By Erica Verrillo
On December 9 and 10, 2014, the Pathways to Prevention (P2P) workshop was held on the NIH campus, in Bethesda, Maryland. The workshop was sponsored by the Office of Research on Women’s Health (ORWH), which is the division of the NIH responsible for ME/CFS. The purpose of the workshop was to identify gaps in ME/CFS research, and to address key questions about incidence, prevalence, diagnosis and subtyping.
The sessions on December 10 primarily focused on subtyping patients with ME/CFS. The presentations included research that has been conducted by Dr. Natelson on objective markers for identifying subgroups, on genetic markers of fatigue in cancer patients (Saligan), as well as Mady Hornig’s work on how immune markers interact with the microbiome and the brain. Dr. Klimas also presented her work on subgroups as well as a comparison between ME/CFS and Gulf War Illness in which two groups of patients presenting with identical symptoms showed very different test results.
The draft report of the meeting is available HERE. Members of the public can make comments, which will then be incorporated into the report, until January 16, 2015.
You can read the program book HERE.
You can read the meeting agenda HERE.
You can watch the December 9 presentations HERE.
You can watch the December 10 presentations HERE.
You can read information about panel members HERE.
You can submit comments HERE.
You can read a summary of the December 9 (Day 1) morning presentations HERE.
You can read a summary of the December 9 (Day 1) afternoon presentations HERE.
Session 4: What tools, measures, and approaches help define individuals with ME/CFS?
Heidi D. Nelson, M.D., M.P.H. – Evidence-based Practice
Center III: Treatment: Medications and
Complementary and Alternative Medicine
(15 minutes) (begins at 2:54)
Medical Director of Cancer Prevention and Screening, Providence Cancer Center, Providence Health and Service; Vice-Chair and Research Professor, Departments of Medical Informatics and Clinical Epidemiology, Pacific Northwest Evidence-based Practice Center, Oregon Health & Science University
Dr. Nelson described the results of the systemic review as regards complementary and alternative medicine. Nine trials met criteria for inclusion in the review, eight of which were fair quality and one of which was poor. All trials addressed the underlying pathology of ME/CFS, and all were small trials conducted in the U.S. and Western Europe. As with other trials, participants were mostly white, mostly female, and ranged in age from 35 to 50 years. Eight of the nine trials used the Fukuda definition.
Discussing the results of two Ampligen trials, Dr. Nelson pointed out that there were statistically significant results, with between 10% and 35% improvement along the parameters of exercise capacity and tolerance, as well as a decreased use of other medications. Trials of isoprinosine, galantamine, fluoxetine, acyclovir, and corticosteroids showed no significant differences. Two trials of rituximab showed significant improvements in physical function.
In the area of complementary medicine, only two trials met the criteria for good quality methods. Four were fair, and one was poor. Five trials used the CDC criteria and one used the Oxford definition. Interventions included homeopathy, acetyl-L-carnitine, pollen, acclydine, melatonin, phototherapy, distant healing (e.g. prayer, visualization), and low sugar/low yeast diet. There were no differences in any of these groups between placebo and intervention.
Drawbacks of these studies were that they were mostly small, various case definitions were used, outcomes measures used dissimilar instruments, harms were not reported, and studies did not report characteristics of responders vs non-responders. There were also limitations of clinical relevance. For example, few medications were tested more than once, and most trials used the Fukuda case definition, which may have selected patients with less impairment, as well as the Oxford case definition which may have selected for patients with chronic fatigue, rather than patients with ME/CFS.
The only treatments that achieved significant clinical improvements were rintatolimod (Ampligen)(immune modulator), rituximab (monoclonal antibody), and valganciclovir (anti-viral agent).
Beth Smith, D.O. – Evidence-based Practice Center IV: Treatment: Counseling Therapies and
(25 minutes) (begins at 17:22)
Associate Professor, Department of Medical Informatics and Clinical Epidemiology, Pacific Northwest Evidence-based Practice Center, Oregon Health & Science University
Dr. Smith’s talk focused on counseling and behavior therapies, as well as exercise. There were 14 trials included in the review, five of which were good quality, six were fair, and three poor. Seven trials used the Fukuda criteria, four used Oxford, and one used both the Fukuda and Oxford definitions. None used the case definition for ME.
On the forest chart, Dr. Smith explained that clinical significance is at 8. (Only four achieved clinical significance.) “Clinical trials favored CBT” she stated. [Editor’s comment: On the chart, only two of the five achieved clinical significance (Deale 1997 and 2001)].
There were six trials of exercise therapies. Half used the Fukuda criteria and half used Oxford. All of these favored exercise. Harms were poorly documented. In one trial, 47% of patients refused to repeat the CPET. More adverse events were reported compared with usual care or pacing. In addition, there were four head-to-head trials of exercise and/or counseling, Two used the Oxford case definition and one used the Fukuda definition.
Withdrawal rates for the studies were high. Total withdrawal was reported in two studies, and two out of four reported deterioration due to CBT. Neither CBT and GET were associated with serious harms, however more adverse events were reported in exercise studies, and drop-out rates were higher.
Who are responders and non-responders? Responders were younger and less impaired at baseline, and stayed within their energy envelope.
Some of the limitations of the research included exposure bias (self-reported responses may have been influenced simply by receiving more attention), expectation bias may also affect outcome (questionnaires may yield positive responses because participants are told that they will improve), thresholds for significant improvement varied across studies, trials did not measure level of activity, harms were inadequately reported, outcomes were not analyzed for subgroup comparisons, types of therapies differed across trials, and there was inconclusive information about responders vs non-responders. As with other studies, the Fukuda definition may have selected for patients with less impairment, the Oxford may have selected for patients without CFS, and no studies used criteria for ME that required PEM (the CCC and the ICC).
Future research needs include:
Reduce patient variability across studies by using consistent case definitions
Consider reporting on one or several agreed upon case definitions and retiring the Oxford case definition
Evaluate diagnostic testing across diverse groups
Report established measures of diagnostic accuracy as outcomes
Consider clinical subgroups of ME/CFS in reporting results of diagnostic tests
Evaluate the role of intermediate outcomes, such as biomarkers and CPET testing to identify subgroups
Future needs of treatment trials:
Enroll larger numbers of patients in trials
Improve design and execution of trials
Include longer follow-up
Report concurrent treatments and timing of treatment in relation to diagnosis
Report participants’ adherence with the treatments
Develop innovative ways to include more severely affected patients
Use consistent patient-centered outcomes across trials such as quality of life, fatigue, employment, improvement and recovery
Evaluate findings by patient characteristics, such as baseline severity, comorbidities, demographics, and symptom predominance (e.g. PEM, neurocognitive status, and autonomic function)
Determine the relationship between intermediate outcomes such as biomarkers, to patient outcomes
Determine differences in intermediate outcomes between patients with ME/CFS compared to patients with overlapping conditions
Review research available in the pediatric population, as well as among men and minorities
Determine the effects of ME/CFS across the lifespan and how treatment effectiveness and harm varies by age
Benjamin H. Natelson, M.D. – Reducing Patient Heterogeneity by Subgrouping
(20 minutes) (begins at 43:00)
Director, Pain and Fatigue Study Center, Mount Sinai Beth Israel; Professor of Neurology, Icahn School of Medicine at Mount Sinai
We would expect that any clearly defined set of criteria will reduce patient heterogeneity, but that any clinical syndrome will have multiple causes. ME/CFS is the tip of the iceberg. If we talk about idiopathic chronic fatigue, this is severe fatigue without other symptoms. Most pain syndromes are associated with fatigue. Severe fatigue is also associated with medical illnesses, such as COPD and heart failure, as well as post-infectious sequelae, breast cancer survivors, and neurologic diseases such as Parkinson’s, MS and stroke. Without the existence of a test we have to use stratification strategies to reduce patient heterogeneity.
Because this was a short talk, Dr. Natelson limited himself to presenting data from patients with and without FM, with no psychiatric diagnoses, and with low NK function.
When comparing patients with major depressive disorder (MDD) to those without MDD, Natelson’s group found differences across a number of brain studies, suggesting that patients with ME/CFS have some ongoing encephalopathy. Dr. Natelson is currently conducting a study which should confirm (or not) the hypothesis that this group of patients has an ongoing encephalopathic process producing the illness.
Those patients with low NK function have less vigor, more fatigue, poor mental processing, and poorer daytime functioning.
Dr. Natelson asked the question, “Is CFS the same as fibromyalgia?” Similarities do exist. There is a 37% comorbidity between CFS and FM. But importantly, if the ailments are the same, there should be no physiological differences. However, there are a number of studies that show substance P to be elevated in FM but not in CFS. CFS is common in people with sleep apnea but not in patients with FM. Patients with CFS do not respond to antidepressants, but FM pain does.
When comparing patients with ME/CFS to those with FM alone, patients with ME/CFS alone score worse on cognitive tests, and when they have comorbid FM, they score worse on every other test. Sleep studies also show marked differences. CFS patients show a remarkable increase in REM to wake transitions. [They wake up after dreaming.] This is very different from the mixed group, which shows sleep disruption. Something is forcing them to go from deep sleep to lighter sleep.
In terms of central serotonergic systems, FM, like MDD has a down-regulated central serotonergic system, while CFS has an upregulated serotonergic system.
In terms of fitness and conditioning, when patients with ME/CFS and with comorbid FM were matched with deconditioned controls, they found no differences in any variables using VO2max. Natelson’s group then did a sub-maximal test and found there was a major increase in stroke volume (blood pumped out of the heart) in the CFS plus FM groups as compared to either CFS only patients and deconditioned controls. In short, the CFS plus FM group has a significant increase in cardiac work despite the same workload.
If you add multiple chemical sensitivity (MCS) to patients with CFS alone and to those with CFS plus FM, other syndromes, such as IBS increase markedly. The majority of CFS patients alone have no comorbid psychiatric diagnoses. But when all three (CFS+FM+MCS) are present, 42% of patients have more than one psychiatric diagnosis.
There are several possible conclusions that may be drawn from the data. One is that pure ME/CFS may be a different illness from ME/CFS with comorbidities. But looked at another way, patients with several comorbidities may have a high likelihood of having others well. The high burden of illness may lead to increased psychiatric diagnoses.
The future is in biomarker research. Natelson’s group has looked at spinal fluid for proteomic [protein] analysis. The patients with CFS had 738 unique proteins relative to post-Lyme patients. Dr. Natelson concluded with a call for volunteers for his study.
Elizabeth R. Unger, Ph.D., M.D. – Measures: Patient-Reported and Physiologic
(20 minutes) (begins at 1:04:05)
Chief, Chronic Viral Diseases Branch, Division of High-Consequence Pathogens and Pathology, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention
Dr. Unger promoted the idea that using common measurements would facilitate research on identifying the natural history of the illness, as well as identifying the efficacy of treatments, and patient populations. Some of the challenges in measuring ME/CFS are the complexities of the illness, comorbidities, and individual patient variables such as type of onset, age, race and medications. Patient-reported outcome measures (PROM) are simple to use, are inexpensive, and can be used to measure change over time. There is also extensive literature on PROMs, because they have been widely used, particularly those that measure pain, fatigue and function.
Even though there are many case definitions, there are symptoms common to other illnesses, such sleep disturbance, pain, malaise, fatigue, and cognition, but post-exertional malaise may be unique to ME/CFS.
PROMIS is an NIH-sponsored initiative to create instruments for broad application across many chronic diseases and conditions. Their goal is to provide comparisons across conditions. Few of these instruments have been rigorously evaluated in the ME/CFS population, which means these instruments may not capture the all the aspects of ME/CFS. There are also limitations to using PROMs.
PROMs rely on self-reporting, and therefore are not objective. It is also difficult to establish what constitutes a clinically significant difference. Self-reporting does not take into account other factors that may affect a single domain (mood may affect function, etc.). Important aspects of physiologic abnormalities are not specifically represented in symptoms. Standardization of physiologic measures is needed as well.
What is needed?
Validation of self-reported measures in multiple well-designed studies
Improved PROMs for ME/CFS to better capture the patient experience
Standardized physiologic measures
Validation of physiologic measures and correlation with disease type and severity
Identification of subgroups
There is some ongoing work in this area. NIH and FDA have a work group on ME/CFS PROM. This is a direct outcome of the FDA Drug Development meeting. There are a couple of multi-site and biobank initiatives which will help generate a critical mass of data. Unger went on to describe the CDC multi-site study, and to share some preliminary results.
Dr. Unger concluded that there is a real need for measures. “Measure is the start of everything… We need measures because case definition alone is going to be insufficient.”
Leorey N. Saligan, Ph.D., R.N., CRNP, FAAN – Innovative Approaches in Fatigue Research:
Phenotyping, Biomarker Discovery, and Statistics
(18 minutes) (begins at 1:26:25)
Investigator, Symptom Management Branch, National Institute of Nursing Research, National Institutes of Health
Dr. Saligan’s research area is the investigation of fatigue. His initial focus was on fatigue in sarcoidosis patients. Later he began looking at fatigue in prostate cancer patients who were receiving radiation therapy. He noticed a significant increase in fatigue midway through treatment, which is around day 18 to day 21 of the patients’ entire course of treatment. The fatigue persisted to the ends of treatment (day 38-42). But 40% of patients did not develop fatigue. After doing gene expression microarray, they found that there was a differential expression of genes, in particular mitochondrial-related genes and T-cell-related genes were down regulated in the fatigued patients. Upon further analysis, most of the genes were found in cancer, neurological diseases and psychological disorders. They also found differences in proteomics, specifically apoliprotein A1 and E were related to high fatigue.
What was fascinating about Saligan’s research is that his group was later able to predict which patients would develop fatigue using 14 genes. (High fatigue patients had lower expression of the genes than low fatigue patients.) Although Dr. Saligan did not relate his research directly to ME/CFS the implications are that using proteomics and gene expression, it may be possible to predict who will develop ME/CFS after viral infections. (This relates to Dr. Jason’s study of university students who contract mono.)
Mady Hornig, M.D., M.A. – The Role of the Immune System in ME/CFS
(20 minutes) (begins at 1:46:06)
Director of Translational Research, Center for Infection and Immunity; Associate Professor of Epidemiology, Columbia University Mailman School of Public Health
Dr. Hornig’s group is interested in examining the microbiome, and, in particular, looking at the metabolites that may be neuroactive, and play an important role in the development of chronic fatigue syndrome. We have an enormous number of neuropsychiatric disorders that are considered to be due to an interaction between the brain and the immune system, ranging from autism, depression and schizophrenia, as well as ME/CFS. [Editor’s comment: Dr. Hornig has placed ME/CFS on the same spectrum as several neuropsychiatric disorders. However, none of these neuropsychiatric disorders occur in epidemic form.] There are a wide range of infectious factors that we have thought of as triggers for this process, not only gut microbes, but also herpesviruses, and a wide range of other viruses.
Microbes are important for normal brain function. Germ-free mice have abnormal cognition. The brain-gut axis can create a variety of effects in the body. Dietary products interact with gut bacteria to produce tryptophan, which is the building block for serotonin, which is, in turn, important in key functions, including sleep, sex drive, and mood. Tryptophan can also be broken down by immune factors like interferon gamma, and TNF alpha, as well as by glucocorticoids. This means that tryptophan can be degraded by the immune system, and also by the stress response (HPA axis). That degradation takes tryptophan down a pathway that can be neurotoxic, one that regulates mood, thought processes, as well as memory. We believe that the degradation of tryptophan may play an active role in the biochemical changes that we see in ME/CFS.
The lining of our intestines is flooded with bacteria that condition our immune system. Segmented filamentous bacteria (SFB), for example, are pro-inflammatory. They increase Th17 cells that are involved in autoimmune and inflammatory disorders like RA, and are implicated in pain syndromes such as FM. The absence of this type of bacteria allows the increase of the type of cells that quiet down the inflammatory process, the T-regulatory cells.
Autoimmunity is also regulated by gut microbes, and is altered by association with genetic factors. Microbes may also create autoantibodies that mimic brain proteins. There are a number of studies that suggest that an autoimmune process may be present in a subset of ME/CFS. Certain microbes are butyrate producers, and when these are reduced you see the development of classical autoimmune disorders like type 1 diabetes. Probiotics can influence this scenario. In animal models, lactobacillus johnsonii will increase tryptophan, and one can also see this in human studies where brain dysfunction in certain conditions can be prevented by using comensal bacteria.
Hornig’s group is working on several studies involving pathogenic agents (e.g. viruses and mycoplasma). There is very little to suggest a specific pathogen in CFS, although some viruses are found in lower numbers among patients than controls. Hornig suggested that perhaps certain viruses are important in calming the immune response, and that patients with ME/CFS may be deficient in them.
What Hornig’s group has found is that early (three years or less) cases of CFS have more pro-inflammatory cytokines than long-term patients. The immune system is highly activated in early stages, whereas in longer cases immune exhaustion may set in. Treatment considerations for the short-term group may involve antibodies to IL 17, which are currently being used for RA. Also in early stage ME/CFS allergy-related cytokines, and Th2 cytokines are up-regulated. So, there is a profound upregulation of anti-inflammatory cytokines as well. There is a 104-fold increase in interferon gamma, which is a very important cytokine in viral immunity, in short-term CFS. Plasma serotonin is also undetectable in short-term patients, suggesting that there is tryptophan degradation.
[Hornig was cut off before she could explain the remainder of her slides.]
Q: Dr. Elk, panelist: This question is for Dr. Smith and Dr. Nelson. In your analysis, were Dr. Natelson’s and Dr. Buchwald’s studies included?
A: Dr. Nelson: Dr. Natelson’s were not. Dr. Buchwald’s, I’d have to look that up. I will put out there that if there are studies we have missed that are not on our exclusion list, we would love to look at them.
Q: Kathleen M. O’Neil, panelist: In response to the same two literature reviewers, with response to a question from yesterday about pediatrics being excluded, is this because you found very little or because it was just the charge?
A: Dr. Smith: It was just the charge. We didn’t look at pediatric literature at all.
Q: Dr. Rasmussen, panelist: Dr. Saligan, in your signature of the genes of the patients that would go on to develop fatigue, did you test that signature in other indications of fatigue as a potential diagnostic marker?
A: Dr. Saligan: That is the next phase of our investigation.
Q: Dr. Green: There are a number of studies that are ongoing. Do we have any of the preliminary results of studies that have been accepted or are under peer review?
A: Dr. Unger: We are preparing a manuscript for the IOM, but it’s not under review yet.
Q: Dr. Green: There are a number of measures that were talked about, and so, the reliability and validity of those measures in an ethnically diverse population, has that work been done?
A: Dr. Unger: That is a critical area. Our studies are relying on tertiary care, so it is a largely white population.
A: Dr. Natelson: I just wanted to mention that one of the groups in the CDC cooperative study we are trying to reach are the housebound bedridden patients. If you in the audience know of anyone we want to find these patients, and we want to include them by going to their bedside.
Q: Dr. Green: Have we reached out to rural Americans?
A: Dr. Unger: There’s been very little done, but our study in Georgia did identify an increased risk in the rural population. That’s another critical need. Patients don’t have access to care anywhere, let alone in rural areas.
Q: Dr. Green: In regards to that, so we’re seeing increased morbidity, are we seeing increased mortality? I’m thinking about the study about older women having decreased longevity.
A: Dr. Unger: We don’t have any data on that.
A: Dr. Hornig: I’m thinking of the Medicare study in which there was an increase in lymphoma in individuals over 65, that was in men. There was an 85% risk of lymphoma in individuals with a history of chronic fatigue syndrome. Alterations in the immune system over a 10-year or longer period is an important consideration.
Q: Dr. Green: Does anyone know of any data that would suggest that there is an increase in mortality?
A: Dr. Unger: No –
A: Dr. Natelson: Yes –
Dr. Unger: But there is, umm…evidence of metabolic syndrome being increased in these patients. So, they have a burden of cumulative chronic illness.
A: Dr. Hornig: There is that one cancer study, and we certainly need more in that area.
Q: Penney Cowan, panelist: Dr. Smith, how long was the time between the original test and the retest (referring to the CPET used in the exercise test)?
A: Dr. Smith: I’d have to go back to the specific trial. It was longer than three months.
Q: Dr. Jason (2:17:17): In patient polls, the CBT studies have been somewhat varied in results from some of results of the more empiric trials. Patients feel that these interventions are not helpful. In addition, the cure rate is often talked about in some of those CBT trials, do you think people are actually cured of this illness? And finally, the methodological issues of some of these trials, for example the PACE trial, of having differential standards coming into the trial, versus recovery, and how those standards have changed. Are those methodological issues considered in the types of reviews that you’ve done? And, real quickly for Beth Unger, could you just mention some of the case definitions in the seven centers you are using to bring patients in, and whether you are still using the operationalization of Reeves in some of the work at the CDC in this trial and possibly others?
A: Dr. Smith: As to why benefit is noted when we looked at the trials but patients don’t sense that, there are multiple trials with different results. In the combination of those trials, we did see improvement. Why patients don’t report this is that many of these different approaches are geared at helping an individual cope and manage and tolerate a chronic disease. Across the spectrum of chronic diseases different forms of therapy show improvement. That is a very different thing from cure. I did not include the evidence on recovery. The reason is that it was poorly evaluated. When it was, the outcomes were inconsistent and were flawed as far as measuring recovery, One large trial [PACE] had an inclusion of an SF-36 of 60. They then changed that criteria to less than 65 to improve recruitment. But when they looked at measures of recovery they used an SF-36 of 60 as one of the measures of recovery. So there is a contradiction on measures of cure or recovery. There are also Chalder fatigue scores of less than 4 as normal, and in another trial they looked at Chalder fatigue scores of 18. You could get to a score of 18, but does that really reflect recovery? In looking at the evidence, recovery was not well measured as an outcome. I would encourage researchers to set some standards for recovery. As far as methodological issues, we scrutinize the quality based on methodology. We used strict appraisal techniques.
Dr. Natelson (2:24): May I just put in a thought. The meta-analysis takes all the studies and looks at them equally but when I looked at those data I saw them as dichotomous. I saw the UK studies, two UK studies producing a very overwhelming positive response, but the US-based studies didn’t. So there seems to be a difference between the United States and the UK and I’ve asked Drs. Friedberg and Jason about this, and we really don’t understand, but I think that that’s another way to look at it. Are they successful or not, and then try to come up with reasons.
Dr. Unger: What case definitions were the clinicians using? They are using both the Fukuda and the CCC. I think in their own work, they don’t always separate out perfectly. Clinical work is guided by what the patients are once you’ve got them in the door and [physicians] are evaluating them. The instruments we used were operationalized from the Empiric case definition. So, we are still looking for the best way to operationalize.
Dr. Green: This brings up an important question. The Oxford definition, does anyone disagree that the Oxford definition needs to be retired? [No response. Laughter]
Q: May Ann Fletcher, NOVA: There was a paucity of men in clinical trials, this would happen naturally because the majority of patients are women. But we have found there are important differences between men and women. We’re looking for 50 men with ME/CFS who will participate in our study. Please contact me at firstname.lastname@example.org
Dr. Green: So, is there a question here?
Q: Joe Landson: I am interested in the role of mitochondria and the role of immunity. Since there is a little literature showing higher rates of cancer in ME/CFS, it seems there is a need for studies.
A: Dr. Hornig: That makes a lot of sense, there is likely to be a dysfunction in mitochondrial function in autoimmunity. There can be various antigens that are presented to the immune system in the context of oxidative stress than can lead to a Th2 autoimmune response. Perhaps we can look at certain markers related to dysregulation of the cell cycle as well as cancer markers in an in vitro setting, and then apply that to a community setting.
Dr. Saligan: My lab scientists are looking at that pathway. Our initial data have found that there is a big link between mitochondria and inflammatory immune response. That might be related to the ability of the body to go back to homeostasis after stress.
Q: Dr. Fred Freidberg (2:33:22): I’d like to think we have much to learn from fatigue in different disorders. This is the essence of collaboration, that we talk to people who are also studying severe, disabling fatigue in other diseases.
Q: Robert Miller, patient: Yesterday I asked the panel to prioritize immune and autoimmune research in their recommendations. I’d like to ask you how important you think immunological research is in finding causation and treatments for this illness.
A: Dr. Hornig: I think that it is likely to be a very essential aspect. We are looking at the disrupted host-immune response. The disruption has many biochemical mediators but the immune system is a key factor that is relating to microbial triggers as well as “stressors” of other sorts, psychosocial stressors. Cytokines increase to as large a degree when you are preparing for an exam as they do when you have just had an influenza vaccine. In terms of the microbiome, I don’t think it’s a panacea, take a probiotic or a fecal transplant and you are cured, but we have not been able to broach the tip of what that will be able to show us. Perhaps different phenotypes will map to different immune signatures.
Q: Dr. Green: There was some discussion yesterday about the MAPP network and as to what type of research we should be thinking about. In a world of finite resources, where would be the best investment if we are to think about this from a scientific standpoint?
A: Dr. Unger: I think the MAPP framework is really worth considering. It brings investigators together, and it allows for standardization and it allows for rapidly testing hypotheses in a broad population.
Dr; Natelson: I sort of agree with what Dr. Clauw said yesterday. While the R021 mechanism is fabulous it only allows very focused, specific questions. With networks like the one we had with NIAID for a dozen years, you have multiple sites, with a number of disciplines, to answer more complex questions. In what I understand of the MAPP network, ideas are multiplied across sites, which allows for a more complex approach. But then the issue is, what are the core studies?
Dr. Saligan: Standardization of phenotypes and case definitions would be ideal, especially from those who are working on biomarker discovery.
Q: Dr. Green: What type of studies would be critical?
A: Dr. Hornig: I think that the approach that helps us integrate across a range of ways to examine an individual – the biochemical, the metabolomic, the immune, the infectious and oxidative stress at the same time, particularly in a larger scale. Exercise studies, and imaging studies will help us. Although we call this chronic fatigue syndrome, we need to look at those early events. We need to get samples early on.
Dr. Unger: We need clinical best practices. We need standardized approaches to patients. So much of the damage occurs in the primary care setting. We need measures and standardization. That could come about with a fatigue research network
Dr. Natelson: We need to expand our patient pool. Right now we have to exclude patients. We need to identify phenotypes, again, focusing on the brain. I’d like to see brain studies that focus on neuro-inflammation. We need more patients in order to sort out the differences in biomarkers.
Q: Dr. Nacul (2:50): Question to Dr. Smith. Your evidence review shows quite a marked effect for psychological and behavioral therapies for ME/CFS. I think there is quite a bit of evidence that CBT is effective in other chronic diseases such as diabetes and cancer but only as an adjunct therapy and only as long as patients take specific medications for their conditions. So my question is whether you have any sense of how the results of CBT in ME/CFS compare with CBT in other chronic diseases.
A: Dr. Smith: I did not look at that data so I am unable to comment.
Q: Mary Dimmock, parent of patient (2:50:57): Dr. Unger, I know that there are differences in the different clinics in terms of the types of patients they see and that leads to some conflicts. My question is, when you do the multi-site study will there be any validation from one site to another, cross validation of how the diagnosis is done, or will you end up with a super-set of everybody that anybody diagnoses with CFS? You indicated in your report that 96-98% of patients have PEM yet doctor Jason mentioned that as few as 25% of patients using the Fukuda definition have PEM. How do you reconcile that and are we ready to finally accept that PEM is a required symptom?
A: Dr. Unger: When we have looked across the clinics, there were very few differences in any of the measures. We are not planning on validating one clinic’s diagnosis versus another. It’s really based on clinical experts that have a lot of experience with ME/CFS patients. The question of post-exertional malaise is a question of how you collect the data and how you apply the other case definition criteria. The problem is that when studies say they use the Fukuda case definition they don’t give any other information as to how they applied it. I think the range of PEM reflects as much how the case definition is applied as the case definition itself. No case definition is going to result in a homogeneous population. PEM is intimately tied with the severity of the fatigue. If you include fatigue and a number of symptoms, it is very difficult to avoid PEM. The patients endorse it very highly. In our population-based studies, over 80% of patients endorsed PEM. We do feel that it is a very important measure.
Mary Dimmock: The only thing I would add to that is to remember Dr. Nacul’s presentation yesterday in which he talked about the 163 permutations. There is any number of ill people who can fall into that bucket. And until we can get to the point that we are really clear on what this disease is, we’re not making any progress.
Dr. Unger: Other case definitions have similar multiple permutations.
Q: Unidentified patient: The cognitive behavioral therapy is not going to address the grief and the loss of every single day when you lose something new. This pressure to use CBT is absurd. I practiced for 15 years. CBT is not the answer to everything, and it needs to get out of the CDC’s recommendations. What are the studies besides CBT that have shown benefits for patients with ME?
Dr. Smith: We’ve done a review of all the studies, all the other counseling and therapeutic studies. All we’ve done is report what the studies show. We saw from the forest plot that the larger study predominated.
Robert Miller: The charge that is being placed on this panel is huge. Over the past few days, this illness has been referred to as CFS, chronic fatigue, pure CFS, acute CFS, and ME/CFS. What are we discussing?
Dr. Smith: I can say that the evidence report used the term ME/CFS and we included any trial that included ME, CFS or ME/CFS.
Dr. Nelson: The approach we took was very inclusive.
Nancy G. Klimas, M.D. – Identification of Subsets of Individuals
(20 minutes) (begins at 3:03:19)
Director, Miami Veterans Affairs Gulf War Illness and ME/CFS Research Program; Professor and Chairperson, Department of Clinical Immunology; Scientific Director, Institute for Neuro-Immune Medicine, Nova Southeastern University
Disclosure, I don’t earn anything and I’m broke. I was asked by the IOM to do a special disclosure that my thoughts are not the thoughts of the IOM.
Subgrouping is important for a lot of different reasons. First, we need to define a group that allows us to define pathogens. For example, you can subgroup by how high viral titers are. This is important, because in Montoya’s Valcyte study, above a certain titer you got an 80% response rate, and below you got a 5% response rate. So, he was able to show subtyping. You can also create subgroups to help target treatment.
Biomarkers can show severity. The FDA encouraged us to use symptom severity as our best marker in our clinical trials. Nonetheless, we should still be pressing on to find those surrogates (e.g. vital titers) because they do make for a tighter design. HIV is the best example of that. Before we had HIV testing, we subgrouped by CD4 count. It became our best surrogate, and our best outcome variable. (Klimas pointed out that animal models have not been explored much in ME/CFS, but that they have been used with success for Gulf War Illness.)
Objective clinical markers can also be used as ways to subgroup patients. These include neuroimaging, autonomic testing, sleep studies, physical exams, cognitive testing, as well as laboratory-based markers. Attention, concentration and cognitive speed are the best markers.
When you subgroup people by symptoms, you have another strategy. This can include symptom clusters, level of severity, duration, mode of onset, age at onset, comorbidities, and the trajectory of the illness. The Chronic Fatigue Initiative found nine factors. Of course, fatigue will always come up first when you define people by fatigue. Inflammation (including neuro as well as inflammatory problems) came up second. GI symptoms came up as the third best factor. Pain was fourth.
Looking at comorbid conditions in 960 patients, we found that the prevalence of any cancer was 8%, which is about twice the national average. Contrary to what we heard in other presentations, psychiatric problems are not the most common comorbidities. The most frequent comorbid condition was FM (61%), followed by menopause (60%). Non-Hodgkin’s lymphoma is also significantly elevated.
In terms of severity, lymph node tenderness tracked best. So, the inflammatory group tracked highest with severity.
Neuroimaging is an important player. Solid, good science has been coming out of this area. The PET scan study by Watanabe showing neuroinflammation points out that sometimes a single paper can be influential. Do we have to argue about neuroinflammation, or can we accept that systemic inflammation is attached to the brain?
Among laboratory studies for biomarkers, blood studies are the most common, but they don’t necessarily reflect every compartment (e.g. brain). Other studies include biopsies. John Chia showed that with a biopsy of the stomach he can identify 80% of chronic fatigue subjects. These patients had evidence of enterovirus in the biopsy. Pathogenesis work on toxic substances, proteomics, detoxification pathways, mitochondrial studies, autonomic and genomic and “big data” methods have been very productive. Exercise challenges are also very productive, because we know that exercise leads to relapse. We can map mediators of relapse after exercise, which allows us to virtually model the illness.
Interestingly, when genomic pathways are compared between GWI and ME/CFS there is only a little bit of overlap, in spite of the fact that these are two identically presenting conditions. As a clinician you cannot tell them apart, but they are different. GWI is a toxic exposure, not a virus exposure. CFS downregulates metabolism in response to exercise, which is the opposite of what it should be doing. Exercise turns off energy, it turns off cell metabolism. In GWI other cell-signaling pathways are used to try to compensate for things that aren’t working. So, what we have in CFS is a turned-off genomic process, and in GWI we have a turned-on genomic process. In both, inflammatory response is increased after exercise. Pro-inflammatory cytokines describe a subgroup of patients with CFS and should be used as a subgrouping strategy. (Klimas mentioned the Younger study on leptin production as a case in point.)
NK function is an excellent biomarker. It is the best biomarker in terms of consistency. Out of 17 studies, 15 saw a group effect. The NK cell dysfunction correlates well with immune activation and with T cell function. This can be completely restored in vitro with IL 15.
Renée R. Taylor, Ph.D. – Subtypes of CFS/ME: New Discoveries and Unanswered Questions
(20 minutes) (begins at 3:24:42)
Vice Provost for Faculty Affairs, Licensed Clinical Psychologist, Professor, Department of Occupational Therapy, University of Illinois at Chicago College of Allied Health Sciences
Dr. Taylor’s focus is on adolescents. This group has cognitive and learning impairments, academic decline, and declining social activities, relationships and extracurricular activities. Rates of association with mononucleosis in retrospective studies are high, and with mono you can control viral exposure. Adolescents also have more consistent activity and less psychiatric comorbidities.
Obviously, there are other infectious agents aside from EBV. We chose EBV because it is highly prevalent in the adolescent population. Adolescents who go on to develop ME/CFS have other symptoms during mono that are more severe. They also have more autonomic symptoms, and greater severity.
Of 301 adolescents with mono, 39 ended up with the diagnosis of CFS after six months. After 12 months, the sample dropped to 22 (7%). After 24 months, 13 (4%) still met the case definition (Jason’s modified CDC.) These figures are consistent with all the adult studies.
Both the CFS and non-CFS group were equally as active. The difference is that the CFS group was sleeping more, and had greater severity of fatigue. Basically, they were pushing themselves to keep up with their peers. These adolescents also had greater severity at mono onset, which is consistent with adult studies. In exercise studies, teens showed the gas exchange abnormality which is also found in adults with ME/CFS. They also found severe autonomic dysfunction in teens with ME/CFS both during the illness, and afterward at all intervals. There was no correlation between POTS and autonomic dysfunction.
There was also an increase in Th17 T cells that corresponded with Dr. Hornig’s findings in adults. Taylor’s group found a clear deviation from the usual role of cytokines, including a dysfunction in responses. Cytokines either over- or under-responding.
Conclusion: CFS is non-linear. There is no developmental trajectory. There is no predictable progression. It is person-specific. For some there is a relapsing and remitting course, for others they are always at the same level of illness. How do we measure this? We need to gather data in the moment, using current technology, because recall is not reliable. Taylor recommended using the Framingham heart study as a model.
Suzanne D. Vernon, Ph.D – What Outcomes Represent Improvement, Recovery,
Prevention, Benefits, or Harms?
(20 minutes) (begins at 3:46)
Scientific Director, Solve ME/CFS Initiative
Dr. Vernon’s talk focused on three key points:
Several outcomes reported by patients as clinically meaningful are not measured in the literature.
The knowledge base needs to be expanded by including data from similar diseases
Defining core concepts will be accelerated by using patients as partners
Dr. Vernon began with a recap of the FDA workshop held in 2012. The FDA defined ME/CFS as a serious illness, with an unmet need for drug therapies.
Vernon stated that when there is confusion and lack of consensus, it is important to go back to the core. In this case, the core is defined by the patient community. The task is to determine whether what the patients were reporting as important could be correlated with the research literature to devise biologically relevant subgroups.
In the research done by the Solve ME/CFS initiative, the “patient voice” ( a million words) was analyzed through a natural language processing approach. Sentences were parsed and broken down into parts of speech. The words were then disambiguated. (Single words can have a variety of meanings.) The words were then classified according to standard concepts.
Four main groups of symptoms clusters emerged from the analysis, with nine small clusters. Vernon’s group then compared these to 12,000 abstracts that shared the same lexical items. There were 10 frequently-reported symptoms that do not appear in the literature. Vernon says that because text mining reveals what is important to patients, it should be used to help direct research.
[Editor’s comment: Linguistic studies have shown that meaning can only be derived from context. It is impossible to disambiguate a word once the context has been removed. In addition, there is, in any language, a plethora of dialects and idiolects with a variety of meanings assigned to common words.]