I have now read the Tomas paper and the Myhill/McLaren Howard response, and think it would help to focus on some of the detail.
Overall, the findings in the new paper appear to be pretty good evidence that the MES test isn't reliable, because it can't separate patients and controls. However, it's worth looking at both the use of frozen and fresh samples, and also the significance of processing delays.
Main result
The translocator components of the MES test couldn't be replicated because reagents listed in the protocol are no longer available. That left 3 elements.
For neutrophils, there was no significant difference (
fig 1) between patients and controls on any of the three remaining tests. One, for ADP to ATP efficiency, came close (p = 0.054), but there was extensive overlap between the effeciency levels patients and healthy controls (see graph). So that means it isn't clinically useful.
The study is quite small, with 10 patients and 13 controls but, as the paper says, "abnormalities in CFS/ME patients should be reproducible even in small samples given the current use of the test for diagnostic purpose".
The situation for PBMC's (white blood cells, or leucocytes) was very similar. In any case, although the MES test has switched from neutrophhils to PBMCs, there is no published data to show that this gives an effective test.
The impact of delayed processing
The authors were concerned that delaying analysis of blood samples for 24 hours or more would affect test results. This is because the status of the cells would change (eg neutrophil activation), cells would deteriorate and ADP/ATP-relevant metabolite levels would change. The study found substantial changes to the neutrophil fraction within 24 hours and dramatic changes to glucose levels (see graph).
Although McLaren Howard said that blood sample were tested the following day, the new study shows there would have been substantial changes in that time.
Critically, all the neutrophil results and almost all the results with PBMCs were significantly different 24 hours on from blood draw (see
fig 6).
By contrast, McLaren Howard said that he "explored heparinised whole blood storage times in relation to patients and controls. Samples were processed within minutes of venepuncture and retested 6, 12, 24, 48 and 72 hours later. Provided the samples were kept in the original "vacutainer" and not subject to extremes of temperature there were only minor changes in test results up to be 48-hour point."
Interesting point from
@Adrian that blood sent through the post might experience considerable changes in temperature as well as a lot of agitation.
The new paper says blood was either processed immediately or processed after 24 hours, suggesting that the blood was left in vacutainers - but this wasn't explicitly stated.
Use of both fresh and frozen samples
The new study deviates from the published protocol in one significant way, using frozen-then-thawed samples as well as fresh ones. "The use of fresh and cryopreserved samples was born out of necessity as samples could not be collected in a short enough time frame to conduct the experiments at the same time for all CFS/ME samples and controls". So that all samples are effectively processed within a couple of hours.
However, the authors showed (
supplementary data) that there was no significant difference in MES results between fresh and frozen samples. You can see the same pattern in Fig 1 (eg the first graph above) where frozen samples are indicated by solid markers and fresh samples by open ones.
McLaren Howard reports a very different experience with freezing/thawing samples.
"That is not in accord with my findings or those of several scientists from other labs who have spent time in my lab exploring these tests."
That raises some doubt, but Tomas et al have published data to support their claim which really carries more weight. But perhaps this area needs exploring.
Perhaps, once a little water has flowed under the bridge, the two parties can sit down and discuss how they got different results. But as things stand I think there hs to be considerable doubt over the orginal findings and continued use of the test.
Footnote
Within his response, McLaren Howard seems to argue that their results mus be correct because they explain fatigue problems on inceased energy demand in ME/CFS patients. "As almost all the energy for such activity is ATP-derived, one would expect to see the kind the differences that we so frequently find in our test." And the new results make it "very difficult" to explain the same fatigue. Strange to use an assumed specific biological abnormality to justify your own result and cast doubt on a null result.
