Obstructive sleep apnea begins with a physical failure in the throat. During sleep, the upper airway repeatedly narrows or closes, airflow falls, oxygen drops and the brain briefly pulls the sleeper toward wakefulness to restore breathing. Yet the damage associated with those events may not remain anywhere near the airway.
A 2026 study in iScience adds a possible intestinal route to the familiar pathways connecting sleep apnea with high blood pressure. In men who had both severe apnea and hypertension, researchers found altered gut bacteria and microbial metabolites, higher blood markers consistent with injury to the intestinal barrier, and more lipopolysaccharide, or bacterial endotoxin, in circulation. The results fit a gut-barrier hypothesis. They do not prove that gut changes caused any participant’s hypertension.
Four groups turned one night of sleep into a biological cross-section
Chih-Yuan Ko and colleagues recruited 47 men at a single hospital in China and divided them into four groups. Twelve had no obstructive sleep apnea, 11 had severe apnea but blood pressure below 130/85 millimetres of mercury, 13 had severe apnea with prehypertension, and 11 had severe apnea with hypertension. Severe apnea meant an apnea-hypopnea index of at least 30 breathing interruptions an hour.
Each man underwent overnight polysomnography, the multi-signal sleep test used to confirm apnea and record its severity. The team collected fasting blood and stool the following morning. They measured acetate, butyrate and propionate in both samples; intestinal fatty acid-binding protein, known as i-FABP, and D-lactate as signs of intestinal barrier injury; lipopolysaccharide, or LPS, as a measure of endotoxin exposure; immune cells and cytokines; and bacterial DNA in stool.
The four groups created a biological cross-section from normal breathing and blood pressure through severe apnea with progressively higher pressure. This design can show that measurements travel together. It cannot establish which change came first, or whether a third factor produced both. That distinction matters throughout the paper.
Endotoxin rose most clearly as blood pressure status worsened
Two short-chain fatty acids showed the first part of the pattern. Fecal acetate and butyrate were lower in the prehypertension and hypertension groups than in controls. These molecules are produced when gut microbes ferment dietary material, especially fiber. They help nourish cells lining the colon and participate in immune and metabolic signalling.
Blood did not simply mirror stool. Plasma acetate and butyrate did not differ significantly among the four groups. Plasma propionate was higher in men with apnea and those with apnea plus prehypertension than in controls, but it was not significantly higher in the hypertension group. The mismatch is useful: a fecal concentration records what remains in the colon after production, absorption and use, while plasma reflects absorption, liver processing and distribution around the body. One sample cannot reveal every step.
The barrier-related measurements moved more consistently. i-FABP, D-lactate and LPS were higher in all three severe-apnea groups than in men without apnea. Among those with apnea, the hypertension group had higher D-lactate and LPS than the normal-pressure group. LPS was also higher with hypertension than with prehypertension. In other words, the endotoxin signal tracked the blood-pressure categories more cleanly than the short-chain fatty acids did.
The hypertension group also had a higher proportion of pro-inflammatory Th17 cells than the control and apnea-only groups. But the Th17-to-regulatory-T-cell ratio did not reach statistical significance, and the measured circulating cytokines did not differ significantly. This was a partial immune signal, not evidence of uniform systemic inflammation.
“Weakened barrier” describes markers, not a view through the intestinal wall
Popular accounts often turn “markers consistent with intestinal barrier injury” into “a leaky gut.” The study did not biopsy the intestine, measure tight junctions directly or watch bacterial material cross the lining. i-FABP can enter blood when intestinal epithelial cells are injured. D-lactate, produced largely by microbes, can rise in circulation when barrier function is impaired. Their elevation supports the interpretation, but it remains an indirect one.
LPS is part of the outer membrane of many Gram-negative bacteria. The researchers measured it with a commercial ELISA. Once in circulation it can activate inflammatory pathways, including signalling through the TLR4 receptor. In the study’s multivariable analysis, LPS was the only measured gut-related variable significantly associated with movement into a higher blood-pressure category after adjustment for age and body mass index.
That result sounds more exact than it is. The reported odds ratio was 43.64, but its 95 percent confidence interval stretched from 1.828 to 1,042.107. A range that wide is the statistical footprint of a small, unstable model. It says the association deserves testing in a much larger cohort; it does not provide a reliable personal risk multiplier.
Sleep apnea already has several direct routes to hypertension. Repeated arousals drive the sympathetic nervous system. Cycles of oxygen loss and recovery promote oxidative stress, endothelial dysfunction, inflammation and hormonal changes affecting vessel tone and fluid balance. Earlier ScienceBlog coverage examined the broader connection between sleep-disordered breathing and cardiovascular disease. A gut route would join that network, not replace it.
Ten CPAP users moved in the predicted direction
Ten participants who used continuous positive airway pressure for at least four hours a night were assessed again after three months. Their average apnea-hypopnea index fell from 61.76 to 44.7 events an hour, minimum oxygen saturation improved from 69.1 to 78.6 percent, and sleep continuity improved. The follow-up sleep study was conducted without CPAP, so that index describes residual underlying disease rather than breathing while the device was operating.
The same men had lower i-FABP, D-lactate and LPS at follow-up. Stool microbial diversity increased and the overall community shifted. Fecal short-chain fatty acids, however, did not significantly change, while plasma propionate fell. The mixed result again resists a simple story in which CPAP restores one missing microbial chemical and everything downstream recovers.
There was no randomized untreated comparison group. Ten people are highly vulnerable to chance, regression to the mean and unmeasured changes in diet or behaviour. The before-and-after movement is consistent with the proposed mechanism, but it cannot show that CPAP caused the gut changes. Other research has found benefits in different tissues, including the earlier ScienceBlog report that CPAP was associated with recovery of brain tissue. Those findings make effective treatment important, but they do not turn this small intestinal follow-up into a trial.
Animal transfers make the route plausible, not proven in people
The proposed sequence has biological logic. Intermittent hypoxia can change the intestinal environment and the microbes able to thrive there. Reduced local butyrate could deprive colon cells of a useful fuel and weaken barrier support. Bacterial products could then reach the circulation, activate immune signalling and add to vascular dysfunction.
A 2025 mouse experiment went farther mechanistically. Transferring fecal microbes from people with apnea, particularly apnea with hypertension, into mice was followed by higher blood pressure and vascular injury involving the LPS-TLR4-NF-kB pathway. That is evidence that a microbial community can transmit part of a phenotype under controlled animal conditions. It is not evidence that the same transfer explains hypertension in a particular person.
The broader idea also predates this apnea study. A 2018 human study of high blood pressure reported altered gut microbes, higher i-FABP and LPS, and gut-homing inflammatory cells. ScienceBlog’s 2025 review of how gut bacteria might reshape sleep-apnea treatment brought together earlier intermittent-hypoxia and microbiome experiments. The new work’s contribution is to measure stool metabolites, barrier markers, endotoxin and immune cells in the same small clinical cohort, then repeat several measures after CPAP exposure.
The next experiment must be larger and harder to fool
The paper calls itself exploratory, appropriately. Its 47 participants came from one centre and were all men; the apnea groups had severe disease. Diet, fiber intake and physical activity were not measured even though all can affect gut microbes, blood pressure and short-chain fatty acids. Blood pressure was recorded in the clinic rather than with 24-hour ambulatory monitoring. Only 10 men completed the CPAP follow-up, devices varied, and there was no control arm.
A stronger test would follow a larger and more diverse group before hypertension develops, repeatedly measure diet and medication, use ambulatory blood pressure, and randomize treatment where ethical. Direct measures of barrier function and microbial genes or metabolites would help distinguish microbes that merely accompany apnea from those participating in its effects. A controlled CPAP study could then ask whether biomarker changes follow treatment exposure and whether they predict a meaningful blood-pressure response.
Nothing here shows that probiotics, prebiotics or butyrate supplements prevent apnea-related hypertension. No reader should substitute a microbiome product for a sleep assessment or established care. The American Academy of Sleep Medicine’s clinical guideline recommends positive airway pressure for adults with obstructive sleep apnea and coexisting hypertension, with treatment chosen and monitored for the individual.
The study is valuable because it puts several pieces on the same clinical map: interrupted breathing, stool chemistry, indirect signs of intestinal injury, bacterial endotoxin and immune cells. The airway is where each apnea begins. The 2026 study asks whether part of the cardiovascular aftermath continues in the gut.