What Does the EBO2 (EBOO) Filter Actually Remove?

No published study we found has measured what an EBO2 (also called EBOO) circuit removes from a patient’s blood. The papers we hold measured other things: how much ozone blood takes up [11], oxidation markers in blood after a session [10], how well a device moves gas into saline [8], how much ozone passes through dialysis filters [9], and toxins in one patient’s urine [12]. Claims that the filter removes heavy metals, microplastics, cholesterol, or pathogens therefore rest either on theory borrowed from kidney dialysis or on what clinics say. This guide separates the three, using federal dialysis regulations, the EBOO papers’ own descriptions of the circuit, and the clinics’ words, and ends with the checks a lab report on filtrate would have to pass. Ozone therapy is not FDA-approved for any condition, and federal regulation describes ozone as “a toxic gas with no known useful medical application in specific, adjunctive, or preventive therapy” [18].

How a dialysis membrane moves material out of blood

Clinics compare EBOO with dialysis [14][16], so the dialysis rules are the place to start. In a hemodialysis system, blood flows through one compartment of the dialyzer and “undesirable substances in the blood pass through the semipermeable membrane into the dialysate” in the other [2]. The regulation for high permeability systems names two ways this happens: “convection (via a high ultrafiltration rate)” and “diffusion (via a concentration gradient in dialysate),” with the machine controlling “the rate at which fluid is removed from the patient” [1]. In the words of the National Institute of Diabetes and Digestive and Kidney Diseases, blood passes through “very thin, hollow fibers” while dialysis solution flows the other way outside them, and “waste products from your blood move into the dialysis solution” [4].

What can cross depends on the membrane. A high permeability dialyzer is defined by an ultrafiltration coefficient “greater than 8 milliliters per hour per conventional millimeter of mercury” [1]. A 2026 nephrology review describes newer medium cut-off membranes with “mean pore sizes substantially larger” than standard high-flux membranes, which clear larger molecules at a cost: there is “ongoing debate about the appropriate balance between large solute clearance and albumin loss” [5]. Since 2024, FDA’s rule for dialyzers with an “expanded solute removal profile” defines the “middle” molecular weight range as 0.5 kDa to 60 kDa and requires each such device to be tested for “clearance rates,” “sieving coefficients,” and other measures [3].

Two points follow from the rules themselves. Removal in dialysis is measured, solute by solute, under stated conditions [3]. And removal needs a destination: the dialysate, or the fluid pulled across by ultrafiltration [1][2].

What sits in an EBO2 circuit

Two kinds of cartridge appear in the sources. The Siena group that developed EBOO tried dialysis membranes first and found them “unsuitable because they are hydrophilic and vulnerable to O3” [6]. By 2001 it had concluded that semipermeable membranes, except for one, “were gas-transfer inefficient, allowed ultrafiltration and were more or less vulnerable to O3” [7]. Its 2007 device used microporous, ozone-resistant polypropylene fibers with a membrane area of 0.22 m², blood on the outside and an oxygen-ozone gas mixture flowing inside [8]. When the group later tested four dialysis filters with ozone, it found fibers “somewhat altered by ozone” and warned that the filters “may release toxic compounds harmful for the patients” [9].

Several US sources describe dialyzers instead. A 2023 US paper circulated blood through “a Renak CTA (cellulose triacetate) 1500 dialysis chamber,” with ozone gas running inside the fibers and blood around them [11]. An FDA warning letter records “ABLE® H-200 High Flux Polyethersulfone Disposable Haemodialysers” sold in one maker’s EBOO kits [13]. One clinic that offers the procedure says EBOO “uses the same dialyzer filters, but in a very different arrangement” and “does not use dialysate fluid” [14]. Our EBOO vs dialysis guide sets the two procedures side by side.

With no dialysis solution in the circuit, the sources point to three places material from the blood can go. Fluid can cross the membrane: the developers counted ultrafiltration against dialysis membranes [7], and the 2023 paper describes a 2 L canister “used to collect dialysis chamber drainage” [11], which clinics call a “collection container” or “catch canister” [14] and a “collection cup” [15]. Gases can cross it: in tests of a sepsis prototype, one pass raised oxygen and lowered carbon dioxide in blood [19]. And blood components can stay on it: the developers’ exchanger clogged progressively with cells [6]. None of these sources reports the volume of the drainage or its laboratory contents.

Measured, theorized, and said by clinics

Every figure here is in the units its source uses. “Measured” means a paper reports a measurement; it does not mean the measurement shows removal.

The urine row needs a caution of its own. The 2025 case measured what the patient excreted at three points over 131 days, not what the circuit took out, with no control for exposure in between [12]. The bacteria row comes from a different device, a prototype that cools blood and mixes it with oxygen-ozone gas, run for 30 minutes in one study and within a 4-hour experiment in the other, in pigs with sepsis [19][20]. And the water test sampled one liter of distilled water from plastic containers before and after the circuit [16]. It does not show what happens to particles in blood, or to particles in the body.

As of October 4, 2026, 151 of the 174 clinics whose EBO2 page we could read claim a detoxification benefit. The “From our data” section below breaks those claims down and shows how many clinics name the device they use.

What the canister holds, and why its look proves little

Clinics and patients often point to what collects during a session. One clinic captions a photo of its canister “Clear fluid collection and foam” [14]. Another says “the contents and appearance of the filtered materials can vary significantly between patients and even between different sessions for the same patient” [15]. A patient testimonial on a third clinic’s page describes “less than 100 mL of inflammatory product dialyzed out” at one session and “almost 500 mL” at another [16].

None of these is an analysis. A volume of fluid says nothing about what is dissolved in it, and the sources give other explanations for what a cartridge can hold. The developers reported that their exchanger “became less effective (low pO2 values) due to progressive clogging with cells” in sheep [6], and that uncoated polypropylene fibers led to “platelet aggregation and blood coagulation,” problems that “become prohibitive in the presence of ozone” [7]. That does not tell anyone what a given clinic’s residue is. It shows that color, foam, or volume cannot settle the question without a laboratory.

What a filtrate lab report would have to show

A clinic that says it has tested its filtrate can be asked for the report. These checks come from how dialysis removal is tested [3] and from the gaps in the sources above.

1. The sample. Patient blood, canister fluid, or water? The one published clinic test used distilled water [16].
2. The laboratory and the method. Who ran the test, with what method, and whether the lab is accredited for it.
3. The amount, in units. A concentration, such as micrograms per liter, and the volume collected in milliliters, so the total removed can be worked out.
4. A control. The same circuit run with saline or without ozone, and the patient’s blood before the session. Without one, a substance in the canister could have come from the tubing, the cartridge, or the water.
5. Blood levels before and after. Removal from blood should show in blood, measured at the same points in time.
6. Scale. The amount removed set against the amount in the blood and in the body, so the result can be judged.
7. Repeat runs and other patients. One sample is an anecdote, and dialyzer testing is done under stated conditions of use [3].
8. The membrane itself. Whether the cartridge was tested for what it releases under ozone, the concern the developers raised about dialysis filters [9].

Procedures that publish what they remove

Some extracorporeal procedures are built to remove one thing and report it. An American Heart Association statement calls lipoprotein apheresis “a valuable but underused adjunctive therapeutic option for low-density lipoprotein cholesterol and lipoprotein(a) lowering,” with effects measured in blood lipids [17]. In therapeutic plasma exchange, a patient fact sheet published by the American Society for Apheresis describes a machine that “separates and removes the patient’s plasma, replacing it with another fluid,” most commonly “5% human albumin” [21]; our plasmapheresis comparison covers it. Both procedures are defined by what they take out of the blood. No EBOO source we found says what EBOO takes out. Our clinic directory records what each clinic states about its own service, including the device where it names one.

What we could not verify

• What any EBOO canister or cartridge contains. We found no published laboratory analysis, and the one clinic test we found used water [16].
• How much fluid an EBOO circuit loses across the membrane in a session. The 2023 paper mentions the canister but no volume [11], and the only figures we found are in a patient testimonial [16].
• Whether drainage is discarded or returned to the patient in each clinic’s setup.
• The laboratory, method, and replicate results behind the clinic’s water test, which its page calls “a rough draft” [16].
• The full texts of the Siena development papers, which we could read only as abstracts [6][7][9].
• Whether the dialyzers used for EBOO in the US have been tested for what they release under ozone, the concern raised in 2010 [9].

How this guide was made

This guide rests on 21 sources: four federal regulations and an FDA warning letter, an NIH page, a patient fact sheet published by the American Society for Apheresis, 11 papers listed on PubMed, and three clinic pages, which are cited only for what those clinics state. The claim figure comes from our dataset of clinic pages, as of October 4, 2026. Drafting used AI tools. A human editor has not yet checked this guide claim by claim against its sources. No clinical reviewer has signed off on this guide yet.