The Imperial College XMRV Replication Study: XMRV in Trouble….or Not?

Posted by Cort Johnson

And if you see some negative papers coming out, don’t be discouraged. It’s going to happen. There are going to be some negative papers. People really jump to do this. And the method is not that easy and getting the right bits and pieces you need together  (is not easy)” Dr. Nancy Klimas

The Brits Smack XMRV…Or Do They? – The first XMRV Replication attempt has been published and its a doozy. Originating from the Imperial College in the UK (with patients supplied by Simon Wessely), the study found zero (that’s zero!) evidence of XMRV in 186 CFS patients. (Here’s a link to an article by the BBC and a link to the original paper .)

The study included a lot of patients (186) who appeared to be pretty ill,  had high rates of disability, about 50% of which had infectious onset. They all met the standard CFS Criteria (1994 Fukuda) which meant they did not have a major psychiatric condition.

They used a positive sample of XMRV to ensure that they could find the virus but remarkably, they didn’t find the virus in any of the samples – a similar finding to an earlier German study that failed to find XMRV in any prostate cancer samples.

In rather strong terms Dr. McClure, the lead retrovirologist in the study suggested that the WPI had moved too quickly stating “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,”

These studies underline how complex situation these efforts are. Earlier the CFIDS Association noted that the German study did not adequately replicate the original XMRV prostate cancer study. Now, Dr. Vernon of the CFIDS Association asserts the same is true with this Imperial College study.

CFIDS Association of America (CAA) States Study Not a “Valid Attempt”

The CFIDS Association of America posted its response within hours. In a CFIDS Link report Dr. Vernon stated that this study ‘should not be considered a valid attempt to replicate the findings” of the Science Study and she listed a series of methodological questions that could have interfered with the Imperial College Researchers ability to find the virus. They included:

    • collecting the virus in different kinds of collection tubes
    • the DNA from the patients was extracted and purified in a different manner
    • they used different amounts of DNA to amplify their assays
    • they looked at different parts of the genome
    • they ran the PCR under different conditions


Based on Dr. Vernon’s experience working with PCR’s any of these could have affected the results. Dr. Vernon  pointed to a larger much more rigorous study that the Department of Health and Human Services is engaged in. (Both Dr. Vernon and Dr. Mikovits are part of a team overseeing that study). Since that study will involve multiple laboratories that will come up with a standardized test first that study will take longer to finish but it will provide more definitive results. She did say that the CFIDS Association is urging that the DHHS study is completed as expeditiously as possible. She urged patients to be prepared for conflicting results’

“The U.S. Department of Health and Human Services Blood XMRV Scientific Research Working Group is conducting a rigorous study to detect XMRV. Multiple laboratories will standardize methods to optimize sensitive detection of XMRV proviral DNA and viral RNA and then, once methods are standardized, these same laboratories will test coded panels of blood samples obtained from healthy blood donors and CFS patients. We look forward to the results of this study and urge that it be completed expeditiously, especially in light of this report from the U.K. In the meantime, be prepared to read about more studies with conflicting findings. Rather than simply accept or dismiss new information, we will help make sense of why discrepant results occur.”

The next day the CFIDS Association of America produced a press release that stated in no uncertain terms their concerns about “many elements” of the study including the rush to publication (three days between submission and acceptance), the different processes used in the two studies and the differing patient groups.

Standardized Test for XMRV – Despite the fact that VIPDx in collaboration with the WPI and Cooperative Diagnostic Labs it’s clear there is no accepted test for XMRV yet. (VIPDx was taken to task a bit in one article for providing a commercial test at this point.)  Dr. Vernon closed the press release again alluding to a need for a ‘standardized test’ stating that potentially millions of dollars may be wasted on studies that are later deemed to be invalid because they are not using the correct test. Developing a standardized test requires having multiple labs test and retest different procedures until their results match and they agree they have found the correct test. Only when a standardized test that meets the acceptance of the research community is developed will it be impossible to determine the true prevalence of XMRV. In his CFSAC talk Dr. Coffin agreed that the first priority is developing a standardized test.

Whittemore-Peterson-Institute Responds – The WPI weighed in the next day and in very blunt language took the Imperial College study to task calling it ‘meaningless’.  Besides the different patient cohorts and different blood sampling procedures they cited

    • the use of a water control rather than a blood control
    • different primer sequences and amplification protocol which was not validated by a clinical control
    • fewer rigorous tests of accuracy

The WPI appears to be asserting that the group should have used an positive sample from a patient rather than the standard XMRV sample they used as a control. They also noted that many more tests of verification the WPI used (involving three different labs).

It’s clear that one thing we’re going to be seeing are studies replicating their results against standardized XMRV samples (eg Imperial College) or against positive samples from the WPI. Theoretically, if both samples contain XMRV they should produce similar results but if the ‘XMRV’ in them is slightly different it could create problems.

There are differences of opinion regarding how exact a lab needs to be in replicating another labs results. There are several different PCR techniques and variations within each kind of technique. Some researchers assert that any valid technique should yield positive results. Others believe slight differences can make big differences. At this step of the game – with such a new finding – it seems pretty clear that the first priority is exactly duplicating the original study’s findings.

Getting into the Nitty Gritty – Comparing the Science and Imperial College Studies.

The two studies were different in a number of ways.  Here’s a closer look . Note the highly speculative assessments by a laymen (who perhaps should know better). For fun only.

The Location – the disconnect between researchers ability to find XMRV in the US and Europe is growing; both attempts to replicate XMRV findings outside of the US have failed; both a large German study prostate cancer and this large UK study found zero evidence of XMRV. Could geography play a role?There were reports of Dr. Mikovits stating there were significant geographical differences in XMRV prevalence in the US;- some regions reportedly had much rates of infected CFS patients than others. Its hard to believe XMRV is missing in Europe – it has, reportedly, been found there and its hard to imagine that the geographic differences in prevalence could be this distinct (zero positive results!).

But what if the XMRV in Europe differs a bit genetically? XMRV was discovered only a few years ago and no research on its geographic genetic variability has been done. The very low genetic variation in XMRV across the US suggests it could be a recent entry to the region. Could XMRV have hit Europe first? Is a different more genetically variable kind of XMRV found there? Different genetic sequences in two areas could cause XMRV to ‘disappear’ in one of them.   Laymen’s assessment – different prevalence rates are a unlikely cause of the two divergent findings but genetically different strains could conceivably be a key issue.

Different Gene SequencesThe Imperial College Group did not search for the same gene sequences as did the WPI. (In PCR researchers don’t look for a whole virus, they look for genetic sequences they believe indicate that the virus is there. Different research groups can and do look for different sequences of the same virus) They may be following the same strategy as CooperativeDiagnostics Lab and we’ll probably see other labs do the same. Because the WPI looked at genetic sequences that reported demonstrate high genetic variability some labs are examining more stable, conservative sequences to determine if XMRV is there. Theoretically both tests should find the bug but if the ‘XMRV’ the WPI has found is a different kind of XMRV than found in standard research samples, they may not be picking up the WPI’s bug (see above).

However, the WPI did fully sequence the XMRV they found in two patients and partially sequenced another and they were 99% similar to a reference XMRV, making it less likely that their XMRV is different. Their ability to layout the genetics of their virus in full was important in convincing the other labs and Science that this was not an endogenous retrovirus (ie contamination). Laymen’s assessment – possibly a key issue although its hard to figure out how; this is why we need exact replications of the original study.
XMRV and CFS
Blinded Study the Imperial College Study was ‘blinded’ (albeit in a strange way) while the WPI study was not. Although the PCR technician didn’t know which samples he/she was assessing all the samples came from either CFS patients or ‘water controls’.  Laymen’s assessment – not a huge issue; differing percentages of positive patients could be attributed to operator error – but not to such huge differences.

Different Procedures Both Dr. Vernon and the WPI noted that several techniques were different in the two studies. Some researchers will argue that techniques don’t have to be replicated exactly for a robust finding to show up. Others argue they do need to be replicated exactly. Inexact replication of the first retroviral finding was a major controversy of the first retroviral finding in ME/CFS and experts differed on whether it mattered or not. Laymen’s assessment – potentially a major issue. We’ll only find out if it is a major issue when the first study is replicated exactly.

Different Patient Groups the Imperial College group used quite ill patients who met the CDC criteria judged (by the investigators) to be ‘typical’ of chronic fatigue syndrome patients in Australia and the US. It’s possible that the WPI group (immunologically challenged with metabolic dysfunction) were a different subset. However, in a large group of CFS patients (168), surely a good portion would fit the WPI profile, and no patients at all tested positive. Laymen’s assessment – if this is an it’s not THE issue.

Laymen’s Overall Assessment – This is a methodological problem rooted in the complexities of PCR analysis not a patient cohort problem.  (Any PCR experts outs there?)

Gird Your Loins ME/CFS Patients – in the meantime Dr. McClure’s comments about more negative studies suggest we may be about see a stream of similar findings. A participant in the Phoenix Rising forums reported a CDC report may be out in a couple of weeks.

Another interesting outcome was Dr. Lombardi’s statement that of about 300 samples tested at VIPDx Labs (only 300?) 36% have tested positive – about half the rate in the paper – and far lower than the 95% number we heard later.

Now is the time to member both Dr. Klimas and Dr. Vernon’s admonitions to sit tight and not take any one (or two) results too seriously at this point. We’ve more or less documented 15 studies/groups we believe are studying XMRV and we’ve undoubtedly. This study is, to paraphrase Winston Churchill, only the beginning of the beginning, not the beginning of the end. It will take the scientific community some time to wade through the different issues; the different types of PCR, the different gene sequences used, the possible location problems, the possible cohort problems.  Sit tight.

Share this!