When Criteria are Misapplied: Misdiagnosing MS
In 2005, the New England Journal of Medicine published the first two cases of Tysabri associated progressive multifocal leukoencephalopathy (PML). The succeeding autopsy, however, revealed that, one of the patients had been misdiagnosed and did not in fact have multiple sclerosis (MS). Such misdiagnoses are, I believe, a growing concern when it comes to MS.
Diagnosing a disease such as MS is often complicated, and at times, the approach doctors take to do so is influenced by factors that are not implicit in the process: the accessibility to certain services, the ubiquity of a diagnostic criteria, or the inertia of a given mindset. I’d like to take a moment to introduce the challenges of diagnosing MS, pointing out some of the pitfalls of our current methods, as well as make some suggestions as to how they could be improved.
The majority of Canadian neurologists have access to more tools and imaging devices to help them in their diagnosis than ever before. That being said, like many neurological diseases, the diagnosis of MS remains primarily clinical: neurologists mostly rely on what a patient says and what is revealed by physical examination in order to formulate their conclusions. This is why the establishment of diagnostic criteria is so important, and perhaps, why, once established, such criteria have the potential to quickly become canon.
When it comes to MS, it is often the MacDonald criteria – revised in 2010 – which helps neurologists determine a conclusive diagnosis. These criteria are clear and, in of themselves, create an appropriate benchmark for most MS diagnoses. Unfortunately, their tendency to be applied universally increases the rate of false positives. This is because the statistical values upon which these criteria are based, were obtained from a population of patients which presented with symptoms highly characteristic of MS. The result is a set of criteria which are extremely sensitive and specific for typical presentations of the disease, but do not allow for the same specificity with regards to patients without this typical presentation. In other words, the application of the MacDonald criteria to both typical and a-typical symptoms, makes it unlikely that patients with uncharacteristic presentations of MS be overlooked, but also more likely that patients that present a-typical symptoms be diagnosed with MS even though they do not have the disease.
Unfortunately, this is not always taken into account in practice. Any explanation as to why this is the case remains conjecture at this point, but I suspect that the widespread adoption of the criteria and the strong pressures doctor’s feel to diagnose MS patients as quickly as possible so as to start treatment early, have a significant role to play. Newly available therapies have the potential to be very effective if applied early in the course of the disease, yet these treatments are also more aggressive, with serious potential side-effects, and should, if anything, emphasize the need for high diagnostic certainty.
No one currently knows the rate of false positive diagnoses which result from the application of the revised 2010 MacDonald criteria, but the answer may be statistically significant. Placebo groups in recent clinical trials have displayed an unusually benign disease evolution, discombobulating statisticians. It’s possible to infer that, just as the 2005 autopsy showed, a small but significant number of patients which do not actually have MS, are being inadvertently recruited in such trials. Making matters worse, it is often difficult to remove a label, even a false one. As things stand, it’s very likely for patients, having previously been miss-diagnosed with MS, to continue on with treatment for years, even though they clearly do not have the disease. Though there’s no easy way to fix this, I believe certain steps can be taken by Neurologists to reduce the risk of misdiagnosis.
How can we begin to fix this?
The false diagnostic certainty the MacDonald criteria provide is likely responsible for several misdiagnoses. Take this certainty away, and it becomes a good idea to buttress these criteria with other tests such as the presence of cerebrospinal fluid (CSF) oligoclonal bands (OCB’s). The later has been reported present in over 85% of cases of MS. This data however dates back to the age of hospitals testing using their own unique laboratory recipe, leading to a high inter-laboratory variability. Recent studies have shown that CSF OCB results using standardized lab kits are now highly reliable within different hospitals. Since the adoption of these new laboratory kits the incidence of CSF OCB’s in MS, is probably much closer to 100% meaning that CSF OCB negative patients are unlikely to have MS or have a less aggressive subtype of MS.
How does this change the diagnostic approach?
An MS diagnosis should be made using the revised MacDonald criteria confirmed with the presence of CSF OCB’s. In clinically typical patients but CSF OCB’S negative, definite clinical and/ or radiological evolution would be required before one could confirm the diagnosis and before any therapy be initiated. Patients, with clinically isolated syndrome (CIS) knowing that 20% will not evolve into MS should be followed and not treated. Once the diagnosis of MS is firmly established however, aggressive therapy aiming for clinical and radiological remission should be instituted from the onset.