Peter Attia MD283 ‒ Gut health & the microbiome: improving and maintaining the microbiome, probiotics, & more
CHAPTERS
- 0:00 – 3:23
Colleen's Background and Why She Started Pendulum
Peter Attia explains why he held off on this topic and introduces Colleen Cutcliffe. She describes her path from a biochemistry PhD through pharma and a DNA sequencing startup to founding Pendulum, premised on DNA sequencing making microbiome science real.
- •Attia cites a lack of rigorous guests and a sketchy product market as reasons for his earlier reluctance
- •Cutcliffe's PhD was in biochemistry and molecular biology, followed by a postdoc and work in Parkinson's drug development
- •She joined a DNA sequencing instrument company before starting Pendulum
- •No new probiotic ingredient had been introduced in over 50 years
- •Sequencing technology let microbiome work be treated as a systems biology problem
- 3:23 – 5:45
What the Microbiome Is and How It Develops Across Life
Cutcliffe defines the microbiome and describes how it is first seeded at birth, grows more diverse through childhood, and declines with age. They also discuss whether the "10 to one" figure matters and how bacterial cells differ from human cells.
- •The microbiome includes bacteria, viruses, fungi and yeasts on and in the body
- •Initial seeding comes largely during vaginal delivery, with diversity then rising with diet and exposure
- •Diversity declines with age, and peak diversity is roughly when you could eat anything as a teenager
- •Exact ratios of microbial to human cells matter less than the functional output of microbes
- •Each bacterial cell is an independent organism that can divide in 10-15 minutes and evolve quickly
- 5:45 – 17:42
Good vs. Bad Bacteria, Microbiome Variability, and Animal Models
Cutcliffe argues that "good" and "bad" bacteria are context-dependent, using C. difficile as an example. She discusses the Human Microbiome Project's finding of large person-to-person differences and the limits of mouse and germ-free models.
- •Almost everyone carries C. difficile at low levels without harm
- •People differ greatly at the strain level, with more redundancy seen at the level of function
- •The Human Microbiome Project was a one-time snapshot with no longitudinal component
- •The role of viruses in the microbiome is still unknown
- •Diet drives animal microbiomes, and germ-free mouse models don't reflect real life
- 17:42 – 27:52
Types of Bacteria, Gut Anatomy, and Rapid Evolution
Attia and Cutcliffe cover how bacteria are divided (aerobic vs. anaerobic, location, mucin-dwelling) and how the microbiome changes from mouth to colon. The conversation then turns to taxonomy, strain definition, and why bacteria's fast evolution makes the microbiome so mutable.
- •The distal colon is strictly anaerobic, while Lactobacillus and Bifidobacterium live nearer the small intestine
- •A story about hyena mouth microbiomes illustrates interest in new antibiotic sources
- •Taxonomy runs phylum, family, genus, species, strain, and strain definition may evolve
- •Pendulum audits manufactured batches by whole genome sequencing and biochemical assays
- •Bacteria evolve toward their environment, chiefly the diet they are fed, making the microbiome highly mutable
- 27:52 – 39:52
Measuring the Microbiome: Sequencing, qPCR and Functional Assays
Cutcliffe walks through shotgun sequencing with long- and short-read methods and qPCR for quantifying strains, and explains why longitudinal data matter. They discuss functional assays of short-chain fatty acids and the difficulty of sampling the gut directly.
- •Shotgun sequencing needs long reads paired with short reads, and gene copy number can matter, as with a digoxin-metabolizing strain
- •qPCR gives quantitative abundance of specific strains
- •One snapshot is not enough because the microbiome changes with diet, travel and antibiotics
- •Full longitudinal characterization costs thousands of dollars per sample
- •Functional assays measure short-chain fatty acid output, and sampling along the gut is prone to contamination
- 39:52 – 47:23
Fiber, Butyrate and Feeding the Microbiome
Cutcliffe explains that microbial carbohydrate metabolism means fermenting fibers into short-chain fatty acids, ultimately butyrate. They discuss why butyrate matters, fiber intake levels, supplement delivery, and which strains are involved.
- •Butyrate is the main energy source for colon cells and triggers GLP-1 release from L cells
- •Insoluble fiber from vegetables and fruit feeds these strains, and a green drink is no fiber substitute per Attia
- •Recommended fiber is around 20-30 grams a day, while the average American eats far less
- •Delivery and proximity matter, which is why Pendulum includes a small amount of inulin in its pill
- •Clostridial strains, Bifidobacterium and Akkermansia contribute to the pathway, mostly in the distal colon
- 47:23 – 53:25
C. difficile and Fecal Microbiota Transplants
Cutcliffe explains how antibiotics let C. diff overgrow and why FMT, with about a 99% success rate, can cure it. They cover FMT methods, safety concerns, and the FDA's attempt to restrict it.
- •Antibiotics remove competitors, letting C. diff propagate unchecked, and the antibiotic cure succeeds about 70% of the time
- •FMT has roughly a 99% success rate and works best via the original delivery route, less so as capsules
- •Risks include unknown donor pathogens and changed metabolism, with case reports of new obesity
- •The FDA tried to shut FMT down but reversed after patient and family protests
- •Attia suggests banking one's own stool as an autologous transplant
- 53:25 – 59:15
The Gut-Brain Connection and Factors Shaping the Microbiome
Prompted by the question of why transplanted microbiomes persist, Cutcliffe describes how the gut produces neurotransmitters, influences cravings, and connects to the brain via the vagus nerve. They discuss what depletes the microbiome: antibiotics, diet, aging, stress and circadian disruption.
- •The gut produces serotonin, dopamine, GABA, GLP-1 and GIP and has its own neurons
- •Microbes may influence food cravings and stress or anxiety
- •Antibiotics and nutrition are the two biggest drivers of change
- •Aging, intense stress, circadian disruption and menopause deplete the microbiome
- •Depletion means loss of specific strains and functions, such as low Akkermansia
- 59:15 – 1:03:48
The Vaginal Microbiome
Attia asks about factors affecting the vaginal microbiome. Cutcliffe describes the protective role of acid-producing Lactobacillus, links to bacterial vaginosis and preterm labor, and notes how little is known about menopause and the microbiome.
- •Lactobacillus strains keep the vaginal environment acidic as a protection
- •Diet may influence the vaginal microbiome, as in the cranberry juice recommendation for UTIs
- •Bacterial vaginosis has been highly associated with preterm labor
- •Menopause-related microbiome research is extremely sparse
- •Gut microbes may remove a modification on estrogen, potentially raising circulating estrogen, but this is very early
- 1:03:48 – 1:06:26
Microbiome and Obesity: Mouse Transplants and Challenges in Humans
They discuss reproducible mouse findings that transplants can make animals lean or obese, and why this hasn't translated simply to humans. The challenges are donor safety and the irreproducibility of stool as a "messy drug," which led Pendulum to seek the specific strains responsible.
- •Mouse studies show lean or obese phenotypes transfer, including from human donors
- •Human FMT studies in type 2 diabetes exist, but donor safety is a concern
- •A donor's stool varies over time, so dose-response studies are impossible
- •Pendulum's premise is to identify and manufacture the active subset of strains
- 1:06:26 – 1:16:34
Probiotic Strains, Regulation and the Akkermansia Observation
Cutcliffe explains why Lactobacillus and Bifidobacterium dominate probiotic labels (easy to grow, grandfathered by the FDA) and how regulatory pathways differ. The conversation then turns to twin studies linking low Akkermansia to metabolic disease and the puzzle of healthy people on very different diets.
- •Lactobacillus and Bifidobacterium are facultative, easy to culture, and grandfathered as safe
- •New strains sold to consumers must go through the GRAS pathway, while drug companies file INDs
- •Twin studies show depleted Akkermansia in the obese or diabetic twin
- •Healthy people are said to have roughly 5-10% Akkermansia across populations
- •Cutcliffe says it is unknown whether diet-independent host factors explain the pattern
- 1:16:34 – 1:31:13
Probiotics vs. Prebiotics, CFUs and Manufacturing Viability
Cutcliffe defines probiotic, prebiotic, postbiotic and synbiotic, then critiques CFU as a metric. She describes the manufacturing process of freeze-drying and encapsulation and how much viability is lost along the way.
- •Probiotics are organisms, prebiotics their food, and postbiotics what they secrete, like butyrate
- •Lactobacillus and Bifidobacterium are best documented for GI symptom relief
- •CFUs count only cells that form colonies, while flow cytometry shows live, dead and in-between cells
- •Freeze-drying can kill about 90% of cells, and capsule type affects delivery
- •Labels may report CFUs at manufacture rather than at end of shelf life, and this is unregulated
- 1:31:13 – 1:39:57
Recovering from Antibiotics
Attia asks what to do after taking antibiotics. Cutcliffe notes that antibiotics are often broad spectrum and that the post-antibiotic period is a chance to reset with a high-fiber, high-polyphenol diet, and she discusses newer evidence for probiotics during antibiotic courses.
- •Childhood antibiotic exposure has been associated with later obesity, type 2 diabetes, ADHD and celiac disease
- •Post-antibiotic is a good time to clean up the diet because the microbiome is a blank slate
- •Studies suggest probiotics during antibiotics or an FMT afterward may support a healthier microbiome
- •Cutcliffe says she would double down on probiotics during antibiotics, then eat a high-fiber diet
- •Antibiotics have become more broad spectrum over time, and phage therapies aim for precision
- 1:39:57 – 1:48:50
Artificial Sweeteners and Individual Responses to Diet
Attia raises non-nutritive sweeteners, and Cutcliffe says the data are conflicting and mostly from animals. They discuss anecdotal individual differences and the case for crossover study designs. Cutcliffe cites a diet study in which higher starting Akkermansia predicted better response.
- •Evidence on sweeteners and the microbiome is conflicting and sweeteners are not all alike
- •Attia describes patients whose weight changes after swapping diet soda for soda water
- •Cutcliffe argues crossover designs suit microbiome research because the individual matters
- •A diet study found higher starting Akkermansia associated with better responses on BMI, A1C and weight
- •Attia frames a hypothesis that low Akkermansia may blunt healthy interventions
- 1:48:50 – 2:01:41
Manufacturing Akkermansia and the Pendulum Business Model
Cutcliffe explains why Akkermansia is hard to produce, requiring a closed anaerobic system with mixed gases and a custom plant. She also discusses fake Akkermansia products, the funding raised, and the decision to sell to consumers.
- •Akkermansia muciniphila is a strict anaerobe, and a single molecule of oxygen can kill a batch
- •Production uses a master freezer stock, mixed gases and anaerobic chambers throughout, including for flow cytometry
- •Freeze-dried Akkermansia is stable but must be kept dry, hence the desiccant packets
- •Pendulum raised $150 million and spent about $10 million on its plant
- •Selling to consumers aimed to democratize access beyond prescriptions, despite the marketing-heavy supplement space
- 2:01:41 – 2:28:07
How Akkermansia Boosts GLP-1 and the Clinical Trial Results
Cutcliffe describes three mechanisms by which Akkermansia stimulates GLP-1 secretion and contrasts it with injected GLP-1 drugs. She then presents the five-strain formulation trial in type 2 diabetes, in which A1C fell 0.6 and glucose spikes fell 34% versus placebo.
- •Mechanisms: surface protein AMUC1100, secreted protein P9, and propionate converted to butyrate
- •The effect is physiologic and meal-linked, not comparable to injected GLP-1 drugs
- •Akkermansia was discovered by Dr. Lee Kaplan, a bariatric surgeon, in the early 2000s
- •The 76-person placebo-controlled trial found a 0.6 A1C drop and 34% lower glucose area under the curve over 90 days
- •The formulation without Akkermansia performed in between but not statistically significantly
- 2:28:07 – 2:48:48
Products, Customer Data and Ongoing Studies
Cutcliffe describes Glucose Control, the lower-dose Metabolic Daily, and standalone Akkermansia, along with pricing and cold shipping. She discusses customer feedback on cravings, trial stool samples, ongoing third-party studies, and future strains.
- •Glucose Control costs $165 a month and Metabolic Daily $49, with the same five strains at lower dose
- •Labels report AFU (active fraction units) rather than CFUs
- •Customer reports of reduced sugar cravings led to a pilot study and plans for a clinical trial
- •Starting microbiome and Akkermansia levels did not predict response in the trial
- •About a dozen third-party trials are underway, including one on bipolar disorder, and Cutcliffe hints at other strains tied to immune response
