Will Rogers, MS Diagnosis, Progression, and Disability

Recently I came across an interesting article (at $31.50 was a bit pricey to access, so I didn't) which discusses whether the criteria by which patients are diagnosed or categorized affects outcomes in clinical trial data. The subject seemed interesting enough that I looked for more information regarding the theory and found that the background research was completed years ago and presented at an ECTRIMS meeting in 2007.

First, the article abstract:

The Will Rogers phenomenon: the effect of different diagnostic criteria

by Maria Pia Sormani in the Journal of the Neurological Sciences, Volume 287, Supplement 1, Pages S46-S49 (December 2009)

The ‘Will Rogers phenomenon’ is an apparent epidemiological paradox named after a remark made by the humorist Will Rogers about migration during the American economic depression of the 1930's: “When the Okies left Oklahoma and moved to California, they raised the average intelligence level in both states.” In 1985, Alvan Feinstein proposed the name ‘Will Rogers Phenomenon’ to describe the ‘stage migration’ he observed in patients with cancer.

Changes in the criteria for assigning patients to the various stages of a disease can produce spurious improvements in stage-specific prognosis, even though the outcome of individual patients has not changed. In oncology, new imaging tools allowed detection of cancer metastases before they became evident clinically. In consequence, more patients are classified into the more severe metastatic disease stage from the less severe single tumour stage. Such a ‘stage migration’ resulted in an improved survival of patients in both the less and the more severe disease stages.

Multiple sclerosis is also subject to the Will Rogers phenomenon since the introduction of the imaging-assisted McDonald criteria. Given the sensitivity of magnetic resonance imaging for detecting disease activity, which is believed to be five to ten times greater than that of clinical assessment, the McDonald criteria are expected to allow earlier diagnosis, hence the ‘stage migration’ relative to the previous Poser diagnostic criteria.

Because of the ethical problems associated with carrying out placebo-controlled trials in multiple sclerosis owing to the wider availability of therapeutic options, the use of historical controls groups remains an attractive option. However, the Will Rogers phenomenon, which is recognised as one of the most important biases limiting the use of historical controls groups in experimental treatment trials, compromises the interest of this approach.

In this context, the use of different diagnostic criteria may generate spurious improvements in the medium-term prognosis of multiple sclerosis, which may be wrongly interpreted as treatment effects.

How was MS diagnosed decades ago versus now?

In 1965 Schumacher established a set of diagnostic criteria which are still commonly used by neurologists to make a clinical diagnosis of multiple sclerosis. The Schumacher criteria relied entirely upon clinical signs and symptoms:

• Neurological examination reveals objective abnormalities of central nervous system (CNS) function.
• History indicates involvement of two or more parts of CNS.
• CNS disease predominately reflects white matter involvement.
• Involvement of CNS follows one of two patterns:
• Two or more episodes, each lasting at least 24 hours and at least one month apart.
• Slow or stepwise progression of signs and symptoms over at least 6 months.
• Patient aged 10 to 50 years old at onset.
• Signs and symptoms cannot be better explained by other disease process.

The Schumacher criteria led to the development of the following designations:

• Clinically Definite MS was made if all the Schumacher criteria are fulfilled.
• Probable MS refers to RRMS symptoms where only one neurological symptom commonly associated with MS is found or if there is only a single attack and there was no better explanation for the symptoms.
• Possible MS refers to RRMS symptoms without documented signs or where the objective signs are insufficient to establish more than one site of CNS involvement.

In 1983, the Poser criteria were developed to incorporate laboratory studies in addition to clinical evaluation in making a diagnosis. Laboratory studies including cerebrospinal fluid analysis, evoked potentials, and imaging studies provide paraclinical evidence to help the neurologist determine the existence and location of lesions.

The Poser criteria and designations are:

• Clinically Definite MS
• 2 attacks and clinical evidence of 2 separate lesions
• 2 attacks, clinical evidence of one and paraclinical evidence of another separate lesion

• Laboratory supported Definite MS
• 2 attacks, either clinical or paraclinical evidence of 1 lesion, and cerebrospinal fluid (CSF) immunologic abnormalities
• 1 attack, clinical evidence of 2 separate lesions & CSF abnormalities
• 1 attack, clinical evidence of 1 and paraclinical evidence of another separate lesion, and CSF abnormalities

• Clinically probable MS
• 2 attacks and clinical evidence of 1 lesion
• 1 attack and clinical evidence of 2 separate lesions
• 1 attack, clinical evidence of 1 lesion, and paraclinical evidence of another separate lesion

• Laboratory supported probable MS
• 2 attacks and CSF abnormalities

In 2001, the International Panel on the Diagnosis of Multiple Sclerosis (headed by McDonald) updated the diagnosis criteria to incorporate a set of standardized MRI criteria into the diagnostic process to support evidence of demyelination and show dissemination in space.

The McDonald criteria helped to establish a new diagnostic category for persons who have experienced only one relapse (or exacerbation) of MS-like symptoms, a Clinically-Isolated Syndrome (CIS). A person with CIS may or may not go on to develop MS which raises some controversy and confusion to the diagnosis and treatment of MS.

In 2005, the McDonald criteria was revised to help demonstrate dissemination in time, to clarify the use of spinal cord lesions, and to simplify diagnosis of primary progressive disease. The 2005 revisions were designed to simplify and further speed diagnosis, while maintaining adequate sensitivity and specificity.

The 2005 revised McDonald criteria include the following:

• At least two attacks with objective clinical evidence of at least two lesions
• At least two attacks with objective clinical evidence of one lesion plus dissemination in space shown on MRI, or two or more MRI lesions consistent with MS plus positive CSF finding or second clinical attack
• One attack with objective clinical evidence of at least two lesions plus dissemination in time on MRI or second clinical attack
• One attack with objective clinical evidence of one lesion, plus dissemination in space shown on MRI, or two or more MRI lesions consistent with MS plus positive CSF finding and dissemination in time shown on MRI or second clinical attack
• Insidious neurologic progression suggestive of MS plus one year of disease progression determined retrospectively or prospectively and two of the following: positive brain MRI result (nine T2 lesions or at least four T2 lesions with positive Visual Evoked Potential), positive spinal cord MRI result with two focal T2 lesions, and positive CSF findings.
  

Back to the Will Rogers' phenomenon in multiple sclerosis and the study presented at the ECTRIMS meeting in Washington, DC in October 2007 which was published in the Annals of Neurology in October 2008.

The availability of disease-modifying drugs has made placebo-controlled clinical trials ethically questionable which in turn makes the use of historical controls attractive (comparing the results to previous trials and observations). However, MS diagnostic criteria have seen dramatic changes over the last twenty years with the introduction of MRI in the diagnostic work up (see Poser vs. McDonald criteria above).

As mentioned in the abstract at the beginning of this post, using different criteria for classifying patients into various stages of a disease can modify the stage-specific prognosis, even though the overall disease course remains unchanged. Known as the 'Will Rogers phenomenon' this precludes the use of historical controls for treatment trials. The researchers assessed whether the Will Rogers phenomenon may affect multiple sclerosis (MS) prognosis when different diagnostic criteria are applied to the same patient population.

In 1995-2001, 322 consecutive patients with a clinically isolated syndrome (CIS) suggestive of MS were studied. They each underwent brain MRI within 3 months of clinical onset, at 12 months, and 5 years later. Number and location of lesions on baseline MRI and the development of new T2 lesions at follow up were evaluated.

After 1 year, each patient was classified as CIS or evolved to MS according to two different diagnostic criteria (Poser and McDonald). The outcome for prognosis was the time to reach an Expanded Disability Status Scale (EDSS) score of 3.0, compared in CIS and MS according to each set of criteria. Concordant and discordant diagnoses between the Poser and McDonald criteria were also studied.

309 patients were followed for a median period of 84 months (7 years). After 1 year, 16% of patients had developed MS according to Poser and 44% according to McDonald criteria. The probability to reach EDSS score 3.0 at median follow-up was 11% in CIS patients according to Poser and 7% according to McDonald criteria; it was 46% in MS patients according to Poser and 27% according to McDonald criteria.

[Statistics presented at ECTRIMS vary slightly from those above. Probability to reach EDSS score of 3.0 at year 8 was 15% in CIS patients according to Poser and 9% according to McDonald criteria; it was 52% in MS patients according to Poser and 34% according to McDonald criteria.]

An improved prognosis was seen in both groups (CIS and MS) for patients classified according to the McDonald criteria as compared to the Poser criteria. The group with discordant diagnosis had worse prognosis than patients classified as CIS by both criteria (p=0.014), but better prognosis than patients classified as MS by both criteria (p=0.01).

These observations support the hypothesis that “stage migration” (caused by different diagnostic criteria) may generate spurious improvements in the medium-term prognosis of MS and call for caution in using historical control groups in MS clinical trials.

Personal thought:
While I'm still digesting all the possible implications of this information, it seems that 16-18% of the 309 patients reached EDSS score of 3.0 in 7 years from the first neurological event.

So many MSers tell a similar story in that they were diagnosed at one point in time, but that their neurologists think that they've had MS years before then. If your prognosis is determined by how quickly you accumulate disability since diagnosis, then it does matter when and how you are diagnosed.

If we are being diagnosed earlier in the normal course of the disease, then our prognosis will certainly look before. It just feeds into the ambiguous questions of - Where would my MS be if I had did nothing? How do I know if my treatment choices are really working? How long have I really had this disease?

On a personal note, I often struggle with that last question. Is it since 2005 when I was diagnosed? Is it since 2000 when I went blind (but did not go beyond a brain MRI and vision fields testing) to determine optic neuritis? Is it possibly since 1993 (1994) when I first when through the MRI tube to rule out brain tumor?

I guess the state of my MS disability looks somewhat different if I could say I've had MS 5 years vs. 10 years vs. 16 years. But regardless, I am where I am right now no matter what. And I have to admit that I've been feeling better since the Rituxan infusions in November-December.