Bruce Patterson MD rolled into a most interesting Solve M.E. event like he didn’t have a care in the world. Patterson has aggressively pushed his diagnostic tests and treatment regimen in the media – including publishing the same opinion piece in the San Francisco Chronicle, the Des Moines Register, and the Tampa Bay Times, and being interviewed on many social media sites.
His purported successes quickly got around. In “Lab discovers root cause of confusion, fatigue experienced by Covid ‘long haulers,” an NBC affiliate last year basically reported that he’d solved long COVID!
It’s been too much for some critics who prefer that Patterson take the time-honored (and time-intensive) approach of publishing first and promoting later. Patterson, of course, has published about the diagnostic test, but the studies are relatively small and need to be replicated in larger cohorts. Crucially, no studies have assessed the most important part of his package – treatment effectiveness. Even so, Patterson with his Stanford research background and his good communication skills has captured the public’s attention.
His promotional style might have gotten Patterson, a former Stanford researcher, bounced out on his ear if he was still in academia, but he’s in industry now and apparently hasn’t crossed any legal red lines.
For his part, Patterson replies, “We’re in a pandemic. I don’t know that these patients can wait 18 months for the results of a trial.” One could make the argument that if he really has the goods he has the moral obligation to get them out to the public and reduce the suffering as quickly as he can.
There’s always, of course, been a to and fro between doctors who push the boundaries of the medical profession and those more comfortable remaining well within them. We’ve seen that story played out many times at different levels in underserved diseases like ME/CFS, fibromyalgia, and now long COVID.
With the medical establishment unwilling to provide the funding needed to understand these diseases and assess treatments, patients have to choose between relegating themselves to few treatment options and generally poor health, and taking a chance with treatments that haven’t received much study.
For their part, doctors who use treatment approaches that haven’t been validated – and couldn’t have been validated through no fault of their own – can run risks. One of my MDs whose office was raided and his practice temporarily shut down felt that that kind of stuff was the cost of helping his patients. (It wasn’t the first time it had happened and he quickly reopened).
So here’s Patterson promoting a treatment that he says his data indicates is working but for which he has provided no hard evidence. He may have the goods but, Stanford background or not, it was inevitable he was going to starting taking some hits for moving forward so quickly, and he has.
In “Long-haul Covid patients desperate for answers turn to a private company,” NBC questioned whether vulnerable patients were taken advantage of, and featured a former patient whose test results returned to normal under his treatment – but she didn’t. Another doctor stated that marketing the diagnostic test/treatment protocol before both had been assessed in a wide variety of patients constituted a red flag.
That was nothing, though, compared to the takedown Mother Jones did in January of this year in their “Big Feature”, “Desperate Patients Are Shelling Out Thousands for a Long COVID Cure. Is it for Real?“.
- Bruce Patterson MD seemed to stride into a recent Solve ME event with the wind at his back. Proclaiming that over 80% of the over 24,000 long COVID patients on his protocol had achieved “profound improvements”, all seemed good.
- Patterson, though, has taken some hits from researchers and medical ethicists who’ve questioned if he’s gotten too far over the tips of his skis and is promising more than his data can deliver.
- A Mother Jones feature piece ” “Desperate Patients Are Shelling Out Thousands for a Long COVID Cure. Is it for Real?” noted that in the past year or so Patterson has been reined in by the FDA for overstating the results of one study, and was taken to task by a judge for exhibiting either “reckless indifference or deliberate gamesmanship” in a court case.
- Mother Jones brought in an outside statistician who skewered the findings of Pattersons’ core long COVID paper stating that it lacked some fundamental statistics, was too good to be true, and was too small a study to draw conclusions from. A prominent figure in the long COVID opined that he was one of the “sketchiest” figures around.
- Mother Jones, however, appeared to have a clear mission – to portray Patterson in a negative light. No patients who had done well with his protocols were featured and its characterization of a patient group as ‘cultish”, and its characterization of a doctor’s bedside manner was cheap and unnecessary.
- For his part, Patterson argues that given all the suffering produced by long COVID he’s morally bound to move forward to relieve it. He might say that he built his hypothesis around the data he gathered and then used it to successfully treat many of his patients. Whatever else anyone says, he must be right – and is going to prove it.
- The proof of his pudding will come with the placebo-controlled long COVID trial slated to start in a couple of months.
- Meanwhile, Patterson announced that he’s validated his biological signature in long COVID using another approach, that he expects his diagnostic kit for long COVID to be approved soon, and that ME/CFS, post-treatment Lyme disease, and other diagnostic kits are coming.
- Patterson’s preprint classification paper was short, lacking in description, and rudimentary. It attempted to use his long COVID cytokine grouping to differentiate ME/CFS. The long COVID cytokine package was derived, though, from an analysis of a wide variety of cytokines in long COVID (not ME/CFS).
- The paper did suggest that Patterson got his long COVID cytokine signature correct but its mediocre results regarding ME/CFS and other conditions suggested he has a ways to go before he can confidently state he’s found the same for these other conditions.
- Still, the paper was the first to attempt to do a difficult thing – differentiate long COVID, ME/CFS, post-treatment Lyme disease, and post coronavirus vaccine syndrome from each other immunologically. As such it could be viewed as a much-needed first step to attempt to differentiate these post-infectious diseases.
- So who is the real Bruce Patterson? Is he the former Stanford researcher who glommed onto long COVID early, quickly deciphered its secrets, and is hell-bent to do whatever it takes to help sick people? Is it just a matter of time before he gets all his ducks in a row and sets the long COVID, ME/CFS, and other diseases’ world afire? Or will the whole thing fall apart at some point?
- The rubber will meet the road with the clinical trial Patterson promised. A big, robust, and well-managed clinical trial will tell us if Patterson got too confident at some point or if he was right all along.
Mother Jones also reported that a judge concluded, in a suit that Patterson lost, that Patterson’s failure to disclose pertinent facts in a takeover bid for the company that produced leonlimab “reflected either reckless indifference or deliberate gamesmanship.” Mother Jones reported that Patterson then began disparaging the same drug he’d gotten in trouble for promoting on social media.
Following a long-COVID patient named Owen, the Mother Jones piece described a potentially costly treatment regime featuring tests and consults running about $600 and treatments up to $1,000 a month or more.
A Stanford bioethicist told Mother Jones, “The red flags are in full force on this one.” The piece noted one prominent member in the long-COVID community, Diana Berrent, founder of the big long-COVID Facebook group Survivor Corps, was not a supporter. She apparently hasn’t seen many good results from Patterson’s protocol, stating that “Bruce Patterson and Dr. Yo are, to me, the sketchiest guys out there.”
The biggest hit, though, came when Mother Jones took a whack at Patterson’s central finding – the cytokine results which underlie his entire diagnostic and treatment regimen. Maarten van Smeden, a statistician from University Medical Center Utrecht in the Netherlands, reported that the sample size – 224 patients was too small for a machine-learning study.
The fact that no-confidence intervals were provided and even the study’s sterling results raised question marks for Van Smeden. (He believed they were simply “too good to be true”.) He concluded that, “Even if we’d ignore the many methodological issues with this study…(it) is a long way from proving there is a cytokine storm going on.”.
Van Smeden presents a cautionary tale for us. While we do know that all studies need to be validated, hopefully by another research group, as patients we don’t have the statistical chops to properly assess these complicated studies. Nor is it clear that we should take one statistician’s word for it. With no consensus present, we’re stuck in the middle.
Patterson, for instance, might say, I have a good reputation. I got my data and developed a hypothesis and treatment plan around it. I even showed that the most potentially most problematic part of the hypothesis – that long-lived monocytes were carrying bits of the coronavirus around with them – was viable. Mostly importantly, my personal data indicates that my patients responded. Therefore I must be right. I’m not going mess around validating my diagnostic signature in an even larger group of patients – I’m going to treat people and and clinical trial will eventually show that I was right all along
The red flags didn’t only apply to Patterson. Mother Jones clearly spent little time finding and interviewing patients who had improved and, if Patterson is right, there are many of them. Describing a Facebook group of Patterson’s long-COVID patients as “weird and cultish” and a patient who got a “strange feeling” from Patterson’s partner Dr. Yogendra didn’t inspire confidence in the magazine’s objectivity.
The fact that Patterson said a randomized clinical trial will begin in a couple of months is a good sign that: a) he’s confident in the efficacy of his treatment protocol; and b) and he’s willing to subject it to a clinical trial. The important thing is that it actually happens and is big enough and rigorous enough to dispel any question about whether his protocol is effective or not.
While the Mother Jones piece raised some red flags, it didn’t indicate that Patterson is wrong. Anecdotal reports are close to useless. Only well-designed trials can provide us with any real certainty about the efficacy of his treatment approach. It’s simply too early to tell whether he’s on the right track or not.
The Solve M.E. Webinar
Beware! This part of the blog was taken from notes…
If Patterson had taken some hits from the media, you couldn’t tell at the Solve M.E. webinar – which provided our first update on his work in 6 months.
He breezed into the Solve M.E. webinar and confidently started announcing some impressive results. With his database now up to a staggering 24,000 patients, Patterson said his treatment protocol had produced “profound improvements” in over 80% of them.
Patterson stated that, using his machine learning technique, he’s been able to identify distinct cytokine signatures for people with chronic fatigue syndrome (ME/CFS), post-treatment Lyme disease, and post-coronavirus vaccination syndrome. If Patterson’s actually been able to do that, he’s just made a profound discovery in each of these diseases – none of which have a biological biomarker.
High VEGF levels (if my notes are right) are seen across the board in both ME/CFS, long COVID, etc. Produced by monocytes, they cause peripheral neuropathy. If you decrease VEGF, Patterson said, the neuropathy goes away, and so does the vasodilation and the headaches, tinnitus, brain fog, and heat and cold sensitivity, that comes with it.
Reducing VEGF also resolves a drop in blood pressure, which increases the heart rate in POTS. (That’s not exactly true. POTS is sometimes associated with a significant drop in blood pressure but is also sometimes associated with a significant increase in blood pressure, or no obvious change in it.)
Post-treatment Lyme disease is more confusing and, oddly enough, shares some commonalities with people with post-vaccination coronavirus syndrome. A cytokine node centered around elevated IL-13 characterizes these patients. That and an elevated IL-8 agree with past Lyme study findings.
In a good sign, Patterson stated that a totally different algorithm and classification system highlighted the same diagnostic factors in long COVID. It too suggests that inflammation of the blood vessels (IFN-y, IL-2) plays a key role.
People with the post-vaccination syndrome have the same symptoms as long-COVID patients. Noting that their prevalence was low compared to the chance of getting long COVID from having COVID-19, Patterson did not suggest staying away from the coronavirus vaccines.
A profound decrease in CD8 T-cells in long COVID and Lyme produces similarly profound immunosuppression, which, in turn, allows herpesviruses as well as two tick-borne pathogens Borrelia (Lyme) and Babesia to become reactivated in some patients.
Long-COVID patients for whom the virus is still active 3 months later can benefit from antivirals. After viral replication stops – leaving perhaps viral ghosts behind – drugs that stop viral replication won’t help.
Patterson also stated he’s seeing a lot of people with “long COVID” who have prior cases of bad EBV or Lyme disease. Even ten years afterward, the old Lyme cases still have distinct cytokines (high IL-13).
Patterson stated he’s producing the first diagnostic kit for long COVID and will produce similar kits for ME/CFS, Lyme, etc. In the middle of May, Patterson stated he was 1-2 weeks away from the first regulatorily approved diagnostic for long COVID. If my notes are correct, he expected European approval for his long-COVID diagnostic test on May 26th. (That doesn’t appear to have happened.)
Patterson reported that a 20-patient clinical trial of pure long COVID (unpublished) using maraviroc and statins (to bring down clotting) found that autonomic symptom scores (including fatigue and shortness of breath) decreased significantly and were associated with changes in cytokine levels. Reductions in TNF-a and IL-1 were also associated with improvements in fatigue.
Patterson noted that cytokine hub-directed therapeutics have been proposed for autoimmune diseases. He said his long-COVID patients get almost completely better. Those who come down with long COVID after Omicron are a bit different: he sees dramatic increases in IL-6 in them.
Focus on Precision Medicine
Patterson’s focus on precision medicine to inform treatments provided an inspiring ending. Stating, “It’s so critical to have precision medicine” in these diseases, in particular, Patterson noted that in his career “there has never been a more complicated disease”. Noting that, “We don’t make a move in cancer without precision medicine,” he asked, “Why not with long COVID?” Why do we stop wit “stupid markers like CRP (c-reactive protein)? Hear-hear!
Patterson’s Classification Preprint Study
Patterson’s preprint paper, “Cytokine Hub Classification of PASC, ME-CFS and other PASC-like Conditions,” was recently published. Note that a warning on the paper stated:
“preprints are preliminary reports that have not undergone peer review. They should not be considered to be conclusive, used to inform clinical practice, or referenced by the media as validated information”.
It’s a very short paper. It’s short in every way. Documentation is sparse (no mention of disease criteria), and the results and discussion sections are rudimentary. Even though the paper is focused on using cytokines as diagnostic tools, it doesn’t state which cytokines were highlighted in ME/CFS. (It also misspelled ME/CFS (as “ME/CSF”) five times). In short, everything in this paper shouts “rush job”. I can’t imagine that it’s going to mollify Patterson’s detractors – quite the opposite. It’s a bit puzzling as Patterson has generally produced full-fledged papers.
The study used a machine learning approach and the same group of cytokines that Patterson used to produce his cytokine signature in long-COVID patients.
64 people with post-acute sequelae of COVID-19 (PASC/long COVID), 26 with mild/moderate acute COVID-19 (MM), 25 with severe acute COVID-19, 50 with myalgic encephalomyelitis /chronic fatigue syndrome (ME-CFS), 29 with post-treatment Lyme disease (PTLD), and 42 post-vaccine individuals with PASC-like symptoms (POVIP).
Note that Patterson first assessed a wide range of cytokines to come up with his cytokine signature for long COVID. He didn’t do that here. Instead, he assumed that the same cytokines would tell a similar tale in these other diseases and plugged in the same cytokine set. The model did a good job at identifying the long-COVID, or PASC, patients. It wasn’t as effective with the other diseases.
The study used an F1 performance metric (0-1.00) to assess the quality of his cytokine-based machine learning model. The F1 score is a convenient score because it combines two different factors: precision and recall. For that reason, though, it’s not considered a particularly rigorous score. One website stated the F1 scores are best used for a “quick, hi-level, comparison“.
A 1.00 score means that every person was put in the correct classification. The score Patterson’s model received (.61) was described as being below “good” and on the lower end of “OK” (with OK ranging from .5-.8) by one site. The authors acknowledged this, though, noting that some conditions were more difficult to correctly classify. Post-treatment Lyme disease was particularly problematic.
How effective the model was at predicting who did and did not have ME/CFS can be seen in the “recall” and “precision” scores:
- Precision refers to “what proportion of predicted Positives are truly Positive?”; i.e. what proportion of the patients in the model predicted to have ME/CFS actually had it? Table I in the paper indicated that 66% of the patients that the model predicted to have ME/CFS actually had it.
- Recall refers to “what proportion of actual Positives are correctly classified?”; i.e. what proportions of the people with ME/CFS were classified correctly? The model identified 70% of the people who had ME/CFS as having ME/CFS.
The model did have high specificity though. That is, while the model missed identifying a significant portion of people who had ME/CF – and when it sometimes incorrectly reported that someone had ME/CFS – when it stated that someone did not have ME/CFS, it was generally spot-on (.90).
For some reason even when the model produced horrible results, otherwise it still did well with regards to specificity. The model, for instance, was simply terrible at classifying people with Lyme disease correctly (recall .17) and was poor (.55) at predicting who had Lyme disease, but it rarely suggested that someone who did not have Lyme disease had it (.98!)
Trying to differentiate these very similar occurring diseases using just 14 cytokines was not an easy task, and Patterson’s the first one I know of to even attempt it. Rather than seeing this as a victory or a defeat I prefer to view it as a good first step and applaud Patterson for taking on this knotty problem. I hope we’ll see many more studies like this as researchers take on post-infectious illnesses.
The model’s lower F1 score and the 30% mistake rate in the model’s recall and precision figures suggest the model was perhaps a start but wasn’t particularly effective either. That wasn’t the message I got from the webinar.
The Patterson saga reminds me just how much at sea we are with all this and how little firm ground there is to stand on. Take the critique the statistician in Mother Jones leveled on Patterson’s first paper. Would another statistician have another opinion? We don’t know.
Patterson has a good pedigree, is confident and convincing, is gathering more patients all the time, and must have tons of data. He’s got a fascinating hypothesis, appears to have validated an important part of it (monocytes with coronavirus proteins), and is planning to do a clinical trial.
Over the past year or so, he’s also gotten in trouble with the FDA, exhibited, according to a judge, “reckless indifference or deliberate gamesmanship” in a legal case, had his core study skewered by an outside statistician, and some researchers and medical ethicists have called his promotional style into question.
Who is the real Bruce Patterson? Is he the former Stanford researcher who glommed onto long COVID early, quickly deciphered its secrets, and is hell-bent to do whatever it takes to help sick people? Is it just a matter of time before he gets all his ducks in a row and sets the long COVID, ME/CFS, and other diseases’ world afire? Or is it all going to fall apart at some point?
Time will tell. The most important step in this saga for all of us will be the clinical trial that Patterson stated will begin in a couple of months. Let’s hope it happens and is big enough and done well enough that we’re finally left with some certainty.
Health Risings Hot, Hot Summer Drive Update
It’s still early summer but it’s gotten blistering hot over at Health Rising as the donations continue to roll in – leaving us 2/3rds the way to our goal after just five days of our 3-week (3-week?) fundraising drive. Thanks to the over 100 people who have contributed over $10,000 thus far.
The Patterson blog exemplifies a core commitment of Health Rising – to try to be as balanced and objective as possible – even in a situation where we would above all love to have good news. If that kind of approach makes sense to you please support us.
HEALTH RISING IS NOT A 501 (c) 3 NON-PROFIT