No survivors over 65 after choking arrest — and the LMA fails in it; plus DCD hearts at five years, a heart transplant volume threshold of 31, and donor lungs that recover

Four items, every one published yesterday, all with deposited abstracts read in full. The index is healthy (2,329 records for 14 September, 568 so far today). Three of the four come from a transplant-heavy Annals of Thoracic Surgery issue that landed the day after this archive closed the ISHLT gap — useful timing.

1. Choking arrest — the laryngeal mask fails, and nobody over 65 survived

Journal · Published: Resuscitation, 14 September 2026 — published yesterday Delardes, Kennett, Ralph, Schwarz, Dantanarayana, McManamny, Nehme

Retrospective analysis of every EMS-attended out-of-hospital cardiac arrest in Victoria, Australia, 1 July 2020 to 30 June 2024, from the Victorian Ambulance Cardiac Arrest Registry, extracting the sequence, type and effectiveness of interventions used to relieve obstruction.

Of 29,684 out-of-hospital cardiac arrests, 223 (0.8%) were due to foreign body airway obstruction.

• Median age 77. 56% in private residences, 30% in residential care homes
• 16% of patients received no bystander intervention at all
• Most common EMS interventions: CPR 84.2%, Magill’s forceps 58.6%, suction 54.7%
• Magill’s forceps and tracheal intubation were the most effective for establishing ventilation — each 73% success when used
• Laryngeal mask airway devices were used in 67 patients (35%) but were effective in 7%
• Foreign body removal achieved in 165 (74.0%); ROSC in 107 (48.0%); survival to hospital discharge in 8 (3.6%)
• Among the 165 patients aged over 65, there were no survivors

Interpretation. This is the most practically alarming paper this archive has carried, and two of its findings should change what you reach for.

The laryngeal mask was used in a third of these patients and worked in seven percent. That is a near-complete failure of the device in this setting, and the mechanism is obvious once stated: an LMA sits above the larynx and ventilates through the obstruction it cannot remove. It is the reflex airway device for a periarrest patient, it is what is in most airway bags, and in choking arrest it is the wrong tool. Magill’s forceps and direct laryngoscopy — the tools that let you see and remove the object — each succeeded 73% of the time. For an anaesthetist this is not a surprising physiological claim; the surprise is how often the LMA was reached for anyway.

“Among patients aged over 65 years, there were no survivors” — none, out of 165. Overall survival was 3.6%, so the eight survivors were all younger. That number sits directly against the 74% foreign body removal rate and 48% ROSC rate: the obstruction was relieved in three quarters of cases and circulation returned in half, and almost nobody left hospital. The hypoxic insult is complete long before the airway is cleared.

The authors name the implication themselves — “assessment of futility” — and it deserves to be said carefully. This is a single-state registry, 223 patients, and an unwitnessed choking arrest in a care home is a different event from a witnessed one at a dinner table; the abstract does not stratify by witnessed status or downtime, which is exactly the stratification that would tell you whether “over 65” is really “long downtime.” I would not read this as grounds for withholding resuscitation from an older choking patient. I would read it as grounds for knowing that the odds after arrest has occurred are close to zero, which matters for how long you continue and for what you tell the family.

The number that is actually actionable is upstream: 16% received no bystander intervention, and 30% of events were in residential care homes. In a care home, trained staff are present. Back blows and abdominal thrusts before EMS arrival are the only intervention in this entire chain that happens early enough to matter.

This lands directly on the ERC 2025 Special Circumstances chapter backfilled yesterday — choking is one of the circumstances that chapter covers, and its abstract gave no content. This paper is the evidence that chapter ought to be reflecting, and the LMA finding in particular is the kind of device-specific detail guidelines rarely state plainly.

Read the paper · PMID 42735815

2. DCD heart transplantation — 19,432 transplants, and five-year survival is equivalent

Journal · Published: The Annals of Thoracic Surgery, 14 September 2026 — published yesterday Cho, Kibel, Raina, Lander, Hadi, Kassis-George, Alpert, Ranganathan, Aldweib, Lee, Tsukashita

UNOS registry, adult heart transplant recipients December 2019 to December 2025, stratified by donation after circulatory death (DCD) versus donation after brain death (DBD). Severe primary graft dysfunction captured September 2023 onward. Kaplan-Meier five-year survival.

19,432 adult heart transplants; 2,588 (13.3%) from DCD donors.

• DCD recipients were less likely to be on ECMO pre-transplant: 3.9% vs 8.1%, P < 0.001
• DCD donors were younger: 33 vs 34 years, P < 0.001
• Severe primary graft dysfunction: 10.3% vs 5.3%, P < 0.001 — nearly double
• 30-day mortality: 2.8% vs 2.7%, P = 0.80
• Five-year survival: 77.3% vs 78.3%, P = 0.90

Interpretation. The headline is genuinely reassuring and the caveat is genuinely important, and they are separable.

Nearly double the severe primary graft dysfunction, and identical survival at five years. That is the finding, and it is the one a cardiac anaesthetist should carry into the room: the DCD heart is more likely to need mechanical support in the first hours and no more likely to fail the patient in the first five years. Severe PGD at 10.3% means roughly one in ten DCD recipients will need VA-ECMO or equivalent immediately post-bypass — a planning number, not a prognostic one.

The confounding runs in the direction that should make you cautious about the equivalence. DCD recipients were less likely to be on ECMO before transplant (3.9% vs 8.1%) — that is, they were less sick going in. Programmes select stabler recipients for DCD hearts, which is rational and also means the two arms are not exchangeable. A registry comparison showing equivalent survival in a healthier DCD cohort is weaker evidence than equivalent survival in matched cohorts would be. The one-year-younger donors are trivial; the ECMO difference is not.

“These findings provide reassurance and support continued expansion” is the authors’ conclusion and it is fair, with that caveat attached. DCD now supplies 13.3% of US adult heart transplants — a substantial share built in six years, and the reason the donor pool has grown at all.

The connection to yesterday is direct: the ISHLT perioperative ECLS consensus exists precisely for the patients in that 10.3%.

Read the paper · PMID 42735884

3. Institutional volume in heart transplantation — the threshold has moved to 31 a year

Journal · Published: The Annals of Thoracic Surgery, 14 September 2026 — published yesterday Roa-Vidal, Sollie, Haran, Zhang, Welch, Inampudi, Kilic

UNOS registry, adult isolated heart transplant, 18 October 2018 to 31 December 2023 — that is, entirely after the 2018 allocation change, which is the paper’s reason for existing. Patient-level tertiles by centre volume; Cox regression; sequential ROC and loess smoothing to estimate a threshold.

139 centres: 92 low-volume, 31 moderate, 16 high-volume.

• 90-day survival 94.3% vs 95.9% (P = 0.0005); 1-year 90.5% vs 93.3% (P < 0.0001); 3-year 86.5% vs 89.5% (P < 0.0001) — low versus high volume
• Risk-adjusted, low-volume transplant predicted 90-day mortality HR 1.48 (1.23–1.78) and 1-year mortality HR 1.51 (1.30–1.74)
• Modelled continuously: HR 0.995 (0.992–0.998) per unit volume increase, P = 0.001
• Inflection point for improved 90-day and 1-year survival: 31 heart transplants annually
• The authors’ own caution: interpret the threshold “cautiously as a reference point and not a discrete optimal threshold due to the continuous association”

Interpretation. Two things, and the second matters more than the number everyone will quote.

The threshold has risen. Historical volume-outcome work in heart transplantation put the inflection lower; this analysis, in the post-2018-allocation era, puts it at 31 a year — and only 16 of 139 US centres are in the high-volume tertile. The allocation change redistributed sicker patients and made the operation harder, which is a plausible reason for a higher bar.

The authors’ caveat is the honest part and will be ignored. “A reference point and not a discrete optimal threshold due to the continuous association” — the continuous model gives HR 0.995 per additional transplant, a smooth gradient with no step in it. A “31 transplant threshold” is an artefact of fitting a breakpoint to a smooth curve, and it will nonetheless be quoted as a standard, because thresholds are quotable and gradients are not. The real finding is that more is better, continuously, with no safe floor.

And the usual volume-outcome confounding applies with force: high-volume centres differ in referral patterns, ICU staffing, mechanical support programmes and the option to decline a marginal organ. HR 1.48 for 90-day mortality at low-volume centres is a statement about institutions, and institutions differ in more ways than volume.

It sits beside the STS-ACSD reoperation risk model sent on 11 September, where an institutional model built at a high-volume centre systematically underestimated national risk because that centre operated on a 5.4%-urgent population against a 26.6%-urgent national one. Same structural fact from the other side: what a high-volume centre does and what it sees are both different.

Read the paper · PMID 42735887

4. Donor oxygenation trajectory and lung procurement — low PaO₂ is not a reason to decline

Journal · Published: The Annals of Thoracic Surgery, 14 September 2026 — published yesterday Bai, Yan, Liu, Yang, Delhi, Chang, Farahnak, Joseph, Hamilton, Heiden, Witt, Guillamet, Byers, Marklin, Nicely, Harmon, Hartwig, Nava, Patterson, Meyers, Kozower, Kreisel, Puri (Washington University, with Duke)

4,854 brain-dead donors from three US organ procurement organisations, January 2014 to June 2020, with their lung recipients where applicable. Donors stratified by initial PaO₂ then by final PaO₂, cutoff 300 mmHg. Multivariable regression for procurement, Cox models for recipient survival.

• Over 75% of donors had a low initial PaO₂ (3,702/4,854)
• Overall lung procurement rate 32.4% (1,574/4,854); 40.0% in the high-initial group vs 30.1% in the low-initial group, P < 0.001
• In the low-initial group, 1,380 (37.3%) reached a high final PaO₂ — and their procurement rate was 67.9%. Associated with recovery: frequent bronchoscopy, inotropes, and antibiotics
• In the high-initial group, 615 (53.4%) declined to a low final PaO₂; shorter length of stay was associated with procurement in that group
• Among recipients of lungs from donors with a high final PaO₂, survival was the same regardless of initial donor PaO₂ — aHR 0.96 (0.80–1.15)

Interpretation. This is a donor-management paper with an unusually clean message, and the key comparison is the last one.

A donor lung that starts poorly and recovers performs like one that never deteriorated. Survival aHR 0.96, confidence interval comfortably spanning 1 — the initial PaO₂ carries no residual signal once the final value is high. That converts a widely used screening variable into what it actually is: a snapshot, not a verdict. Over three quarters of donors in this cohort started below the threshold, and more than a third of those recovered to above it, with a procurement rate more than double the group average.

The recovery is associated with things the ICU team does — bronchoscopy, inotropes, antibiotics. That is an observational association and not a trial, and it is obviously prone to confounding by prognosis: donors who look recoverable get worked on. But the combination is mechanistically coherent — atelectasis and secretions are reversible causes of a low PaO₂ in a brain-dead donor, and bronchoscopy addresses both.

The neglected half is the 53.4% of good-looking donors who deteriorated. More than half of donors with a high initial PaO₂ fell below 300 by the end, which is a much larger attrition than most would guess and reframes the accepted donor as needing active protection, not just acceptance. The finding that shorter length of stay was associated with procurement in that group says the same thing in the other direction: time in the ICU is when good lungs are lost.

It is the direct companion to the two-surgeon donor review paper sent on 2 September, which found a second look at declined lungs added 8.6% transplant volume. One paper says look again at the lungs you declined; this one says the reason you declined them may have been reversible.

Read the paper · PMID 42735885

---

Notes for the next run

• No abstract deposited: Annals of Thoracic Surgery, 14 September — “The Role of Guideline-Directed Medical Therapy on Outcomes after Coronary Artery Bypass Grafting” (Ganduboina, Dell’Aquilla, Iribarne, DOI 10.1016/j.athoracsur.2026.09.003, PMID 42735886). Almost certainly an invited commentary rather than original data. Directly on the CABG thread this archive has been following since the IACTS position statement — retry, or it needs a PDF. Not characterised from its title.
• Also seen 14 September, worth a slot: Resuscitation on structured diagnostic strategies for psychological problems in cardiac arrest survivors (a named knowledge gap, and it pairs with the post-intensive care and driving-resumption papers already logged); Shock on admission whole-blood transcriptomics showing a neutrophil-predominant systemic immune response in acute traumatic brain injury; Annals of Thoracic Surgery on the heart as a unit — pericardial interactions in constrictive pericarditis and pericardiectomy; and a JTCVS commentary on dynamic risk assessment in thoracic endovascular aortic repair, which is relevant to the EACTS/STS aortic organ guidelines gap but is a commentary.
• Still blocked, fifth attempt: the Annals of Thoracic Surgery pulsatility commentary (PMID 42716270). Adding a rule for this archive: a commentary or editorial in Annals of Thoracic Surgery essentially never deposits an abstract — four now in ten days. Recognise the pattern early and do not spend retries on them.