Puzzle Pieces

Let’s play a game. Imagine you have a large puzzle that makes an Impressionist picture of a colorful cottage-style garden. You can put it together as long as you have the picture on the box. First you assemble the lower left corner, all lavender and yellow flowers. Another section of red roses sits somewhere in the middle. Near the upper right corner is a section of white and gray paving stones, and you also put together an area of green herbs although you are not sure where it goes yet. This puzzle will take a lot of time to solve, but with the finished image on the box you know that you’ll put it together eventually.

Now imagine the box is gone. All you have is a white/gray blob, and lavender/yellow section, the red rose section and another green blob. The rest of the pieces are all mixed up together, and while you can separate out some edge pieces and consolidate others by color, without the box you cannot even be certain what the final picture should look like. It’s frustrating, isn’t it, to have all those pieces on the table and not see how it fits together or even know for certain that you have all the pieces. That’s the feeling I got reading the American Family Physician’s article on Chronic Fatigue Syndrome: Diagnosis and Treatment. I dissected the AAFP patient information sheet on CFS in a recent post, but now I think it’s important to examine this review article by the same authors. The article attempts to present a finished picture of CFS for family practitioners, but so many pieces are missing that the paper bears little resemblance to the CFS I live with.

Generally Speaking

“Chronic Fatigue Syndrome: Diagnosis and Treatment” by Dr. Joseph Yancey and Dr. Sarah Thomas gives an overview of CFS for family physicians. They review the Oxford and Fukuda criteria, the basic lab workup recommended by CDC, and a list of exclusionary conditions. In a section on etiology of CFS, the authors quickly cover the immune system, genetics, psychosocial, adrenal system, and sleep/nutrition. Finally, the treatment section focuses on cognitive behavioral therapy (they say it works), graded exercise therapy (this works too), nonpharmacological (nothing really helps) and pharmacological treatments (these don’t work either).

In the authors’ defense, there are significant space limitations in the American Family Practitioner journal: 1,500 to 1,800 words in the case of clinical review articles like this one. There is no way to include all the information about CFS that family doctors need in such a limited space. It also appears that neither Yancey nor Thomas are CFS experts, based on the very limited information I could find online. I emailed Dr. Yancey, the corresponding author for the paper, on October 24th with a few questions but to date I have not received a response.

Method Madness

Drs. Yancey and Thomas describe their research methods as follows:

A PubMed search was completed using the MeSH term chronic fatigue syndrome. The search included randomized controlled trials and clinical trials in English from the past 10 years. We also searched the Cochrane database, Essential Evidence Plus, the National Institutes for Health and Clinical Excellence guidelines, and the Centers for Disease Control and Prevention Web site. Search date: August 26, 2011.

This methodology accounts for some of the missing pieces. First, anything published after August 26, 2011 was not captured in the search. That includes the IACFS/ME Primer, NCI’s paper on the risk of cancer among elderly CFS patients, the ME-ICC criteria, and the Rituximab trial. But before we forgive the authors’ oversight of these papers based on the date of their literature search, consider a curiosity I found in the paper references. The authors cite one paper published after August 26, 2011: The FITNET trial of internet based CBT for adolescents with CFS is included as reference Number 27. Does that strike you as odd? If the authors truly limited themselves to the references found on August 26, 2011 then this paper should not be included. Furthermore, of all the papers published after August 2011 to include in a review of CFS treatment and diagnosis, why was a CBT paper the one cherry-picked by the authors?

Even within the boundaries of the search methodology, the authors missed some papers that would have been helpful in their overview sketch of CFS. I attempted to recreate the authors’ search in PubMed, and found more than 1,300 clinical study papers alone. These include all of the letters critical of the PACE study and Tom Kindlon’s many letters and papers on the potential harms and inaccuracies in CBT/GET studies. Other important papers such as the spinal fluid proteome by Schutzer, et al., the differential gene expression post-exercise paper from Light, et al., and the cytokine network modeling by Broderick, et al. were captured in the PubMed search but did not make it into this review paper.

Finally, there are several seminal papers that are not returned in the PubMed search. The Journal of Chronic Fatigue Syndrome published the Canadian Consensus Criteria by Caruthers, et al., in 2003. This case definition is gaining broad acceptance among policy makers and researchers, but it does not show up in a PubMed search because the journal was never indexed in Medline. Another example is the Van Ness, et al. study showing the significance of two-day exercise testing in differentiating CFS patients from controls. This is a critical paper, suggesting a possible diagnostic test (albeit an extremely unpleasant one) for CFS. But because the journal was never indexed, these papers do not show up in a PubMed search and so non-experts like Yancey and Thomas never see them.

Cognitive Bias

I do not know what Dr. Yancey and Dr. Thomas believe about CFS, including whether they believe the illness is primarily psychological in origin. After reading this paper, however, I fear this may be the case. I can best illustrate this through examples.

In the opening paragraph of the article, the authors say “CFS is often mentally and emotionally debilitating, and persons with this diagnosis are twice as likely to be unemployed as persons with fatigue who do not meet formal criteria for CFS.” What about physically debilitating? If the authors recognized the physical disability experienced by many CFS patients, and the physical suffering of all of us, wouldn’t they mention it in this paragraph? This simple omission is a very subtle way to communicate that people with CFS are not physically ill.

There is a brief discussion of the case definition in the paper. According to the authors, the 1988 CDC definition focused on physical symptoms, and the 1991 Oxford definition “emphasize mental fatigue over physical symptoms.” But the criteria, printed as Table 1 in the article, require fatigue to be “severe, disabling, and affects physical and mental functioning.” I’m no fan of the Oxford definition, but even I can see the requirement of physical disability. Again, Yancey and Thomas gloss right over the fact that CFS has serious, physical symptoms.

In discussing the biopsychosocial model of CFS etiology, the authors say: “CFS is often associated with depression, which has led many physicians to believe that CFS is a purely psychosomatic illness. Evidence supporting this conclusion is lacking.” Fair enough. But then they say, “Strong evidence suggests that childhood trauma increases the risk of CFS by as much as sixfold.” Sigh. I covered this in my dissection of the patient information sheet. Childhood trauma may have physical systemic affects, but I am not aware of any evidence showing that CFS patients have higher rates of trauma compared to patients with other illnesses like MS or lupus or diabetes or cancer. In my opinion, it is misleading to single out childhood trauma as a risk factor for CFS in the absence of such evidence.

The authors devote space and attention to CBT and GET studies, and this is understandable given the fact that CBT and GET treatments have received the most study in CFS. CBT “can help persons with CFS recognize how their fears of activity lead to behaviors that ultimately cause them to feel more fatigued and disabled.” It is true that CBT can help patients correct activity avoidance behavior, but in my experience this is a very small minority of patients. Even the CDC, target of so much criticism, does not describe CBT this way. The CDC says: “CBT can be useful by helping them pace themselves and avoid the push-crash cycle in which a person does too much, crashes, rests, starts to feel a little better, and then does too much once again.” This is a more appropriate description of CBT that acknowledges the importance of self-management and the prevalence of the push-crash cycle, as opposed to the activity avoidance highlighted by Yancey and Thomas.

Graded exercise therapy is very controversial for CFS patients, mainly because traditional GET uses a scheduled increase process as opposed to a patient-driven increase process based on symptoms. Not surprisingly, this issue is not discussed in the paper. The authors do mention that a heart rate monitor can be used to avoid overexertion during exercise, but there is no mention of the body of evidence on CFS exercise testing and pacing methods. They even cite a study that suggests improvements in GET do not correlate with increases in exercise capacity, suggesting that GET may actually work by “decreasing symptom-focusing behavior in persons with CFS.” Pacing, the only behavioral technique that truly helps CFS patients, is not mentioned by name, although the authors do say:

Patients should be encouraged to take rest periods as necessary, and to practice relaxation techniques. Although there is no evidence these modalities are effective, they are unlikely to be harmful and may be helpful.

Neither CBT nor GET is curative because it does not target the underlying mechanism of illness. CBT is not curative for cancer or heart disease either, for the same reason. Drawing the conclusion that these therapies are not curative because of the patient is a fallacy, but this is the conclusion that Yancey and Thomas suggest:

Despite the positive results of CBT and graded exercise therapy, the effects are usually moderate and rarely lead to resolution of CFS. Patients with poor social adjustment, a strong belief in an organic cause for fatigue, or some sort of sickness benefit (i.e. financial incentive) tend to have worse responses to therapy. Unlike with many other illnesses, membership in a CFS support group was associated with worse outcomes.

The study cited by the authors in support of these statements is this one from 2002. That study points out its own limitations: it uses the Oxford definition, lost 17% of the patients to follow-up, and did not actually measure the exercise capacity of the patients. But this is the kind of evidence that is sufficient for Yancey and Thomas.

The overall tone, selective quotation, and reference choices give me the impression that the authors believe CFS to be a psychological condition, at least in part. I do not know this for a fact, but if I read only this article about CFS and nothing else, I would believe that it is an emotional problem. It’s not just the amount of space devoted to the psychosocial research. The authors focus on the psychological elements to the exclusion of discussion of physical disability, post-exertional malaise, and the well-documented physiological findings in this illness.

Missing Pieces

There are huge gaps in this paper. Orthostatic intolerance, an issue for most CFS patients, is not mentioned at all. Post-exertional malaise is not explained, and no CFS exercise studies are referenced. The importance of medications and other treatments in managing sleep and pain issues is ignored, and pain is barely discussed at all.

This article illustrates a few pieces of the puzzle, mainly CBT, GET and the psychosocial model of CFS. A family physician reading only this article would not be able to separate chronic fatigue from CFS patients, and would understand almost nothing about the complexity of CFS. I found the tone to be generally hopeless: try therapy and exercise but it probably won’t help you much. Maybe a motivated physician would visit the CDC website (and this illustrates the importance of fixing problems in those materials).

No one will be able to assemble the CFS puzzle using the pieces in this article. Too much evidence is ignored, too much emphasis is placed on the psychosocial pieces, and there is very little information about how to manage the other symptoms of the illness. I know the full picture exists and I can identify the gaps. But a family physician who does not have the picture of the box will not recognize all that is missing and will never be able to assemble the pieces in a way that will help CFS patients.

I fear that doctors will rely on this article to provide the same kind of advice I received from doctors in 1994: keep going to the gym, staying in bed is the worst thing you can do, get some counseling, there is nothing else we can do to help you. This bad advice and hopelessness did not help me, and may have even hurt me by keeping me much more active than my body could tolerate. It was years before I found and received adequate care for pain, sleep, and orthostatic intolerance, and even more years before I found expert help for pacing and activity management. This article will do nothing to change the way doctors treat CFS, and will reinforce the destructive pattern already in place.