Why include patients with phobias at all? Unless the assumption is that CFS is an exercise phobia:
We excluded those with comorbid psychiatric disorders (with the exception of simple phobias) ....
How sure are they that the allowed meds don't have an impact on the immune system? Paracetamol certainly acts as an anti-inflammatory, which would implicate immune involvement. And I have to wonder why their non-depressed patients are on antidepressants:
We excluded participants (patients and controls) if they had regularly been taking any prescribed medications in the past two weeks that might affect the immune system or exercise challenge. We allowed the use of selective serotonin reuptake inhibitors, tricyclic antidepressants and paracetamol.
How do you judge that the thyroid disease is truly in remission, and not the cause of ongoing fatigue?
Participants could have certain co-morbid medical conditions if in remission (e.g. hypothyroidism if biochemically euthyroid on thyroxine replacement therapy) [1].
This is actually pretty good:
An aerobic exercise challenge was completed on day 14. On this day blood was drawn before, immediately after, and 3 hours after exercise. On day 16 (two days after the exercise) a final blood sample was taken at the hospital and follow-up questionnaires were administered. Participants were studied 48 hours after the exercise, rather than 72 hours afterwards as in the pilot study [11], since our clinical impression suggested that CFS patients perceive most post-exertional symptoms at this time.
All of this seems rather dodgy:
Ad hoc self-rated five item Likert scales were used to measure the effect of activity/exercise on delayed fatigue, pain and malaise (unwell) at the same times. Response options ranged from ‘strongly disagree’ through ‘neither agree nor disagree’ to ‘strongly agree’. Fear of exercise was determined on day 14 using the Tampa scale for kinesiophobia for fatigue [28], with healthy volunteers being asked to recall the last time they felt extreme tiredness that was not related to a medical condition, being pregnant, or dehydration.
Why such a big gap? And why such a big delay in publishing?
The TGF- samples were analysed in three batches by two laboratory technicians, with one technician analysing the first batch of samples in 2009 and the other technician analysing the second and third batches of the samples in 2011
Why use the mean instead of the first result, for one batch?
For quality assurance, in the first batch, where an analyte looked to be a possible outlier we re-analysed the sample and took the mean of the two assays. These values did not alter the results so we did not repeat this with the remaining batches.
If symptoms didn't increase after the exertion, ME/CFS is an unlikely diagnosis:
Cases who agreed that their fatigue, malaise or pain increased after the exercise test did not have significantly greater relative changes in TNF, IL8 or IL-6 protein or RNA either immediately, 3 hours or 2 days after the exercise, compared to those who disagreed.
Perhaps the reason for the 2 year delay in batch analysis was driven by a need to find some fatigue patients to water down the big differences found in 2009:
This shows that batch one TGF- concentrations (both for cases and controls) were significantly higher than the concentrations for batches 2 and 3.
They made dozens of comparisons with few participants, so it's impressive anything was statistically significant with the deck (deliberately?) stacked against finding significance:
After correcting for multiple analyses, TGF-â protein was the only cytokine protein or RNA that showed significantly different values between CFS and control groups.
I would suggest that this study is too much of a mess to support any conclusions:
A systematic review of the association between circulating cytokines and CFS showed that only TGF- was elevated in the majority of case control studies [10]; a finding we were unable to replicate. Another systematic review concluded that cytokine concentrations were not abnormal after exercise in CFS [9]; a finding we replicated. We suggest that circulating levels of cytokines are unlikely to be important in the pathophysiology of CFS.
So Peter White can share the blame for some poor design aspects, and providing fatigue patients to the study. But his involvement otherwise seems very peripheral:
We thank the Barts Charity for funding this work. We would also like to thank Professor Anthony Pinching for co-leading the grant application. PDW, LVC, MB and VV designed the study, which was run by LVC. PDW and GM recruited patients from their clinics. CM and EW undertook RNA analyses. MB oversaw cytokine analyses. MS, LVC and NT analysed the study. All authors contributed to and approved the manuscript. We are particularly grateful for discussions with Megan Roerink regarding an earlier draft of this paper, and we are also grateful to our reviewers for their wise advice.
