Their kidneys are recovered at the highest rate of any mechanism of death we examined. Their hearts, well above average. Their lungs, at 0.39 times the rate of everyone else — a deficit that survives age standardisation and survives conditioning on the donation having gone ahead.
Source: OPTN/UNOS STAR, Dec 2025 release · 316,248 deceased donors, of whom 3,051 died by drowning (median age 6)
95.5% vs 7.2%kidney vs lung recovery, same donors
Age-standardised, 220 drowning donors gave a lung against 431 expected — an observed/expected ratio of 0.51.
01 · The finding
These are unusually productive donors, except for one organ
Share of donors from whom each organ was recovered Drowning All other
Kidneys 95.5% against 92.4%. Hearts 43.5% against 33.1%. By the evidence of their other organs these are among the best deceased donors in the file. The shortfall is confined to the lung: 7.2% against 18.7%.
A low lung recovery rate could mean two very different things — an organ-specific judgement about the drowned lung, or these donors being less suitable in general. The cross-organ comparison rules out the second.
02 · The two obvious explanations
Neither donation failing, nor age
A companion brief partitions these donors by water type, age and sex — water type turns out not to be recorded anywhere in the file. Restricting to donors from whom a kidney was recovered — donors in whom donation demonstrably proceeded — leaves the deficit untouched: 7.4% of those drowning donors also gave a lung, against 19.5% of others. So it is not donors being declined outright.
Lung recovery by donor age band Drowning All other
Nor is it age, though age is the obvious suspect. Drowning skews overwhelmingly paediatric — median donor age 6 — and paediatric lung recovery is low for every mechanism. But the gap is present in every band, and direct standardisation over the drowning age distribution removes only about a third of it: 220 observed against 431 expected, a deficit of 211 donors across the file.
03 · What the evidence says
Recipient outcomes were settled five years ago
A propensity-matched UNOS analysis of 1,016 paediatric bilateral lung transplants, 128 from drowning or asphyxiation donors, found comparable survival at 90 days, one year and two years, and concluded such donors should be considered for use (Seese et al., Transplantation 2021;105:620–7).
Among lungs that were recovered
Drowning
All other
p
Recovered lungs transplanted
90.2%
92.1%
0.14
Severe PGD at 72 h
23.6% (n=55)
32.8%
0.15
1-year graft survival
81.9% (n=127)
85.2%
0.30
The honest limit, stated plainly. Those 220 donors are a heavily selected group — a screen that turns away most candidates would produce exactly these numbers. Lungs that are never recovered generate no outcome, so we cannot say the declined organs were usable. Severe PGD being numerically lower here is consistent both with accurate selection and with excess caution, and this analysis cannot separate them.
04 · Practice is already moving
Narrowing, not closed — and the right tool is already in hand
Era
Drowning donors
Lung recovered, drowning
Lung recovered, other
Ratio
2000-2010
662
5.4%
16.8%
0.32
2011-2019
996
8.0%
22.9%
0.35
2020-2025
829
10.7%
20.9%
0.51
This is not simple conservatism, and we should not report it as such. Lung recovery from drowning donors rose from 5.4% to 10.7%, and the ratio to other mechanisms improved from 0.32 to 0.51. Machine perfusion — the established way to assess a lung of uncertain quality — is already applied to these donors more often than to any other mechanism (10.2% against 6.5%). The concern is recognised and the right instrument is being reached for.
What remains is scale: perfusion was recorded for 14 drowning donors in the entire file.
05 · What would settle it
A question registry data can size but not answer
Where a drowning donor’s other organs are already being recovered, a protocolised lung assessment — inspection, bronchoscopy, oxygenation challenge, and ex-vivo perfusion where the result is equivocal — would convert an unobservable counterfactual into data. Registry analysis can establish that the gap is real, organ-specific and roughly half the expected rate. It cannot establish who was right.
Why this one is different. Most briefs in this series report that a variable fails to predict an outcome — death mechanism, HLA mismatch, every CAS component. This one reports a decision rather than a prediction, and the decision is measurable even where its correctness is not. The registry records what was done with each organ, which makes the gap visible; it does not record what would have happened otherwise, which is why it stays open.