Can the Giardia lamblia ME/CFS outbreak in Bergen explain why the phase II rituximab trials were positive, but the phase III trial negative?

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━​

SUMMARY: Could the positive results of the early Fluge & Mella 2011 blinded phase II rituximab trial and 2015 open label phase II study be due to inadvertently including some Giardia lamblia ME/CFS patients in the cohort, from the 2004 Bergen Giardia lamblia outbreak?

The later Fluge & Mella negative 2019 phase III multi-center rituximab study would have had much fewer Giardia lamblia patients (because it was conducted in several other cities, not just Bergen).

On the assumption that rituximab works for Giardia lamblia ME/CFS, if there were Giardia lamblia patients in the two phase II studies, this might explain why these studies were positive, whereas the later phase III study was negative.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━​

In 2004 in Bergen Norway, there was a major outbreak of giardiasis (Giardia lamblia infection) resulting from sewage contamination of the drinking water supply, which affected an estimated 2500 people (with 1300 actual laboratory-confirmed cases of giardiasis).

Bergen of course is the city where Oystein Fluge and Olav Mella conducted their phase II rituximab studies.

From March 2009 to March 2010, five years after the giardiasis outbreak, 253 patients who had been suffering chronic fatigue symptoms ever since their giardiasis were recruited to participate in a 2013 study at Haukeland University Hospital in Bergen. The study concluded that 41.5% of the participants had ME/CFS by the CDC Fukuda criteria.

So lots of patients with Giardia lamblia-triggered ME/CFS were attending the Haukeland University Hospital at the time Fluge & Mella were conducting their phase II rituximab trials at Haukeland. The Fluge & Mella 2011 phase II trial recruited ME/CFS patients referred to the Department of Neurology, Haukeland University Hospital, from June 2008 to June 2009.

So the question is, how many of these Giardia lamblia ME/CFS patients would have been included in the rituximab phase II trials? Giardia lamblia ME/CFS is a rare form of ME/CFS, so if lots of these patients got onto the phase II trials, it would have created an unusual patient cohort.

By contrast, the phase III multi-center rituximab study was conducted in various university hospitals in different cities in Norway (Bergen, Oslo, Notodden, Trondheim and Tromsø judging by the authors' affiliations in the study). So you would thus have had far fewer Giardia lamblia ME/CFS patients in the phase III study, because the giardiasis outbreak only occurred in Bergen.

Thus if rituximab were effective for Giardia lamblia ME/CFS, it might explain why the two phase II studies showed such positive results, but the phase III study had a negative result.



Just to remind you of Fluge & Mella's rituximab trial results:

2011 blinded phase II study on 30 ME/CFS patients given rituximab or placebo.
Response rate: 67% in the rituximab group, and 13% in the placebo group.

2015 open label phase II study on 29 ME/CFS patients all given rituximab. No placebo group.
Response rate: 64%.

2019 blinded phase III study on 151 ME/CFS patients given rituximab or placebo.
Response rate: 26% in the rituximab group, and 35% in the placebo group.

 

That was not the result. The primary endpoint was negative. These figures relate to a post- hoc six month time point only.

I reviewed the phase 2 results in detail, including a three hour meeting with Fluge and Mella. We agreed that the six month data were enough to proceed to further work but they were not 'the result of the trial'. Moreover, the kinetic profiles for all the component measures were extremely inconsistent, which made it likely that the six month data were a chance finding.

 

Intriguing question. I know a Giardia ME patient who had a positive response to Rituxan. Measures of improvement included change in total fitbit activity and also hours of upright activity in a day and on subsequent days as well as patient reported improvement in sleep quality. Cognitive impairment was not helped

 

I am a bit confused then why F&M would include the response rate in the results section of the abstract of their 2011, 2015 and 2019 rituximab papers, if this response rate is not really that meaningful.



Anyway, I always understood that the phase II results were considered positive enough to proceed to phase III, so the question of a possible Giardia lamblia confounding factor may still remain.

I could not find any mention in the phase II studies about the giardiasis outbreak, and whether F&M had taken steps to exclude such giardiasis ME/CFS patients from their studies.

 

I don't remember exactly what was in the abstracts but my memory is that they made it clear that the primary endpoint failed to show efficacy. The six month endpoint may well have been included but I don't think as anything more than a subsidiary analysis.

The proceeding to phase III was on the basis that the six month result was plausibly the most relevant and that since it looked positive there was a serious risk that the Phase 2 study might have missed the key measure of an effect. Since treatments for ME are so hard to find the justification for going ahead seemed adequate.

However, in the light of the Phase 3 result, which showed no effect at all, and in fact higher improvement rates in the placebo group I think the argument for hoping the Phase 2 six month result was real disappears. The Phase 3 study also showed that spurious positive results (erroneously called responses) can be expected in a substantial proportion of patients in trials of this sort. So we have no reason to think anyone with ME who feels better after rituximab is better because of the rituximab.

As an aside, it is maybe worth mentioning that the reason Roche went to the major proof of concept trial of rituximab in RA that I published in NEJM was that they wanted to be sure that my pilot study result was spurious. They had no belief that the results would replicate. But again, the trial was justified on the basis that it would be wrong to discount a less than even chance of an effect.

 

I appreciate that unknown factors can contribute to improvement following treatment and that we dont have a handle on the complexity of this disease. But in the patient I mentioned, the only change in treatment at the time was Rituxan, the degree of improvement was much greater than this patient experienced in to waxing and waning, and the improvement was reflected in objective measures.

Admittedly, I cant explain how this fits with the Norwegian Phase III trial. But as @Hip suggests, I'd think it worth considering whether and how the nature of the trigger might contribute to subtype differences in patients overall, even if that's not the answer in the case I mentioned.

 

But we hear that happening with the Lightning Process. For a condition with such unpredictable kinetics these observations are no use to us. The one thing that stuck in my mind reviewing the thousands of raw data points from the Phase II rituximab study was just how unpredictable the kinetics are, with or without a possibly active treatment.

The only way one can begin to draw a causal inference with treatments is if you know there is a stereotyped kinetic pattern that the current case fits with. I spent thirty years watching the kinetics of responses to gold, methotrexate anti-TNF and rituximab for RA. We had hard objective measures like CRP but it was still often difficult in an individual case. I think the idea that one can attribute cause in individual cases in ME without having any stereotyped kinetics is just wishful thinking I am afraid.

As an aside, I cannot think of any reason why rituximab should specifically benefit people who have had giardia infection. Autoimmunity has nothing to do with prior infection as far as we know.

 

The phase 3 trial has spoken. No evidence of improvements on the highest standards of clinical trials.
This is unfortunate as we so desperately need treatments, but we need treatments that are effective, and we need biomarkers to prove that the pathology is altered by said treatments. Rituximab did not do that.

We would be better off continuing piecing the puzzle, and carry on the leads we already have. i am looking forward to hear about the Cyclo trial completed but not published. But Cyclo is definitely more toxic than Rituximab. It needs to be administered with caution and with regular follow-up to mitigate the potential complications. I would doubt very much it is a magic pill, but hopefully it provided insight for Drs Fluge and Mella.

 

Agree with all of that, though I am more cautious around the Cyclophosphamide trial: we see a massive placebo response* in the Rituximab Phase 3 trial and I cannot see how one achieves decent blinding in a drug as grim as Cyclo.

*Not massively surprising - take a patient community that is given very little hope from existing treatments and that is often subject to their symptoms being dismissed, then start taking their symptoms and treatment seriously and a strong placebo response is all but guaranteed.



----

Edited to add: I don't think the theory around Giardia being particularly different is likely, which is a shame as my own ME/CFS diagnosis followed contracting Giardia on a business trip to India. Aside from a couple of tracking studies following the Bergen outbreak, is anybody aware of any decent studies on post-Giardia fatigue and whether there's any contrasting symptoms with ME/CFS? I'm not aware of any, and my post-Giardia symptoms would meet all the major diagnostic criteria for (generally mild/moderate) ME/CFS.