Children do die from circumcision — documented, not disputed — but the deaths are rare, under-recorded, and easy to either dismiss or inflate. Here is what the primary literature and official data actually establish.
The strongest source is official US data: Earp et al. (Clinical Pediatrics 2018;57(13):1532–1540) found 200 early in-hospital deaths among 9,833,110 circumcised newborns, 2001–2010 (~1 per 49,166). The authors stress this is a circumcision-RELATED count, NOT proven causation — it may both under- and over-count, and the deceased infants had vastly higher odds of pre-existing illness.
A bigger figure — ~117 deaths/yr (Bollinger 2010, Thymos) — is a contested MODELLING ESTIMATE, derived not counted, and criticised as overstated. We flag it as an estimate, never a fact.
When circumcision is fatal, the mechanisms are haemorrhage and sepsis (Schröder et al., European Urology Focus 2022) — the same pathways behind documented non-clinical traditional-circumcision deaths in South Africa (Douglas et al. 2018, ~40 deaths in one Eastern Cape season).
Cited from the primary sources, never from any third-party archive. Open the full article for the detail and the hedging.
This is the hardest part of the circumcision record to write about honestly, because it is the easiest part to get wrong. Children do die from circumcision — that is documented, not disputed. But the deaths are rare, hard to count, and easy to either dismiss or inflate. What follows is drawn from the primary medical literature and official data, not from any third-party archive, and it is written to a single standard: state what the sources actually establish, hedge it exactly as they do, and never reach for a bigger number than the evidence supports.
What the official data show
The strongest single source is not an advocacy estimate but an analysis of government data. In 2018, writing in Clinical Pediatrics, Brian Earp and colleagues examined the United States Kids’ Inpatient Database — a large, official record of paediatric hospital stays — for the decade 2001 to 2010. Among 9,833,110 newborns who were circumcised during their birth hospitalisation, they found 200 early in-hospital deaths: roughly one death for every forty-nine thousand circumcised infants.
It matters enormously how that number is read, and the authors are careful to read it correctly. This is a count of deaths that followed circumcision, not a count of deaths caused by it. The figure, they write, "may include both undercounting and overcounting of deaths attributable to circumcision." Crucially, the infants who died were not typical: they had dramatically higher odds of serious pre-existing illness — congenital cardiac disease, bleeding disorders, fluid and electrolyte problems. Many of these babies were already gravely ill. The honest reading is that circumcision-related death is a real but uncommon event, often entangled with underlying disease, and almost certainly under-recorded because death certificates rarely name an elective procedure as a cause.
The estimate that goes further — and why we flag it
A larger figure circulates widely: that more than a hundred American infants die from circumcision-related causes every year. It comes from one paper — Dan Bollinger’s 2010 "Lost Boys," published in Thymos: Journal of Boyhood Studies — which estimated about 117 such deaths annually, or 9.01 per 100,000. We mention it because it is influential, and we flag it plainly: this is a modelling estimate, not a counted total. It was built by extrapolation rather than drawn from a death register, it was published in an advocacy-adjacent venue, and it has been openly criticised as overstated. It sits uneasily beside the official-data approach of Earp and colleagues. We do not present ~117 deaths a year as a fact. It is a contested estimate, and it should always be labelled as one.
How circumcision kills, when it does
When circumcision does prove fatal or nearly so, the medical literature is consistent about the mechanism. In a 2022 case series in European Urology Focus, Annette Schröder and colleagues reviewed previously-healthy newborns admitted as emergencies after circumcision in the Greater Toronto Area. The pattern was two-fold: massive haemorrhage and infection, with several infants progressing to haemorrhagic or septic shock. Their conclusion was blunt — the risk of serious complications and death from neonatal circumcision "is greater than generally assumed." These are not exotic failure modes. A small organ in a small body can bleed dangerously, and an open wound can become infected; both can, rarely, be fatal.
The same mechanisms dominate where circumcision happens outside the hospital. In South Africa, traditional male circumcision performed during initiation rites has a documented and recurring death toll. A peer-reviewed 2018 study in the American Journal of Men’s Health by Mbuyiselo Douglas and colleagues reported official Eastern Cape figures: in a single 2013 initiation season the province recorded roughly forty deaths and a series of penile amputations, concentrated in one district. These deaths arise in non-clinical settings — a different context from a hospital neonatal circumcision — but they confirm the same lethal pathways of bleeding, infection, and dehydration, and they are counted, named events, not estimates.
What the record does not let us say
Two temptations have to be resisted. The first is to wave the deaths away as vanishingly rare and therefore irrelevant; the official US data alone document two hundred of them in a decade, and under-recording means the real figure is likely higher, not lower. The second is to brandish a dramatic annual death toll as settled fact; the most-quoted such figure is a contested model, and using it as a hard number damages the very case it is meant to support. Honesty cuts both ways here.
The honest bottom line
Circumcision-related deaths are real, documented, and under-recorded. The best official data put them at roughly one per forty-nine thousand US infant circumcisions over 2001–2010 — a circumcision-related count, not a proven causal rate, and entangled with underlying illness. A higher annual estimate exists but is a contested model, not a count, and we treat it as such. When circumcision is fatal, it is almost always through haemorrhage or sepsis, the same mechanisms that drive the documented deaths in non-clinical settings abroad. For a procedure that is, for most infants, elective and non-therapeutic, even a rare and uncertain risk of death belongs squarely in the consent conversation — stated plainly, neither dismissed nor inflated.
Compiled June 2026 from the primary literature and official data: Earp et al., Clinical Pediatrics 2018;57(13):1532–1540 (DOI 10.1177/0009922818790060); Bollinger, Thymos: Journal of Boyhood Studies 2010;4(1):78–90 (treated as a contested estimate); Schröder et al., European Urology Focus 2022;8(5):1500–1505 (PMID 34973956); and Douglas et al., American Journal of Men’s Health 2018;12(2):453–462. Deaths are presented as circumcision-RELATED (association, not proven causation); the Bollinger figure is flagged as an estimate. No third-party archive text was copied; all sources are cited directly.
