May 05 2011

XMRV Not Associated with Chronic Fatigue Syndrome

Chronic fatigue syndrome (CFS) is an enigmatic disorder. The primary symptom is debilitating fatigue that does not resolve with rest. Fatigue, however, is a very non-specific symptom, meaning that it can potentially result from many underlying causes. Anything that saps the energy our body uses to function will cause fatigue. In part CFS is a diagnosis of exclusion – it is based upon the presence of fatigue in the absence of any identifiable underlying cause. Therefore not everyone with chronic fatigue has CFS.

The non-specific nature of the clinical syndrome also frustrates our attempts to find the underlying cause or causes (it may, in fact, be many diseases all with a similar clinical presentation). So many things can potentially cause chronic fatigue – where do we begin. However, a chronic viral infection has long been suspected as a likely cause, at least in some cases.

Recent studies have found a potential association between CFS and a retrovirus called XMRV (xenotropic murine leukemia virus-related virus). This has led some desperate patients with CFS to take anti-retroviral drugs (used mostly to treat HIV – another retrovirus) off label in order to treat their CFS. But these studies, while intriguing, were considered preliminary.

Now a larger, more rigorous and comprehensive study has been published and has found no association at all between XMRV and CFS.

They report:

We collected blood samples from 100 CFS patients and 200 self-reported healthy volunteers from the same geographical area. We analyzed these in a blinded manner using molecular, serological and viral replication assays. We also analyzed samples from patients in the original study that reported XMRV in CFS. We did not find XMRV or related MLVs, either as viral sequences or infectious virus, nor did we find antibodies to these viruses in any of the patient samples, including those from the original study. We show that at least some of the discrepancy with previous studies is due to the presence of trace amounts of mouse DNA in the Taq polymerase enzymes used in these previous studies.

This looks like a very thorough study – they looked for the virus, or pieces of the virus, or antibodies to the virus, in their new subjects plus in samples from a previous positive study. They further identified possible evidence of contamination that could have caused the prior study to be a false positive. This seems like a pretty solid nail in the coffin of the XMRV hypothesis.

This episode reveals many of the features of medical research that are worth pointing out. First – much preliminary evidence will turn out to be wrong. That is the nature of exploratory research. This is partly due to the fact that preliminary studies tend to be small and less than rigorous. This is not due necessarily to being sloppy – the point of preliminary studies is to see if there is any potential to more elaborate research. These studies are testing the waters, to see if further investment in money and effort is warranted.

Preliminary studies are also used to learn how to better study a question. The most useful data to come out of such studies is often how to conduct better studies. This partly flows from having the broader community criticize the preliminary studies, probe for weaknesses and potential error. Then all of that criticism can be taken into consideration when designing later confirmatory studies.

But as I have discussed often on this blog, we now live in the age of mass media and the internet. Preliminary exploratory studies are made available to the public at large, including patients and their families. The media often does not do a good job of putting such preliminary evidence into context, so every study is a “breakthrough” or “may lead to a cure.” More often than not the preliminary research does not pan out, as seems to be the case with XMRV and CFS.

It is human nature to focus on the potential for benefit. People who are sick see a potential cure, and may not appreciate the potential harm of an experimental treatment. I am not aware of any comprehensive studies, but there is at least the reasonable concern that all of the medical treatments based upon preliminary studies likely cause more harm than good. Patients are playing the lottery, and everyone thinks they will be a winner, when in fact most people lose (almost by definition).

There is some threshold of evidence beyond which benefit starts to outweigh harm. That line is not sharp, and honest and informed experts can disagree (and do) about where exactly the line should be placed – but there is a line. It is the responsibility of the medical community and of regulatory bodies to minimize and avoid practicing medicine below that line (causing more harm than good). While the purpose of research is to clarify where the line actually is.

There are numerous examples today of medicine being practice below the line – taking anti-retroviral therapy for CFS is just one example. Another prominent recent example is using the liberation procedure for alleged CCSVI as a cause of multiple sclerosis. Stem cell clinics for a host of neurological disorders are another example, although that also appears to be an example of fraud, not just premature treatment.

I would also argue that the vast majority of what passes for so-called complementary and alternative medicine (CAM) also exists below the line, and is causing more harm than good. This is the very definition of CAM – treatments that have not be adequately established to be safe and effective, and many of which are so implausible they are likely to never be.

The XMRV story is just another example of why we need good science to guide medical treatment, and why we need to be patient to let the process work itself out to an adequate degree before we jump on a new treatment.

17 responses so far

17 thoughts on “XMRV Not Associated with Chronic Fatigue Syndrome”

  1. Joey H. says:

    I’m still on the fence about XMRV (like a good skeptic) but with this particular study I’m now listing to port. So many of the issues levied at other negative studies have been addressed here, such as the criteria used (patients had to meet both the Canadian and the Fukuda) and the inclusion of patients from the 2009 study.

    Unfortunately, this won’t convince the XMRV hardliners. (Sample comment: “I thought Singh was on OUR side.”) They’ll pick the study to death like all of the negative studies — funny, they rarely seem to apply such rigor to the more positive studies — and go on believing.

    Granted, patients have been subject to a lot of mistreatment individually and as a group, such as the CDC’s documented misappropriation of CFS research funding. So some paranoia isn’t totally unwarranted. But eventually, as the evidence against XMRV as a causative agent in CFS mounts, I hope we can all turn our “belief” back to the science, rather than any *single* study, institute, or researcher.

  2. SimonW says:

    Have seen a large number of studies where a particular group was found to have antibodies to, or other evidence of infection from a particular virus (whilst none or fewer such infections in control groups). Mostly flagged as causing various autoimmune, cancers, or disorders of unknown origin.

    Such things immediately have “plausible mechanism” and indeed some of these studies result in great advances like the HPV vaccine. So what aside from replication should we look for in such studies to help us know if it is really a promising advance in medicine or just noise.

    I mean as a Layman I think I understand what antibodies to a particular virus means, although I suspect like most things the medical meaning probably involves a lot of technology and a lot of possibility of error or cross reactivity when it is measured in the lab. But I’ve seen enough such papers from which nothing ever came, to think there may be systemic errors being made by the people writing these kinds of papers.

  3. I think the key is this – is there a firm consensus? Always look for the dissenting opinion. Consensus will emerge from replication and rigorous studies.

    There are multiple types of false positives in these types of studies. A correlation may be real, but not causative. We all are exposed to numerous viruses, and have antibodies to many viruses even if they never caused a recognized infection. So the mere presence of antibodies does not always tell us much.

    There are also technical errors. Even the tiniest amount of contamination can lead to false positives.

    For this reason, in addition to replications, we like to see the whole package put together. Not only evidence of a virus infection, but some indication that it is active and causing problems. An animal model to back it all up is nice also. And finally – clinical evidence that treating the infection improves the disease.

  4. HHC says:

    Would it be helpful to assist the client to restructure their daily routines they perform in spite of their multiple medical diagnoses?
    Recently, I spoke casually to one gal who told me while working that CFS is the basis of all her problems. I found out that her lack of sleep, coupled with her own medical problems, and her family’s problems seemed to me to explain a lot. Depression does result from prolonged illness.

  5. vbalbert says:

    On a side issue here, I’m curious as to what makes a statistically significant group size. I’ve read about studies that use anywhere from tens of thousands of people to this one which used only 100 for the non-control group down to as little as 12 or 13 which Andrew Wakefield used. (I’m not claiming that Wakefield’s study was good, merely pointing out how many people he used in it.)

  6. SARA says:

    It seems to me that some standardized definitions and disclaimers being required on all reported studies might help with the public misconceptions.
    If you have a preliminary study – that is stated in all public reports of the study, along with a definition of what a preliminary study is. In layman’s language – not the language commonly used in studies. (ie – it should be 6th grade level of understanding)

    I don’t know what sort of either voluntary or governmental body could oversee such a thing. If it were governmental – I would include some extension of the law that stated that media and indeed any blog with a following of greater than 100 people would be required to include the disclaimers.

  7. vb – it depends on the magnitude of the clinical effect you are looking for, as well as the background variability, and the precision/objectivity of the thing being measured.

    Generally we refer to the power of a study, which is dependent upon the number of subjects and determines statistically what size effect can be detected to what P-value.

    Some studies have thousands of subjects because we are looking for statistically tiny (although real) effects. For example – aspirin to prevent strokes. Strokes are a relatively rare event, so thousands of patients need to be followed for years in order to have the power to detect a 20% decrease in stroke risk.

    In this case the study appears to have been appropriately powered, and the results were fairly unambiguous.

  8. etatro says:

    This is very interesting, thanks for posting about it.

    I was recently designing an experiment with antiretroviral compounds and I wanted to see at what concentration other researchers use raltegravir in cell culture models and I came across this paper: “Raltegravir Is a Potent Inhibitor of XMRV, a Virus Implicated in Prostate Cancer and Chronic Fatigue Syndrome.”

    My reaction to that was: “Huh …. I didn’t realize that CFS was an accepted medical diagnosis.” The article was fine and the science was sound, indeed, Raltegravir seems to inhibit XMRV, as would be expected. However, the implications beyond the actual experiments, toward CFS are totally unfounded. In this particular journal, PLOS, readers can read comments. The comments so far are generally glowing reviews of the article for >>finally<< finding a potential cure for CFS.

    The very same journal published later articles that indicate a lack of evidence linking XMRV with Chronic Fatigue Syndrome. This cursory literature search took a whole of five minutes.

    "Investigation into the Presence of and Serological Response to XMRV in CFS Patients" –

    "No Evidence for XMRV in German CFS and MS Patients with Fatigue Despite the Ability of the Virus to Infect Human Blood Cells In Vitro" –

    "Failure to Detect the Novel Retrovirus XMRV in Chronic Fatigue Syndrome" –

  9. daedalus2u says:

    HHC, inability to sleep is one of the major symptoms of CFS. That sleep is unrefreshing is another.

    I think this is quite a good study and that it was done so rapidly is an indication of how much the science has progressed since Wakefield. They took extreme precautions to avoid contamination because of earlier problems with false positives and used multiply redundant tests which were validated for sensitivity and contamination at every step along the way.

    14 of the subjects were patients who had been reported to be positive in earlier studies.

    I hope that this will let people consider other causes of CFS, such as low basal NO/NOx/RSNO status. 😉

  10. Steven Novella,

    I’m not sure it’s correct to characterize the October 8, 2009 Science paper (Lombardi et. al), which presented the first association of XMRV with CFS, as a “preliminary exploratory” study.

    That study involved the NIH, the Cleveland Clinic, the University of Nevada and the Whittemore Peterson Institute in Reno, and nine months later the finding of MLV-like human gammaretrovirus in CFS patients was validated by scientists at FDA and the NIH (see Lo. et. al., PNAS). An equivalent number of patients and controls were studied in the Science paper, and the experiments took two years. I don’t see that you can call a major study in Science, which I believe is tied with Nature for the highest impact rating of any scientific journal in the world, and which was peer reviewed by top U.S. and British retrovirologists, a “preliminary, exploratory” study. And to suggest it might be wrong simply because it was first seems rather unscientific.

    Also, talk of nailing coffins shut seems premature when the disease in question afflicts an estimated 20 million around the globe, has been an emerging disease for three decades, and FYI, has nothing to do with “fatigue” (though an oft-repeated cliche) but does have myriad biological findings that are consistent with retroviral infection. Given the rigor of the original paper, and the fact that the association has been replicated by other groups, not just Lo et. al., why so eager to shut down the work? Tired of hearing about it?

    It’s the 21st century–time to solve the problem rather than seeking to lump and split the disease, or to declare one negative study the defining blow to what remains a strong hypothesis. I recommend holding the applause–it’s only been 17 months since the Science study.

  11. Hillary,

    Where a study is published does not determine whether it is exploratory or confirmatory. High-impact journals will often publish new and provocative ideas. And – this is a continuum, not a dichotomy. The follow up study was more rigorous, and that is why I think the results are likely to be more reliable.

    The fact that 20 million people are affected with CFS is a non-sequitur to this question. All the more reason why we do not want to settle prematurely on the wrong answer.

    I also never said anything about shutting down research on either CFS or XMRV. Clearly we need more research in this area.

  12. SteveA says:


    Steve said: “This seems like a pretty solid nail in the coffin of the XMRV hypothesis.”

    One nail. Nothing about nailing the lid shut.

    And your comment: ‘Tired of hearing about it?’ Why?

  13. SquirrelElite says:

    For a slightly different view of preliminary research, I will mention the following article from the current issue of Neurology Now, which I noticed yesterday on the table in a doctor’s waiting room.

    I was curious because I like computers and cellphones and other electronic gadgets. But I read further because I have been under treatment to control seizures for many years and one section of the article discussed using cellphones, especially smartphones, in conjunction with web-based monitoring and reporting sites to gather more consistent and accurate data on seizure episodes and other varying conditions such as psychological mood, blood sugar, etc.

    A later part of the article talks about participants in online health communities for specific diseases (in this case ALS), using this method to track their progress and gather and share data on the treatments being used and their effectiveness or lack thereof.

    In one case, a group of ALS patients recorded results from an unorganized trial of lithium treatment for ALS.

    With this new technology, patients participating in virtual health communities can even run their own “clinical trials” of a sort. For example, when a scientific study suggested that lithium—a drug typically used to treat bipolar disorder—might slow the progression of amyotrophic lateral sclerosis (ALS, or Lou Gehrig’s disease), 400 PatientsLikeMe members with ALS began taking the drug. Within months, more than 2,000 members—all posting their reactions and blood test results online—were in the experiment, creating the first grassroots, patient-driven drug trial. (See “Hype or Hope?” in the September/October 2008 issue of Neurology Now at

    “The goal of the lithium study on Patients-LikeMe was to report on a trend that was already happening and to help ensure that any benefits or harms could be identified quickly,” says Heywood.

    With data from the PatientsLikeMe study and four traditional clinic-based trials, researchers came to the unfortunate conclusion that lithium does not help ALS patients. Unlike professional studies, PatientsLikeMe shared its results right away without spending millions of research dollars. The Web site is now in the process of publishing its methodology to raise awareness about how health-related social networks can help identify promising treatments that patients are already using experimentally.

    Obviously, this trial suffers from the participants being self-selected and unblinded and lacks a well-defined control group, but it looks like a promising avenue for improving the usefulness of standard clinical trial results as well as gathering preliminary data to discover future areas for research.

    Two questions occurred to me:

    1. What ethical rules should guide participation in these semi-organized trials?

    2. What methodology needs to be in place to assure the results have scientific validity and possible medical usefulness?

  14. ccbowers says:

    “HHC, inability to sleep is one of the major symptoms of CFS. That sleep is unrefreshing is another.”

    I think his point was that depression and life events was the likely cause of many of her problems, and the the patient’s identifying it as CFS may not have been helpful. An analogy is someone who says they can’t sleep well and needs medication, and you find out they work third shift and drink coffee all day… they don’t need sleep medication, they need to alter their habits where possible.

  15. Joey H. says:

    HHC: Correct, people should not self-diagnose. That’s a given. Insomnia and unrefreshing sleep can be symptoms of many things. There are two major sets of criteria for diagnosing CFS, the Fukuda and the Canadian. A patient must have a certain number of symptoms, as well as some of a whole range of secondary symptoms to be diagnosed (e.g. post-exertional malaise, orthostatic intolerance, frequent or constant fever from 90 to 98 degrees F, etc.) If the patient doesn’t meet either of those criteria sets, depression or other issues are probably the cause.

    The problem of incorrect self-diagnosis exists, but that has nothing to do with the fact that people with actual gosh-darn CFS exist, and while still a vague diagnosis it is acknowledged by most sane physicians to be a somatic problem — one that often comes with *secondary* depression.

    In fact some antidepressants like Cymbalta (can’t remember the real name) can provide some pain relief in CFS patients. If a patient claims to be all better after going on ADs, however, that patient never had CFS.

  16. daedalus2u says:

    Joey, you are correct, if a patient gets better after taking antidepressants, then they don’t have CFS. CFS is a diagnosis of exclusion, if there is any known cause for the symptoms of CFS, then it is not CFS. That is the definition of CFS, a constellation of unexplained symptoms.

    That the definition of CFS is a certain constellation of unexplained symptoms does not tell us that the physiology of CFS is unrelated to the physiology of other conditions that lead to the same symptoms. Actually it is pretty likely that the physiology of CFS is related to other conditions that have the same symptoms, those symptoms being mediated by physiology too.

    We know that depression is not “caused” by a deficiency of a xenobiotic SSRI. That a xenobiotic SSRI relieves some of the adverse symptoms of depression is not proof that a deficiency of that xenobiotic SSRI is what caused it. There are many drugs that relieve some of the adverse symptoms of depression. That does not mean that the physiology of each case of depression improved by one antidepressant is unrelated to the physiology of every other case of depression that is relieved by a different antidepressant.

  17. Joey H. says:

    Totally agree about the possible relationships between CFS and other illnesses; people with early MS, lupus, fibromyalgia and other autoimmune annoyances are often diagnosed with CFS incorrectly or concurrently. My own doctor is pretty convinced that fibro and CFS are different manifestations of the same illness.

    Good doctors will go through the process of elimination (which has been described in the literature) before checking one of the two main sets of diagnostic criteria for CFS.

Leave a Reply