For decades, newborns were circumcised with no anaesthetic, on the belief that babies didn't really feel pain. Two 1997 studies ended that belief.
Lander et al. (JAMA 1997;278(24):2157–2162) ran an RCT comparing ring block, dorsal penile nerve block, EMLA, and a placebo (= no anaesthetic). The untreated infants showed sustained tachycardia and high-pitched crying, and two became ill (choking, apnoea) — so that arm was stopped early on ethical grounds. All three anaesthetic methods reduced the pain response; the ring block worked best.
Taddio et al. (Lancet 1997;349(9052):599–603) then found a significant linear trend at later vaccination: intact infants reacted least, EMLA-circumcised more, placebo-circumcised most — an association consistent with neonatal pain sensitising the later response (hedged by the authors; not proof of lifelong causation).
Cited from the primary journals (PMID 9417009; PMID 9057731), never from any third-party archive. Open the full article for the detail.
For most of the twentieth century, doctors circumcised newborns with nothing for the pain. The justification was not cruelty but a belief — that babies did not really feel pain, or could not remember it, so anaesthesia was an unnecessary risk. Two studies published in 1997 dismantled that belief so completely that one of them had to be stopped partway through. This is what the primary literature actually says about the pain of circumcision, drawn from the journals themselves rather than any third-party archive.
The trial that had to be stopped
In December 1997, JAMA published a randomised controlled trial by Janice Lander and colleagues (JAMA 1997;278(24):2157–2162). Its design was simple and, in hindsight, damning: newborns due for circumcision were randomly assigned to a ring block, a dorsal penile nerve block, a topical anaesthetic cream (EMLA), or a topical placebo — which is to say, nothing. The placebo arm was not a contrivance. It represented the standard of care at the time: circumcision with no anaesthetic at all.
What happened in that untreated group settled the question. The infants who received no anaesthetic showed a sustained, homogeneous response — heart rates that climbed and stayed up, and a high-pitched cry that ran through the whole procedure and beyond. Two of them became ill, with choking and apnoea. The distress was severe enough that the untreated arm was discontinued early on ethical grounds: the researchers concluded it was no longer defensible to keep circumcising babies with nothing for the pain in order to measure how much it hurt.
The contrast with the anaesthetised infants was stark. All three of the treated groups cried less and had lower heart rates than the untreated group. The ring block was the most effective, working across every stage of the procedure; EMLA was the weakest, doing little during the moments of foreskin separation and incision. The practical lesson was twofold — circumcision hurts, and the way it is anaesthetised matters a great deal.
A pain that did not end with the procedure
The second study reached further forward in time. Earlier the same year, The Lancet published work by Anna Taddio and colleagues (Lancet 1997;349(9052):599–603) that followed infant boys to their routine four- and six-month vaccinations and measured how they reacted to the needle. The infants fell into three groups: those who had not been circumcised, those circumcised with EMLA, and those circumcised with placebo — no effective analgesia.
The pain scores at vaccination rose in a clear, statistically significant line across those three groups. Intact infants reacted least. Infants circumcised with EMLA reacted more. Infants circumcised with no effective pain relief reacted most. The authors interpreted this as evidence that a painful neonatal experience, left under-treated, can shape how an infant responds to pain months later — and, notably, that pre-operative analgesia attenuated this later reaction without erasing it.
It is important to state what this study is and is not. It is a prospective observational comparison, not a randomised causal experiment, and the authors framed their conclusion accordingly — as an association consistent with early pain sensitising the developing nervous system. It does not prove a lifelong, fixed effect. But the direction of the finding is hard to dismiss: the babies who hurt most as newborns were the babies who hurt most at the doctor's office half a year later.
Why this matters for consent
Put together, the two papers retire an old defence of routine infant circumcision — that the patient does not really feel it. Lander shows the acute pain is real, measurable, and severe enough to make an untreated study arm unethical to continue. Taddio shows the experience can echo forward into how a child meets pain later. Neither finding requires any ideological reading; both come straight from JAMA and The Lancet.
For a bodily-autonomy lens, the implication is direct. If a non-therapeutic procedure performed on a newborn is painful enough that researchers cannot ethically withhold anaesthesia to study it, then pain is no longer a footnote to the circumcision debate — it is part of the harm being weighed against a benefit the infant did not choose. Adequate analgesia mitigates that harm; it does not make the underlying question of consent go away.
The honest bottom line
Neonatal circumcision is acutely painful, and infants register that pain in their bodies, not just their cries. For decades it was performed with no anaesthetic at all, until a 1997 trial had to halt its no-anaesthetic arm on ethical grounds. And there is reasonable evidence — an association, hedged by its own authors — that under-treated newborn pain can sensitise an infant's later pain response. Those are the facts the primary literature supports. They are enough.
Compiled June 2026 from the primary journals: Lander et al., JAMA 1997;278(24):2157–2162 (PMID 9417009); and Taddio et al., The Lancet 1997;349(9052):599–603 (PMID 9057731, DOI 10.1016/S0140-6736(96)10316-0). The Taddio finding is treated as an association, as the authors framed it. No third-party archive text was copied; both studies are cited directly.
