# 10 Things I Have Learned About Long COVID From Client Data and Research

> I have not written in depth about Long COVID for a while, although it has continued to appear repeatedly in the complex cases I analyze.

- **Author:** Mohammed Attallah (BiomeLogic)
- **Published:** 2026-09-30
- **Category:** SIBO
- **Tags:** long-covid, sibo, pots, dysautonomia, bile-acids, hormones, mitochondria, butyrate, microbiome, serotonin, dopamine, glp-1, cardiovascular, autoantibodies
- **Canonical URL:** https://biomelogic.net/articles/10-things-i-have-learned-about-long
- **License:** CC BY-NC 4.0 — please cite "Mohammed Attallah, BiomeLogic" with link to https://biomelogic.net/articles/10-things-i-have-learned-about-long.

---

I have not written in depth about Long COVID for a while, although it has continued to appear repeatedly in the complex cases I analyze. Over time, working through clinical histories, longitudinal laboratory data, microbiome profiles, metabolomics, mitochondrial testing, immune markers, autonomic findings and treatment responses has changed how I think about the condition. I no longer find it useful to describe Long COVID simply as persistent inflammation, mitochondrial dysfunction, dysautonomia, viral persistence, mast-cell activation or microbiome disruption. Each of those can be relevant, but none adequately explains the entire phenotype.

What has become far more informative is following the relationships between these systems. I have reviewed cases involving severe post-exertional malaise, POTS, gastrointestinal dysmotility, recurrent SIBO and IMO, bile-acid abnormalities, mast-cell-type reactivity, neuropathic symptoms, endothelial and coagulation abnormalities, marked cytokine disturbances, intestinal inflammation, altered mucosal immunity, unusual amino-acid patterns and profound mitochondrial abnormalities. These findings rarely sit in isolation. A patient with severe orthostatic intolerance may simultaneously have impaired mitochondrial respiration, disturbed nitric-oxide biology, extreme intestinal fermentation, inflammatory cytokine activation and altered mucosal immunity. Another may have thrombosis, persistent antiphospholipid antibodies, bile-acid malabsorption, dysautonomia and recurrent intestinal overgrowth. Another may have conventional bloodwork that appears relatively unremarkable while exercise physiology, small-fiber testing or metabolomics reveals a very different biological state.

The strongest lesson from these cases has been that Long COVID often behaves less like a single lesion and more like a disturbance in communication between metabolic, vascular, immune, neural and gastrointestinal systems. The abnormalities begin to make much more sense when they are reconstructed as a physiological sequence rather than interpreted as independent diagnoses.

These are ten observations that now shape how I think about Long COVID.

**1. Long COVID is not one disease state. It contains multiple biological endotypes that can converge on similar symptoms.**

Fatigue is one of the clearest examples. Two people can both describe profound fatigue while the physiology producing that symptom may be completely different. One may have impaired oxygen extraction and endothelial dysfunction. Another may have inadequate mitochondrial oxidative phosphorylation. Another may have severe orthostatic intolerance with reduced effective cerebral perfusion. Another may have an inflammatory state that redirects amino acids and cellular energy toward immune metabolism. Another may be nutritionally depleted because of chronic gastrointestinal dysfunction. In many patients, several of these mechanisms coexist.

This heterogeneity is increasingly visible in multi-omics research. Long COVID cohorts have demonstrated abnormalities involving immune-cell populations, platelet activation, arginine metabolism, methionine metabolism, taurine metabolism, lipid handling and the tricarboxylic-acid cycle. Integrated analyses have also identified different molecular patterns among neurological, cardiopulmonary, musculoskeletal and multisystem phenotypes rather than one universal biochemical signature.

I have seen the same pattern in client data. One particularly severe post-COVID case had a resting heart rate around 72 beats per minute that increased to approximately 140 with standing. The same patient had an extraordinary gastrointestinal phenotype, with breath testing reaching approximately 126 ppm hydrogen and 90 ppm methane. Mitochondrial testing reported Complex IV activity around 22 percent of the laboratory reference and Complex II plus III around 16 percent, while citrate synthase activity was reported at approximately 563 percent.

Citrate synthase is commonly used as an indirect marker of mitochondrial abundance because it is a relatively stable matrix enzyme. A pattern of very high citrate-synthase activity alongside severely impaired respiratory-chain activity is therefore interesting. It raises the possibility of a compensatory increase in mitochondrial machinery while the functional quality of that machinery remains poor. It does not prove mitochondrial biogenesis, since citrate synthase can change for reasons other than mitochondrial number, but the relationship between mitochondrial content and respiratory function becomes considerably more informative than either measurement by itself.

The same case demonstrated substantial intestinal and immune abnormalities. Fecal calprotectin changed over time from approximately 25 to 120 to 80.9 to 57 to 202 and later 122 µg/g. Lactoferrin became positive. Secretory IgA exceeded 7,500 µg/g at one point and remained around 5,310 on another measurement. The immune profile was equally striking, with IFN-γ reported at 183.2 pg/mL, IL-4 at 61.2, IL-13 at 53.3, IL-6 at 32.8, IL-8 at 26.4, IL-10 at 13.9 and TNF-α at 12.1. Intermediate monocytes were approximately 14.27 percent.

None of these measurements, taken individually, proves a Long COVID mechanism. Specialized assays have analytical limitations and reference ranges vary substantially between laboratories. What matters is the convergence. Autonomic dysfunction, respiratory-chain impairment, intestinal inflammation, exaggerated mucosal immune activity, dysbiosis and systemic immune signaling were occurring in the same biological system.

That is the level at which these cases become much easier to understand.

**2. Arginine metabolism in Long COVID is more complicated than simply having too little arginine. The important variable may be how effectively arginine is converted into functional nitric oxide.**

Nitric oxide is one of the most important molecules connecting vascular physiology, platelets, mitochondria, immunity and autonomic function. Endothelial nitric-oxide synthase, or eNOS, uses L-arginine together with oxygen, NADPH and tetrahydrobiopterin, BH4, to generate nitric oxide and citrulline. Nitric oxide then activates soluble guanylate cyclase in vascular smooth muscle, increases cyclic GMP and promotes vasodilation. It also restrains platelet activation and leukocyte adhesion and helps regulate microvascular blood distribution.

Long COVID research increasingly suggests that this system can become dysregulated. Human studies measuring arginine, citrulline, ornithine, monomethylarginine, ADMA and SDMA have found impaired indices of nitric-oxide bioavailability in Long COVID. Ratios such as arginine to ADMA and arginine to ornithine can therefore be more informative than arginine concentration by itself.

This is important because an elevated or normal arginine level does not necessarily mean nitric-oxide biology is functioning normally. Arginine has competing metabolic destinations. Arginase converts arginine into ornithine and urea, while nitric-oxide synthases use it to generate nitric oxide and citrulline. ADMA competitively inhibits NOS. Oxidative stress can oxidize BH4, and when BH4 availability becomes insufficient relative to BH2, eNOS can become uncoupled. Electrons that should participate in nitric-oxide formation begin reducing molecular oxygen directly, generating superoxide.

Superoxide then reacts extremely rapidly with nitric oxide to form peroxynitrite. At that point, the system can have adequate substrate availability while effective endothelial nitric-oxide signaling remains impaired. The endothelial problem can therefore coexist with considerable reactive nitrogen chemistry elsewhere.

Inflammatory nitric-oxide synthase biology adds another layer. Inducible NOS can generate far greater amounts of nitric oxide during immune activation than constitutive endothelial NOS. In a highly oxidizing environment, increased nitric oxide does not necessarily indicate healthy vascular signaling. It may instead increase nitrosative stress.

This is why I think the stronger model is not simply low arginine or high arginine, but dysregulated arginine flux with impaired compartment-specific nitric-oxide signaling. When I see vascular symptoms, cold extremities, headaches, exertional intolerance, orthostatic symptoms or evidence of endothelial dysfunction, I think beyond serum arginine and look at the relationships among arginine, citrulline, ornithine, ADMA, SDMA, arginase competition, BH4 biology, oxidative stress and functional endothelial physiology.

The pathway becomes particularly important when it is connected to mitochondrial biology. Peroxynitrite and other reactive species can modify mitochondrial proteins and lipids, while mitochondrial respiratory dysfunction can itself increase reactive oxygen species production. This creates the possibility of a self-amplifying redox loop in which impaired endothelial nitric-oxide signaling and mitochondrial dysfunction reinforce one another.

**3. Tryptophan metabolism may be one of the most important intersections between immunity, the gut, the nervous system and cellular energetics.**

Tryptophan is often discussed as though its primary purpose were serotonin production, but only a relatively small portion of systemic tryptophan follows that pathway. A much larger fraction is metabolized through the kynurenine pathway. Inflammatory signaling, particularly interferon-associated signaling, can increase indoleamine 2,3-dioxygenase activity and redirect tryptophan toward kynurenine metabolism.

Human Long COVID data support this pattern. A systematic review and meta-analysis including 14 studies and 1,167 participants found lower tryptophan in Long COVID and a significantly increased kynurenine-to-tryptophan ratio. That ratio can be more informative than either molecule alone because it reflects substrate partitioning and immune-regulated tryptophan catabolism.

The pathway then branches further. Kynurenine can move toward kynurenic acid, 3-hydroxykynurenine and quinolinic acid, while downstream metabolism ultimately contributes to de novo NAD synthesis. The physiological consequences therefore depend on where the metabolic flux goes rather than simply whether kynurenine is elevated.

I have seen this complexity in longitudinal client data. In one case, a specialized assay reported serotonin falling from approximately 43.56 to 4.39 while dopamine fell from 3.59 to below 0.1. Urinary quinolinic acid, however, remained around 1.6 rather than becoming markedly elevated. I would not use urinary neurotransmitter measurements to diagnose central serotonin or dopamine deficiency because peripheral and urinary measurements cannot be assumed to reflect synaptic concentrations in the brain. The broader lesson is that a low downstream molecule does not automatically reveal where the upstream substrate has gone.

Tryptophan can enter host kynurenine metabolism, serotonin synthesis or microbial metabolism. Intestinal organisms can convert tryptophan into indole compounds including indole-3-aldehyde, indole-3-acetate and indole-3-propionate. Several of these interact with the aryl-hydrocarbon receptor and influence epithelial integrity, IL-22 signaling and mucosal immune function.

A widely discussed 2023 Cell study proposed a pathway in which persistent viral signaling and type-I interferon activity impair intestinal tryptophan absorption, reduce peripheral serotonin, alter platelet serotonin handling and decrease vagal signaling. That model is important, but the serotonin component should not be treated as settled science because later human work failed to reproduce peripheral serotonin reduction consistently.

The broader tryptophan story is considerably stronger than the claim that all Long COVID patients have low serotonin. The more useful framework is tryptophan partitioning. Plasma tryptophan, kynurenine, the kynurenine-to-tryptophan ratio, downstream kynurenines, nutritional status, intestinal absorption, microbial indole production and the inflammatory environment controlling IDO activity together provide a much more informative mechanistic picture.

**4. ACE2 deserves much more attention in Long COVID gastrointestinal biology because intestinal ACE2 is not merely a viral receptor. It is part of amino-acid transport and mucosal homeostasis.**

ACE2 is usually introduced through the renin-angiotensin system, where it counterbalances angiotensin II signaling partly through conversion toward angiotensin 1-7 and the Mas-receptor axis. This role is important for vascular tone, inflammation and tissue signaling, but the intestine adds another layer.

ACE2 forms a functional complex with the neutral amino-acid transporter B⁰AT1, encoded by SLC6A19. B⁰AT1 transports several neutral amino acids, including tryptophan, and ACE2 is important for its normal expression at the intestinal epithelial surface. This creates a mechanistic bridge between SARS-CoV-2 biology, intestinal amino-acid absorption and microbial ecology.

Reduced intestinal ACE2 or B⁰AT1 activity can impair tryptophan transport. Lower intracellular tryptophan availability can affect mTOR signaling, and mTOR contributes to regulation of antimicrobial peptides. Antimicrobial peptides in turn influence which microorganisms can successfully occupy the mucosal ecosystem.

This pathway is well established physiologically and experimentally. What remains less certain is how strongly persistent ACE2-B⁰AT1 dysfunction contributes to established human Long COVID months or years after infection. I therefore consider this a compelling mechanistic model rather than a proven universal Long COVID mechanism.

One post-COVID case made this pathway particularly interesting because β-defensin-2 was low on repeated measurements in the setting of substantial dysbiosis and gastrointestinal dysfunction. Low β-defensin-2 does not prove impaired ACE2-B⁰AT1 signaling, but it changes the direction of investigation. Instead of treating the microbiome as an isolated collection of organisms, it raises the possibility that the host is failing to impose normal ecological control over the mucosal surface.

In that setting, neutral amino-acid profiles, plasma tryptophan, kynurenine, citrulline, antimicrobial-peptide production, intestinal inflammation and microbial composition become much more informative when interpreted together rather than separately.

**5. Bile acids may be one of the most underappreciated systems connecting Long COVID gastrointestinal dysfunction with metabolism, immunity and the microbiome.**

Bile acids are often treated clinically as digestive detergents, but their biology is much broader. The liver synthesizes primary bile acids, principally cholic acid and chenodeoxycholic acid, which are then conjugated primarily with glycine or taurine and secreted into bile. Once they reach the intestine, bacterial bile-salt hydrolases can deconjugate them, and a smaller group of anaerobic organisms carrying the bai pathway can perform 7α-dehydroxylation, producing secondary bile acids such as deoxycholic acid and lithocholic acid.

These compounds are not metabolically inert. Bile acids activate FXR, and ileal FXR signaling induces FGF19. FGF19 then travels through the portal circulation and suppresses hepatic CYP7A1, reducing new bile-acid synthesis. Bile acids also signal through TGR5, influencing enteroendocrine function, GLP-1 secretion, energy metabolism, immune signaling and gastrointestinal physiology. At the same time, bile acids exert antimicrobial pressure and influence microbial selection.

This creates a reciprocal system in which the microbiome modifies the bile-acid pool while the bile-acid pool shapes the microbiome.

Several cases I have reviewed after COVID have shown abnormalities consistent with disrupted enterohepatic physiology, including bile-acid malabsorption, abnormal fecal bile-acid patterns, postprandial gastrointestinal symptoms and persistent diarrhea. One particularly complex post-COVID patient had recurrent SIBO and IMO together with bile-acid malabsorption, severe nutritional depletion, leukopenia and neutropenia, dysautonomia, two upper-extremity venous thromboses and persistent isolated anticardiolipin IgM. That is not a simple dysbiosis phenotype. It involves the intestine, liver, vascular system, coagulation, immunity and nutritional status at the same time.

The direction of bile-acid abnormalities also matters. One patient may have inadequate secondary bile-acid generation because 7α-dehydroxylating organisms have been depleted. Another may have excessive secondary bile acids. Another may have excessive hepatic synthesis because ileal FXR-FGF19 feedback is inadequate. Another may synthesize bile acids normally but fail to reabsorb them efficiently.

These represent different physiological states and should not be collapsed into one label. When bile-acid diarrhea is suspected, serum C4 can provide an indirect estimate of hepatic bile-acid synthesis while fasting FGF19 can provide information about ileal feedback signaling. In established bile-acid diarrhea, higher C4 and lower FGF19 are well described.

Long COVID metabolomics has also identified disturbances in lipid-associated and bile-acid-related pathways, although the evidence is not yet sufficient to define one characteristic Long COVID bile-acid profile. The larger point is that some post-COVID intestinal cases may be approached as microbiome disorders when part of the ecological disturbance is actually being driven by abnormal bile physiology.

**6. Mitochondrial abundance and mitochondrial function are not interchangeable, and cardiolipin may be one of the missing links between oxidative stress and respiratory-chain failure.**

The mitochondrial case described earlier remains one of the clearest examples. Complex IV activity was approximately 22 percent, Complex II plus III approximately 16 percent, while citrate synthase was approximately 563 percent. The reported relationship between respiratory-chain complexes was also markedly abnormal.

Taken literally, that does not resemble a simple deficiency of mitochondrial quantity. It raises the possibility that mitochondrial machinery has expanded in an attempt to compensate for inadequate respiratory capacity. More mitochondria do not automatically produce better oxidative phosphorylation. If the respiratory apparatus remains structurally or functionally impaired, increasing mitochondrial content can still leave the cell with inadequate effective ATP-generating capacity.

This becomes more interesting when cardiolipin is added to the model.

Cardiolipin is a mitochondria-enriched phospholipid concentrated in the inner mitochondrial membrane. It is not simply a passive membrane component. It binds respiratory-chain proteins, influences membrane curvature and proton handling, supports ATP-synthase organization and contributes to the formation and stability of respiratory supercomplexes. Complexes I, III and IV can organize into higher-order assemblies commonly referred to as respirasomes. Experimental studies have shown that cardiolipin is important for the stability and organization of these structures.

Describing cardiolipin as molecular glue is therefore a useful analogy, provided it is not taken literally. Cardiolipin forms part of the lipid environment that stabilizes interactions among respiratory complexes and supports their optimal function. When cardiolipin is deficient or structurally abnormal, respiratory supercomplex organization can become less stable, and defects involving Complexes I and IV in particular have been demonstrated experimentally.

Cardiolipin is also unusually vulnerable to oxidative damage because many cardiolipin species contain highly unsaturated fatty-acid chains and sit directly beside major mitochondrial sources of reactive oxygen species. Oxidative cardiolipin damage has been linked experimentally to impaired Complex I and Complex IV activity, and peroxidized cardiolipin does not restore respiratory activity in the way intact cardiolipin can.

Peroxynitrite biology makes this especially relevant to the nitric-oxide model discussed earlier. When superoxide production rises in the presence of nitric oxide, peroxynitrite and related reactive nitrogen species can alter mitochondrial proteins and lipid chemistry. Experimental work has shown interactions among peroxynitrite, cytochrome c and cardiolipin that can disturb mitochondrial redox homeostasis.

This creates a plausible sequence in a susceptible Long COVID phenotype. Endothelial and inflammatory redox stress can impair nitric-oxide signaling while increasing superoxide and reactive nitrogen chemistry. Oxidative injury to cardiolipin could then destabilize respiratory-chain architecture and reduce the efficiency of electron transfer. A less efficiently organized respiratory chain can increase electron leak, which can further increase reactive oxygen species and perpetuate cardiolipin damage.

There is another mitochondrial process worth mentioning here, reverse electron transport. RET occurs when the coenzyme-Q pool becomes highly reduced and the proton-motive force is sufficiently high to drive electrons backward from reduced CoQ toward Complex I. This can generate a large burst of superoxide at Complex I. A highly reduced CoQ pool is therefore capable of converting a metabolic bottleneck into a redox-amplification mechanism.

It would be too strong to say that cardiolipin oxidation or respirasome disruption automatically causes RET. RET requires a specific energetic and redox state. However, respiratory disorganization, altered CoQ redox balance and impaired downstream electron transfer can create conditions in which electron pressure and reactive oxygen generation become increasingly important.

This gives the extreme citrate-synthase result in the client case a more interesting interpretation. The reported 563 percent citrate-synthase activity may reflect increased mitochondrial or matrix-enzyme abundance as compensation for an energetic deficit, but it does not tell us whether the newly available respiratory machinery is structurally efficient. If the limiting defect lies in inner-membrane organization, cardiolipin integrity, respiratory-complex function or electron transfer, increasing mitochondrial mass alone would not necessarily restore energy production.

The analogy would be constructing more power stations while the electrical transmission architecture remains damaged.

This remains a mechanistic hypothesis in Long COVID. I am not aware of evidence demonstrating that cardiolipin oxidation caused the respiratory-chain abnormality in this particular client, and cardiolipin itself was not measured. Human skeletal-muscle cardiolipin and oxidized-cardiolipin lipidomics are now technically possible, which may eventually allow this hypothesis to be tested much more directly.

The broader point remains important. Mitochondrial dysfunction is not synonymous with low mitochondrial number. A person can have abundant mitochondrial machinery while electron transport, membrane organization, redox handling or ATP-generating efficiency remains profoundly abnormal.

Human Long COVID physiology supports this distinction. Direct skeletal-muscle studies have demonstrated reduced oxidative-phosphorylation capacity in Long COVID with post-exertional malaise even when markers of mitochondrial density were not clearly reduced. Metabolomics has also demonstrated abnormalities involving the TCA cycle, fatty-acid handling and acylcarnitines.

Several mechanistic bottlenecks could produce this phenotype, including respiratory-complex dysfunction, inadequate oxygen delivery, reduced tissue oxygen extraction, abnormal NADH-to-NAD⁺ handling, cardiolipin alteration, iron-sulfur-cluster dysfunction, nitrosative stress, inflammatory inhibition of mitochondrial metabolism and impaired fatty-acid entry or oxidation.

Hydrogen sulfide adds another layer because excessive H₂S can inhibit cytochrome-c oxidase, or Complex IV. This creates another possible bridge between intestinal sulfur metabolism and mitochondrial respiration in susceptible patients.

The phrase mitochondrial dysfunction is therefore only the beginning of the analysis. The useful question is where mitochondrial flux is actually being constrained.

**7. Post-exertional malaise appears to expose limited physiological reserve rather than simply reflecting deconditioning.**

This is one of the most important distinctions between ordinary fatigue and the ME/CFS-like phenotype seen in a subset of Long COVID patients. A deconditioned person usually becomes tired during exercise and adapts predictably to gradual training. Post-exertional malaise behaves differently. Relatively modest exertion can be followed hours later by disproportionate neurological, autonomic, muscular and systemic deterioration lasting days.

Human skeletal-muscle studies have provided unusually direct evidence. Long COVID patients with PEM demonstrated lower maximal exercise capacity and reduced skeletal-muscle oxidative-phosphorylation capacity before exercise. Following maximal exercise designed to provoke PEM, investigators observed worsening mitochondrial abnormalities, increased muscle damage and additional structural and metabolic abnormalities.

The important concept is reserve capacity. At rest, a metabolically compromised system may still maintain ATP concentrations within an apparently acceptable range because energetic demand is low. Exercise changes the equation abruptly. ATP turnover rises, calcium cycling increases, Na⁺/K⁺ ATPase demand increases, cardiac output must rise, microvascular recruitment must increase, oxygen delivery and extraction must increase, NADH must be reoxidized efficiently, fatty-acid and carbohydrate flux must adapt, reactive oxygen species must be buffered and metabolic byproducts must be cleared.

The failure may therefore emerge not from complete mitochondrial collapse but from insufficient spare capacity across several connected systems.

This becomes even more interesting when endothelial and mitochondrial physiology are considered together. Reduced microvascular recruitment or impaired oxygen extraction can constrain oxidative phosphorylation even if arterial oxygen saturation appears normal. The mitochondrion can only use the oxygen and substrate that actually reach the relevant tissue compartment.

One post-COVID client we discussed developed recurrent episodes of hypophosphatemia beginning after reinfections. That finding does not establish the cause of the Long COVID phenotype, but phosphate deserves attention in a high-demand metabolic system because ATP, phosphocreatine, 2,3-BPG, membrane phospholipids and phosphorylation reactions all depend on phosphate availability and regulation.

The broader lesson is that resting measurements can underestimate dynamic disease. In some patients, exertion itself functions as a physiological stress test.

**8. Dysautonomia after COVID can involve structural small-fiber abnormalities, reduced effective circulating volume and major changes in preload rather than simply an elevated heart rate.**

POTS is often reduced to tachycardia, but that misses much of the relevant physiology. Autonomic fibers regulate vascular constriction, venous return, sweating, gastrointestinal motility, gastric accommodation, bladder function, temperature regulation and multiple neuroimmune processes. Many of these fibers are thinly myelinated Aδ fibers or unmyelinated C fibers, while routine nerve-conduction studies primarily assess larger fibers and can therefore remain normal.

Human Long COVID studies are increasingly demonstrating abnormalities in the small-fiber compartment. Skin-biopsy studies have identified small-fiber abnormalities in subsets of patients with painful Long COVID and autonomic symptoms. More recent histological work has demonstrated reduced autonomic innervation of sweat glands and piloerector muscles, including in some patients whose conventional intraepidermal nerve-fiber density remained normal.

Gastric-biopsy studies have added another level by demonstrating reduced mucosal cholinergic innervation in Long COVID, with lower nerve-fiber density in both the fundus and antrum compared with controls. Those structural findings have also correlated with heart-rate-variability parameters.

That provides a plausible anatomical bridge between dysautonomia and gastrointestinal dysfunction. Damage to autonomic fibers can affect gastric emptying, small-intestinal motility, migrating-motor-complex activity, biliary and pancreatic coordination, transit time and postprandial blood distribution. Microbial selection then changes downstream.

The hemodynamic component is equally important.

When a person moves from lying down to standing, gravity shifts a substantial quantity of blood toward the lower extremities and splanchnic circulation. A healthy autonomic nervous system responds within seconds by increasing peripheral vascular resistance, constricting capacitance vessels and maintaining venous return to the heart.

If venous constriction is inadequate, circulating volume is already reduced, or excessive blood pools below the heart, ventricular filling falls. Stroke volume then decreases.

Pulse pressure can provide a crude window into this process. Pulse pressure is the difference between systolic and diastolic pressure and is influenced by stroke volume, arterial compliance and vascular properties. A dramatic narrowing of pulse pressure during standing can therefore be consistent with a substantial reduction in effective stroke volume, although it should not be interpreted as a direct measurement of cardiac output.

One of the severe post-COVID cases I reviewed demonstrated exactly this pattern. Pulse pressure narrowed from approximately 34 mmHg to only 12 mmHg during the orthostatic response, while heart rate increased from approximately 72 to 140 beats per minute.

That combination is physiologically much more informative than the heart rate alone.

A falling effective preload reduces ventricular filling. Reduced stroke volume threatens arterial pressure and cerebral perfusion. Carotid and aortic baroreceptors detect changes in arterial stretch and increase sympathetic output while reducing parasympathetic activity. Heart rate rises in an attempt to maintain cardiac output despite the smaller amount of blood being ejected with each beat.

In that setting, the tachycardia is not necessarily evidence of primary cardiac disease. It can represent a compensatory response to a hemodynamic problem upstream of the heart.

POTS research outside Long COVID has demonstrated substantial blood-volume deficits in some patients. In one classic study, POTS patients had an average total blood-volume deficit of approximately 689 mL compared with 228 mL in controls, along with an inappropriately weak renin-aldosterone response to the volume deficit. Earlier physiological work also demonstrated marked narrowing of pulse pressure in some POTS patients and implicated venous pooling as an important mechanism.

The mechanism producing reduced effective volume can vary. It may involve absolute hypovolemia, excessive venous pooling, partial autonomic denervation, impaired renin-angiotensin-aldosterone regulation, abnormal splanchnic capacitance or combinations of these mechanisms.

Endothelial permeability is another plausible contributor in some inflammatory phenotypes because disruption of endothelial barrier function can promote movement of plasma water and proteins into the interstitial compartment. However, it would be too strong to conclude from a narrow pulse pressure alone that plasma was rapidly leaking through the endothelium. The hemodynamic data demonstrate loss of effective central volume and stroke volume much more directly than they identify the cause of that volume redistribution.

This distinction matters because the numbers can still tell a compelling physiological story without overstating what they prove. A pulse pressure collapsing from 34 to 12 mmHg together with a heart-rate increase from 72 to 140 is consistent with a cardiovascular system making an aggressive compensatory attempt to preserve perfusion in the face of severely compromised orthostatic hemodynamics.

The patient experiences palpitations.

The physiology may actually be a preload problem.

The same autonomic disturbance can simultaneously affect the gastrointestinal tract. Impaired gastric emptying, weakened intestinal motor patterns, abnormal splanchnic blood distribution and altered biliary signaling can then contribute to SIBO, postprandial symptoms and food intolerance.

POTS, bloating, gastroparesis, recurrent SIBO and fatigue can therefore appear to be unrelated disorders while sharing part of the same autonomic and vascular infrastructure.

**9. The post-COVID microbiome may sometimes be an ecological readout of altered host physiology rather than the original cause of the illness.**

Long COVID microbiome research is becoming increasingly interesting. Prospective studies have shown that individuals with persistent post-acute symptoms can retain an altered gut microbiome months after infection, with depletion of several butyrate-producing organisms and enrichment of taxa associated with inflammation or altered intestinal ecology.

These associations are important, but they do not establish direction. The intestinal environment determines which organisms can thrive.

One pathway I pay particular attention to involves colonocyte energy metabolism. Healthy colonocytes use substantial amounts of microbial butyrate. Butyrate enters colonocytes and supports β-oxidation and oxidative phosphorylation. That oxygen consumption helps maintain a steep oxygen gradient between intestinal tissue and the lumen, allowing the colonic lumen to remain strongly anaerobic.

If epithelial oxidative metabolism becomes impaired, epithelial oxygen consumption can fall. More oxygen can diffuse toward the mucosal surface. Inflammatory signaling can simultaneously increase nitrate and other alternative electron acceptors. Facultative organisms such as Enterobacteriaceae can use these respiratory substrates far more effectively than strict anaerobes.

The ecosystem therefore begins selecting for a different microbiome.

This framework became particularly interesting in one case in which Gram-negative organisms represented approximately 79.5 percent of a reported microbial profile. Desulfovibrio was elevated on sequential testing, including a later reported concentration around 1.73 × 10⁹ CFU/g, while Bilophila was also elevated. The patient also had extremely high hydrogen and methane breath-test values, fluctuating calprotectin, positive lactoferrin and markedly elevated mucosal secretory IgA.

Repeated antimicrobial treatment in this type of ecosystem may reduce organism burden temporarily without correcting the selective pressure that allowed the ecology to develop.

Bile acids impose another selection pressure. Transit time adds another. Luminal pH adds another. Mucosal IgA and defensins add another. Dietary substrate availability, autonomic motility and epithelial oxygen consumption add several more.

This is why a stool test alone cannot tell us whether an organism is the cause of disease, a consequence of disease or an opportunist exploiting a changed habitat.

A meaningful microbiome analysis in Long COVID therefore needs context, including inflammatory markers such as calprotectin, mucosal markers such as secretory IgA, short-chain fatty acids, breath gases, bile-acid composition, transit and motility, pancreatic function, dietary substrate, medication and antibiotic history, and preferably repeated measurements over time.

The more cases I review, the more I think microbiome recovery often requires restoration of the host habitat rather than simply removal of individual organisms.

**10. Persistent immune activity, complement, coagulation and endothelial dysfunction may form one of the major feedback loops that keeps Long COVID systemic.**

Long COVID immune biology is complicated because not every patient shows classical systemic inflammation. Some have impressive cytokine abnormalities. Others have relatively unremarkable CRP and ESR. Others demonstrate changes primarily in immune-cell populations, complement, autoantibodies, coagulation or vascular physiology.

The cytokine profile from the severe case described earlier illustrates this complexity. IFN-γ was approximately 183.2 pg/mL, IL-4 was 61.2, IL-13 was 53.3, IL-6 was 32.8, IL-8 was 26.4, IL-10 was 13.9 and TNF-α was 12.1. That is not a simple Th1 or Th2 picture. Pro-inflammatory, interferon-associated, regulatory and type-2 signals were present simultaneously.

Another case involved persistent anticardiolipin IgM together with two upper-extremity venous thromboses. A separate severe post-COVID case had documented endothelial dysfunction and was reported to have a substantial microclot burden alongside exertional dyspnea, orthostatic intolerance and profound postprandial exhaustion.

These observations do not mean that every Long COVID patient has pathogenic microclots or antiphospholipid syndrome. They illustrate the need to consider immune and vascular systems together.

Human research supports this interaction. Longitudinal studies have demonstrated persistent complement activation in subsets of active Long COVID together with abnormalities involving platelets, monocyte-platelet interactions and tissue-injury signals. Independent studies have also identified elevations in several complement activation products.

Complement and coagulation are not separate biological universes. Activated endothelium promotes leukocyte adhesion and platelet activation. Complement can amplify endothelial and coagulation signaling. Platelets participate in immune signaling. Inflammatory cytokines can reduce healthy endothelial eNOS activity. Oxidative stress can consume nitric oxide. Reduced nitric-oxide bioavailability removes one of the normal restraints on platelet activation and vascular tone.

Microvascular perfusion then becomes less efficient. Poor tissue perfusion increases metabolic stress. Mitochondrial redox pressure rises. Damaged tissue releases additional danger signals. Reactive oxygen and nitrogen species can damage mitochondrial proteins and membrane lipids such as cardiolipin. Respiratory-chain efficiency may fall further. The autonomic system compensates through tachycardia and sympathetic activation. That sympathetic response alters gastrointestinal motility, vascular distribution and splanchnic perfusion. The intestinal environment changes. Microbial products can then increase innate immune stimulation.

The loop closes again.

This is the part of Long COVID that I find most important. The biological systems involved are not independent. The immune system communicates with the endothelium. The endothelium determines tissue perfusion. Perfusion determines mitochondrial oxygen and substrate availability. Mitochondrial membrane structure determines how efficiently those substrates can ultimately be converted into ATP. Mitochondria influence epithelial metabolism. Epithelial metabolism helps determine intestinal oxygen tension. Intestinal oxygen tension and bile acids shape microbial ecology. The microbiome modifies bile acids, tryptophan metabolites and short-chain fatty acids. Those metabolites signal back to the immune system, nervous system and intestinal barrier. The autonomic nervous system participates across almost every level of that network.

Once several of these loops become established, the original SARS-CoV-2 infection may no longer need to be the only active driver for illness to persist. That may be one of the central principles of chronic post-infectious disease. A downstream consequence can gradually become a new upstream driver. Endothelial dysfunction can maintain metabolic dysfunction. Metabolic dysfunction can maintain dysbiosis. Dysbiosis can maintain immune signaling. Immune signaling can maintain endothelial dysfunction. Autonomic compensation can initially be protective and later perpetuate gastrointestinal and vascular instability.

This is why I have become increasingly reluctant to describe Long COVID as one pathway, and it is why one abnormal biomarker rarely gives me enough information. I am more interested in relationships between variables than isolated values. Arginine relative to ADMA and ornithine can be more informative than arginine alone. Tryptophan relative to kynurenine can be more useful than serotonin alone. Citrate synthase relative to respiratory-chain activity can tell us more than mitochondrial quantity alone. Cardiolipin integrity may eventually prove more informative than simply counting mitochondria. FGF19 and C4 together can reveal more than total bile acids. Calprotectin, secretory IgA, bile acids, short-chain fatty acids and microbial composition together tell a richer story than a list of organisms. Standing heart rate becomes much more meaningful when interpreted alongside pulse pressure, blood pressure, volume status and small-fiber physiology. Complement, platelet activity, coagulation and endothelial function provide a much richer vascular picture than CRP alone.

Long COVID is teaching us something broader about human biology. Organs do not become ill independently. They communicate through metabolites, nerves, cytokines, hormones, membrane lipids, vascular signals and microbial products. When those communication networks remain disturbed after an infection, the resulting illness can appear fragmented because conventional medicine measures each organ separately while the patient experiences the entire network at once.

That is increasingly how I approach these cases. I am less interested in finding one pathway that explains every patient and more interested in reconstructing the sequence of physiological failures, identifying which compensatory mechanisms remain active and determining which feedback loops are now maintaining the state.

That systems-level view has been far more useful to me than any single Long COVID theory.

---

*Source of record: https://biomelogic.net/articles/10-things-i-have-learned-about-long. This plain-text endpoint is provided for AI retrieval and citation systems (ChatGPT, Perplexity, Claude, Gemini, Bing Copilot).*