MESSAGE FROM THE EDITOR: Balancing risk and reward: the question of natalizumab September 2009

Balancing risk and reward: the question of natalizumab

In the pursuit of doing everything possible for all of our patients, we cause considerable harm to some.  Preventable medical errors, inexcusable causes of harm, have appropriately engaged the attention of all of us as the magnitude of this problem has become recognized (1).  A far more common source of harm, however, results from the known, i.e. “expected”, complications of our treatments.  These are not caused by obvious errors in oversight or judgment, but rather from adverse outcomes resulting from standard-of-practice care inherent in the complication rates of interventions.  Examples include toxic reactions to medications or poor outcomes from surgical or interventional procedures.  Such complications are particularly unsettling whenever the decision to treat in the first place is debatable.  What factors determine whether a patient undergoes an endarterectomy for asymptomatic carotid stenosis, an intervention for a vascular malformation that has not bled, or treatment with natalizumab for multiple sclerosis?   Although evidence-based medicine strives to reduce medical decision-making to standardized, codified principles, when applied to many real-life situations it is the judgment of the clinician and the personal choice of the patient that ultimately determine the treatment plan.  This is especially true when the decision involves accepting the risk of an unlikely but potentially catastrophic adverse event.    Often, the treatment decision is driven by how the risk and reward of a proposed treatment is perceived, both by the patient and by the clinician.

There is nothing unique about decisions that we make about our health – we often make “qualitative” decisions at odds with the “quantitative” reasoning of experts.  Some of us are risk-takers and others are risk-averse.  Whether or not a patient feels lucky can profoundly influence a health care decision (2), especially when in a gray zone of medical practice.  Some patients place greater value on near-term than long-term outcomes, and not surprisingly our cultural and religious backgrounds, prior experiences, and family responsibilities all may play roles. In one study, a high risk for a complication that is not particularly frightening – a swollen leg from phlebitis – was preferred by many over a tiny but much more frightening risk – cerebral hemorrhage; patients who weigh the rare adverse event heavily are, not surprisingly, more likely to refuse prophylactic anticoagulation  (3).  A variety of risk analysis methods have been developed to quantitate the relative values (“utilities”) that individuals apply to health care decision making.  As examples, the standard gamble method weighs the importance of the range of possible outcomes plus willingness to accept risk; the time trade-off method considers the benefit of early compared to late outcomes, and willingness to pay model estimates the financial value for a specified improvement in health (4-6).  Although none of these scales is readily applied at the bedside, they do serve to focus our attention on the underlying and often subconscious factors that can profoundly influence medical decision-making and compliance to recommended treatment plans.

In many situations, patients prefer to leave medical decisions to their physicians’ best judgment (7). For many, such informed trust may be more important than a wish to have control over treatment decisions.  Some delegate decision-making responsibility because they have confidence in the professional expertise and opinion of their physicians, and others do so because they are overwhelmed by the situation or unable to grasp the nuances of the factors that come into play.  Whatever the underlying motive, implicit in the patient’s transfer of decision-making responsibility is the assumption that all actions will be in his or her best interest, and in such situations the responsibility for risk-reward analysis falls entirely to the physician.

Risk-reward decision making is brought into sharp focus with the alpha-4 integrin receptor antagonist therapy natalizumab (Tysabri).  Natalizumab is a wonderfully effective therapy for relapsing forms of multiple sclerosis (MS); it is approximately twice as effective as the beta interferons or glatiramer acetate.  However, use of natalizumab carries with it a small risk of progressive multifocal leukoencephalopathy (PML), estimated at approximately 1:6000 annually (11 cases have been identified to date).  The risk is probably much lower during the first year of therapy, and for this reason some experts advocate that its use be limited whenever possible to 6-12 month treatment cycles.  Not surprisingly, the PML cases have sensitized the neurologic community to potential risks of natalizumab, and this could result in the over-reporting of unrelated conditions (for example melanoma [8]) in natalizumab-treated patients.  In this issue of the Annals, Bethele (9) reports a case of primary CNS lymphoma associated with natalizumab therapy, and in our editorial pages we’ve asked several experts to offer their perspectives on this anecdotal observation (10-13).  All concur that the current case (and a single previously identified one) offer insufficient evidence to implicate natalizumab at this time as a risk factor for primary CNS lymphoma.  Bozic et al (representing the manufacturer, Biogen/IDEC) argue that the number of cases identified to date conforms to the expected rate of primary CNS lymphoma, given more than 56,000 patients have thus far treated with natalizumab (10).  This said, the case report is notable – and deserving of publication – given the preexisting concern about natalizumab’s safety profile, especially with respect to immunodeficiency-associated CNS complications.

As physicians, we are taught from the first days of medical school that the first obligation is to do no harm; risk-averse decision-making is encouraged.  We are also subject to many qualitative biases influenced by our individual personalities, values, and perspectives.  One potential bias that we bring to our daily professional lives has been categorized as the availability heuristic, in which the recall of a recent case or experience – for example an adverse outcome – can unduly influence subsequent decision-making (13).   For one of us (SLH), an unforgettable personal experience with one of the original patients with natalizumab-associated PML led subsequently to a fear of prescribing and less use of the drug than may have been appropriate.  Unfortunately, there is no current way to predict individual risk for any natalizumab-associated complication.

In the absence of science to rationalize and simplify decision-making, we must practice the art of medicine whenever the choice is not clear-cut.  In doing so, we need to remember that, whereas all therapeutic decisions are ultimately in the hands of our patients, their decisions are heavily influenced by the manner in which we present the choices to them.  Awareness of our own biases may allow us to present the alternatives more accurately thus making us better practitioners.  Although the first precept may be “do no harm”, if this were taken too far to mean that we should eliminate the risks of all complications we would be selling short many interventions that on average produce benefits.

Stephen L. Hauser MD

S. Claiborne Johnston MD, PhD

1. Preventing Medical Errors: Quality Chasm Series.  Institute of Medicine Report, 2006.
2. Teigen KH. Hazards mean luck: counterfactual thinking in reports of dangerous situations and careless behavior. Scand J Psychol 1998;39:235.
3. O’Meara JJ, McNutt RA, Evans AT, et al. A decision analysis of streptokinase plus heparin as compared with heparin alone for deep-vein thrombosis. N Engl J Med 1994;330:1864.
4. Kassirer JP. Incorporating Patients’ Preferences into Medical Decisions N Engl J Med 1994;330:1895.
5. Torrance GW, Thomas WH, Sackett DL. A utility maximization model for evaluation of health care programs. Health Serv Res 1972;7:118.
6. Thompson MS. Willingness to pay and accept risk to cure chronic disease. Am J Public Health 1986; 76:392-6.
7. Mangset M, Berge E, Førde R,et al. ‘‘Two per cent isn’t a lot, but when it comes to death it seems quite a lot anyway’’: patients’ perception of risk and willingness to accept risks associated with thrombolytic drug treatment for acute stroke. J. Med. Ethics 2009;35;42-46
8. Mullen JT, Vartanian TK, Atkins MB. Melanoma complicating treatment with natalizumab for multiple sclerosis. N Engl J Med 2008;358:647-648.
9. Berthele A et al Primary Central Nervous System Lymphoma in a Patient treated with Natalizumab Ann Neurol, this issue
10. Bozic et al Natalizumab and CNS Lymphoma – No Clear Association Ann Neurol, this issue
11. Kappos L. Another complication of Natalizumab treatment? Taking the challenge Ann Neurol, this issue
12. DeAngeles L Ann Neurol, this issue
13. Ransohoff R Ann Neurol, this issue
14. Salem-Schatz SR, Avorn J, Soumerai SB. Influence of clinical knowledge, organizational context, and practice style on transfusion decision making. Implications for practice change strategies. JAMA 1990; 264:476.