The PACE Study Results Exposed as Meaningless and Harmful [UPDATED]

The overwhelming weight of scientific opinion in the U.S. is that ME/CFS is a biomedical disease, and important strides toward identifying a biomarker have recently been made at Stanford and Columbia.  It is a different story in the U.K, unfortunately, where a coterie of psychiatrists have continued to push the theory that ME/CFS is the result of “unhelpful illness beliefs” that cause patients to become “deconditioned.”  The primary evidence for this theory has been the PACE study, which purported to show that Cognitive Behavior Therapy (CBT) and Graded Exercise Therapy (GET) could lead to “improvement” and even “recovery” for ME/CFS patients.

The PACE study was deeply flawed, however, and sharply criticized by the patient community, almost from its first publication in 2011.  Last year, the Virology Blog published a lengthy article by David Tuller detailing the problems in the study, most notably that the investigators had actually weakened their criteria for “recovery” midway through the trial – so much so that a participant could actually have deteriorated during the study, and yet still be counted a “recovered” at the end.

When challenged about this discrepancy, the PACE investigators refused to release their underlying data for review by independent experts.  Fortunately, there is a strong Freedom of Information Law in Britain, and it applied to the publicly funded study.  After some years of litigation, the original data have now been made public, and it turns out that the PACE investigators had overstated their results on two key determinants by factors of three and four.  In other words, there is no statistically significant support for the theory that CBT or GET can lead to improvement or recovery. 

The details of the re-analysis are reported on the Virology Blog here.

The PACE trial, and the attendant claims of “recovery,” have done enormous damage to ME/CFS patients, and not only by diverting funding and effort away from biomedical research.

As the U.S. Institutes of Medicine report established, ME/CFS is an “exertion intolerance” disease, meaning that even moderate exertion exacerbates the symptoms.  Every ME/CFS patient will tell stories of “blowback” or “crashing” brought on by too much physical activity.  Thus, the recommendation of Graded Exercise Therapy is actually counterproductive and harmful to many patients. 

Unfortunately, the initial PACE results were widely circulated, and they found their way into treatment recommendations by the CDC and Mayo Clinic, among others.  Although medical practices tend to change slowly, the recent reanalysis clearly shows that GET is not an effective therapy for ME/CFS.  In the words of Stanford’s Ronald Davis, Columbia’s Vincent Racaniello, and over 40 other signatories on an open letter to The Lancet, the PACE results “have no place in published research.”  

UPDATE: A more complete account of the PACE saga, and its harmfulness to patients, is here.

NOTE: This post has been slightly edited for syntax.

Comments

[DrSpeedyandME]

Thx !! See also
The PACE Trial Invalidates the Use of Cognitive Behavioral and Graded Exercise Therapy in Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome: A Review https://www.researchgate.net/publication/301202396_The_PACE_Trial_Invalidates_the_Use_of_Cognitive_Behavioral_and_Graded_Exercise_Therapy_in_Myalgic_Encephalomyelitis_Chronic_Fatigue_Syndrome_A_Review

[Grumpy Ninja]

Thank you for this.

One small correction. You said:

"In other words, there is no statistically significant support for the theory that CBT or GET can lead to improvement or recovery. "

The re-analysis from Matthees et al deals with claims of recovery, not of improvement.

The PACE team have recently self-published a protocol-defined re-analysis of improvement [1,2] which showed a similar reduction in the percentage of patients who report improvement as Matthees et al report for the recovery rates, compared to the much weaker post-hoc revised protocol [3]. These new results from PACE for improvement are clearly marginal and still leaves patients way below anything resembling a practically useful degree of functional improvement, let alone good health.

Furthermore, this was an unblinded trial using only subjective self-report measures for the primary outcomes and the reported effect size is well within the range of standard confounders for such measures. There is also no support for this effect from any of the objective measures used in PACE (besides a very marginal improvement in the GET arm on the Six Minute Walk Distance test of just 35m after a year-long intervention, which does not reach clinical significance, and still leaves patients scoring around half the healthy average) [3].

Lastly, the 2.5 year long-term follow-up results were released last year, and even with the benefit of using the post-hoc revised protocol, it still delivered a null result [4].

It is long past time this matter was properly dealt with by the medical scientific community.

1. http://www.wolfson.qmul.ac.uk/images/pdfs/pace/PACE_published_protocol_based_analysis_final_8th_Sept_2016.pdf

2. http://www.biomedcentral.com/1471-2377/7/6

3. http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)60096-2/abstract

4. http://www.ncbi.nlm.nih.gov/pubmed/26521770

[Justin Reilly]

Thanks for this great piece, Steve!

[Steve L.]

test