New Math for MS Treatment
Improving predictions for who will develop a rare natalizumab side effect
Researchers have published an algorithm for evaluating the likelihood that MS patients taking natalizumab (Tysabri) will develop a potentially fatal brain infection. Although many MS specialists were already applying the algorithm in practice, they say that its publication allows for a critical look at the underlying data and gives them a tool to use when discussing risk with patients. The study, published 17 May in the New England Journal of Medicine, was funded by Biogen Idec and Elan Pharmaceuticals, which comarket the drug.
Decisions about using natalizumab in MS have challenged patients and clinicians alike since the discovery that the therapy can, in rare cases, lead to the sometimes deadly disorder progressive multifocal leukoencephalopathy (PML). One risk factor for treatment-associated PML is evidence of infection with the common JC virus (JCV) that causes PML by going rogue and attacking brain tissues. Previous immunosuppression therapy (with agents such as mitoxantrone, azathioprine, methotrexate, cyclophosphamide, or mycophenolate mofetil), and a relatively long natalizumab treatment duration, especially beyond 2 years, are also associated with PML.
The paper’s authors used several information sources, including clinical studies and a Swedish patient registry, to estimate the incidence of PML among MS patients using natalizumab. They compared this incidence to patient JCV antibody status (an indicator of exposure to the virus), prior immunosuppressant use, and length of natalizumab treatment to classify risk based on these factors. As of February 2012, the investigators identified 212 cases of PML among almost 100,000 patients on natalizumab treatment.
According to the newly published algorithm, patients are at the highest risk—having a 1.1% chance of developing the complication—when all three factors are present and at an intermediate risk when they test positive for the antibodies and have had either prior immunosuppressant use (0.16%) or 25 to 48 months of natalizumab exposure (0.46%). Patients are at the lowest risk (0.009%) if they test negative for JCV antibodies. A positive JCV antibody result alone confers a less than 0.1% risk. The authors write that the algorithm will help clinicians make “more informed, individualized treatment decisions.” MS specialists observe that they and their colleagues have been applying the algorithm for a while, as the information was widely disseminated before its publication. But Lauren Krupp, a neurologist at Stony Brook University School of Medicine in New York and scientific consultant for Biogen Idec, says that the point of this “important” paper is that it “is consistent with a growing trend in personalized medicine, which is to find the right medicine for a particular individual.”
Although its reception has largely been positive, the new stratification system is complicated by the fact that the test for JCV antibodies is not standardized among different labs for negative-result cutoff values, says Avindra Nath, a neurologist at the U.S. National Institute of Neurological Disorders and Stroke (NINDS). In addition, the virus can trigger the development of different antibodies depending on the viral particle the immune system targets, and labs also vary in the type of antibody evaluated. Nath says that the data in the paper are very reliable but that “the gray areas are with the [JCV antibody] assay itself.” However, Gary Bloomgren, first author of the new paper and vice president of drug safety and risk management at Biogen Idec, says by email that the antibody test used to derive the algorithm “is available to clinicians through much of the world,” giving as examples Focus Diagnostics in the United States and Unilabs elsewhere. But Eugene Major, a senior investigator at NINDS, says, “As everyone knows, there is no standardized assay for this [antibody testing], and it would be helpful for other labs to run those same samples using their assays.”
Major also expresses a concern about the relevance of a second risk factor, immunosuppression history. He notes that immunosuppressant therapy was stopped long before patients developed PML and observes that the rate of PML in patients using immunosuppressive therapies alone for other disorders is extremely low. Given the timing disconnect and the rare PML event rate among patients using immunosuppressant therapy alone, “I’m not sure what exactly the relationship is between immunosuppressant drugs and PML,” he says.
Bloomgren, however, contends that the immunosuppressant link to PML risk is a plausible one despite the delay between treatment and development of PML. He points out that prior immunosuppressant use has been linked to a long-term increased risk of cancer, even when a patient has discontinued the therapy. Noting that one possible explanation for the link to malignancy is a lasting negative effect on the immune system’s surveillance capability, he speculates that “a similar mechanism could be at play here” with the development of PML in MS patients.
The third risk factor, length of time of natalizumab treatment, could also be less clear-cut than the algorithm indicates, Major says. He notes that in his experience, patients might be on treatment for, say, 26 months but receive 24 infusions during that time. For this reason, Major suggests that researchers should focus on the overall number of natalizumab doses rather than treatment duration.
Another issue that the study did not address is a possible link between changes in blood-borne JCV antibody concentration and risk of developing PML, says R. Philip Kinkel, a neurologist at Beth Israel Deaconess Medical Center in Boston and a scientific consultant to Biogen Idec. With the current antibody testing, the result can be arbitrary because of the interlab variability, Kinkel says. But monitoring a patient who tests positive for JCV antibodies for an upward shift in antibody concentrations could be more informative. “A virus replicating more often will probably be exposed to the immune system more often and be associated with a higher titer of antibodies to the virus, which will probably be associated with a higher risk of PML.” Major echoes that observation, saying, “We need to be more rigorous when looking for stratification of risk factors, and we think that the rise in antibody titers is an important factor.” Bloomgren noted that although the data don’t answer this question, Biogen Idec’s ongoing studies might address this gap.
Despite the potentially lethal complication of PML, natalizumab remains a viable treatment option because it can have a “dramatic” positive effect, Krupp says, reflecting the expert consensus. It’s this potential benefit that leaves patients and their physicians considering natalizumab in spite of the PML risk. Even if the algorithm were thoroughly confirmed, however, treatment decisions would remain complex and, as the paper authors note, require consideration of other factors, such as MS severity, previous treatments, and what the patient wants to do.