“Future refinement of case definitions may wish to focus on requiring a small set of core symptoms such as post-exertional malaise, neurocognitive symptoms and possibly unrefreshing sleep”. Jason et. al.
There’s something right about doing a birthing a definition blog on Christmas. Disease definitions are about new beginnings – about clarity, about really zeroing in on a disorder so that researchers and drug manufacturers can be more effective. A good definition that works makes everything easier.
We’ve heard a lot recently from the P2P panel and earlier from the IOM project about finding or creating a definition for chronic fatigue syndrome (ME/CFS). Earlier this year fifty advocates and many in the ME/CFS signed a petition requesting the Dept. of Health and Human Services (DHHS) accept the Canadian Consensus Criteria at least in the short term as both the clinical and the research definitions for ME/CFS.
That idea was rejected as the federal government brought in outside researchers and in the case of the IOM panel, ME/CFS experts, to do their own analysis of the situation. Neither the AHRQ report nor the P2P draft report settled on a definition; instead the P2P report called for an international network to come to a consensus on a definition and then lay out a roadmap to refine and clarify it.
Make no doubt about it – the Fukuda definition that has ruled the roost for the past twenty years in ME/CFS research is on its way out. There’s too much momentum for change for a new definition not to occur. The big question is what it will be.
The Biggest Need – A Research Definition
In this blog we’re talking about a research definition – not a clinical one. Research definitions do not describe the disease – clinical definitions do. Usually short and concise, research definitions ensure that people with the core aspects of a particular disease – and only that disease – make their way into research studies.
Unlike a clinical definition that seeks to include all types of patients in it, a research definition will exclude some people who have the disorder. It may not look particularly representative. It may not include some prominent symptoms and may include less prominent symptoms. The makeup of a symptom-based research definition will depend on not on the most recognizable symptoms, but what symptoms are unique to it relative to those found in other disorders.
The research definition is the greatest need and the IOM panel is not tackling that – perhaps for fear of opening a can of worms. But what if we’re already pretty close to a research definition – one that needs little modification? One that was derived not through a consensus approach –something the AHRQ panel specifically warned about – but using a statistically rigorous approach? What if the basic components of a new definition are sitting right under our noses? If Lenny Jason’s latest study is correct, we very well may be.
A Little History – A Little Mystery
Leonard Jason, M. Sunnquist, A. Brown, M. Evan et. al. 2014. V Examining case definition criteria for chronic fatigue syndrome and myalgic encephalomyelitis. Fatigue: Biomedicine, Health and Behavior. Vol 2, No. I, 40-56.
Jason has been comparing the different definitions for several years now and in the first part of his paper he went over the definitions and what his studies have found.
Holmes Definition (1988) – The first definition, the Holmes definition, required a person to have 8/11 symptoms. On the face of it, the definition with its sweeping symptom set (fever or chills, sore throat, lymph node pain, muscle weakness, muscle pain, post-exertional malaise, headaches of a new or different type, migratory arthralgia, neuropsychiatric complaints, sleep disturbance and sudden onset) looks quite good, but it was discarded when it was found that requiring 8 symptoms inadvertently selected out patients with high rates of psychiatric diagnoses.
Fukuda (1994) – the collapse of the Holmes definition lead to the Fukuda definition which required, along with life-altering fatigue, a patient have four of eight symptoms. By not requiring core symptoms such as post-exertionaly malaise, however, the Fukuda definition ran the risk of allowing people without core ME/CFS symptoms into ME/CFS research studies.
Canadian Consensus Criteria (CC) (2003) – created by physicians and researchers this definition focused on core symptom domains found in ME/CFS and required that patients experience post-exertional malaise plus seven core symptoms.
International Consensus Criteria ICC) (20 ) – an updated version of the CCC, the ICC again highlighted post-exertional malaise and required ME/CFS patients to have eight symptoms.
A Little Mystery – Jason’s comparisons of people meeting the CCC and ICC and Fukuda criteria indicate that the CCC and ICC definitions do select out patients who have the core symptoms of ME/CFS (Fukuda does not). They also more accurately identify people who have ME/CFS, and they identify patients who are more severely ill than the Fukuda definition does.
Jason’s studies also indicate, however, that the CCC and ICC select out patients with significantly higher rates of psychiatric disorders than Fukuda does probably, because, like the Holmes definition of 1988, they require patients to have more symptoms.
It’s perhaps telling that Jason, our resident expert on definitions and characterizing symptoms did not participate in the creation of either the CCC or the ICC. The two definitions look good on paper and they provide superb characterizations of the illness, but the studies suggest they’re better off being used as clinical definitions for doctors than as research definitions.
Importantly, none of these definitions assess symptoms according to their frequency/severity – they simply require that a symptom be present – and that will end up making a big difference.
Appearing in the IACFS/ME’s “Fatigue: Biomedicine, Health and Behavior journal earlier this year, Lenny Jason’s study “Examining case definition criteria for chronic fatigue syndrome and myalgic encephalomyelitis” took a big step towards producing a statistically derived research definition for ME/CFS.
The study used data from over 300 participants in the SolveCFS BioBank and healthy controls. Each took the DePaul Symptom Questionnaire that had been developed specifically for ME/CFS by Jason at DePaul University. The questionnaire uses frequency (from 0=none to 4=all the time) and severity (from 0=not present to 4=very severe) measures to assess the importance of 54 symptoms. Someone, for instance, who infrequently has mild headaches will have a much lower score (n=2) than someone who often has severe headaches (n=8).
The symptoms in the questionnaire include everything from fatigue and extreme tiredness to pain to stiffness in the joints to sensitivity to noise to night sweats to alcohol intolerance to fever.
Since having healthy controls make it into ME/CFS research studies as patients is not a good thing, Jason first used the DePaul Questionnaire results to determine what percentage of ME/CFS patients and healthy controls met the Fukuda, Canadian and International Consensus Criteria for ME/CFS. First he assessed the definitions as they are – simply requiring that a symptom be present or not. Then he introduced severity into the equation by requiring that for a symptom to count it had to be present at least half the time and be of moderate severity.
He also used an algorithm that identified which symptoms best differentiated the ME/CFS patients from the healthy controls.
The Fukuda Definition Goes Down – Hard….
The Fukuda definition has taken some blows over the years, but probably none greater than Jason’s finding that a third of the healthy controls meet the Fukuda criteria for CFS. The Fukuda definition’s criteria that anyone having 4/8 symptoms plus fatigue apparently didn’t prove much of a bar.
When Jason required that a symptom be moderately severe and present at least half the time to fulfill the Fukuda criteria, however, only 5% of the health controls were diagnosed with having ME/CFS.
Canadian Consensus Criteria (CCC) and International Consensus Criteria Take Their Hits
Before you laugh too much at the poor Fukuda definition, get this – the CCC and ICC fell into the same trap – albeit to a lesser degree. Twenty-one percent and 16% of the healthy controls meet the CCC or ICC criteria for ME/CFS if the symptoms in those definitions were simply required to be present.
If, however, the symptoms were required to be present at least half the time and be of moderate severity just 4% of the healthy controls met either criteria for having ME/CFS.
The moral of the story is that adding frequency and severity characteristics to a symptom-based definition is critical.
Jason also found that the four core symptoms that occurred with the highest prevalence in ME/CFS patients were fatigue, post-exertional malaise, neurocognitive problems and unrefreshing sleep. Pain, autonomic, immune and neuroendocrine symptoms were not as common. Muscle pain was moderately common in the ME/CFS patients but since it was fairly common in the healthy controls it would not be a good distinguishing factor.
The Perils of Consensus
It’s clear that the two consensus criteria (CCC,ICC) do pluck out more severely ill patients than Fukuda, but they fail on two aspects:
- significant percentages of healthy controls meet their criteria for ME/CFS
- ME/CFS patients that meet the criteria tend to have higher rates of psychiatric disorders.
It’s findings like this that explain why consensus definitions or findings are viewed with some skepticism. Consensus findings that appear on the face of it to make perfect sense, as the CCC and ICC do, don’t always work.
The Next Research Definition?
Creating a definition that incorporates frequency and severity into it and doesn’t require a lot of symptoms is clearly a priority. Jason’s data mining analysis indicated, however, that so long as you require that a symptom be present at least half the time and have moderate severity, just three symptoms are needed to correctly differentiate almost all the ME/CFS controls from the healthy controls.
Three symptoms below
- fatigue or extreme tiredness
- inability to focus on more than one thing at a time
- experiencing a dead or heavy feeling after starting to exercise
correctly identified 95% of ME/CFS patients and controls.
Fatigue, Cognition and Post-Exertional Malaise Highlighted
Note that the symptoms correspond with three of the four core symptom domains – fatigue, cognition and post-exertional malaise – Jason identified in this study. Symptoms from these domains are likely going to make up the next research definition for ME/CFS. Unrefreshing sleep might make it or it might not; apparently it’s on the cusp. Although pain is ‘moderately common’ in ME/CFS it’s not going to make it into the definition. Sensitivity issues are apparently not common enough to make into the core definition for ME/CFS but Jason suspects they will identify subsets. (He has a formula for identifying subsets as well.)
It’s interesting that fatigue, a common symptom, made the cut. This is presumably because while fatigue is common, more severe fatigue is not that common and fatigue that’s associated with problems in multitasking (problems focusing on more than one thing) and experiencing a heavy feeling after exercise is very uncommon in healthy people.
Some of the fatigue or exertion-based symptoms that didn’t make the cut include “next day soreness or fatigue after non-strenuous, everyday activities”, “being mentally tired after the slightest effort” and “minimum exercise makes you physically tired and feeling physically drained or sick after mild activity”.
Either these symptoms were too common in the healthy group or too uncommon in the ME/CFS group for them to differentiate the two groups. My guess is that heavy feelings after exercise are so common and so noticeable that that symptom easily made the cut despite (or maybe because of) the fact that it’s kind of odd. Who, after all, hasn’t experienced heavy legs after they over-do it?
New Era in Definition Creation
This large study of rigorously identified ME/CFS patients opens a new era in how to produce a definition. The well-loved consensus definitions played an important role in clarifying the key symptoms in ME’/CFS, but they are not appropriate as research definitions. Statistically determined definitions, on the other hand, will likely elicit trust from all the important stakeholders (drug companies, researchers, funders) in the medical world.
Next up Jason and company need to compare the symptoms found in ME/CFS with those in closely related disorders such as fibromyalgia and multiple sclerosis. Jason got an NIH grant to do this study. Hopefully with the P2P report calling for a new definition, he’ll get another grant to do the next one. That study is going to be interesting.