirin is being studied for its potential cancer prevention effects.
The story states that Celebrex is the only COX-2 inhibitor remaining on the market in the U.S., so readers should know it’s available. The story could have been more clear that the drug is available only by prescription and can’t be purchased over the counter.
There is appropriate acknowledgment of previous research on the potential for COX-2 inhibitors to prevent cancer.
There was enough original reporting for us to be certain that this was not based on a press release.