The First of many studies attempting to replicate the Whittemore Peterson Institute’s October 9th XMRV paper has been released. The Whittemore Peterson Institute paper found that 67% of those with CFS were XMRV positive compared with 3.7% of controls being XMRV positive. The replication attempt has been made by the Imperial College London which includes the notorious Simon Wessely.
The paper can be found here http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0008519
They failed to find XMRV in any of the 186 patients with CFS. I will analyse this paper and attempt to determine the validity of it.
Difficulty finding it
XMRV is hard to find even when you are looking for it carefully. Two German studies and an Irish study failed to replicate the US study that linked prostate cancer to XMRV. The original US study found XMRV present in the cells of tumours of men with prostate cancer. These 3 subsequent studies failed to find this link. It is suspected that a difference of methodology was used that resulted in a failure of replication here. What can be concluded from this is that we would expect many studies attempting to replicate the WPI study to show mixed results and some to find zero CFS patients with XMRV. This points to a methodological error. 3.7% of the healthy control group population is suspected to be XMRV positive and this number was not reflected in the Imperial College London study.
In the UK?
The authors of the ICL study have said “We are confident our results show there is no link between XMRV and CFS, at least in the UK” This is an inaccurate and unscientific comment from the authors of the ICL paper. Mikovits from the WPI has said that the WPI has tested over 500 samples from London and the same numbers are holding up, ie 95% + positivity for XMRV. To therefore make a comment about a small preliminary study that used a dubious technique and conclude “confidence” that XMRV isn’t present in the UK is fallicious. If the authors do have confidence in their ICL study, they can at most conclude that “more studies are neccesary to draw a link between XMRV and CFS in the UK.”
The ICL paper says “All patients had undergone medical screening to exclude detectable organic illness, including a minimum of physical examination, urinalysis, full blood count, urea and electrolytes, thyroid function tests, liver function tests, 9 a.m. cortisol and ESR…… All subjects met the CDC criteria.”
The ICL group were attempting to replicate the WPI findings and hence would need to use the same cohort as the WPI. A different cohort may yield different results. The WPI cohort used the Canadian CFS guidelines. These differences in cohorts are therefore potentially significant in terms of ICL’s claims of a replications of the WPI’s study. I don’t believe this was the reason that they found zero XMRV in the CFS group however this “poor science” is worthy of note. Also it is dubious to rule out “detectable organic illness” from the ICL cohort and exclude an array of other signs that many CFS patients have in common such as abornmal cortisol levels.
Dr Suzanne Vernon from the CFIDS Association of America has pointed out the difference in methodology between the WPI study and the ICL study.
- The blood was collected from CFS patients in different types of blood collection tubes.
- The genomic DNA was extracted and purified using different techniques.
- The amount of genomic DNA included in the amplification assay was different.
- Different primer sequences were used that amplified different regions of the XMRV proviral DNA.
- The conditions of the PCR amplification assay were different – from the numbers of cycles, to the type of polymerase used.”
Dr Suzanne Vernon’s full article is well worth reading and is found at http://www.cfids.org/cfidslink/2010/010603.asp
These criticisms are the most damning aspect of the ICL study and highlight how technical testing for XMRV is. They also highlight the need to follow the exact steps when attempting a replication. The ICL have gone off on their own path and done their own thing with no regard to precision-based replication methodology. Dr Vernon has emphasised that this study “should not be considered a valid attempt to replicate the findings.”
ICL Authors’ comments
Professor Myra McClure from the ICL study said “When you’ve got such a stunning result you want to be absolutely clear that you are 1,000 per cent right and there are things in that [previous study] I would not have done. I would have waited. I would have stalled a little,”
This is not a very well thought out comment from McClure. She should not have just stated that there were things in the WPI study she wouldn’t have done without elaborating on what these things were. It is like saying to someone that you are wrong without justifying it with examples of why they are wrong or with evidence alluding to why they are wrong.
McClure also says that she would have delayed releasing the WPI study. This is an ignorant comment. The WPI had replicated their original study in two other labs, the Cleveland Clinic and the National Cancer Institute. They were published in the most prestigous scienctific journal in the world ‘Science magazine’ which closeley examines the scientific method used and only accepts the highest class science. The WPI had already delayed releasing the XMRV study, as the original XMRV study was done over a year ago. The WPI also needed funding to continue to investigate the presence of XMRV hence another reason to release the study. The world’s blood supply is also potentially contaminated and millions of people around the world are suffering from a devastating illness that this study may have unearthed the cause of. Scientists all over the world are now studying XMRV because of when the WPI released the study. I don’t know who would have benefitted from a delay in publication. What else would have the WPI have to wait for? McClure will argue that the CFS patients will get false hope however the WPI have done everything humanly possible to avoid this false hope and have checked their science continually.
McClure went on to say “If it had been there, we would have found it. The lab in which we carried out the analysis had never housed any of the murine (mouse/rat) leukaemia viruses related to XMRV, and we took great care to ensure there was no contamination.” This is an arrogant comment that subtly condescends the WPI study by implying laboratory contamination caused their XMRV results. The WPI study occurred at 3 different locations and the same numbers held up. It is extremely unlikely that all three locations were contaminated. If all 3 locations were somehow contaminated, the odds of having the CFS group of blood samples infected and the controls not getting infected in all locations is minimal. The nail in the coffin for the contamination theory is the other XMRV tests performed by the WPI. These included growing XMRV in cells from CFS patient blood samples. All of these sub tests confirming XMRV had a control group that found zero presence of the retrovirus. These sub tests can be read about here https://livingwithchronicfatiguesyndrome.wordpress.com/2009/10/15/xmrv/
The better study
The ICL study wouldn’t have met the rigorous scientific standard that publication in ‘Science’ requires. It is therefore poor science and ignorance for the ICL to dismiss the WPI study. The WPI study was the higher quality study. The WPI spent months carefully choosing their cohorts, performing XMRV studies, performing sub-tests and getting their work replicated by two reputable institutions- the Cleveland Clinic and the National Cancer Institute. The ICL came along and essentially rushed in order to be the first group out to ‘attempt’ to replicate the WPI findings.’ In less than 6 weeks the ICL finished their study. When such precise methodology is necessary to replicate something as difficult to find as XMRV, rushing to be the first for publication isn’t good science.
Due to the large methodological differences and the different cohort used, the ICL study can’t be listed as an attempt to replicate the WPI study. Any replication attempt must do just that- ‘replicate’ the variables. Many more XMRV replication studies are due out soon. I expect many more to use dubious methods to replicate the WPI study. These should not negate the validity of the WPI study. There are also some studies coming out that the WPI helped in replicating their results. These are the ‘good science’ studies and when these are released we will have a good idea about the causal link between XMRV and CFS. Whether it’s the root cause or a spurious similarity will be the next question….