At least 30 high-sensitivity C-reactive protein assays are on the market, and all of them have come into routine use by clinicians performing cardiovascular risk assessment. But a recent policy shift by the Food and Drug Administration, requiring manufacturers to submit a new 510(k) submission for that particular use, is unsettling some laboratories and clinicians alike.
The confusion stems from new FDA guidelines, or “updated recommendations,” issued Sept. 22, 2005. They explain the FDA’s decision to require manufacturers of hsCRP assays, if they wish to market the assays for use in cardiac risk assessment/stratification, to apply for a new labeling and include evidence of efficacy for that use under the new name “cardiacCRP,” or cCRP.
Laboratories can choose to continue testing for cardiac risk assessment using assays that have not received approval for the new label, but if they do, the FDA says, “Such use would constitute an off-label or unapproved use of the device.” That means, under the provisions of the Clinical Laboratory Improvement Amendments of 1988, that “the laboratory must validate the test performance.”
At issue: new guidelines the FDA issued Sept. 22, 2005, titled “Review
Criteria for Assessment of C-Reactive Protein (CRP), High Sensitivity C-Reactive
Protein (hsCRP) and Cardiac C-Reactive Protein (cCRP) Assays”.
C-reactive protein assay is a test traditionally used to diagnose and manage active infection and inflammation, because the level of protein is usually quite high in these cases, explains Nader Rifai, PhD, professor of pathology at Harvard Medical School and director of clinical chemistry at Children’s Hospital, Boston.
But more than a decade ago, CRP was shown to reflect a very low level of inflammation that is associated with the development of atherosclerosis. “We now know that atherosclerosis, at least a major component of it, is inflammatory in nature. But in order to measure CRP for purposes of cardiovascular disease risk prediction, you need an assay that is capable of measuring this protein at very low concentration, such as levels you would see in normal individuals. That’s why high-sensitivity methods were developed.”
The FDA has approved a number of these assays, and the indication was to detect the low level of inflammation associated with cardiovascular disease, Dr. Rifai says.
However, with its September 2005 guidelines, the FDA is saying, “If we want to add the claim that this CRP is used for cardiovascular disease risk predicting, we are introducing another term, which is cardiac CRP, for these assays,” he says.
In Dr. Rifai’s view, the FDA move has spread confusion, not cleared it up. “The only confusion you see in the medical community is that physicians are using regular instead of high-sensitivity CRP. The FDA has increased confusion by adding terminology that nobody is familiar with.”
He contends there is no precedent for this FDA policy. It’s standard practice in the laboratory to use tests for more than one indication. “When we develop another clinical application, we don’t give the test another name; it is the same test,” he says.
“It took a long time for physicians to get used to ordering hsCRP and not regular CRP when they want to assess risk in their patients, and all the literature uses the term hsCRP when referring to cardiac disease risk prediction, and introducing another term for exactly the same assay seems to be counterproductive.”
Also puzzling, he says, is that if clinicians are not going to use hsCRP for cardiac risk, “I am at a loss as to what kind of clinical application it could be used for, because there is none.”
The “bridging study” that the FDA is requiring of any manufacturer that wishes to have its hsCRP assay labeled as cCRP is also questionable, in Dr. Rifai’s view. “When hsCRP methods were first introduced, the FDA required demonstration of comparability of the new assays to a predicate device, in this case the Dade Behring assay.” But the required bridging study “entails a repeat of the method-comparison study,” and will be an unnecessary burden on manufacturers, he says.
Dr. Rifai does not question the need for additional standardization of the assay. “That is ongoing at the Centers for Disease Control and Prevention right now, and we are in total agreement that additional standardization is needed. But their effort has nothing to do with the FDA policy, and it doesn’t mean that you have to introduce a new name.”
Another problem is that the FDA’s term “cardiac CRP” is similar to the trademarked term “Cardio CRP” used by Quest Diagnostics (which uses Dade Behring’s CardioPhase assay) for its hsCRP assay. Says Dr. Rifai, “We believe the use of such a term by the FDA will inadvertently appear to be an endorsement by the agency of one particular assay, a stance we believe inconsistent with agency policy and one likely to further confuse the clinical community.”
For its part, the FDA stresses that the difference between hsCRP and cCRP is not in the analyte itself, but in the analytical performance, which has been documented through submitted evidence.
Since the FDA is responsible for approving device labeling, the agency says, it requires appropriate evidence to support each clinical purpose. “FDA could have chosen to approve ‘High-Sensitivity CRP-with-evidence-of-efficacy-for-identification-and-assessment/stratification-of-individuals-at-risk-for-future-cardiovascular-disease,’ but it chose instead to use the shorthand ‘Cardiac CRP (cCRP),’” the FDA said in a March 2006 statement intended to address some of the concern expressed by the laboratory community. “No hsCRP assay has been allowed to have this indication (cCRP) without supporting clinical studies or bridging studies to a device which was the subject of the clinical studies.”
“There has been a whole series of articles published in the last five or six years, demonstrating the association between CRP and the presence of risk for cardiovascular disease, presumably mediated by inflammation,” says Steven I. Gutman, MD, MBA, director of the FDA’s Office of In Vitro Diagnostic Device Evaluation and Safety.
“The FDA has appreciated this link probably since the day of the first article, and we’ve slowly accommodated, by allowing sponsors to include language about the association. Then a company came in and wanted to make that specific claim, and they presented a specific data set linking their assay to that particular risk.”
“And we do what we always do when we review. We get a claim, we get supporting evidence to support the claim, and we clear the product,” he explains. “So it’s not as startling as people seem to think it is.”
The whole basis of the misunderstanding, Dr. Gutman says, is “the belief that assays with cardiac claims are the same as assays that have just general claims suggesting an association. They’re not actually the same; they’re really quite different. When an assay has a clear link to cardiac end points, we’re very anxious for the company to make that claim. And when the link to cardiac end points is more ambiguous, we don’t object to some description of an association, but we don’t like people to make claims when there is no data to support the claim.”
The requirement for a bridging study is not as onerous for manufacturers as has been alleged, Dr. Gutman says. “If the claim is actually that the test result is associated with cardiac end points, then there has to be a very rigorous demonstration that that claim is supported. It can be clinical information that could show a direct association with cardiac risk points, or it could be indirect by showing the assay is virtually equivalent to one that has already been pedigreed.”
“Our process is based on showing ‘substantial equivalence,’ and there isn’t a magic number at which equivalence occurs. So if it’s very good, we delightfully rush it through quickly. If it’s really, really terrible, then we might actually say it’s not substantially equivalent and we would actually not clear it for the market. And when it’s in an intermediate zone, we’ll give the manufacturer the benefit of the doubt, but we’ll label it saying it’s intermediate. It may be faster or cheaper or more portable than another assay, so the fact that its performance isn’t quite as good, in some clinical circumstances, might be acceptable.”
The point the FDA is trying to make, he says, is there is a hierarchy of data, and companies can make choices. “If they either want to do clinical studies or analytical studies that link themselves strongly with clinical trials, they should get credit for that, and the credit is they get to make very explicit claims for cardiac risk. If they wish to do some more quick and dirty studies, they may still have a useful marker, but certainly that doesn’t provide the same certainty of association of cardiac risk as the well-credentialed assay.”
In response to fears that the name cardiac CRP closely resembles a privately trademarked term, Dr. Gutman says the FDA has no jurisdiction. “People play around with names all the time. We have no authority over that. People can say they have First Choice Pregnancy Tests, or SuperGreat Tests, or Heart-Smart Cholesterol. We don’t have any control over names.”
The FDA would like to be transparent, he stresses. “The FDA position is that a claim should be based on data that support that claim, and all C-reactive protein assays are not equal. Some are well-credentialed for cardiac end points, some moderately credentialed, and some are not well-credentialed at all. Our purpose is to allow for variable credentialing.”
As to fears that the FDA action could have a negative impact on clinicians, Dr. Gutman downplays the possibility. “I would hope it actually would be the opposite. We hope if people educate clinicians and promote their products well, there would be a better understanding of how predictive the assay would be.”
But Dr. Rifai, who is urging the FDA to consider withdrawing this particular guidance document, thinks this is unlikely. “Does not the decision about introducing this terminology undermine the effort that has been taken by several other federal agencies like the National Institutes of Health, CDC, and scientific organizations like the American Heart Association and American Association for Clinical Chemistry in educating clinicians and laboratorians about the use of hsCRP? And does the FDA believe it is really important to give it a different name rather than just an indication in the insert sheet?”
Anne Paxton is a writer in Seattle.
The FDA notes that written comments and suggestions for its guidance
document can be submitted at any time for agency consideration to the
Division of Dockets Management, Food and Drug Administration, 5630
Fishers Lane, Room 1061 (HFA-305), Rockville, MD 20852. Questions can
also be submitted to James V. Callaghan at firstname.lastname@example.org.
Readers are referred to the “Point” (Rifai N, Ballantyne CM, Cushman M,
et al) and “Counterpoint” (Callaghan JV, Gutman SI) published last month
in Clinical Chemistry; 52(7):1254–1257.