27 years ago, I was diagnosed with ME/CFS. Even as far back as the early 1970s, while at university, I experienced severe cognitive difficulties—including a sudden decline from managing triple honors to simply trying to finish my degree. At the time, my physician correctly attributed these issues to stress, though the biological mechanism was unclear. Today, it’s recognized that stress can readily disrupt the microbiome, contributing to dysbiosis.
During these 27 years, I’ve thoroughly reviewed most of the scientific literature on ME/CFS. I was a subscriber to the Journal Of Chronic Fatigue Syndrome throughout its publication from 1995 to 2007. I’ve experienced multiple relapses, often triggered by stress, leading me to adopt a preventative approach: avoid stress when possible. My guiding philosophy now is “Que Sera, Sera (Whatever Will Be, Will Be)”—which can be especially challenging given the current political climate in the US.
ME/CFS and Long COVID (LC) are both highly heterogeneous, varying widely in symptoms and duration; specific symptoms of LC can change or resolve unpredictably over time. Studies—including those involving twins and across genders—demonstrate that such diversity complicates the search for a universal treatment that provides consistent symptomatic relief.
My Criteria
Recently, friends have sent me several new studies asking for my feedback. Given years of seeing the field marked by repeated announcements of “breakthroughs,” my essential question remains:
- “Do these papers suggest immediate, actionable clinical steps? “
- “Is there robust evidence that these interventions benefit a substantial proportion of people with ME/CFS?”
To see if the papers do not met those criteria, I will quickly review them. I will neither get hopes up nor excited about them if they fail…. they are speculation. Speculation is awesome for getting grants for research but not for improving patients. Getting excited about them and the subsequent disappointment is not healthy for the microbiome.
Two Example Papers
To me the two papers that met these criteria are:
- From Journal Of Chronic Fatigue Syndrome
- Hypercoagulation hypothesis
Treatments primarily involved rotating courses of appropriate antibiotics and administering suitable anticoagulants. Both methods can profoundly affect the microbiome; for example, anticoagulation improves oxygen levels, creating conditions that affect bacterial growth patterns. My guiding principle—KISS (Keep It Simple, Stupid)—leads me to focus on the belief that the underlying issue is persistent microbiome dysbiosis.
This conviction is why I dedicated substantial time and some personal resources to the development of the free Microbiome Prescription website, which was originally designed with a sole focus on ME/CFS.
BioMapAI: Artificial Intelligence Multi-Omics Modeling of Myalgic Encephalomyelitis / Chronic Fatigue Syndrome [2025]
Being a former University Instructor at Chapman University for Artificial Intelligence, my first item was check cross validations. “including ‘omics altogether (AUC=82.3%), immune (78.5%), KEGG (69.1%), species (71.5%), and metabolome (76.4%), while Glmnet excelled in Quest data (74.8%)“. The study did use Shotgun sampling but ignored The taxonomy nightmare before Christmas… (the blue whale in the room).
Two years ago, using my model, I posted “Cross Validation of AI Suggestions for Nonalcoholic Fatty Liver Disease“, I got 92% correct for to take and 83% correct for to avoid. IMHO, their model should have perform better.
Key Bottom line — There was no treatment suggestions or protocols
A ton of issues were identified …. ” This map uncovers disrupted associations between microbial metabolism (e.g., short-chain fatty acids, branched-chain amino acids, tryptophan, benzoate), plasma lipids and bile acids, and heightened inflammatory responses in mucosal and inflammatory T cell subsets (MAIT, γδT) secreting IFNγ and GzA.” but no discussion of a unified treatment for all of them. ME/CFS is complex.
My own approach, the microbiome dysbiosis, is very treatable as demonstrated by Microbiome Prescription, some examples of treatment suggestions are here. Since it is likely that the microbiome drives most of the above issues — life is simpler and improvement can happen in weeks in many cases. All of the issues cited in this paper, listed below, has evidence that the microbiome is a significant contributor.
ME/CFS is characterized by persistent fatigue, post-exertional malaise, multi-site pain, sleep disturbances, orthostatic intolerance, cognitive impairment, gastrointestinal symptoms, and other issues. This complexity not only hinders timely diagnosis but also poses significant challenges for effective treatment.1,2,3. The pathogenesis of ME/CFS is not well understood, with some triggers believed to include viral infections such as Epstein-Barr Virus (EBV)4, enteroviruses5 and SARS coronavirus6, in addition to bacterial infections and other causes7.
A Perspective on the Role of Metformin in Treating Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) and Long COVID [2025]
I went over to my Metformin page on microbiome prescription and saw the most often reported impacts,
Looking at studies on ME/CFS
- Akkermansia is low with Long COVID,Gulf War Syndrome, Irritable Bowel Syndrome
- Bifidobacterium ⬇️ ⬇️ ⬇️ ⬇️
- Lactobacillus ⬇️ ⬇️ ⬇️
- Roseburia inulinivorans ⬇️
- Bacteroides ⬆️ ⬆️ ⬇️ ⬇️
- Prevotella ⬆️ ⬇️
My conclusion is simple, it is a viable candidate for treatment…. BUT….
The BUT with Single Item Treatments!
If the root issue is microbiome dysbiosis, no single item is likely to be sufficient to be a “magic cure all”. We are not talking about eliminating a single virus or bacteria. A single virus or bacteria belief has been rampant with ME/CFS researchers for decades. Why? Simple, treating a dozen shifts at the same time is too complex given the methods available. But it is not too complex with a suitable fuzzy logic expert system running of several million facts.
On the last article the reader wrote:
I find it interesting because they found the correlation with the microbiome. I wonder why researchers don’t look more at it.
The reason is simple – they lack the skills and training to deal with it. Additionally, there is the absence of data in a suitable form. IMHO what is needed?
- Training in the full range of Artificial Intelligence methods. I have found fuzzy logic is essential. It is rarely taught in AI classes today, everything has shifted to the hottest tech: Large Language Models (LLM, ChatGPT) that is well known for hallucinations.
- A suitable database, a summary of the data used for the fuzzy logic engine is here. Some analysis uses over 13 million facts encoded into a database. Most of these facts would not “be seen” using the bots that provides data to LLM due to paywalls and other factors.
Suggested reading (intro book with a sample of the text)
CFS Remission 