Wednesday, November 28, 2012

Dissecting a hepatitis C screening recommendation

After being banished from the public spotlight in 2010 and 2011 due to health reform politics, the evidence-generating operation known as the U.S. Preventive Services Task Force is finally making waves again. On Monday, the USPSTF got on board with a CDC guideline to screen all "baby boomers" born between 1945 and 1965. In their draft recommendation statement, the Task Force (which in 2004 found insufficient evidence for screening even patients with self-reported risk factors) assigned a level "C" rating to birth cohort screening for hepatitis C, which means they believe there is moderate certainty of at least a small net benefit. Although newspaper reports described this endorsement of the CDC's guideline as lukewarm, it is anything but that. For clinicians who have never given a thought to routinely screening my parents' generation for this virus, it is a sea change.

How beneficial a medical test or intervention would be for society is ultimately a subjective assessment. The CDC has argued that the prevalence of undiagnosed hepatitis C in baby boomers is higher than in the general population, and affected persons (who contracted the virus via intravenous drug use or tainted blood transfusions in the 1970s and 1980s) are more likely than younger people to be at risk for life-threatening problems such as cirrhosis and liver failure. Plus, treatments are more effective and less toxic than before. I'd still argue that the magnitude of benefit is less than either the CDC or Task Force say it is, but ultimately it's anyone's best guess.

The "moderate certainty" part bothers me more. Based on the USPSTF's extensive body of work and the evidence contained in two systematic reviews in the Annals of Internal Medicine, I don't see how they managed to get there. Generally, the first piece of evidence the Task Force looks for is a randomized clinical trial that compares a screened population to a non-screened population and follows them up long enough to see if fewer people in the screened population end up dying from the disease. It takes decades for hepatitis C to lead to premature death, though, so it's no surprise that the evidence review turned up no such trials.

Given the absence of "direct" evidence that screening for hepatitis C saves lives, the Task Force was forced to rely on an indirect chain of evidence, namely: screening tests detect hepatitis C, hepatitis C leads to bad outcomes for patients, treatment prevents these bad outcomes (or surrogate measures of these outcomes) without causing too much collateral damage in the process. This type of chain has been used in other recommendations; this is why, for example, the USPSPTF recommends screening older women for osteoporosis. For hepatitis C, this chain of evidence wasn't there in 2004. Now, the Task Force says it is.

The key point in the USPSTF's argument seems to be that two drugs that the FDA recently approved for hepatitis C infection, boceprevir and telaprevir, lead to a sustained virologic response (SVR) in substantially more patients (70 to 80 percent) than the previous treatment combo, pegylated interferon alfa plus ribavirin, which was not only poorly tolerated but led to SVR in fewer than half of patients. But how do we know that SVR from these new drugs is associated with fewer liver failures and deaths? Strictly speaking, we don't. The Task Force extrapolated from cohort studies that show that SVR was associated with lower mortality in patients who had been treated with interferon and ribavirin.

To summarize: there's no direct evidence that screening for hepatitis C saves lives. New treatments are more likely to lead to SVR, but there's no evidence that they save lives, either; the studies have not been done. We know that if patients stick out therapy with interferon and ribavirin and get a SVR, they seem to do better than patients who don't, but the Task Force knew that in 2004 and still declined to make a recommendation. To me, this chain of evidence doesn't hold up to scrutiny. In fact, it resembles screening for osteoporosis in men (which received an "I" rating from the USPSTF in 2011) more than screening in women. So for the second week in a row, I find myself saying that the Task Force missed the boat.