A 64-year-old woman re-presents to the ED with persistent GI symptoms. She was seen four days prior with nausea, non-bloody diarrhea, and abdominal pain. A CT of her abdomen showed mild small-bowel wall thickening. Vitals were normal. After receiving IV fluids, she felt somewhat better and was discharged home from the ED.

She returns with worsening abdominal pain that is now severe, and diarrhea that is now bloody. She is febrile, with a pulse 110/min in new-onset atrial fibrillation, blood pressure 85/55, and respirations of 22/min.

She reports regularly taking aspirin-containing analgesics for the pain. Salicylate level is 35 mg/dL (upper limit of therapeutic level = 30 mg/dL).

On exam, her abdomen is soft with diffuse mild tenderness.

Labs show:

Nephrology are consulted. She receives 30 cc/kg of crystalloid resuscitation fluids.

Read more

Microaxial flow pumps (Impella™ being the only available product) were shown in the Danger-Shock trial (NEJM 2024) to reduce mortality in patients with cardiogenic shock due to ST-elevation myocardial infarction. The survival benefit came with the cost of significant risks for limb ischemia, bleeding, and acute kidney injury.

Almost as soon as Danger-Shock was published, pressure grew for the expansion of Impella beyond its evidence-based use case for STEMI-related cardiogenic shock.

Cardiologists commonly face gray-zone situations in which patients deteriorate during or after PCI, without frank shock. Impella was FDA-approved for mechanical circulatory support during procedures way back in 2012. After Danger-Shock, interventional cardiologists used Impellas much more frequently during high-risk procedures (and maybe some not-so-high-risk ones).

Reimbursements in the tens of thousands of dollars per Impella (to the hospital) incentivized the expansion. Cardiologists don’t receive big per-device payments for placing Impellas, but they gain massive leverage in salary negotiations with their centers based on the revenue they bring in.

Abiomed, acquired by J&J’s Medtech division in 2024 for $16B, surfed this wave well. Under the slogan “Impella Protected PCI,” the device was marketed as a safety hedge to cardiologists traumatized by patients dying on their cath lab tables, and sales grew dramatically. The PROTECT I, II, and III trials were cited to support the practice (but did so weakly, if at all).

However, to really secure the expansion into this market, i.e., patient population, higher-quality randomized trial data would be needed to establish the benefit of Impella in MI without shock.

In a refreshing (or depressing, or both) abandonment of any pretense, J&J MedTech’s own chief medical science officer served as a lead investigator of a randomized trial (‘Door to Unload’) that, if successful, could expand Impella’s FDA indication (or just clinician willingness) to MIs without shock.

Contemporaneously, the U.K. funded its own smaller trial (‘Chip’) testing Impellas for MI without shock—presumably as a due diligence counterweight against the massive expenditure that would be required of the NHS if pressure rose to include Impellas as standard care in this much larger patient population.

The ‘Door to Unload’ and ‘Chip’ Trials

In the “Door to Unload” trial, 527 patients with acute anterior STEMI were randomized to Impellas for 30 minutes before PCI vs. immediate PCI alone.

(Yes, you read that correctly: in patients with STEMI, PCI was deferred to place an Impella and run it for 30 minutes before revascularizing. Patients remained ischemic for about 47 minutes on average.)

In the Chip trial, Impellas were placed immediately before PCI in the intervention arm. Three hundred patients with STEMI and severe LV dysfunction (<35%) were randomized.

Neither trial showed a benefit in the Impella arm over usual care:

• In Chip, there was a concerning trend toward harm (hazard ratio for death 1.54 in the Impella arm).

• In DTU, Impellas did not improve infarct size (the primary outcome) or secondary endpoints (win ratio 1.04). All-cause mortality and cardiovascular mortality at 30 days did trend toward a benefit from Impellas (P=0.08).

A Takeaway

The Chip trial argues strongly against routine use of Impellas for patients with STEMI without shock.

Door-to-Unload’s trend toward a mortality benefit left the door cracked open for use cases for Impella in this population, which aligns with interventionalists’ strong anecdotal experience of Impella’s benefits in selected patients.

Identifying who will benefit, though, remains as difficult as ever.

An editorialist noted:

[FDA] approval and the commercial framing of these devices under the “Protected PCI” designation has reinforced a perception of procedural safety for the operator…[but] the criteria for the selection of appropriate patients remain poorly defined despite years of use.

These [trial] findings force the interventional cardiologist to occupy an uncomfortable position between two unsatisfying options — the potential burden of perceived undertreatment when a patient’s condition deteriorates or recognition that the act of doing more may itself constitute harm. Any honest appraisal must begin with that tension.

This dilemma pervades medicine, but lands hardest on proceduralists. Even if financial considerations were eliminated, this tension would remain.

References

Perera D, Ryan M, Ezad SM, et al. Left Ventricular Unloading in High-Risk Percutaneous Coronary Intervention. New England Journal of Medicine. Published online March 29, 2026. doi:https://doi.org/10.1056/nejmoa2515704

‌Kapur NK, Mangner N, Aghili N, et al. Left Ventricular Unloading in Anterior STEMI without Shock: The STEMI Door to Unload (DTU) Randomized Controlled Trial. JACC. Published online March 2026. doi:https://doi.org/10.1016/j.jacc.2026.03.071

‌

For most patients who’ve been in the ICU more than a few days who then develop a new suspected bacteremia, empiric treatment with very broad-spectrum antibiotics like piperacillin-tazobactam, cefepime, or carbapenems is almost a clinical reflex. After that, an organism’s sensitivity usually isn’t known for at least a few days, during which broad coverage is continued in most patients.

A new generation of rapid antimicrobial sensitivity tests may reduce that timeframe. Evidence is emerging that these assays can direct optimized antibiotic therapy earlier, in either direction (broader, for resistant organisms, or narrower, when resistance is unlikely).

There’s no proof they improve hard outcomes— but could insisting on that level of certainty be a disservice to patients and the antibiotic stewardship project?

FAST: It’s Faster (But Not That Fast)

The status quo

First, a recap of traditional sensitivity testing. After standard blood cultures have incubated in broth media (in the bottles) and “turned positive” (usually by detection of high CO2 production inside the bottle), the broth is inoculated onto agar plates. Once enough colonies grow, the organism can be identified through conventional methods and complementary molecular techniques.

Antibiotics can then be applied directly to the agar to determine sensitivity, with other more advanced techniques often used (e.g., dilution of the broth into multiple wells containing antibiotics).

Conventional sensitivity testing takes about 24 to 48 hours after the blood cultures turn positive, often two to five days after blood cultures are drawn.

Rapid susceptibility testing

Rapid susceptibility tests are performed on the blood culture broth in the inoculated bottles after the cultures have returned as positive—rapidly, without the necessity of waiting for growth on agar or other media.

They use techniques not dissimilar to conventional susceptibility testing (e.g., diluting broth across a matrix of antibiotic-impregnated microwells), but identify signs of resistance faster (e.g., detecting volatile metabolites from the wells, or microscopic optical changes, rather than waiting for the traditional slower signals).

Rapid AST assays can return susceptibility results in less than 8 hours.

It’s important to distinguish rapid AST (which test an actual cultured bacteria for observable resistance) from inference-based rapid molecular genetic tests (e.g., BioFire™, Cepheid Expert™) that identify bacterial DNA and resistance genes without actually observing resistance or susceptibility.

Six randomized trials have tested rapid AST in patients with bacteremia:

• Banerjee et al JAMA 2026, (the FAST trial, n=850)

• Banerjee et al Clin Infect Dis 2015 (n=617)

• Beuving et al Eur J Clin Microbiol Infect Dis 2015 (n=129)

• Bannerjee et al Clin Infect Dis 2021 (n=500)

• Kim et al, Clin Microbiol Infect 2021 (n=116)

• Caspar et al Eur J Clin Microbiol Infect Dis 2024 (n=112)

All found that use of rapid AST identified organisms and antimicrobial resistance faster, resulting in faster delivery of optimal antimicrobial therapy—usually by 12 to 48 hours faster than conventional sensitivity testing.

This usually led to more accurate targeting of antibiotics and less use of broad spectrum agents.

Rapid testing had high agreement (e.g., ~94%) with conventional testing.

However, none of the six RCTs found that use of rapid AST reduced mortality, hospital lengths of stay, or other “hard” clinically relevant outcomes.

The FAST Trial

As the first five trials were conducted in countries with relatively low rates of antibiotic-resistant organisms, the FAST trial was performed specifically to test the use of rapid AST at hospitals in nations with high antibiotic resistance.

At 7 medical centers in 4 countries with a high prevalence of multidrug-resistant gram-negative bacteria (Greece, India, Israel, and Spain), 899 patients with confirmed gram-negative bacteremia were randomized to standard susceptibility testing or fast AST using Biomerieux’s system (VITEK Reveal).

There was even less support for rapid AST in this trial than its predecessors: similar proportions of patients received appropriate antibiotics at enrollment, 24 hours, and at 3 days.

However, patients with resistant organisms had a shorter time to optimal targeting of their antimicrobials (escalation or de-escalation). For cephalopsporin-resistant organisms, rapid AST led to changes in ~12 vs. 25 hours with conventional testing; for carbapenem-resistant bugs, tailoring occurred after 13 vs. 47 hours.

Why No Benefit?

Rapid susceptibility testing occurs well after the critical decision and time window has passed for treatment of severe infections with antibiotics. Most patients with positive blood cultures have already received antibiotics for the two days while their cultures were incubating. Empiric therapy, not downstream tailoring does almost all the work in saving lives from sepsis.

Patients in clinical trials are often already receiving high-quality care at advanced centers, which for patients with suspected bacteremia in the ICU includes broad-spectrum antibiotics; this likely led to good outcomes in the control arms of these trials.

It’s also possible that most patients don’t suffer bad outcomes from being on suboptimal antibiotics for 1-2 days, on average.

Mortality is driven by many other factors besides antibiotic susceptibility. Anecdotally, the patient whose death is intuitively attributable to undertreatment with broad spectrum antibiotics is a rarity; overtreatment is by far the more common scenario in modern ICUs.

It’s Worth Being Bullish on Rapid AST Anyway

Accumulated evidence strongly suggests that rapid susceptibility testing can lead to faster appropriate antibiotic de-escalation, and likely leads to more rapid escalation to cover resistant organisms as well. It’s rare to find a new test that works so consistently well on an important, clinically-relevant measure.

Although pricing is opaque, rapid susceptibility testing appears to be expensive, both on a platform and per-unit basis, and in the absence of demonstrable clinical or financial benefit, adoption may be slow. This is an area of ongoing innovation and competition among vendors of the competing diagnostic platforms and assays.

For the practicing clinician, an important distinction at this stage of the game is between assays that test actual resistance on live, cultured bacteria (e.g. the rapid AST test we’ve been discussing and its competitors), which have been demonstrated to have a high concordance with conventional susceptibility testing — versus the molecular DNA-based tests that only infer or predict resistance based on genetic profiles. The latter class of tests can provide a lot of value in early targeting decisions, but must be complemented with phenotypic resistance testing (rapid and/or conventional AST).

References

Fast Antimicrobial Susceptibility Testing for Gram-Negative Bacteremia: Research Summary. JAMA. 2026 Apr 18:e265536. doi: 10.1001/jama.2026.5536. Epub ahead of print. PMID: 41999288.

• Banerjee et al JAMA 2026, (the FAST trial, n=850)

• Banerjee et al Clin Infect Dis 2015 (n=617)

• Beuving et al Eur J Clin Microbiol Infect Dis 2015 (n=129)

• Bannerjee et al Clin Infect Dis 2021 (n=500)

• Kim et al, Clin Microbiol Infect 2021 (n=116)

• Caspar et al Eur J Clin Microbiol Infect Dis 2024 (n=112)

A 46-year-old man who drinks large amounts of alcohol daily and who has a history of MRSA bacteremia presents to the ED after falling down stairs while intoxicated. Ethanol level is 80 mg/dL. He is evaluated by the trauma team and “cleared” after CT and exam show no fractures or hemorrhage. He is agitated and is considered “too much for the floor,” and is admitted to the ICU on a dexmedetomidine infusion and symptom-triggered therapy for potential alcohol withdrawal delirium.

Two hours later, he complains of severe neck pain. On exam, vitals are normal (BP 100/55, pulse 90, respirations 18/min, afebrile), but he has diminished (4/5) strength in the deltoids, biceps, and handgrip bilaterally. Plantar flexion and hip flexion are both 5/5 bilaterally.

Cervical spine CT shows degenerative cervical spondylosis without fracture. He is placed in a cervical immobilization collar. An MRI of the cervical spine is pending. Neurosurgery has been consulted.

Read more

Acute kidney injury is associated with worse outcomes after surgery and acute illness generally. Preventing it has proven to be much harder than predicting it, however.

There’s an irresolvable problem at the heart of AKI research today: it’s impossible to reliably differentiate which patients have true injury from those with a transient, essentially benign reduction in creatinine clearance.

A recent randomized trial found a huge reduction in the incidence of moderate to severe AKI after major surgery, with the application of a biomarker-guided supportive care bundle. This was remarkable in and of itself.

But after congratulating the authors, we should also ask: were any kidneys saved?

Bigpak-2 Trial

At 34 centers in Europe, 1,180 patients undergoing major surgery (90% elective, usually abdominal/general, cardiac, or vascular) and at high risk for AKI were randomized to receive guideline-compliant preventive care (advanced hemodynamic monitoring, optimized volume status and hemodynamics, avoidance of nephrotoxins and contrast, and prevention of hyperglycemia), or usual care.

Intervention patients’ angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers were stopped for at least 48 h postoperatively. Usual care patients’ were as well, but less often.

They got passive leg raises every 3 hours for at least 12 hours, with fluid boluses for predicted fluid responsiveness.

They also got invasive hemodynamic monitoring and/or regular ultrasound exams.

Less Moderate AKI in the Intervention Arm…

Within 72 hours after surgery, 8% fewer patients in the intervention group (14.4%) than in the usual care arm (22%) had developed moderate or worse AKI (at least a doubling of creatinine or urine output <0.5 mL/kg/hr, or about 840 mL/day in a 70 kg patient).

The positive finding was most often triggered by a temporary drop in urine output below 0.5 mL/kg/hour; the creatinine-doubling criterion was met in 6% (intervention) vs 10% (controls).

There was no difference in severe (stage 3) AKI: 16% in the intervention group vs 15.8% in controls.

…But No Difference in RRT

However, there was no difference in clinically relevant endpoints, most importantly the use of renal replacement therapy (dialysis or CRRT) at 30 days (5.1% intervention vs 5.8% control), or 90 days (5.1% vs. 5.9%). Neither was there a difference in deaths at 30 days (5.1% vs 4.6%) or 90 days (7% vs 7%).

There was no difference in the MAKE30 or MAKE90 rates (in-hospital death, new receipt of renal replacement therapy (RRT), or persistent renal dysfunction at 30 and 90 days), which were virtually identical between groups.

Patients in the bundle arm had a 2.4% absolute greater rate of renal recovery at day 90 (non-significant).

The Biomarker Angle

Bigpak-2 was funded by Biomerieux, a family-owned multi-billion-euro diagnostics company (e.g., bacterial cultures, procalcitonin, BioFire, et al) headquartered in France.

Biomerieux markets the NephroCheck™ test, which is promoted as an early-warning system for severe AKI. Nephrocheck measures serum levels of the biomarkers TIMP-2 and IGFBP7. Tissue inhibitor of metalloproteinases-2 (TIMP-2) and insulin-like growth factor-binding protein 7 (IFGBP7) are signaling proteins secreted by renal tubular cells under metabolic stress that lead them to slow mitosis as a conservation strategy.

Nephrocheck multiplies TIMP-2 and IGFBP7 and returns the product as a value indexed to purported renal stress. Results >0.3 /1000 ng/mL were used as an inclusion criterion in Bigpak-2—ostensibly “enriching” the population for patients at high risk for progressing to severe AKI and dialysis.

But patients receiving the intervention guided by Nephrocheck did not have a lower rate of renal replacement therapy, nor of severe AKI (16% vs 15.8%, non-significant).

Nor did they have a greater reduction in their assay level over time (the product of TIMP-2 and IGFBP), compared to the usual care group.

That would seem to pose a problem for Nephrocheck’s use-case story, which is to identify patients at high risk for dialysis and extended stays so these outcomes can be prevented. When the interventions are working, Nephrocheck levels should fall, representing reduced risk for severe AKI.

Earlier randomized trials, including the first Bigpak trial and others, showed the same general finding: bundle-driven care in postop populations at high risk for AKI after “enrichment” with TIMP-2 x IGFBP7 (and clinical factors), led to a reduced rate of AKI but no significant reduction in need for dialysis.

But based on those earlier trials showing similar findings as Bigpak-2, use of Nephrocheck was recommended in the Enhanced Recovery After Surgery (ERAS) guidelines.

Surgeons and their hospitals have signed on, in part because it’s easy to run the seemingly advanced Nephrocheck assay on a machine platform (VIDAS) they already own.

And there’s no strong evidence that it doesn’t help in a clinically meaningful way.

Discussion

“Our findings show that the prevention of hypotension and discontinuation of ACE inhibitors and ARBs had the strongest association with the primary outcome.” —Bigpak-2 authors

The consistent relative improvement in creatinine clearance seen with biomarker-guided AKI prevention could either be from the reduction of genuine injury or better hydration and volume management (and a higher rate of ACEI/ARB discontinuation).

Intervention patients were better monitored and had much higher compliance with KDIGO bundle guidelines. They had half the rate of “hypotension for longer than 5 minutes with no intervention performed to prevent this” (an unusual, post-hoc circular metric).

This might be meaningful, but could alternatively represent a transient improvement in urine output or creatinine clearance by kidneys that were destined for the same outcome, either way.

The lack of a signal showing a reduction in the need for renal replacement therapy—even in an “enriched” population at high risk for AKI—suggests the latter.

All that said, there is a vaguely persistent signal that seems to pervade these sorts of trials. Patients who get more clinical attention do better, in one modest way or another.

That mysterious and scarce resource—focused time and cognitive effort at the bedside—may not always be measurable, but it always matters.

In terms of observable outcomes, however, in Bigpak-2 it resulted in better-managed volume status and reduced transient prerenal azotemia, without saving kidneys or preventing dialysis.

Reference

Zarbock A, et al ; BigpAK-2 study group. A preventive care strategy to reduce moderate or severe acute kidney injury after major surgery (BigpAK-2); a multinational, randomised clinical trial. Lancet. 2025 Dec 13;406(10521):2782-2791. doi: 10.1016/S0140-6736(25)01717-9. Epub 2025 Nov 13. PMID: 41242333.

Göcze I, et al. Biomarker-guided Intervention to Prevent Acute Kidney Injury After Major Surgery: The Prospective Randomized BigpAK Study. Ann Surg. 2018 Jun;267(6):1013-1020. doi: 10.1097/SLA.0000000000002485. PMID: 28857811.

Enhanced Recovery After Surgery (ERAS) guidelines.

von Groote et al, meta-analysis in ICM

After their arrival in the early 2010s, novel anticoagulants progressively displaced warfarin as first-line agents for long-term anticoagulation for stroke prevention from atrial fibrillation, and later, treatment and prevention of venous thromboembolism (pulmonary embolism and deep venous thrombosis).

The new drugs were equally or more effective than warfarin but had lower rates of bleeding in randomized trials, especially dreaded intracranial hemorrhages.

(Within the decade, makers of two of the blockbuster drugs had paid over $1.4 billion to settle multidistrict lawsuits alleging that they didn’t warn patients about the still-significant bleeding risks.)

Because the drugs were compared individually against warfarin, not each other, their comparative efficacy and bleeding risk could only be extrapolated from their individual noninferiority trials.

While rivaroxaban held the early lead in sales for several years, observational data suggested that apixaban had a lower bleeding risk with comparable efficacy. Clinicians began to choose apixaban far more often; since 2020, its sales have exceeded those of all its major competitors combined (~$14 billion for apixaban in 2025).

Apixaban’s marketers (BMS and Pfizer) have protected their patents and massive revenues, but generic versions of competitors dabigatran and rivaroxaban have become available in various markets.

The Canadian government had a financial interest in learning whether pricey apixaban was truly superior, and funded a large randomized trial.

The COBRRA Trial

At 32 centers in Canada, Australasia, and Ireland, 2,760 patients with acute venous thromboembolism were randomized to treatment with either apixaban or rivaroxaban at usual doses.

Within 3 months, clinically relevant bleeding had occurred less than half as often with apixaban as with rivaroxaban (3.3% vs. 7.1%).

Major bleeding occurred in 32 patients (2.4%) taking rivaroxaban, vs. 5 patients with apixaban.

Major bleeding was defined as in a critical site, associated with a decrease of at least 2 g/dL in hemoglobin, transfusion of two or more units of packed red cells, or contributing to death.

There was no difference in recurrent venous thromboembolism (~1% in both groups).

What Now

Rivaroxaban still retains patent protection in multiple markets, with over $5 billion in sales in 2024.

However, it’s hard to see clinicians voluntarily choosing rivaroxaban over apixaban after the COBRRA trial results, at least for acute VTE treatment.

Some clinicians will be compelled to prescribe rivaroxaban due to preexisting negotiated contracts (manifesting on the clinical side as pharmacy formulary decisions). From a patient safety standpoint, renewing such deals will be difficult to defend, however.

Dabigatran was already running a distant third to apixaban and rivaroxaban in sales. It requires several days of initial injections of enoxaparin (other agents don’t), and has fewer approved indications. A major RCT showed higher rates of GI bleeding with dabigatran than with warfarin, and its renal clearance is slower than with other agents.

Edoxaban is used in Asia and Europe at a lower cost than apixaban. In the absence of proof of higher bleeding rates, its use might still be defended. In the U.S., edoxaban carries a black box warning for patients with high-normal renal function (CrCl >95 mL/min); this subgroup taking edoxaban had higher stroke rates in a major RCT.

In other words, after 16 years of the NOAC wars, apixaban wins.

Expect rivaroxaban to go on sale, especially in the secret discounting deals the makers cut with massive health systems and pharmacies.

This will set up some awkward conversations on pharmacy formulary meetings, in which executive leadership pressure will require middle management to defend the financial decision to use a less-safe anticoagulant. “No margin, no mission” will be said a lot.

In the U.S., Eliquis™ (apixaban) retains patent protection until 2031, according to reports. Although BMS-Pfizer only projects sales out ~2 years, continuing at 2025’s levels would produce an additional $70 billion in revenue for the clot-and-competitor-busting little pill. BMS-Pfizer also have options to extend Eliquis’s patent protection beyond 2031.

Reference

Castellucci LA, Chen VM, Kovacs MJ, et al. Bleeding Risk with Apixaban vs. Rivaroxaban in Acute Venous Thromboembolism. The New England Journal of Medicine. 2026;394(11):1051-1060. doi:https://doi.org/10.1056/NEJMoa2510703

‌

An 81-year-old man arrives in the emergency department after experiencing dark stools for a week. He takes warfarin for a mechanical mitral valve, and also for atrial fibrillation and a history of massive pulmonary embolism treated with thrombolytics six months ago.

His blood pressure is 110/55; his pulse is 110/min. Hemoglobin is 6.3 g/dL. INR is 3.4.

Vitamin K and packed red blood cells are ordered. Both fresh frozen plasma and 4-factor prothrombin complex concentrate are available.

Read more

Hemoperfusion and hemadsorption are so-called “blood purification” technologies that can remove endotoxin, cytokines, or other potentially deleterious proteins from the bloodstream in sepsis or other medical conditions.

These are extracorporeal circuits (often running on CRRT machines) that pump blood through a matrix that attracts or adsorbs the offending proteins in bulk before returning the “purified” blood to the patient.

Blood purification technologies have regulatory approval in multiple countries in Europe, Asia, Canada, and elsewhere, but not yet in the U.S.

That’s primarily because although they do remove the proteins they are designed to, blood purification has never been demonstrated to improve outcomes in sepsis.

Spectral Medical, makers of an endotoxin-removal device using polymyxin B and marketed as Toraymyxin™, would like you to believe otherwise.

The TIGRIS trial’s results were published in Lancet Respiratory Medicine in March 2026.

It concluded that among a selected group of sepsis patients with endotoxin levels in a “Goldilocks” zone (not so low or high that removal wouldn’t help), the likelihood that polymyxin B hemoperfusion saved lives was >99%.

However, the path to this conclusion was more winding (and fanciful?) than any fairytale, relying on p-hacking, aggregating pools of patients across clinical trials, and Bayesian probabilities rather than standard (“frequentist”) statistical calculations.

In other words, “precision medicine.”

It went down like this:

The EUPHAS (2009) trial was stopped early at n=64 when it seemed that polymyxin hemoperfusion (PMX) reduced mortality in sepsis.

But ABDOMIX (2015) suggested harm among patients undergoing emergency surgery for peritonitis.

EUPHRATES (2018) then became the lodestar of hemoperfusion. The trial did not show a mortality benefit from PMX over sham treatment in a multicenter, randomized trial of 450 patients with severe sepsis and high endotoxin activity (EAA level ≥0.60) at 55 tertiary hospitals in North America, from 2010 to 2017.

Nor was there a benefit in the sickest patients with high organ failure scores. Nor did patients treated with hemadsorption in EUPHRATES experience a reduction in endotoxin activity compared to sham-treated patients.

However, post-hoc analysis found that the 194 patients in EUPHRATES with endotoxin activity levels between 0.60 and 0.89 had a non-significant reduction in 28-day mortality with polymyxin hemoperfusion (26% vs 37%, p=0.11, reduced to 0.047 after adjustment).

This endotoxin zone of 0.60 to 0.89 units was then used as the “precision” in the future “precision medicine” TIGRIS trial design.

TIGRIS Trial: Can I Borrow a Cup of Patients?

TIGRIS enrolled 194 septic patients, or 1% out of an enormous ~15,000 screened. They were all very sick and had endotoxin activity levels between 0.60 and 0.89. They were randomly assigned 2:1 to receive PMX hemoperfusion or usual care (treatment assignment was not blinded).

The “real” (legible, familiar, frequentist) results were:

• No mortality difference at 14 days (34% PMX vs 37% control)

• 90-day mortality: A trend toward benefit with PMX (HR 0.68 [0.43 - 1.07]), with separation occurring around 28 days.

The authors also used an unusual technique of combining the TIGRIS patients with the (post-hoc fished-out) EUPHRATES cohort. This is called “historical borrowing,” but it is not simply pooling the patients and then performing traditional statistical analysis on the dataset. Rather, the earlier patients are used as weighting factors in a complicated Bayesian modeling scheme. This weighted Bayesian analysis produced the reported 99% likelihood of an improvement in 90-day survival with PMX.

Guidelines came out before TIGRIS

The Surviving Sepsis Campaign (PulmCCM has no affiliation) in its latest guideline update, gave polymyxin B hemoperfusion (along with related technologies) one of its rare raspberries: a thumbs-down for patients with sepsis:

We ‘suggest against’ using blood purification techniques, including hemoperfusion, high-dose hemofiltration, or plasma exchange. … we ‘suggest against’ using polymyxin B hemoperfusion.

They noted the evidence was weak either way and this was a conditional (weak) suggestion (like nearly all of the other recommendations).

However, the committee did not have full access to the published TIGRIS data— although preliminary data was available prior to the guidelines publication.

The “Theranostics” Business Model

It’s likely that even if a benefit were confirmed, only a fraction of patients with sepsis would be eligible for treatment with PMX hemoperfusion (note the 1% enrollment rate from ~15,000 screened).

But Spectral Medical also licenses/sells the endotoxin activity assay that would serve as a screening test for PMX hemoperfusion eligibility. FDA approval of Toraymyxin™ could allow Spectral Medical to market the endotoxin assay as an appropriate test for a large proportion of patients presenting with severe sepsis.

Controlling both the treatment and the test for its eligibility in this way is dubbed “theranostics” on the business side (i.e., diagnostics + therapeutics, but “diagpeutics” didn’t roll off the tongue as well).

Spectral Medical has estimated the total annual U.S. market for PMX hemoperfusion plus its endotoxin assay at about $1.5 billion. A possible price for the PMX cartridge was ~$7,500 in an investor presentation, with about $300 million in annual revenue for both PMX + EAA with strong market penetration. (The endotoxin test was only expected to bring in ~$10 million per year.)

The company’s financials suggest a small startup dependent on investor capital and debt: they reported having about $1.5 million in cash and ~$9 million in debt in March 2024, with only about $1.6 million in annual revenue (USD).

Baxter controls about half the U.S. market for CRRT machines, which would allow for efficiencies in the marketing and distribution of PMX cartridges.

FDA Unconvinced So Far

Toraymyxin™ received a breakthrough medical device designation from the FDA in 2022. Spectral Medical and its major investor, Vantive (a spinoff from Baxter), reported that the FDA had agreed to their Bayesian analysis of the combined EUPHRATES and TIGRIS data rather than a larger conventional trial. The company previously announced it would seek full FDA approval as soon as Q1 2026.

However, their subsequent press releases suggest that FDA has requested additional data (e.g., on longer-term survival in TIGRIS). FDA approval will not occur until January 2027 at the earliest, the company has signaled to investors.

There do not seem to be any randomized trials in process or planned testing polymyxin hemoperfusion for sepsis, and the endotoxin/polymyxin theranostics platform is Spectral Medical’s major offering outside its small reagents business. It seems likely that the firm will live or die by the FDA’s approval decision.

Imprecise Enthusiasm

In any industry-funded randomized trial with n<200, a non-significant trend toward a mortality benefit in an experimental therapy should prompt skepticism, and TIGRIS is no exception.

But they put in the work. The TIGRIS trial was a successful execution of a very difficult trial design, and an achievement worth recognizing.

Its use case may be extremely narrow. It might not, in fact, improve outcomes at all. Hundreds of millions or billions of dollars may be spent on endotoxin tests and PMX cartridges only to enrich Vantive’s and Spectral’s stockholders, before the therapy is finally disproven.

Ten years ago, that would have prompted reflexive scorn in this newsletter.

But the limited data for 90-day mortality in TIGRIS suggest there’s a chance that it might help. And in the sprawling junkyard of failed treatments for sepsis, the plausible hope that something might eventually be shown to work is still worth rooting for. With less than 99% enthusiasm, to be sure—but some.

The question is whether that hope justifies launching what would be in effect a taxpayer- and insurer-funded, population-level postmarketing observational study on patients with endotoxin-predominant sepsis.

Does it?

Reference

Neyra JA, Legrand M, Tidswell MA, et al. Polymyxin B haemoadsorption in endotoxic septic shock (Tigris): a multicentre, open-label, Bayesian, randomised, controlled, phase 3 trial. The Lancet Respiratory Medicine. Published online March 2026. doi:https://doi.org/10.1016/s2213-2600(26)00047-0

‌Spectral Medical 2024 investor presentation

A 56-year-old man has been admitted to the ICU with sepsis due to pyelonephritis. He developed gram-negative bacteremia, with worsening shock, respiratory failure, and encephalopathy, and was intubated six days ago.

He was weaned off vasopressors, but continues to require mechanical ventilation. He is receiving enteral nutrition through an orogastric tube at 50 mL/hour. He had an episode of emesis three days ago, after which enteral feedings were interrupted for several hours and then resumed.

The nurse informs you that his gastric residual volumes are 420 mL. It has been five days since his last bowel movement. His abdominal exam shows unchanged mild distension, with occasional audible bowel sounds. An abdominal film shows mildly dilated loops of small bowel, unchanged from previous films.

For many years, gastric feeding was considered risky in critically ill patients. When postpyloric feeding was infeasible, gastric residual volume checks were used to reduce the perceived risk, with >100 mL considered to be a sign of intolerance of feedings. The 100 mL number, arbitrary to begin with, crept up over the years, but gastric volumes are still commonly checked in ICUs; if “elevated” (e.g., >250 mL, >400 mL, or >500 mL), enteral feedings are interrupted.

Two mid-sized randomized trials concluded that in patients receiving enteral feedings, regularly checking gastric residual volumes does not affect outcomes:

• NUTRIREA1, n=452, France (JAMA 2013): Numerically similar VAP rates using >250 mL residual volume as a trigger to stop feedings in the control group, vs. not checking; more calories were delivered to patients in the unchecked arm

• REGANE, n=329, Spain (Int Care Med 2010): No difference in calories delivered or VAP in patients using >500 mL residuals as feeding-stopping trigger, vs 200 mL

In NUTRIREA1 and some other trials, patients whose gastric volumes were not checked received more total calories.

As a result, U.S. and European nutrition guidelines advised not withholding enteral feedings for gastric volumes <500 mL, and U.S. guidelines advised not checking gastric volumes at all.

Checking gastric volumes is an ingrained nursing practice, however, and is performed in many if not most ICUs. About one-third of patients will have gastric volumes >200 mL during their ICU stays.

But Why Force-Feed Ventilated Patients?

NUTRIREA had an interesting finding of significantly more vomiting in the intervention patients (whose gastric volumes were not checked). About 7% of the intervention patients vomited 5 or more times during the trial—twice as many as in controls, whose feedings were stopped for gastric volumes >250 mL.

Authors seem to consider this to be insignificant, because the (supposedly) more important goal was achieved: more total calories delivered in the intervention arm.

But caloric targets in mechanically ventilated patients have no basis in evidence. They were arbitrarily invented by nutritionists based on theoretical metabolic assumptions.

In multiple large randomized trials, providing as little as half the nutritionist-defined goal to ventilated patients had no detectable influence on outcomes. This has been one of the stronger signals in critical care research.

Not checking gastric volumes in critically ill patients is guideline-compliant. It follows the evidence.

It also probably results in force-feeding a lot of very ill people who aren’t hungry, who don’t need a healthy person’s full day’s worth of calories (as anorexia is a natural aspect of illness and may be adaptive), and making many of them nauseous.

Assuming mechanically ventilated patients are receiving at least 50-75% of a healthy person’s daily calories, either checking gastric volumes or not is equally reasonable.

Reflect to earn CME with Learner+

Sample reflection: I reviewed and reflected on the evidence of and against the use of gastric residual volume checks to manage enteral feedings in mechanically ventilated critically ill patients.

References

Reignier J. Effect of Not Monitoring Residual Gastric Volume on Risk of Ventilator-Associated Pneumonia in Adults Receiving Mechanical Ventilation and Early Enteral Feeding. JAMA. 2013;309(3):249. doi:https://doi.org/10.1001/jama.2012.196377

Montejo JC, Miñambres E, Bordejé L, et al. Gastric residual volume during enteral nutrition in ICU patients: the REGANE study. Intensive Care Medicine. 2010;36(8):1386-1393. doi:https://doi.org/10.1007/s00134-010-1856-y

Zoeller S, Bechtold ML, Burns B, et al. Dispelling Myths and Unfounded Practices About Enteral Nutrition. Nutrition in Clinical Practice. Published online January 29, 2020. doi:https://doi.org/10.1002/ncp.10456

‌Marr AB, McQuiggan MM, Kozar R, Moore FA. Gastric Feeding as an Extension of an Established Enteral Nutrition Protocol. Nutrition in Clinical Practice. 2004;19(5):504-510. doi:https://doi.org/10.1177/0115426504019005504

‌Lindner M, Padar M, Merli Mändul, et al. Current practice of gastric residual volume measurements and related outcomes of critically ill patients – a secondary analysis of the iSOFA study. Journal of Parenteral and Enteral Nutrition. 2023;47(5). doi:https://doi.org/10.1002/jpen.2502

‌ASPEN guidelines

ESPEN guidelines

‌‌

After the publication of HI-PEITHO, which concluded that catheter-directed thrombolysis for acute PE prevents “cardiorespiratory decompensation or collapse” (a semantically exaggerated result the industry-funded authors cleverly parsed out of the observed improvement in physiology)—

—and the endorsement of CDT (but moreso thrombectomy) for PE by multiple professional societies—

—intensivists around the world (but especially the U.S., where financial motivations drive a large proportion of care decisions) are asking:

“How can I help my patients with PE avoid undergoing unnecessary procedures?”

Most critical care folks know of Dr Josh Farkas and his famed PulmCrit blog and online textbook. If not, you should check them out.

Farkas recently gave an enlightening take on recent trials and CDT generally, arguing that catheter-directed thrombolysis is (in all important respects) really just peripherally-delivered low-dose tPA with a useless and expensive plastic apparatus attached, whose purpose is to treat the prescribing clinician’s anxiety about causing bleeding with “systemic tPA” (which, due to the circulatory nature of the circulatory system, catheter-directed tPA is, too).

Others made comments about the trial in the same spirit: I wish there had been a low-dose tPA arm in HI-PEITHO.

As it turns out, the recently published STRATIFY trial included just such a comparison.

Instead of reinventing wheels, let’s take Farkas’s arguments in these great posts on HI-PEITHO (here) and STRATIFY (here) out for a spin.

What Are Low-Dose Thrombolytics for PE?

“Full-dose” thrombolytics include alteplase (tPA), which is typically given as 100 mg IV for massive, life-threatening pulmonary PE, in a 10 mg bolus followed by a 90 mg infusion over two hours. Tenecteplase (TNK) is not FDA-approved for PE, but its full dose is often given as a 50 mg bolus. TNK dosing can also be weight-based up to 50 mg (with ranges of ~30 to 50 mg in the PEITHO trial, e.g.)

“Low-dose” tPA has no standard definition, but 10 mg IV bolus followed by 40 mg IV infused over 2 hours was used in one widely mentioned RCT.

Yet nothing bars careful and thoughtful physicians from using lower doses of tPA (e.g., 25 mg), or slower infusions of any dose. Anecdotally, a minority do.

Reduced-dose TNK has not been well-described in the published literature.

Evidence for Lower-Dose tPA for Acute PE

As with most other treatments for acute pulmonary embolism, there is very little evidence from randomized trials to support the use of reduced-dose tPA.

The MOPETT trial (Am J Cardiol 2013) enrolled 121 patients at one center in Arizona, randomizing them to 50 mg tPA with anticoagulation or AC alone. Although all patients were symptomatic and had elevated estimated pulmonary artery pressures on echocardiogram, the authors did not report how ill the patients were at enrollment (not even their BP), nor after treatment. Low-dose tPA resulted in lower estimated PA pressures and recurrent PE. Four patients died (one in the tPA, three in the AC only arm). Authors reported zero bleeding in either arm (this is unusual). They dubbed their intervention “safe dose tPA” in their methods (before getting results). Investigators were unblinded. The risk of bias is high. MOPETT is minimally informative by today’s standards, and hard to compare to other trials in PE.

A RCT by Wang et al. (n=118), performed in China and published in Chest in 2010, concluded that 50 mg tPA produced outcomes equivalent to 100 mg tPA, with less bleeding.

The STRATIFY trial (n=210) (Cardiovascular Research 2026) suggested 20 mg of peripherally infused tPA over 6 hours worked better than heparin and as well as CDT at reducing radiographic clot burden. Bleeding was numerically higher (equally) with either low-dose lytics or CDT as compared to heparin alone. Mortality was low, few PEs became severe, and the trial was not powered to detect differences in clinical outcomes.

PEITHO-3 is testing low-dose tPA (0.6 mg/kg IV up to 50 mg over 15 minutes) against placebo (plus heparin in both arms) in intermediate-high risk PE. It doesn’t include a CDT arm. Results are expected around 2027.

Guidelines Endorse Low-Dose tPA, Anyway

As we pointed out recently, guidelines co-signed by multiple societies (AHA, ACC, ACCP, ACEP, etc.) endorsed virtually all non-insane approaches to treating acute pulmonary embolism, regardless of the evidence base (which was low for all therapies).

Low-dose tPA got its seat at the table with a 2b (weak) recommendation for many patients with intermediate-high risk PE.

PulmCCM is not affiliated with any specialty society.

Which brings us to Farkas’s most interesting argument.

The Claim: Catheter-Directed tPA ≈ Very Low-Dose Systemic tPA

First, let’s compare apples to apples:

During catheter-directed thrombolysis using Boston Scientific’s ultrasonic-tipped EKOS™ catheter in HI-PEITHO (n=544), about 8 mg tPA was instilled into each pulmonary artery (usually both, ~16 mg total) over about 7 hours. The control arm got heparin, not peripheral tPA.

In STRATIFY, 20 mg tPA was administered over 6 hours, systemically or using the EKOS catheter (or neither in the heparin-only arm). Clot dissolution was statistically equivalent between the CDT and systemic low-dose tPA arms (with methodologic problems noted here). With fewer than 80 patients per arm who were not that sick, the study lacked statistical power to identify clinically meaningful differences.

Bleeding rates were numerically and statistically similar between peripherally-infused and centrally-infused tPA.

Sub-Claim #1: Instillation of tPA in the pulmonary artery does not lyse clot any more effectively than peripheral tPA.

This is supported by a study by Verstraete et al in 1988, showing equivalent results after either form of tPA administration, although in only 34 patients.

This was followed by an in vitro experimental study in dogs by Schmitz-Rode (Cardiovasc Intervent Radiol 1998), which showed what you would intuitively predict: intrapulmonary artery-instilled liquid simply flowed down the path of least resistance, away from the PA clots, into the systemic circulation.

Subsequently, Boston Scientific added an ultrasound tip to its catheter that purports to force tPA sideways into the clot.

Sub-Claim #2: Ultrasound “assistance” doesn’t increase penetration into the clot or provide any advantage over conventional catheters (which aren’t better than peripheral tPA, either)

From PulmCrit:

Over time, it became increasingly clear that ultrasonic energy is ineffective. Some studies in DVT showed this, as well as a registry trial in pulmonary embolism. (25856269, 27630267, 28827014, 25593121, 26993702, 30915914, 30915912) More recently, the SUNSET trial compared USAT to conventional catheter-directed thrombolysis and found no differences. (34167677)

Thus, the entire theoretical rationale for using catheter-directed thrombolysis collapsed. If Verstraete et al. showed that catheter-directed thrombolysis was worthless, and ultrasonic energy is worthless then… the entire intervention makes no sense.

The Verdict:

It is as likely as not that peripherally-infused tPA is equally effective to catheter-directed tPA for acute PE, at equal doses and infusion rates. I.E. there is equipoise on this question.

This is in no small part because device manufacturers have (shrewdly) shied away from testing their catheters directly against peripherally infused tPA.

With STRATIFY providing the only recent and relevant data, until larger trials testing CDT against peripheral low-dose tPA are performed, there is no good evidence that CDT is superior to peripheral tPA at equivalent doses.

The Implication:

It seems medically, ethically, and professionally appropriate for clinicians to advocate against catheter-directed thrombolysis in favor of peripherally-infused low-dose tPA at similar doses and infusion duration, because it spares patients with intermediate-high risk PE a potentially unhelpful procedure, while providing likely comparable efficacy at a similar (or lower) risk.

This approach would bring up yet another wrinkle, though.

Assume low-dose tPA (e.g. 18 mg over 7 hours) carries a higher rate of bleeding than thrombectomy (which seems likely), but with the advantage of being noninvasive.

If that is the case, and thrombectomy is at least as effective as low-dose tPA (as also seems likely, in patients with accessible clot)—at what rate of tPA-induced bleeding does a “thrombectomy first” approach—in which many people with intermediate-high risk PE would be unnecessarily catheterized to improve a few and avoid bleeding in a few more—make sense?

This line of thought tends to align with the recent guidelines’ subtle support of thrombectomy over CDT.

Another possible approach: low-dose tPA for most patients with intermediate-high risk PE with abnormal physiology; mechanical thrombectomy as first-line for the more severe cases or as rescue for those who deteriorate. Their bleeding risk for the procedure after tPA might not be markedly higher than if they had gotten catheterized for CDT in the first place.

A significant additional risk from this option? Hearing “I told you so” from the proceduralist who wanted to do it earlier at 2 pm, not now at 3 am.

PEERLESS II (an industry-funded thrombectomy trial) is expected to wrap up in mid-2026.

References

Avgerinos ED, et al; SUNSET sPE Collaborators. Randomized Trial Comparing Standard Versus Ultrasound-Assisted Thrombolysis for Submassive Pulmonary Embolism: The SUNSET sPE Trial. JACC Cardiovasc Interv. 2021 Jun 28;14(12):1364-1373. doi: 10.1016/j.jcin.2021.04.049. Erratum in: JACC Cardiovasc Interv. 2021 Oct 11;14(19):2194. doi: 10.1016/j.jcin.2021.07.010. PMID: 34167677; PMCID: PMC9057455.

Wang C, et al ; China Venous Thromboembolism (VTE) Study Group. Efficacy and safety of low dose recombinant tissue-type plasminogen activator for the treatment of acute pulmonary thromboembolism: a randomized, multicenter, controlled trial. Chest. 2010 Feb;137(2):254-62. doi: 10.1378/chest.09-0765. Epub 2009 Sep 9. PMID: 19741062; PMCID: PMC7126994.

Kjaergaard J, Bang LE, Sonne-Holm E, et al. Randomized trial of low-dose, ultrasound-assisted thrombolysis or heparin for pulmonary embolism. Cardiovascular Research. Published online January 30, 2026. doi:https://doi.org/10.1093/cvr/cvag038

‌Rosenfield K, Klok FA, Piazza G, et al. Ultrasound-Facilitated, Catheter-Directed Fibrinolysis for Acute Pulmonary Embolism. New England Journal of Medicine. Published online March 28, 2026. doi:https://doi.org/10.1056/nejmoa2516567

Sharifi M, Bay C, Skrocki L, Rahimi F, Mehdipour M. Moderate Pulmonary Embolism Treated With Thrombolysis (from the “MOPETT” Trial). The American Journal of Cardiology. 2013;111(2):273-277. doi:https://doi.org/10.1016/j.amjcard.2012.09.027

‌Schmitz-Rode T, Kilbinger M, Günther RW. Simulated flow pattern in massive pulmonary embolism: significance for selective intrapulmonary thrombolysis. Cardiovascular and interventional radiology. 1998;21(3):199-204. doi:https://doi.org/10.1007/s002709900244

‌Sanchez O, Anaïs Charles-Nelson, Ageno W, et al. Reduced-Dose Intravenous Thrombolysis for Acute Intermediate–High-risk Pulmonary Embolism: Rationale and Design of the Pulmonary Embolism International THrOmbolysis (PEITHO)-3 trial. Thrombosis and Haemostasis. 2021;122(05):857-866. doi:https://doi.org/10.1055/a-1653-4699

And of course PulmCrit

Catheter-directed thrombolysis has been used for more than 30 years as a therapy for acute pulmonary embolism.

Until March 2026, the evidence for its use has been limited to studies showing it improves right ventricular function, not meaningful clinical endpoints.

A well-conducted randomized trial adds some evidence that CDT works better than the standard treatment of anticoagulation alone. But make sure to read the data, not just the marketing materials—I meant manuscript.

HI-PEITHO Trial

Patient selection and enrollment has been a major challenge for randomized trials of invasive treatments for acute PE.

Investigators have reported reluctance among patients with more severe PEs to be enrolled in anticoagulation-only arms. They want the procedures.

Further, more severe (intermediate and high-risk) PEs are far less common than milder PEs. But if the more abundant but less-ill patients are enrolled, it becomes far less likely to find a benefit of the invasive treatment (if one exists).

A major strength of HI-PEITHO is that most enrolled patients had clinically “real” intermediate-high risk PE. Besides evidence of right ventricular strain/failure and elevated troponin, they showed at least two signs of physiologic stress: systolic blood pressures ≤110, heart rates ≥100, or respiratory distress (>20 breaths/min or hypoxemia).

How sick were they? The average NEWS score was 6—e.g., a patient with:

• normal BP 120/80 and HR 95 but breathing 25/min with SpO2 92% despite supplemental oxygen

• a soft BP of 95/65, HR 115, RR 22, SpO2 96% on room air

Over 4,300 were screened at 59 U.S. and European sites, to find 544 to be randomized to treatment with either the EkoSonic™ ultrasound-assisted thrombolysis catheter (plus anticoagulation) or anticoagulation alone. Wires in this catheter’s tip emit low-intensity, high-frequency ultrasonic energy continuously along the clot. This reportedly drives thrombolytics deeper, aiding lysis.

About half of the control group received unfractionated heparin infusions, while the other half received low molecular weight heparin injections.

The protocol advised placement of the EkoSonic catheter within two hours of randomization, and this was achieved in 73% of patients (see? it can be done!), and no later than 6 hours.

Randomization was computer-generated and concealed, and adjudicators of outcomes were blinded to treatment assignments. Boston Scientific, makers of EkoSonic, funded the trial.

HI-PEITHO Results

Patients undergoing catheter-directed thrombolysis had a lower rate of clinical deterioration.

Did CDT-treated patients have a lower incidence of:

• Cardiac arrest (2.2% vs 1.5%)?

• Intubation with mechanical ventilation (2.2% vs 1.8%)?

• Recurrent PE (0.4% vs 0.4%)?

• PE-related death? (1.1% vs 0.4%)

No! All these numbers numerically favored anticoagulation or equipoise (but they are too small to impute significance).

What, you thought you were going to get clinically meaningful hard endpoints? This is an industry-funded trial. The composite endpoint is described further below.

Notably, though, five catheter-treated patients showed signs of shock (hypotension plus end-organ hypoperfusion), compared to 9 in the anticoagulation arm (non-significant but close).

The trial turned positive for EkoSonic because catheter-treated patients had a markedly lower rate of physiologic deterioration within 7 days as measured by the NEWS score.

Only two patients in the catheter arm (0.7%) experienced a NEWS score ≥9, one-tenth of those in the anticoagulation-only arm (19, or 7%).

Big NEWS?

A NEWS score ≥9 generally requires at least moderate hypoxemia and tachycardia, usually with a soft blood pressure or hypotension.

Examples of NEWS = 10:

• RR 30, SpO2 90% on oxygen, normal BP, HR 120

• RR 22, SpO2 95% on O2, BP 95/60, HR 135

• RR 22, SpO2 95% on O2, BP 85/50, HR 125

Fiddle with the knobs on the NEWS calculator yourself here.

This outcome (NEWS≥9) was lumped into “cardiopulmonary decompensation or collapse” along with intubation and cardiac arrest, and drove the difference in the composite outcome, which also included pulmonary embolism–related death (no difference) and recurrence of pulmonary embolism (no difference).

No Mortality Difference, But?

We mentioned the difficulty investigators have reported in enrolling patients into PE trials, due to patients’ reluctance to agree to receive anticoagulation alone.

One way to address that is to promise them they’ll receive the intervention if they deteriorate.

That was built into this trial.

More patients in the anticoagulation arm (25 or 9.2%, vs 8 patients or 2.9% in the intervention group) underwent rescue therapy.

Digging into the supplemental appendix table S15, almost all these crossovers were driven by the higher NEWS scores. Very few patients in either group required vasopressors (1%), intubation (~1%), CPR (1.5%), in a “rescue” context.

What About Bleeding?

There was about double the incidence of major bleeding by ISTH criteria in the catheter-directed thrombolysis arm (~4% vs ~2%) at 72 hours and 7 days. This did not reach statistical significance.

One patient in the intervention group died of an inguinal hemorrhage more than 7 days after catheterization.

How to Train Chatbots (and the Docs Who Rely On Them)

Take a moment to appreciate the achievement of HI-PEITHO, which appears to have been rigorously and fairly conducted, without any overt sketchiness in its design and choice of endpoints.

Unlike a certain maker of mechanical thrombectomy catheters we could mention.

Even more brilliant than HI-PEITHO’s design and execution, however, may be its choice of phrasing for its outcomes and manuscript.

How do you get from a notable but underwhelming finding of significantly fewer patients with transiently worsened physiology (NEWS ≥9), but no reduction in badness like cardiac arrest, intubation, or death (which seldom occurred), to a topline conclusion like this:

“Catheter-directed fibrinolysis plus anticoagulation led to a lower risk of the composite of pulmonary embolism–related death, cardiopulmonary decompensation or collapse, or recurrence of pulmonary embolism than anticoagulation alone, with a relative risk of 0.39 (95% CI, 0.20 to 0.77).”

You craft the frightening-sounding “cardiopulmonary decompensation or collapse,” and then smuggle into it the commonplace, wobbly (but not collapsing) physiology in PE captured by NEWS≥9. Then you put the table showing NEWS driving the study’s findings in the supplemental appendix.

So now, if you ask a widely-used AI chatbot “does CDT prevent cardiac arrest in PE”, it dutifully generates the output

“there is evidence suggesting catheter-directed thrombolysis may reduce cardiopulmonary decompensation or collapse in intermediate-risk PE patients with signs of impending deterioration.”

Mission accomplished!

Besides the two Boston Scientific-employed authors, credit their medical writer, whom the authors thanked for “helping coordinate the manuscript writing process.”

Direct Payments to Investigators

Boston Scientific made more than $528,000 total in direct payments to several of the U.S.-based HI-PEITHO investigators between 2018 and 2024, according to CMS Open Payments data. These were reported as being for consulting, speaking, travel, etc (research payments are reportedly separately). Four physicians received almost all the payments; most received little or none. Many of the investigators are based in Europe, where no payments could be identified.

Inari made over $700,000 in direct payments to investigators for PEERLESS II, according to CMS Open Payments.)

Conclusions

The HI-PEITHO trial provides support for catheter-directed thrombolysis for intermediate-high-risk pulmonary embolism.

It showed that physiologically stressed patients with PE had less physiologic deterioration when they were quickly (<6 hours, usually <2 hours) treated with catheter-directed ultrasonically-assisted thrombolysis.

Because patients in the control arm were also treated with CDT if they had physiologic deterioration, it’s impossible to know if their deteriorations were transient or would have led to an increased incidence of bad outcomes in the control arm. “Hard” events like death, intubation, cardiac arrest, or recurrent PE occurred rarely and without differences between groups.

And that’s another stroke of brilliance to the design and reporting: after building the composite around the scary words “death” and “collapse,” which happened too infrequently to compare, the manuscript proposed that it was the CDT “rescues” themselves (which were not performed in crashing patients, overall) that prevented “collapse” from occurring. This is non-falsifiable. Bravo.

Boston Scientific experienced a significant marketing setback when the AHA, ACC, ACCP, and multiple other societies co-signed a set of guidelines in 2026 that can be interpreted as warning clinicians about the supposed excess bleeding risks of CDT over mechanical thrombectomy:

Major bleeding at 7 days occurred at ~4% vs ~2% in controls in HI-PEITHO; although not statistically significant, the figure doesn’t help CDT. But the top-line conclusions of this unequivocally positive trial certainly do.

It’s good to know CDT can improve cardiopulmonary physiology, not just measures of RV function. Are harder clinical endpoints too much to hope for? Given the limitations of research on people with more severe PEs, there will probably never be a large RCT testing invasive therapies vs anticoagulation without high crossover. It’s ethically dicey and unreasonable to ask patients to take that kind of risk today. By now, most intensivists believe that these treatments help sometimes; it’s patient selection that remains a challenge.

Until that hypothetical trial is performed, interventionalists will continue to have wide discretion in the use of catheter-based therapies for acute PE—and in selecting the patients they believe might benefit.

Reference

Rosenfield K, Klok FA, Piazza G, et al. Ultrasound-Facilitated, Catheter-Directed Fibrinolysis for Acute Pulmonary Embolism. New England Journal of Medicine. Published online March 28, 2026. doi:https://doi.org/10.1056/nejmoa2516567

‌

For much of the 20th century, blood transfusion meant whole blood transfusion, which was the standard both for replenishing ordinary anemia and as resuscitation for hemorrhagic shock.

There was no safety or efficacy reason why whole blood transfusion faded away. Rather, as technological and logistical advancements in the 1960s allowed for the efficient separation, storage, transport, and transfusion of the various components of whole blood (red cells, plasma, and platelets), transfusing components simply became the default.

Type-O whole blood (lacking AB antigens, a.k.a. “universal donor” blood) has continued to be used in emergencies, i.e., massive hemorrhages from trauma and gastrointestinal bleeding, with a large body of observational cohort data suggesting it might have advantages over component transfusions.

Recent research has focused on the use of so-called low-titer type O whole blood, which has minimal amounts of A and B antigen in its included plasma. This reduces the risk of transfusion reactions in recipients with blood types A, B, or AB.

Type-matched whole blood has also been extensively used in forward-deployed combat military units, particularly by the U.S. in Afghanistan.

Why Transfuse Whole Blood?

Whole blood is considered to be inherently more physiologic than blood components.

Transfusing whole blood seems to neatly solve the cumbersome ratio-calculating pertaining to massive transfusion of blood components (e.g., RBC:plasma:platelets). It can also be easier logistically and operationally.

Clotting factors are transfused in the same solution with red cells (not sequentially), theoretically enhancing coagulation.

Whole blood may maintain better oncotic properties, which could enhance resuscitation and perfusion for patients in shock.

Evidence for Whole Blood Transfusion in Hemorrhage

Excitement about whole blood transfusion originated from military experience.

In a small retrospective observational study (Spinella J Trauma 2009), 100 U.S. military combat casualty patients who were transfused whole blood had >12% absolute improvement in 24-hour and 30-day survival compared to 254 transfused with component products.

This signal seemed to grow stronger over time.

Nonrandomized data from U.S. casualties in Afghanistan aligned with the Spinella data, followed by a larger observational cohort by Gurney et al (Surgery 2022) finding an enormous survival benefit (an adjusted odds ratio of 6-hour mortality of 0.27) for soldiers who were transfused warm fresh whole blood, as compared to soldiers transfused component products.

Warm fresh whole blood is typically collected from fellow pre-screened soldiers and transfused within hours, without any refrigeration. It has major logistical advantages over refrigerated blood in combat and forward-deployed scenarios.

In the observational studies, there was no apparent excess risk (such as for transfusion reactions) associated with whole blood, including type-O or recipient-matched whole blood.

Based on the Afghanistan data, warm fresh whole blood came to be considered by some experts to be the gold standard for hemorrhagic transfusion.

Whole Blood Transfusion for Trauma in Civilian Populations

Based on the published military experience, efforts began in the U.S. trauma community to adapt blood banking operations to permit transfusion of whole blood in the prehospital and hospital settings.

Like the military studies, observational studies in the civilian trauma literature found associations between whole blood transfusion and improved survival:

• In a registry of 501 participating U.S. centers, 2,884 trauma patients transfused whole blood had lower mortality at 4 hours (adjusted odds ratio: 0.81), 24 hours, and 30 days compared with component-only (Dorken-Gallestegi et al Ann Surg 2024),

• 840 trauma patients receiving whole blood in southwest Texas had a four-fold increase in survival odds, compared to component blood recipients (Brill et al J Am Coll Surg 2022),

• 1,180 whole blood recipients had twice the survival odds as component blood recipients at 14 U.S. centers (Hazelton et al Ann Surg 2022),

• And a 37% lower risk for mortality with whole blood among 435 trauma patients in North American centers (Torres et al JAMA Surgery 2023),

• Survival was also greater than predicted among patients transfused with whole blood pre-hospital (Rajesh et al, Am J Surg 2025).

A meta-analysis of 38 cohort studies (and two RCTs) including ~50,000 patients (JAMA Surgery 2022) found a large survival boost associated with whole blood transfusion in trauma (odds ratio 0.76 for death).

Of course, due to systemic, operational, and secular factors, the patients receiving whole blood in these studies were different from those receiving component blood.

And the totality of care provided to them was different.

In both measurable and unmeasurable ways.

That’s called confounding.

And we all know where this is going, don’t we.

Then Someone Had To Go And Run Randomized Trials And Ruin Everything

Whole blood has been tested directly against component blood in at least two randomized controlled trials. Both were firmly negative.

U. Pitt pilot study

Among 86 patients in a pilot study at U. of Pittsburgh, mortality at 28 days was virtually identical (25% vs 26%) with either whole blood or components transfused prehospital (Guyette et al J Trauma Acute Care Surg 2022)

SWIFT trial (NEJM 2026)

Among 616 patients transfused prehospital in air ambulances in England, those who were randomized to be transfused whole blood had numerically equal or higher mortality at all time points measured, from 6 hours up to 90 days after randomization. They also had near-identical rates of massive transfusion as those receiving component blood. Death and massive transfusion were combined into a composite outcome, which did not differ between groups. (Smith et al NEJM 2026)

Patients only received up to two units of red cells prehospital, which leaves open the possibility of an inadequate dose to achieve a benefit.

Conclusion

Whole blood transfusion has a strong biologic rationale and has the potential to reduce the logistical complexity of hemorrhagic resuscitation. Although observational data from military and civilian settings strongly suggested advantages, randomized trials have thus far shown no benefit from whole blood over conventional blood component transfusions.

More data is coming. Randomized trials planned or in process for whole blood in trauma include the TROOP trial (U.S. trauma centers, n=1,100), MATIC-2 (in 1,000 U.S. children with traumatic hemorrhage), and a small RCT at Loma Linda.

References

Smith JE, Cardigan R, Sanderson E, et al. Prehospital Whole Blood in Traumatic Hemorrhage — a Randomized Controlled Trial. New England Journal of Medicine. Published online March 17, 2026. doi:https://doi.org/10.1056/nejmoa2516043

‌Guyette FX, Sperry JL. Prehospital low titer group O whole blood is feasible and safe: Results of a prospective randomized pilot trial. J Trauma Acute Care Surg. 2022 May 1;92(5):839-847. doi: 10.1097/TA.0000000000003551. Epub 2022 Jan 25. PMID: 35081595; PMCID: PMC9038638.

Hazelton JP, et al. Use of Cold-Stored Whole Blood is Associated With Improved Mortality in Hemostatic Resuscitation of Major Bleeding: A Multicenter Study. Ann Surg. 2022 Oct 1;276(4):579-588. doi: 10.1097/SLA.0000000000005603. Epub 2022 Jul 18. PMID: 35848743.

Brill JB, et al . Impact of Incorporating Whole Blood into Hemorrhagic Shock Resuscitation: Analysis of 1,377 Consecutive Trauma Patients Receiving Emergency-Release Uncrossmatched Blood Products. J Am Coll Surg. 2022 Apr 1;234(4):408-418. doi: 10.1097/XCS.0000000000000086. PMID: 35290259.

Dorken-Gallastegi A, et al . Whole Blood and Blood Component Resuscitation in Trauma: Interaction and Association With Mortality. Ann Surg. 2024 Dec 1;280(6):1014-1020. doi: 10.1097/SLA.0000000000006316. Epub 2024 May 6. PMID: 38708894; PMCID: PMC11538373.

Spinella PC, Perkins JG, Grathwohl KW, Beekley AC, Holcomb JB. Warm fresh whole blood is independently associated with improved survival for patients with combat-related traumatic injuries. J Trauma. 2009 Apr;66(4 Suppl):S69-76. doi: 10.1097/TA.0b013e31819d85fb. PMID: 19359973; PMCID: PMC3126655.

Gurney J, Staudt A, Cap A, Shackelford S, Mann-Salinas E, Le T, Nessen S, Spinella P. Improved survival in critically injured combat casualties treated with fresh whole blood by forward surgical teams in Afghanistan. Transfusion. 2020 Jun;60 Suppl 3:S180-S188. doi: 10.1111/trf.15767. Epub 2020 Jun 3. PMID: 32491216.

Nessen SC, et al . Fresh whole blood use by forward surgical teams in Afghanistan is associated with improved survival compared to component therapy without platelets. Transfusion. 2013 Jan;53 Suppl 1:107S-113S. doi: 10.1111/trf.12044. PMID: 23301962.

Torres CM, Kenzik KM, Saillant NN, Scantling DR, Sanchez SE, Brahmbhatt TS, Dechert TA, Sakran JV. Timing to First Whole Blood Transfusion and Survival Following Severe Hemorrhage in Trauma Patients. JAMA Surg. 2024 Apr 1;159(4):374-381. doi: 10.1001/jamasurg.2023.7178. Erratum in: JAMA Surg. 2024 Apr 1;159(4):470. doi: 10.1001/jamasurg.2024.0324. PMID: 38294820; PMCID: PMC10831629.

Hanna K, Bible L, Chehab M, Asmar S, Douglas M, Ditillo M, Castanon L, Tang A, Joseph B. Nationwide analysis of whole blood hemostatic resuscitation in civilian trauma. J Trauma Acute Care Surg. 2020 Aug;89(2):329-335. doi: 10.1097/TA.0000000000002753. PMID: 32744830.

Ibrahim W, Meza Monge K, Menzel J, et al. Whole-Blood vs Component Therapy in Adult Trauma: An Updated Systematic Review and Meta-Analysis. JAMA surgery. Published online November 2026:e260197. doi:https://doi.org/10.1001/jamasurg.2026.0197

‌

Please enjoy this next guest Real-World Boards question from Dr. Lloyd Tannenbaum, who authors the fantastic ECG Teaching Cases on Substack. —Ed.

You’re about halfway through your shift when a nurse comes to grab you. She tells you that Mr. Jackson in room 12 does not look well. She’d like you to come look at him.

As you walk to room 12, she briefs you on the patient: “Mr. Jackson is a 57-year-old male admitted to the ICU after a 4-vessel coronary artery bypass grafting. He is 72 hours out of surgery and had been doing well. He was making an uneventful recovery when he suddenly started to become very short of breath and complained of intense chest pain. He’s very pale and diaphoretic, so I asked one of the other nurses to stay in the room with him while I ran to go get you.”

You walk into the room and see an unwell patient who is now unresponsive. Looking at the monitor, this is what you see. You’re unable to find a pulse.

Cardiac arrest after recent cardiac surgery tends to be witnessed, monitored, and due to discrete, rapidly reversible causes like tamponade, bleeding/hypovolemia, graft-related ischemia, or pacing failure, as compared to undifferentiated in-hospital arrests.

For these reasons, post-cardiac surgery patients in arrest are often best managed with a more flexible and situation-driven approach, not rigidly tethered to the standard advanced cardiac life support (ACLS) protocols.

Cardiac advanced life support, or CALS, was brought to international attention by Drs. Dunning et al in 2009 in their landmark paper, “Guidelines for resuscitation in cardiac arrest after cardiac surgery,” which proposed a set of suggested modifications to standardized society guidelines.

They advocated for specific resuscitation practices in patients in cardiac arrest up to 10 days after open-heart surgery. (Note: In the USA, CALS is also known as Cardiac Surgical Unit Advanced Life Support or CSU-ALS. For simplicity, we’ll just be referring to it here as CALS.)

In patients who arrest after recent open cardiac surgery, who are being managed in ICUs specialized in cardiac post-surgical care, the authors argued for deviating from the standard Advanced Cardiac Life Support (ACLS) algorithm in specific ways:

• For ventricular fibrillation, asystole, or severe bradycardia, delay chest compressions until multiple attempts at defibrillation (in ventricular fibrillation) or pacing (in asystole or extreme bradycardia) have been performed.

• Avoid epinephrine, which can cause unsafe increases in blood pressure.

• Proceed to emergency re-sternotomy to definitively address potential causes for arrest (e.g., tamponade).

They argue for delaying chest compressions for defibrillation or pacing when reversible dysrhythmias are present:

• In ventricular fibrillation, they advise three sequential defibrillation attempts, then starting chest compressions if cardiac arrest persists.

• For asystole or pulseless bradycardia, they recommend pacing before starting chest compressions.

• For non-bradycardic pulseless electrical activity, or if these measures fail, chest compressions should be performed.

This advice assumes defibrillation or pacing is immediately available and will not delay CPR.

While chest compressions (external cardiac massage) may be necessary initially, they recommend that emergency re-sternotomy (i.e., within five minutes) be considered.

Re-sternotomy provides for direct access to the heart, for internal cardiac massage (which is much more effective than CPR), as well as identification and management of potential causes for the arrest (e.g., tamponade or hemorrhage).

Although controversial when published in 2009, some version of the CALS framework has become standard in modern cardiac ICUs, and is formally or indirectly endorsed by major professional societies (including AHA, in the “Special Circumstances” section of its ACLS protocols).

AHA’s document advises spending no more than 1 minute shocking, pacing, or determining if the patient is in true PEA. Then, if necessary, start external CPR, recognizing that this may cause harm to the heart while compressing a split sternum, while the rest of the team prepares to re-open the chest.

One of the most interesting (and controversial) topics is not giving adrenalin (epinephrine or epi) to these patients if they arrest. This is mostly based on the data for giving epi in a code not being particularly strong, and a letter to the editor by Dr. Webb describing a situation where 1mg of epi was given in accordance with ACLS protocol when a post-op CABG patient arrested from a tension pneumothorax. Once a chest tube was placed, ROSC was rapidly achieved; however, a large amount of blood started coming out of his chest drains. When the patient had an emergent re-sternotomy, they found that the graft had burst open, likely secondary to the epi causing a huge spike in blood pressure. Dr. Webb encouraged the European Resuscitation Council to consider post-cardiac surgery patients a special population and recommend against giving epi if they arrest. That was followed by a formal 2017 guideline advising that “full doses of epinephrine should not be routinely given [during post-cardiac surgery arrests] owing to the danger of extreme hypertension if a reversible cause is rapidly resolved.”

Post-Cardiac Surgery Arrests: Different Etiologies, Different Approach

The major point that this paper makes and that these guidelines are based on, is that most causes of cardiac arrest in this patient population are reversible and related to their cardiac surgery. Getting back into the chest as quickly as possible (emergent re-sternotomy) leads to better outcomes.

In 2017, the Society of Thoracic Surgeons wrote an Expert Consensus for the Resuscitation of Patients who Arrest After Cardiac Surgery that re-examined these guidelines. They built on the initial flowsheet provided by the initial guidelines:

Walking through this figure, let’s start with early defibrillation (the orange arrow). Assuming that the team can get the machine set up to defibrillate under a minute, external CPR is not indicated before shocking 3 times back-to-back if the rhythm doesn’t convert. Why 3 shocks? Well, the authors pooled data from 15 papers and found that the chance of successfully cardioverting out of a dysrhythmia goes from 78% after the first shock to 35% after the second to 14% after the third. After 3 unsuccessful shocks, external CPR and emergent re-sternotomy are indicated, as your chances of converting out of a dysrhythmia by defibrillating after 3 failed attempts are extremely low.

Now take a look at the green and yellow arrows. If the patient is arresting and the rhythm isn’t amenable to defibrillation, consider pacing if the patient has severe bradycardia or asystole.

If the patient is in asystole or severe bradycardia (yellow arrow) and pacing wires are available, set the pacer to DDD mode between 80-100 beats per minute at max output and attempt pacing.

If the patient is in PEA (green arrow) and being paced, briefly turn off the pacemaker to evaluate for fine ventricular fibrillation, which may be masked by the pacer. If ventricular fibrillation is present, immediate defibrillation is indicated.

This paper also reexamined epinephrine in the peri-arrest state and arrest state. If the senior physician decides to use epi, push-dose epinephrine, which is a lower concentration (0.01mg/ml=10mcg/ml), is recommended. This is prepared by withdrawing 1 mL from the code epinephrine (1 mg/10 mL syringe) and diluting it with 9 mL of normal saline. The code leader will decide how much epi to give, but doses between 50 and 300 mcg are what the authors consider reasonable.

CALS Recap

1. Ventricular fibrillation: CALS recommends three rapid sequential defibrillation attempts (when immediately available) before initiating chest compressions. ACLS also endorses this in their “Special Circumstances” section.

2. Asystole or severe bradycardia: CALS prioritizes immediate pacing (if available) before external cardiac massage (chest compressions). ACLS endorses this as well.

3. Refractory arrest / true Pulseless Electrical Activity: CALS advocates chest compressions while preparing for emergency re-sternotomy to address reversible surgical causes. ACLS agrees.

4. Epinephrine: CALS does not recommend routine early epinephrine administration. When indicated, the standard ACLS code-dose concentration (1 mg/10 mL) should not be administered as a 1 mg bolus. Instead, push-dose epinephrine, which is a lower concentration, (0.01mg/ml=10mcg/ml) has been recommended. This is prepared by withdrawing 1 mL from the code epinephrine (1 mg/10 mL syringe) and diluting it with 9 mL of normal saline. ACLS does not comment on epinephrine for cardiac surgery patients. In any case, the senior physician running the code should be the one deciding if epi should or should not be given.

For more ECG Teaching Cases, please visit:

ECG Teaching Cases

New case every 2 weeks diving into an interesting emergency cardiology topic!

Reflect to earn CME with Learner+

Sample reflection: I reviewed and reflected on the clinical category of cardiac arrest occurring after recent cardiac surgery, and the deviation from standard resuscitation practices that may be indicated and beneficial in these special circumstances.

References

Part 10: Adult and Pediatric Special Circumstances of Resuscitation: 2025 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2025. Cao D, Arens AM, Chow SL, et al.

Part 3: Adult Basic and Advanced Life Support: 2020 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2020. Panchal AR, Bartos JA, Cabañas JG, et al.

The Society of Thoracic Surgeons Expert Consensus for the Resuscitation of Patients Who Arrest After Cardiac Surgery. The Annals of Thoracic Surgery. 2017.

2020 EACTS/­ELSO/­STS/­AATS Expert Consensus on Post-Cardiotomy Extracorporeal Life Support in Adult Patients. The Journal of Thoracic and Cardiovascular Surgery. 2021. Lorusso R, Whitman G, Milojevic M, et al.

European Resuscitation Council Guidelines 2021: Cardiac Arrest in Special Circumstances. Resuscitation. 2021. Lott C, Truhlář A, Alfonzo A, et al.

Dunning J, Fabbri A, Kolh PH, Levine A, Lockowandt U, Mackay J, Pavie AJ, Strang T, Versteegh MI, Nashef SA; EACTS Clinical Guidelines Committee. Guideline for resuscitation in cardiac arrest after cardiac surgery. Eur J Cardiothorac Surg. 2009 Jul;36(1):3-28. doi: 10.1016/j.ejcts.2009.01.033. Epub 2009 Mar 17. PMID: 19297185.

Sepsis has an ontological problem—after hundreds of years, we still can’t coherently say what it is, and clinical intuition remains the primary criterion for its diagnosis.

The heterogeneity of sepsis and the shortcomings of diagnostic testing complicate clinical research and sap its rigor. Owing to these inherent limitations, there are very few high-quality clinical trials in sepsis.

Some commentators have been so outlandish as to suggest that sepsis care can (and even should) be distilled to providing antibiotics promptly, ensuring euvolemia, and maintaining perfusion with vasopressors, if needed, while pursuing source control and following culture data.

With more than 5,400 studies on sepsis published in the past five years, though, surely there have been important recent discoveries ready for dissemination from the halls of academia to bedside clinical care in the community?

The Surviving Sepsis Campaign is the premier group that has tasked itself with periodically wrangling this expanding corpus of errant and unruly information into a set of statements that convey an impression of continuous evidence-based progress in sepsis care. Since the Campaign’s inception in 2001, its panelists have regularly analyzed sepsis research and issued recommendations governing multiple aspects of care, whose authority is undiminished by a lack of strong supporting evidence.

The most recent tranche of guidelines, issued in March 2026, includes “129 statements, with 46 being new statements not previously addressed, and more definitive recommendations” than the 2021 iteration.

We’ll put these under the microscope in future posts, but here are a few of the major takeaways from the new guidance.

PulmCCM is not affiliated with the Surviving Sepsis Campaign or any specialty society.

Read more

Beginning in the 2010s, high-flow nasal cannula oxygen became widely used in acutely ill patients in developed countries, driven by clinicians’ perceptions and strong intuition that it delays or prevents the need for intubation in many patients.

There were good reasons to believe this.

You see (stay with me, now), in the context of supplemental oxygen therapy, the phrase “high-flow” means flow that is higher than some other flow, viz., that of standard oxygen therapy, whose flow is lower by comparison.

Recapping that: high-flow is high. Regular flow is lower (not high).

Wait: it gets more complicated. Flow is volume per unit of time; ergo, higher flow delivers a greater volume of a fluid (gaseous oxygen in this case) per unit of time, into the respiratory system of spontaneously breathing, acutely ill people in this instance. More flow. More oxygen. Oxygen good. Illness bad. Are you getting all this?

Tedious blogsplaining aside, HFNC O2’s higher flow rate conveys some experimentally demonstrated additional advantages over conventional nasal cannula or facemask oxygen.

High-flow nasal cannula oxygen can reliably deliver high FiO2 close to 100%, while also providing low-dose PEEP (~3-5 cm H2O), and washing out nasopharyngeal dead space with its high flow rate. Nonrebreather oxygen, by contrast, does not provide PEEP and is subject to room air entrainment into the breathing circuit, diminishing the delivered FiO2 to an unpredictable degree.

For a walk further into the weeds, read Jon’s excellent review here:

Although many physicians consider the benefits of HFNC O2 to be self-evident, it’s easy to imagine a counterfactual: perhaps, above some threshold of acute hypoxemia, the events of progressive respiratory failure, the need for mechanical ventilation, and mortality occur equally often, such that marginal improvements in oxygenation (e.g., from SpO2 85% to 90%) have no real influence on those outcomes.

You could even go Full Contrarian and wonder if by supporting oxygenation so well in deteriorating patients, HFNC O2 might inappropriately delay intubation and lead to worse outcomes.

If HFNC O2 does have benefits in acute respiratory failure, it would be helpful to understand their magnitude, which could inform the strength of the indications for HFNC O2 in specific clinical situations.

Considering the success and speed with which HFNC O2 conquered the field, there’s surprisingly little evidence to support its use. That said, the overall body of data points in the same direction as clinical intuition and common sense.

FLORALI and HOT-ER Trials

In the FLORALI trial (NEJM 2017), among 313 patients with acute hypoxemic respiratory failure without hypercapnia or cardiogenic pulmonary edema, the 106 who were randomized to HFNC O2 had about half the mortality at 90 days than those randomized to noninvasive ventilation or standard/facemask oxygen (statistically significant). They also had a lower intubation rate (38%) than those treated with NIV (50%) or standard oxygen (47%), although this did not reach statistical significance.

In HOT-ER (n~320), patients with hypoxemic respiratory failure in the ED randomized to HFNC required numerically fewer intubations within 24 hours compared to patients treated with standard oxygen, just missing significance (p=0.053).

Colombian RCT and Meta-Analysis

During the Covid-19 pandemic, it seemed clear that HFNC oxygen prevented intubations, and this intuition was supported by the results of a small multicenter RCT in Colombia (JAMA 2021), in which HFNC O2 reduced intubation rates by about one-third (34% vs 51%) in patients with Covid-related respiratory failure.

A meta-analysis of nine RCTs (n=2,093) also suggested that HFNC O2 reduced the need for intubation (relative risk 0.85), but noted a high risk of bias in the included trials. There was no signal that HFNC O2 reduced mortality (RR 0.94).

Some trials have also suggested a benefit of HFNC at preventing the need for reintubation after an episode of mechanical ventilation.

SOHO Trial

In March 2026, the large and rigorous SOHO trial, conducted at 42 French ICUs by the same group as FLORALI (Frat et al), randomized 1,110 patients with de novo respiratory failure (mostly due to pneumonia, without hypercarbia or cardiogenic pulmonary edema) to either HFNC O2 or standard oxygen therapy (e.g., nonrebreather mask).

The mortality rate at 28 days was precisely equal at 14.6% in each arm.

HFNC O2 patients had an absolute 6% reduced incidence for intubation within 28 days (42% vs 48%). This corresponds to a number needed to treat with HFNC O2 of 17 to prevent one intubation, in this patient group.

An excess of only about one in 40 patients treated with conventional oxygen had severe hypoxemia during intubation (SpO2 <80%). Adverse events (e.g., cardiac arrest during intubation and in the ICU) occurred at rates too low for valid comparisons.

Besides its size, SOHO had the strength of methodological rigor, which included objective criteria for intubation. This standardization should have reduced the risk for bias (e.g., unblinded clinicians intubating more often in standard oxygen groups in previous trials, inflating the apparent benefit of HFNC O2).

U.S. critical care society guidelines are largely silent on the use of high-flow nasal cannula oxygen. European guidelines endorse the use of HFNC O2 in a variety of settings.

Discussion

High-flow oxygen delivers, well, a higher flow of oxygen into the respiratory system than lower-flow conventional oxygen. Oxygen good. More oxygen better?

HFNC O2 intuitively should improve mortality over conventional oxygen, but doesn’t seem to.

It should prevent the need for intubation, and it does seem to—but only occasionally, in about one in 17 patients with de novo respiratory failure (without hypercarbia or cardiogenic pulmonary edema).

It’s hard to reconcile the huge reduction in 90-day mortality (and numeric intubation rates) with HFNC O2 in the first Frat et al trial (FLORALI) with the finding of zero difference between groups in SOHO and in pooled results of multiple other randomized trials … other than to simply believe SOHO over FLORALI.

But that take itself casts significant doubt on the supposed harms of noninvasive ventilation in de novo respiratory failure (without hypercarbia or pulmonary edema), a belief among clinicians which derives in large part from FLORALI.

Because of SOHO’s size, rigor, and recency reflecting current practice, the most plausible takeaway from the trial (in the context of the larger body of research) is that high-flow nasal cannula has marginal but meaningful benefits in preventing the need for intubation and mechanical ventilation.

High-flow nasal cannula oxygen’s current widespread use in developed countries as an effective, safe, and beneficial method of noninvasive respiratory support is fully justified.

References

Frat JP, Quenot JP, Guitton C, et al. High-Flow or Standard Oxygen in Acute Hypoxemic Respiratory Failure. New England Journal of Medicine. Published online March 17, 2026. doi:https://doi.org/10.1056/nejmoa2516087

‌Frat JP, Thille AW, Mercat A, et al. High-Flow Oxygen through Nasal Cannula in Acute Hypoxemic Respiratory Failure. New England Journal of Medicine. 2015;372(23):2185-2196. doi:https://doi.org/10.1056/nejmoa1503326

Ospina-Tascón GA, Calderón-Tapia LE, García AF, et al. Effect of High-Flow Oxygen Therapy vs Conventional Oxygen Therapy on Invasive Mechanical Ventilation and Clinical Recovery in Patients With Severe COVID-19: A Randomized Clinical Trial. JAMA. 2021;326(21):2161-2171. doi:https://doi.org/10.1001/jama.2021.20714

‌Rochwerg B, et al. High flow nasal cannula compared with conventional oxygen therapy for acute hypoxemic respiratory failure: a systematic review and meta-analysis. Intensive Care Med. 2019 May;45(5):563-572. doi: 10.1007/s00134-019-05590-5. Epub 2019 Mar 19. PMID: 30888444.

Drake MG. High-Flow Nasal Cannula Oxygen in Adults: An Evidence-based Assessment. Annals of the American Thoracic Society. 2018;15(2):145-155. doi:https://doi.org/10.1513/annalsats.201707-548fr

‌Jones PG, Kamona S, Doran O, Sawtell F, Wilsher M. Randomized Controlled Trial of Humidified High-Flow Nasal Oxygen for Acute Respiratory Distress in the Emergency Department: The HOT-ER Study. Respir Care. 2016 Mar;61(3):291-9. doi: 10.4187/respcare.04252. Epub 2015 Nov 17. PMID: 26577199.

Oczkowski S, Ergan B, Bos L, et al. ERS Clinical Practice Guidelines: high-flow nasal cannula in acute respiratory failure. European Respiratory Journal. 2021;59(4). doi:‌

A 57-year-old man presents to the emergency department with three days of fevers, chills, malaise, and worsening dyspnea. He has no history of heart failure.

He is placed on non-rebreather oxygen. Vitals 10 minutes later are:

On exam, work of breathing is elevated but appears sustainable. There are crackles and diminished breath sounds at the left lung base. Neck veins are not elevated. There is no edema.

White blood cell count is 16,000. Arterial blood gas shows pH 7.38, pCO2 30, paO2 55. Other labs are unremarkable.

Chest radiograph shows a left lower lobe consolidation. Ceftriaxone and azithromycin are infused.

Read more

In a long-ago age, before most of you were born (as doctors), there was a magical and hopeful time in critical care, when everyone but everyone believed that ARDS, and sepsis too, would soon be dramatically mitigated by pharmacologic treatments made possible by the impressive and regular basic science breakthroughs emerging from NIH-funded research.

Many compounds were instilled, infused, or sprayed into patients with lung infiltrates and severe hypoxemia receiving mechanical ventilation: aspirin, budesonide, beta-agonists (both IV and inhaled), corticosteroids, statins, N-acetylcysteine, nitric oxide, prostacyclin, ketoconazole, lisofylline, keratinocyte growth factor, stem cells, interferon, and surfactants, to name a lot of them.

Except for corticosteroids, which might have maybe helped and thus ended up on some-but-not-all critical care societies’ recommended treatments for ARDS, no medicine demonstrated significant benefits in mortality or other clinically important endpoints.

But there was another agent: sivelestat. An inhibitor of neutrophil elastase (aka leukocyte elastase) that protected twelve 1990s-vintage hamsters from a specific type of endotoxin-induced ARDS.

Neutrophil elastase has many effects in humans, some of which are destructive (e.g., direct tissue damage) and others likely beneficial (e.g., regulation of inflammation) during critical illness. Its role in ARDS is unclear, but there’s abundant circumstantial evidence to implicate it in at least some subtypes of acute lung injury/ARDS.

STRIVE Trial

In the early 2000s, at 105 ICUs in the U.S., Europe, and Australasia, 492 mechanically ventilated patients with ARDS were randomized to receive either sivelestat at a dose of 0.16 mg/kg/hr or placebo for up to 14 days of mechanical ventilation. Eli Lilly, makers of sivelestat, funded the trial.

The trial was stopped early (before target enrollment of n=620) because of the statistical near-impossibility of a reduction in ventilator-free days or 28-day mortality (the co-primary endpoints, which occurred virtually identically between groups).

More concerningly, at six months, there was significantly higher mortality observed in the sivelistat arm (40%) than placebo (31%; p=0.006).

Reported in 2004, the STRIVE trial seemed to end sivelistat’s prospects in the U.S., where the drug was never FDA-approved.

However, STRIVE left open the possibility that although sivelistat might be ineffective at treating ARDS, it might prevent lung injury, especially if given earlier.

Looking East: Sivelistat as Periop Prevention

Sivelistat had already received regulatory approval in 2002 in Japan, and later in South Korea and China, based on earlier low-quality studies showing improvements in lung function. The neutrophil elastase inhibitor has continued to be produced and marketed in Asia as Elaspol® by Ono Pharmaceuticals and its partners.

Two cohort studies conducted in China and published in 2023 and 2024 suggested that sivelistat administered preventively to cardiac bypass patients significantly reduced ARDS (to zero, vs 55% in propensity-matched controls in one study!).

A 2023 meta-analysis of 15 studies, including STRIVE, but mostly in Asia, many non-randomized, and at moderate risk of bias, concluded that sivelistat clearly improved mortality, ARDS, and ICU stays.

A randomized trial followed.

Nanjing Hospital sivelistat trial

At a single tertiary cardiac bypass surgery center (Nanjing Drum Tower Hospital) in China, 424 patients were randomized to receive sivelistat or placebo within 6 hours of ICU admission in 2024-2025. Rigorous trial processes of randomization, blinding, and outcome measurement were described in the manuscript.

The authors report that sivelestat roughly halved the incidence of postoperative ARDS (16.8% vs 31.2%) and significantly reduced 90-day all-cause mortality (1.1% vs 5.2%), compared with placebo. There was no increase in mortality at 180 days.

Cases of ARDS in the sivelestat arm that did occur were milder, and there were no severe cases. Sivelistat-treated patients also had lower rates of adjudicated pneumonias and lower levels of numerous inflammatory markers. The results were published in JAMA Network Open in 2026.

Discussion

Keep in mind, this is not your grandfather’s ARDS, nor your fellowship attending’s. Retrospective U.S.-based cohorts estimate the incidence of postoperative ARDS after cardiopulmonary bypass to be 1-5%, and 15% in the very highest-risk cohorts. Those represent the more clinically significant, unequivocal cases.

More recent prospective studies, especially out of Asia, commonly found ARDS rates to be much higher (20-30%), and this trial’s was among the very highest (31% in the placebo arm). The higher prospective incidences may be due to closer surveillance and stricter application of definitions, including the inclusion of patients on HFNO immediately post-extubation, for example (endorsed by the newest Berlin definitions).

The prophylactic postoperative infusions of sivelistat in the Nanjing Drum Tower Hospital trial can’t be compared with STRIVE, which delivered sivelistat to medical patients who had already developed ARDS.

That said, dramatic reported reductions in a condition like ARDS that has already proved to be highly resistant to intervention should raise skepticism, especially in a single-center trial. Triple the skepticism when it’s accompanied by an 80% mortality reduction. Add in a skepticism multiplier when it comes from a nation that has produced a disproportionate share of research findings that were later retracted.

There is nothing in the NDTH trial that suggests fraud or misconduct, and it is deeply unfair that the majority of honest, good-faith researchers in any country should labor under a cloud of skepticism based on their countrymen’s past behavior. But the truth is that neither we (nor the editors and reviewers at JAMA, for that matter) have any reliable way to vet the research coming out of China.

Suffice it to say that the finding that preventively-administered sivelistat dramatically reduces ARDS and mortality after cardiopulmonary bypass surgery would be a remarkable and practice-changing development—if later replicated in multicenter randomized trials conducted in the U.S., Europe, and Australasia.

Sivelistat is not approved in the U.S. today for any indication. At least one other randomized trial testing sivelistat further for the prevention of ARDS is in process (in China).

References

Pan T, Xu C, Wang YP, et al. Sivelestat and Incidence of Acute Respiratory Distress Syndrome After Cardiovascular Surgery. JAMA Network Open. 2026;9(3):e260390. doi:https://doi.org/10.1001/jamanetworkopen.2026.0390

‌Lee WL, Slutsky AS. Preventing ARDS, Not Treating It—Lessons From Sivelestat After Cardiopulmonary Bypass. JAMA Network Open. 2026;9(3):e261201. doi:https://doi.org/10.1001/jamanetworkopen.2026.1201

Neutrophil Elastase Inhibition in Acute Lung Injury: Results of the STRIVE Study. Critical Care Medicine. 2004. Zeiher BG, Artigas A, Vincent JL, et al.‌

Sivelestat for Septic Patients With Acute Respiratory Distress Syndrome: A Systematic Review and Meta-Analysis of a Deadly Duo. Frontiers in Medicine. 2025. Zheng WH, Hu YG, Yu DX, Huang HB.

Effect of Sivelestat Sodium in Patients With Acute Lung Injury or Acute Respiratory Distress Syndrome: A Meta-Analysis of Randomized Controlled Trials. BMC Pulmonary Medicine. 2017. Pu S, Wang D, Liu D, et al.

Therapeutic Effect and Mechanism of Sivelestat Sodium on Acute Lung Injury: A Randomized Controlled Trial. Medicine. 2025. Zhou Y, Chen G, Xu J, et al.

Neutrophil Elastase Inhibitor (Sivelestat) in the Treatment of Acute Respiratory Distress Syndrome Induced by COVID-19: A Multicenter Retrospective Cohort Study. Respiratory Research. 2025. Li Y, Zhao J, Wei J, et al.

From the 1990s through, let’s say, 2011 (when Xigris™ was removed from the market), sepsis research meant molecule-hunting. Drug companies backed hundreds of investigations led by the Indiana Joneses of academic medicine as they sought the Cure for Sepsis in the heart of darkness. With the level of brainpower and resources being deployed, a cure seemed to be just around the corner.

Dozens of distinct compounds were tested in more than 100 phase II/III clinical trials: afelimomab, eritoran, various anti-TNF monoclonal antibodies and receptor fusion protein, anti-endotoxin antibodies, and many more.

None of them ever showed a replicable mortality benefit in phase III trials. After Xigris’s™ withdrawal from the market, it seemed like the search for the sepsis cure had reached a dead end, or run out of road, or maybe its horse had become very tired. The metaphor doesn’t matter. What mattered was making sure science kept moving forward, and NIH’s annual sepsis research budget had somewhere to flow.

There was a ray of hope: in those negative trials, there were often intriguing suggestions of benefits in subgroups (e.g., those with elevated IL-6 or some other cytokine). Could the benefit in these subgroups have been diluted to non-significance in the larger pool of undifferentiated septic patients?

As the classic sign-off from a hundred thousand manuscripts goes, more research would be required.

Sepsis Phenotypes: Some Assembly Required

Sepsis produces tremendous variation and complexity in physiology and immune response, but also significant overlap between patients.

So the first pass at defining sepsis phenotypes consists of collecting massive amounts of data through high-throughput systems like transcriptomics, proteomics, metabolomics, as well as clinical data from RCTs, and commanding computers to find patterns in it.

Unsupervised learning algorithms mine datasets for statistical correlations between constellations of patient factors and outcomes. Next, the phenotype construct is applied to other available trial data or prospectively collected data from registries or new RCTs (“validated”). If the correlation persists in the new data set, a new phenotype is christened and imbued with meaning. Rarely, however, have phenotypes been prospectively evaluated in clinical trials, largely because the methods to identify them are so cumbersome.

Broadly, sepsis phenotypes have been constructed from clinical or immunologic variables (although this choice is arbitrary and convenience-based, and should disappear as research methods evolve).

Clinical/Physiology Sepsis Phenotypes

The prototype for “clinical phenotypes” of sepsis is an analysis that was dubbed SENECA (JAMA 2019). Among 20,189 patients meeting Sepsis-3 criteria at 12 UPenn hospitals, 29 clinical variables (vitals, labs, vasopressor use, etc) were identified and fed to machine learning tools, which found clusters of the variables that could be stratified by mortality. Greek letters were added as a garnish.

Voilá, the apparent discovery of four clinical phenotypes of sepsis: α, β, γ, and δ. Mortality rises from 5% to 40% from alpha to delta. Gammas have more inflammation and lung dysfunction, while deltas have more liver dysfunction, shock, and mortality.

You may wonder what this adds to clinical gestalt (i.e., shocky patients with DIC die more often), or whether the patterns reflect any real categories or are just assembled by a computer largely out of epiphenomena and chance associations, or question how the schema could be used either for research or clinical purposes.

Clinical phenotypes did not “scale”, but a vast new research enterprise seeks to identify sepsis phenotypes tied to more fundamentally causative factor(s) at the biochemical or immunologic levels.

Biochemical/Immunologic Sepsis Phenotypes (Endotypes)

While clinical sepsis phenotypes don’t claim to be more than hypothesis-generating or as potential filters for clinical trial enrollment, this category of sepsis phenotyping seeks to identify subsets of patients with distinctly different operative immunologic processes.

These phenotypes get their own name: endotypes, to emphasize that there is Something Inside that is producing the observed manifestations, a Something we might precisely identify and influence for the patient’s benefit. (Using endotype also ditches the baggage of the word phenotype as used in phrenology, eugenics, etc.)

And investigators really are measuring something inside, a whole lot of it. Endotypes are proposed/constructed out of the patterns computers discover inside staggeringly complex “multi-omics” combinatory datasets that can include the expression of 20,000 genes, thousands of proteins, and metabolites per patient, each of which evolves over hours during sepsis.

Human researchers then work with the computer-identified patterns, each of which may implicate one known sepsis pathway over another, or suggest potential biochemical targets.

For example, two sepsis patients might both be hypotensive with elevated lactate, but one has sepsis driven primarily by hyperinflammation (an endotype detectable by elevated IL-6, e.g.) while the other by immunosuppression (an endotype which could be identified by low HLA-DR expression on monocytes on flow cytometry, e.g.).

If these two patients could be quickly differentiated at presentation, the argument goes, steroids could be given to the hyperinflammatory patient and withheld from the immunosuppressed patient, tailoring treatment beneficially to each’s need. (Keep in mind, this is all theoretical, today.)

Here’s just a slice of the research in sepsis phenotypes-slash-endotypes:

SRS and MARS

Unsupervised machine learning on transcriptomics signatures (gene expression) of peripheral blood leukocytes in patients with sepsis led Davenport et al and Scicluna et al to propose distinct endotypes of sepsis, with variable immune response patterns.

Davenport et al’s group proposed the SRS1 and SRS2 endotypes:

• Immunosuppressed (SRS1) pattern: T-cell exhaustion, downregulation of HLA class II

• Hyper-inflammatory/immune-competent (SRS2), with more normal inflammatory signaling

SRS1 (immune-suppressed) had higher mortality than the hyperinflammatory SRS2 group, and this was replicated in a cohort of 106 patients.

Seven genes were implicated in the differential gene expression between clusters. This 7-gene screen was then applied post hoc to data from 176 patients enrolled in the VANISH trial (JAMA 2016), suggesting that hydrocortisone may have harmed patients with immune-competent sepsis: hydrocortisone use was associated with an odds ratio of ~8 for increased mortality in those with the immunocompetent/hyperinflammatory SRS2 phenotype.

However, the trail stopped there. A bedside PCR test for the 7-gene expression profile has never been tested prospectively, and no test is commercially available.

A separate group, the MARS consortium identified four patterns (MARS 1 through 4) using expression from 140 genes: an immune competent/adaptive, immunosuppressed, hyperinflammatory, and interferon-gamma driven response to sepsis.

Subsequent groups (e.g., Chenoweth et al) have more or less replicated similar broadly sketched patterns, often with nuances or sub-subtypes added. This research has thus far defied systematization because of the protean variability in patients and their illness states, the differences in biological samples, and methodologies between research labs, among other factors.

Macrophage-activation-like syndrome (MALS) and “immunoparalysis”

An uncommon but particularly lethal observed sepsis pattern is macrophage activation-like syndrome (MALS), which produces cytokine storm predominantly mediated by IL-1, similar to hemophagocytic lymphohistiocytosis (HLH). Only about 3% of sepsis patients display features of MALS, but they have >60% mortality.

The ImmunoSep trial represents the single instance to date in which investigators have selected patients based on prospectively defined sepsis endotypes, and applied targeted interventions in randomized, blinded (partially) fashion.

MALS patients were identified as those with sepsis criteria and very high ferritin levels (>4420 ng/mL) and were randomized to receive an intervention (IL-1 blockade with anakinra vs placebo).

Patients in this trial with a different endotype, so-called immunoparalysis (<5000 human leukocyte antigen DR receptors on CD45/CD14 monocytes, as computed by flow cytometric analysis) could receive gamma-interferon or placebo.

Organ dysfunction (SOFA score) was reduced in the intervention groups, but 90-day mortality was virtually identical at ~68%.

And Proteomics, Cytokinomics, Metabolomics, (Freakonomics?) … All Together Now

The same essential move—perform a tricorder scan on the patient, collect an enormous volume of data, and feed it into a pattern-finding computer algorithm—can be performed with any high-throughput stream of biologic data.

This has produced claims that there are four “proteomic” subtypes, three “cytokine signature clusters,” at least two flow cytometry subtypes, and three “metabolomic subtypes” in sepsis, according to just a few recent publications.

The next frontier: feeding all these various streams (-omics) into one computationally gargantuan analysis (multi-omics). This is in its relative infancy.

There’s a sense that after the identification ten years ago of a handful of immunologic patterns, or endotypes, progress has slowed. There hasn’t been a process of continual refinement, but rather a rhyme-and-repeat vibe, describing the same associations in slight variations or riffs.

This is highly reminiscent of the early complex-genomics era, when a firehose of associations between SNPs and disease states seemed to show that geneticists were just about to crack this whole human disease thing wide open. Eventually, it became clear they were just reporting on the emissions, not understanding the engine, never mind showing us how to fix it.

Ontology Before Epistemology

The high-throughput technologies and computational processes that create the sepsis phenotypes are indeed impressive. They provide near-miraculous X-ray vision into the real-time immunologic and physiologic processes unfolding in this notoriously heterogeneous syndrome, identifying real, possibly vital differences between patients that can’t be detected at the bedside.

It remains to be seen, though, whether any of these suddenly visible and astoundingly complex pathophysiologic phenomena are modifiable to patients’ benefit (beyond what we already routinely do).

It’s possible (and in fact likely) that endotypes mistake phenomenology for mechanisms: endotypes can be real statistical clusters of phenomena, and also have no practical use.

The real answers most certainly lie deeper, and are beyond the capacities or the role of this newsletter to ascertain. The most intriguing thing I read (or skimmed) was also the last: this 2025 Nature Medicine SUBSPACE consortium paper, which plumbs the existing datasets to propose a consensus immune-dysregulation framework in critical illness more broadly, drawing a throughline between sepsis, ARDS, and trauma.

Although high-throughput techniques won’t hand us the Unified Theory of Sepsis anytime soon, they currently seem like the only shot we have of wrestling its heterogeneity complexity into more orderly and compliant shapes. This would at the very least allow for more rationally designed clinical trials like ImmunoSep.

The new catchphrase is “precision medicine,” but that label seems a bit premature: first, we need to see the “medicine” part in action.

References

Moore AR, Zheng H, Ganesan A, et al. A consensus immune dysregulation framework for sepsis and critical illnesses. Nature Medicine. Published online September 30, 2025:1-13. doi:https://doi.org/10.1038/s41591-025-03956-5

‌Giamarellos-Bourboulis EJ, Kotsaki A, Kotsamidi I, et al. Precision Immunotherapy to Improve Sepsis Outcomes: The ImmunoSep Randomized Clinical Trial. JAMA. Published online August 2025:e2524175. doi:https://doi.org/10.1001/jama.2025.24175

Classification of Patients With Sepsis According to Blood Genomic Endotype: A Prospective Cohort Study. The Lancet. Respiratory Medicine. 2017. Scicluna BP, van Vught LA, Zwinderman AH, et al.

Derivation, Validation, and Potential Treatment Implications of Novel Clinical Phenotypes for Sepsis. The Journal of the American Medical Association. 2019. Seymour CW, Kennedy JN, Wang S, et al.

Predicting Sepsis Severity at First Clinical Presentation: The Role of Endotypes and Mechanistic Signatures. EBioMedicine. 2021. Baghela A, Pena OM, Lee AH, et al.

The Immune Landscape of Sepsis and Using Immune Clusters for Identifying Sepsis Endotypes. Frontiers in Immunology. 2023. Tang G, Luo Y, Song H, et al.

Classification of Patients With Sepsis According to Immune Cell Characteristics: A Bioinformatic Analysis of Two Cohort Studies. Frontiers in Medicine. 2020. Zhang S, Wu Z, Chang W, et al.

Defining Critical Illness Using Immunological Endotypes in Patients With and Without Sepsis: A Cohort Study. Critical Care. 2023. Balch JA, Chen UI, Liesenfeld O, et al.

A Transcriptomic Classifier Model Identifies High-Risk Endotypes in a Prospective Study of Sepsis in Uganda. Critical Care Medicine. 2024. Cummings MJ, Bakamutumaho B, Tomoiaga AS, et al.

Profiling the Dysregulated Immune Response in Sepsis: Overcoming Challenges to Achieve the Goal of Precision Medicine. The Lancet. Respiratory Medicine. 2024. Cajander S, Kox M, Scicluna BP, et al.

Sepsis Phenotypes, Subphenotypes, and Endotypes: Are They Ready for Bedside Care?. Current Opinion in Critical Care. 2024. Scherger SJ, Kalil AC.

Through its periodically updated guideline statements, the American College of Chest Physicians has traditionally defined standard care for acute pulmonary embolism. (PulmCCM is not affiliated with any specialty society.)

ACCP’s most recent 2021 guidelines were quite conservative, recommending anticoagulation (low molecular weight heparin) as a standard therapy, and endorsing systemic thrombolysis (e.g., tPA or TNK) for PE with hypotension.

Catheter-directed thrombolysis and mechanical thrombectomy were barely mentioned and not endorsed, owing to the minimal evidence base comparing these treatments to the reference standard of anticoagulation.

This reticence was growing disconnected from clinical practice, where catheter-based treatments are being increasingly performed for many, if not most, significant PEs at some centers.

In 2026, the American Heart Association, ACCP, and seven other specialty societies relieved this tension by issuing a new guideline that recommends catheter-based therapies as reasonable treatments for virtually all patients requiring hospitalization.

The guidance goes further, designating both CDT and MT for the first time as acceptable substitutes for systemic thrombolysis in hemodynamically unstable patients with PE.

And it goes further than that, informally endorsing mechanical thrombectomy over catheter-directed thrombolysis for its perceived advantages.

The only indication reserved solely for systemic thrombolysis was as rescue therapy in patients with cardiac arrest or cardiogenic shock refractory to vasopressors.

Why the abrupt change? Did a tsunami of evidence arrive in the past five years, washing away any vestige of doubt as to the benefits of invasive therapies for pulmonary embolism?

Let’s take a look. And be sure to reach all the way to the bottom of the cereal box to get a special secret decoder ring to let you read the hidden messages embedded in the guidelines.

Patients in shock (E1)

Read more

A 33-year-old man experienced a cardiac arrest after an opioid overdose. He is obese (BMI 45) but has no known medical history. He was resuscitated but experienced anoxic encephalopathy. Four days later, he is unresponsive to pain and has absent brainstem reflexes. Sodium is 148 mmol/L, BUN is 48 mg/dL (2x ULN), creatinine is 2.1 mg/dL (2x ULN), serum bicarbonate is 29 meq/L (=ULN), pH is 7.44, pCO2 40 mmHg, paO2 120 mmHg. Fentanyl infusion was discontinued 48 hours ago. Mean arterial pressure is 68 mmHg and pulse 108/min.

After preoxygenation with FIO2 1.0 for ten minutes, an apnea test is performed by changing the ventilator settings to CPAP mode (pressure support 8 cm H2O) with 100% FIO2. Mean arterial pressure is maintained between 65 and 70 mmHg throughout, on low-dose norepinephrine.

There are no respiratory efforts observed. After five minutes, arterial blood gas shows pH 7.31 and pCO2 of 61 mmHg (a 21 mmHg increase).

The organ procurement organization representative has been nearly omnipresent in the ICU, and makes a sad face, but you (an amateur poker player) notice her pupils dilate.

Death by neurologic criteria (brain death) was a new medical-legal concept created in the 1960s in response to two new simultaneous societal phenomena: neurologically devastated patients could be physiologically maintained on ventilators for prolonged periods, and emerging organ transplantation programs (and their growing lists of desperate patients) faced a scarcity of viable organs. The new construct allowed for the legal and ethical removal of organs from physiologically “alive” persons who, upon meeting stringent objective criteria, would henceforth be designated as neurologically and therefore legally “dead”.

After various medical, sociopolitical, and legal controversies regarding DNC arose in the early 21st century, the American Academy of Neurology, other U.S. societies, and the World Brain Death Project embarked on a project to shore up the concept of brain death, standardize its diagnosis, and bolster public trust.

The effort ran into challenges, many of which derive from the variability of human physiology, the persistence of edge cases, and the limitations of relatively crude bedside tests to infallibly guarantee “the irreversible cessation of all functions of the entire brain, including the brainstem” (the legal standard) in every patient with the necessary 100% sensitivity and specificity.

The AAN, SCCM, and pediatric societies nonetheless produced an updated guideline in 2023 that addressed many of the controversies and borderline cases of brain death.

Let’s examine how these complicate just two of the criteria for BD/DNC declaration, as applied to the case in the above board question.

Correction of Confounders

Confounding factors for encephalopathy must be corrected before declaring DNC/brain death. But most critically ill patients present with or develop confounding factors for encephalopathy. Which abnormalities are considered non-barriers to the determination of brain death, and how abnormal can they be? This is often the most vexing challenge in making the diagnosis of brain death with the required certainty.

The AAN guideline specifies some of these conditions, which include:

• Hypotension: Systolic blood pressure must be ≥100 mmHg, and mean arterial pressure ≥75 mm Hg (with vasopressor use if needed). Patients with chronic high or low blood pressure should have their BP maintained near that mark.

• Hypothermia: At least 24 hours is required after rewarming to 36°C before a BD/DNC evaluation, if core body temperature has been ≤35.5°C. Yet brain death has been demonstrated to be reversible more than 24 hours after normothermia has been restored.

• Sedation: The evaluation should be delayed by at least 5 half-lives (e.g., fentanyl: <7 hours), while also factoring in liver or renal dysfunction, body mass index, age, and slowed metabolism from hypothermia. For example, in the elderly, fentanyl’s half-life is estimated to be five times as long as a young person’s. This could imply waiting longer than a week before brain death evaluation.

• Metabolic/lab abnormalities: the panel specifies numerous lab values to be corrected before BD/DNC evaluation: sodium should be 130-160; glucose 70-300; pH 7.3 - 7.5; BUN <75; etc.

• This list is not exhaustive, though it is exhausting: see the documents for even more caveats and cautions.

Apnea Testing

A complete guide to apnea testing can be found here. Some important points from the 2023 guideline include:

• Patients may remain on the ventilator circuit in CPAP mode during apnea testing; the traditional method of instilling oxygen through a nasal cannula positioned above the carina is optional.

• Although “8 to 10 minutes” is advised, an apnea test can be stopped earlier if the criteria of pCO2 ≥60 mmHg and ≥20 mmHg increase over baseline and pH <7.30 are met (e.g., on serial ABGs).

• The pre-test ABG should be collected after preoxygenation (not before, as in the stem).

• The apnea test must be aborted if at any point systolic BP falls below 100 mmHg, or MAP <75 mm Hg (with or without vasopressors).

• For patients with baseline hypercarbia, the pCO2 increase must be ≥20 mmHg above their “known chronic elevated premorbid baseline level” (if there is a single known level!), with pH< 7.30.

• For those with suspected baseline hypercarbia (like the patient in the stem), they must meet the usual criteria of pCO2 ≥60 mmHg and ≥20 mmHg increase over baseline and pH <7.30, and undergo ancillary testing (EEG, 4-vessel cerebral angiogram, radionuclide perfusion scintigraphy, or transcranial doppler ultrasonography).

EEG, CT angiogram, and MR angiogram should not be used as ancillary tests, as they have not been validated for this purpose, according to the panel.

Transcranial doppler and 4-vessel angiogram had very high specificity (low false positives), the most important test characteristic, in this review.

As we reviewed here, qualitative brainstem CT perfusion did not perform well in prospective testing:

Like many aspects of medicine, the concept of brain death seemed simple in an earlier era. Over decades, an evolving understanding has revealed cracks in the façade of its objective certainty, and of our ability to achieve perfection in its identification.

While most cases of death by neurologic criteria are clear-cut, clinicians should maintain extreme caution, a high vigilance for edge cases, and a stubborn resistance to pressure from financially conflicted actors, to avoid becoming implicated in near-catastrophic situations like this one.

Reflect to earn CME with Learner+

Sample reflection: I reviewed and reflected on the evolving understanding of death by neurological criteria (brain death), its increasingly complex definition and diagnosis, and the medical and legal implications for clinical practice.

References

Greer DM, Kirschen MP, Lewis A, et al. Pediatric and Adult Brain Death/Death by Neurologic Criteria Consensus Guideline: Report of the AAN Guidelines Subcommittee, AAP, CNS, and SCCM. Neurology. 2023;101(24). doi:https://doi.org/10.1212/WNL.0000000000207740

Lewis A, Kirschen MP, Greer D. The 2023 AAN/AAP/CNS/SCCM Pediatric and Adult Brain Death/Death by Neurologic Criteria Consensus Practice Guideline: A Comparison With the 2010 and 2011 Guidelines. Neurology: Clinical Practice. 2023;13(6). doi:https://doi.org/10.1212/CPJ.0000000000200189

‌Webb AC, Samuels OB. Reversible brain death after cardiopulmonary arrest and induced hypothermia. Crit Care Med. 2011 Jun;39(6):1538-42. doi: 10.1097/CCM.0b013e3182186687. PMID: 21494112.

Diagnostic accuracy of ancillary tests for death by neurologic criteria: a systematic review and meta-analysis. Can J Anaest 2023