ResusDoc

Foreign bodies

Swallowed, inhaled, in the nose or ear, or retained in soft tissue — in adults and in children. Answer the questions and the tool gives you the investigation, the action, whether this patient can go home, and what has to happen afterwards.

Adults & children UK sources first Printable advice leaflet
Three things that do not wait for this tool. A button battery in the oesophagus, two or more ‘super strong’ magnets (or one magnet with any other metal object), and complete oesophageal obstruction — drooling, unable to swallow their own saliva — are each a time-critical emergency. Escalate now and work through the rest afterwards.
STEP 1

Where has the foreign body gone?

If you are not sure between swallowed and inhaled, treat as both until imaging tells you otherwise.

Foreign bodies: the thinking behind the pathway

Most swallowed objects are a non-event. The job in the emergency department is not to manage the eighty per cent that will pass — it is to find, quickly, the few that will not, and the two that are burning a hole while you decide.

The whole subject in one sentence. Ask four questions — what is it, where is it, how big is it, and is the patient symptomatic — and every decision in this tool falls out of the answers. Everything below is an expansion of those four.

A case to hold on to

A 14-month-old is brought in by her father at nine in the evening. She has been off her food since teatime and gagged on a drink. No-one saw her swallow anything. She is alert, playing, afebrile, chest clear, abdomen soft. She looks, on any reasonable assessment, well.

The instinct is reassurance and a safety-net leaflet. Hold that instinct until the end of this page.

01The four questions, and why they are in that order

The order matters. What it is comes first because two objects — a button battery and more than one magnet — override everything else about size, site and symptoms. Where it is comes second because the oesophagus is the dangerous place and the stomach usually is not. How big comes third because two measurements predict almost all failures to pass. Symptoms come last, deliberately, and that ordering is the single most important idea on this page.

Symptoms come last because a well child is not a reassuring child. In the largest series of button battery ingestions, 92% of fatal cases and 56% of major-outcome cases were never witnessed, and at least 27% of major-outcome and 54% of fatal cases were misdiagnosed — usually because the presentation was non-specific.[1] The child who looks well and has no story is the child these series are describing. COHORT

This is why the tool asks you to image on suspicion rather than on symptoms, and why “she seems fine” appears nowhere in the discharge criteria of any UK guideline quoted here.

02The button battery — a burn, not a poisoning

The injury is not chemical leakage and it is not heavy-metal toxicity. NHS England’s alert is explicit: severe tissue damage results from a build-up of sodium hydroxide generated by the electrical current discharged from the battery, and not, as commonly supposed, from leakage. The hydroxide burns the oesophagus, and the burn can fistulate into a major blood vessel and cause catastrophic haemorrhage. NHS/PSA/W/2014/017[2]

Three consequences follow directly, and all three are in the alert:

  • A flat battery is still dangerous. “Even apparently discharged (‘flat’) batteries can still have this effect.” A residual voltage is enough to electrolyse tissue water.
  • It is not only about swallowing. “Button batteries pushed into ears or nostrils can also cause serious injuries.” The same electrochemistry works in a nostril.
  • Removal is not the end. Symptoms of tissue damage — haematemesis, haemoptysis, respiratory difficulty — “can manifest up to 28 days after ingestion”, and the alert describes a child who died of late complications after being treated and sent home.
How fast, exactly? The figure worth carrying is not a round “two hours”. Litovitz’s analysis of 56,535 National Poison Data System cases found severe burns with sequelae occurred in just 2 to 2.5 hours.[1] Outcomes were worst for lithium cells of 20 mm or more and children under four. The 20 mm lithium cell — the one in a car key fob and a set of bathroom scales — was implicated in most severe outcomes. COHORT

That series is also the origin of the point the NHS England alert makes independently: injuries extended after removal, with unanticipated delayed perforations, tracheo-oesophageal fistulas and fistulisation into major vessels. Getting the battery out stops the hydroxide being made. It does not undo the burn, and the burn keeps evolving.

On the radiograph, a button battery shows a double ring or halo on the AP film and a step at the edge on the lateral — the features that distinguish it from a coin. If you cannot confidently call it a coin, treat it as a battery.

03Magnets — why one is nothing and two are surgery

A single magnet is an inert lump. Two magnets in different loops of bowel are a clamp. They attract across the bowel wall, compress the tissue between them, and produce pressure necrosis, perforation, fistula, volvulus and death. The UK has a National Patient Safety Alert and an RCEM Best Practice Guideline on exactly this.

The NatPSA is unambiguous about the mechanism and about who is at risk: “If more than one ‘super strong’ magnet is swallowed separately, or a ‘super strong’ magnet is swallowed with a metallic object, then they can be strongly attracted together from different parts of the intestines… The risks of damage apply to children and adults of any size.” NatPSA/2021/002[3]

It also names the four failure modes it found in incident reports, which read as a list of things to avoid doing this shift:

  • Missed diagnosis in a young child where ingestion was not witnessed.
  • Confusion between discharge to ‘watch and wait’, appropriate for other swallowed items, and the urgent surgical referral required for super strong magnets.
  • Reluctance of radiology staff to do the lateral film as well as the AP.
  • Delay in surgical review and intervention.
Why the lateral film exists. Two magnets stuck face to face look like one magnet on an AP film. The RCEM guideline therefore asks for a lateral abdominal X-ray specifically when only a single magnet is seen — the lateral is not a routine extra, it is the test for the thing that will kill you. RCEM BPG[4]

Two further points from the RCEM guideline that are easy to get wrong. First, do not use a metal detector to assess a child with suspected rare earth magnet ingestion — it is recommendation 2. Second, and against the instinct built by every other foreign body: “passage of rare earth magnets into the stomach must not be used as an indication that a child is free from any potentially catastrophic underlying injury”. For magnets, what matters is not that it reached the stomach but that it keeps moving.

The guideline also records that perforation had already occurred in 50 to 75% of symptomatic patients at presentation, and that fistula formation may occur within 2 to 5 days. A symptomatic child with a magnet history is very likely to need an operation.

04Where ESGE and RCEM disagree — and why both are right

This is the sharpest edge in the whole topic, and the tool handles it by asking your patient’s age before it answers.

A single magnet, patient wellWhat it saysPopulation
ESGE 2016 EUROPEAN Urgent endoscopic removal within 24 hours “even if only one magnet is evident on radiographs or the patient’s history suggests ingestion of only one magnet”, because undetected additional magnets may be present Adults, written for endoscopists
RCEM 2021 UK A single accidental magnet in a well child meeting the discharge criteria may go home, with a repeat abdominal X-ray at 6–12 hours to confirm progression Children, written for emergency departments

These are not contradictory so much as differently scoped. ESGE is advising a service that can put a scope in within 24 hours and is reasoning about adults, in whom intentional and repeated ingestion is common. RCEM is advising an emergency department managing a well toddler, and has chosen serial imaging over anaesthetising every child who swallows one bead. Use the guideline that matches your patient, and say which one you used in the notes.

One more ESGE quirk not to ‘correct’. In ESGE’s own timing table, a battery or a sharp-pointed object in the oesophagus is emergent, but a magnet in the oesophagus is only urgent. That is what the table says.[5] In UK practice a magnet in the oesophagus alongside any second magnet is being treated far more urgently than that, and the NatPSA is the reason why.
05Sharp, long and wide — the geometry of what will not pass

Once past the oesophagus, whether an object will pass is mostly a question of shape against four narrow points: the pylorus, the duodenal sweep, the ileocaecal valve and the anus. ESGE gives two numbers that do nearly all the work: objects wider than 2–2.5 cm will not usually pass the pylorus or ileocaecal valve, and objects longer than 5–6 cm will not negotiate the duodenal sweep. ESGE[5]

For scale, a UK £2 coin is 28.4 mm across and a 2p is 25.9 mm — both above the threshold. A 5p is 18 mm and a 10p 24.5 mm. A coin is not automatically a small object.

  • Sharp-pointed objects carry a complication rate up to 35%, which is why they are removed from the oesophagus emergently and from the stomach or proximal duodenum urgently if it can be done safely.
  • Long objects beyond 5–6 cm lodge in the duodenal curve with a 15–35% risk of perforation and are removed urgently.
  • Blunt and small objects in the stomach, in an asymptomatic patient, need no endoscopy at all — clinical observation, and outpatient management if feasible.

Most objects that clear the oesophagus pass within 4–6 days, occasionally taking up to four weeks. If a blunt object has not left the stomach by 3–4 weeks, ESGE advises endoscopic extraction.

A quiet exception worth knowing. ESGE explicitly recommends against endoscopic retrieval in asymptomatic body packers who have swallowed drug packets, because rupture can be fatal; the conservative approach fails in only 2–5% of cases. Surgery is for suspected rupture, failure to progress, or obstruction. Grasping a cocaine packet with a snare is an error with a named mechanism.
06The food bolus — the presentation that is really a diagnosis

This is where UK gastroenterology does have a guideline, and it is emphatic. The BSG and BSPGHAN joint consensus guidelines on eosinophilic oesophagitis treat food bolus obstruction not as a plumbing problem but as a presentation of an undiagnosed oesophageal disease.[6]

The central recommendation is about biopsies, and it carries 100% panel agreement:

“Oesophageal biopsies should be taken at index endoscopy in patients with food bolus obstruction to diagnose eosinophilic oesophagitis.” BSG/BSPGHAN — GRADE moderate, STRONG, 100% agreement

Specifically, six oesophageal biopsies at two levels at the index endoscopy. The numbers behind that recommendation explain its force:

  • EoE is found in up to 46% of patients presenting with food bolus obstruction, and food bolus obstruction is the first presenting symptom in 30% of everyone eventually diagnosed with EoE.
  • Biopsies were not taken at the index endoscopy in 73% of those patients. In a more recent study 55% had no biopsies, and of those biopsied, 66% had too few to exclude EoE reliably.
  • Where biopsies were not taken, 79% were lost to secondary care follow-up — and where a repeat endoscopy did happen, diagnosis was delayed by at least two years.
  • Biopsy at that endoscopy does not increase perforation risk. In 511 EoE patients there were 10 perforations (2%), none attributed to biopsy.
The bolus passing is not a discharge. The guideline has a separate recommendation for exactly that situation: after spontaneous resolution, patients should still be booked for endoscopy and outpatient review before they leave. STRONG, 83% agreement The commonest error in this whole topic is the relieved patient who swallows, feels better, and walks out undiagnosed.

The guideline is also clear about what does not work: “There is no evidence that conservative treatments such as fizzy drinks, baclofen, salbutamol or benzodiazepines are helpful.” ESGE adds that glucagon’s effect is variable, commonly causes nausea and vomiting, and must not delay endoscopy.

07Imaging — what the X-ray will not show you

Plain radiography is the right first test for a suspected radiopaque object, and a poor test for most of what actually gets swallowed. ESGE’s own figures are sobering:

SettingPerformanceWhat to do instead
Plain film, any foreign bodyFalse-negative rate up to 47%Biplanar radiography if not seen on the single view
Non-bony food bolusFalse negatives about 87%Do not image — ESGE recommends against it in the uncomplicated case
Fish boneRadiography sensitivity 32% (358 patients)CT — sensitivity 90–100%, specificity 93.7–100%
Suspected perforationFree subdiaphragmatic gas is almost never seenCT. Absence of free air does not exclude perforation

The last row deserves emphasis because it inverts a rule most of us learned elsewhere. Foreign body perforation is caused by slow erosion, so the site seals with fibrin, omentum or adjacent bowel before much air escapes. Free intraperitoneal air is a poor sign here — its absence means very little.

Wood, plastic, glass, thin metal, and fish and chicken bones are the objects ESGE lists as not regularly identifiable on radiographs. A negative film in a symptomatic patient is not a negative result.

Barium is contraindicated: aspiration risk, and it coats the mucosa and ruins the endoscopic view.

08The inhaled foreign body and the choking patient

Choking is the one arm of this subject with a clean UK algorithm. Resuscitation Council UK separates mild from severe airway obstruction on a single question — is the cough effective? RCUK[7]

  • Effective cough (mild): encourage coughing. Do nothing else. Keep checking for deterioration.
  • Ineffective cough, conscious (severe): five back blows, then five abdominal thrusts, alternating. In an infant, chest thrusts replace abdominal thrusts.
  • Unconscious: start CPR. In the paediatric algorithm, open the airway and attempt rescue breaths, then continue paediatric BLS.
  • No repeated or blind finger sweeps — the paediatric algorithm states this explicitly.
The bit that gets forgotten after the drama. Anyone who has received abdominal thrusts needs assessment for the injury the thrusts themselves can cause — and anyone whose obstruction was relieved by coughing may still have a retained bronchial foreign body. A child with a witnessed choking episode, then a period of calm, then a persistent cough or unilateral wheeze, is the classic missed inhaled foreign body; the quiet interval is part of the illness, not evidence against it.

A retained bronchial foreign body most often lodges on the right, because the right main bronchus is wider and more vertical. Suggestive findings are unilateral wheeze, reduced air entry, and — because most inhaled objects are organic and radiolucent — air trapping and hyperinflation rather than the object itself. The referral is to a team who can do rigid bronchoscopy.

09Nose and ear

Most of this is minor and satisfying. One thing in it is not.

A button battery in the nose is the same emergency as a button battery in the oesophagus. The NHS England alert says so directly: batteries “pushed into ears or nostrils can also cause serious injuries”.[2] The mucosa is moist, the battery is in contact with both walls, and the same hydroxide is produced — with septal perforation and saddle deformity as the results. A metallic-looking disc in a nostril is a battery until an ENT surgeon says otherwise.

For everything else, the general principles are practice rather than national guidance PRACTICE:

  • The first attempt has the best chance. Each failed attempt swells, bleeds and pushes the object deeper, and converts a two-minute procedure into a general anaesthetic. Prepare properly or do not start.
  • Unilateral foul nasal discharge in a child is a retained nasal foreign body until proven otherwise, however long it has been there.
  • The parent’s kiss is a reasonable, well-tolerated first manoeuvre for an anterior nasal object.
  • Do not syringe the ear if the object is vegetable matter (it swells) or if the drum may be perforated. An insect should be killed — commonly with oil — before removal.
  • Refer rather than persist: an uncooperative child, a deeply impacted object, any battery, or a failed first attempt.
10Soft tissue — the one where the evidence genuinely disagrees

There is no UK national guideline for retained soft-tissue foreign bodies. What there is, is a small experimental literature that is more consistent — and more pessimistic — than it is usually given credit for.

Plain radiography is good for the radiopaque and useless for the rest: in a blinded model study it detected gravel, glass and metal about 98% of the time and never detected wood, plastic or cactus spine.[8] EXPERIMENTAL

Ultrasound is the test people over-trust. Two independent blinded studies agree it performs poorly for wood: overall sensitivity 43% (specificity 70%) in the first,[8] and 48.4% (95% CI 42.1–54.8, specificity 67.6%) across 500 examinations by fifty emergency physicians in the second.[9] Critically, that second study set out to show that bigger objects would be easier to see, and found the opposite: sensitivity did not improve as object size rose to 10 mm. EXPERIMENTAL

A correction worth making explicitly. It is often said that the evidence here “conflicts”, with high sensitivities quoted for small wooden fragments. On the primary studies that can actually be retrieved, it does not conflict — two blinded experiments, using different models and different operators, land within five percentage points of each other, and both conclude that ultrasound should not be used to rule a foreign body out. Treat a negative scan as uninformative, not reassuring.
The myth worth retiring. “Glass only shows on X-ray if it is leaded.” It is not true — most glass is radiopaque and shows on a plain film. If you suspect retained glass, get the radiograph.

Practical position: image any wound that might contain glass or metal; for wood and thorns accept that neither radiography nor ultrasound reliably excludes anything, and escalate to CT if the clinical concern persists. Tell the patient explicitly when a foreign body has not been excluded, document that conversation, and safety-net for the delayed presentation — persistent pain, a discharging sinus, or a mass weeks later.

11Discharge, follow-up, and the late catastrophe

The user-facing point of this tool is that discharge is a decision with criteria, not a default. Three different sources give three different follow-up regimens, and they are not interchangeable.

SituationFollow-upSource
Single magnet, well child, dischargedRepeat AXR at 6–12 hours, same position each time, progression confirmed by a radiologist; repeat until past the stomach and progressingRCEM
Sharp-pointed object not retrievedDaily radiographs, inpatient observationESGE
Battery beyond the duodenumPlain film every 3–4 daysESGE
Blunt object, asymptomatic, in stomachWeekly films; extract if still in the stomach at 3–4 weeksESGE
Food bolus, any outcomeEndoscopy with six biopsies at two levels, plus outpatient review booked before dischargeBSG/BSPGHAN

ESGE also names who should be admitted rather than discharged after successful endoscopic removal: a technically difficult extraction, multiple objects, a high-risk object (sharp, battery, magnet, anything over 5–6 cm), or extensive mucosal injury.

The 28-day tail. After a button battery, haematemesis, haemoptysis or breathing difficulty up to 28 days later may be the first sign of a fistula into a major vessel.[2] A herald bleed is a surgical emergency, not a reassurance that the bleeding stopped. Every battery patient goes home knowing this, and it is why the leaflet in this tool exists.

Ten pitfalls

  1. Treating “the child looks well” as evidence. 92% of fatal battery ingestions were never witnessed and over half were initially misdiagnosed.
  2. Calling a battery a coin. Look for the double ring on AP and the step on lateral. If unsure, it is a battery.
  3. Not asking whether there was a second magnet. The question is not “did they swallow a magnet” but “how many, and was there any other metal”.
  4. Accepting a single AP film for a magnet. Two magnets face to face look like one. The lateral is the test.
  5. Applying ‘watch and wait’ to magnets. Named explicitly in the NatPSA as a cause of harm.
  6. Treating arrival in the stomach as safety in a magnet case. RCEM says the opposite — progression, not position, is what reassures.
  7. Discharging a food bolus without biopsies or a booked endoscopy. 79% of un-biopsied patients are lost to follow-up.
  8. Reading a normal chest film as excluding an inhaled foreign body. Most are organic and radiolucent; look for air trapping.
  9. Excluding perforation because there is no free air. Free gas is almost never present in foreign body perforation.
  10. Persisting with a fourth attempt at a nasal foreign body. The first attempt is the best one; after that you are making an ENT problem worse.

Seven red herrings

  1. “It was a flat battery.” Residual voltage still electrolyses tissue.
  2. “The X-ray was normal.” False-negative rates reach 47% generally and 87% for food bolus.
  3. “Glass needs to be leaded to show up.” It does not.
  4. “The pain settled, so it has passed.” Foreign body sensation persists for hours after passage, and the symptom-free interval is classic for an inhaled foreign body.
  5. “Where the patient points is where it is stuck.” ESGE notes the area of discomfort often does not correlate with the site of impaction.
  6. “A fizzy drink will shift it.” No evidence, per BSG/BSPGHAN — and it must not delay endoscopy.
  7. “We removed it, so we are done.” The battery injury extends after removal, and the vascular fistula can be weeks away.

The case, revisited

The 14-month-old who was off her food and gagged on a drink, with no witnessed ingestion and a normal examination, has the exact profile of the cases in the Litovitz series: under four, unwitnessed, non-specific. She does not need reassurance first. She needs a radiograph of neck, chest and abdomen — and if there is a double ring on it, she needs ENT or surgery now, not in the morning.

That is the whole argument for imaging on suspicion rather than on symptoms, and it is the reason this tool asks what happened before it asks how the patient looks.

Evidence and sources

Every figure below was read from the primary document, not from a summary. Where the source is a UK national alert or guideline it is tagged UK; where UK guidance is silent and a European guideline is doing the work it is tagged EUROPEAN, because those are not the same thing and should not read as if they were.

An honest statement about what does not exist. There is no UK national guideline on foreign body ingestion as a whole. The British Society of Gastroenterology hosts a paper titled Foreign body ingestion: dos and don’ts, but it is an Education in Practice review in Frontline Gastroenterology written by Becq, Camus and Dray of Hôpital Saint-Antoine in Paris. It is BSG-hosted, not BSG-authored, and this tool does not cite it as UK guidance. What the UK does have is excellent but hazard-specific: national alerts on batteries and magnets, an RCEM guideline on magnets in children, a BSG/BSPGHAN guideline that governs the food-bolus aftermath, and RCUK algorithms for choking. The gaps between those are filled by ESGE, labelled as such.

Litovitz 2010 — the study behind the clock

Litovitz T, Whitaker N, Clark L, White NC, Marsolek M. Pediatrics 2010;125(6):1168–77. PMID 20498173 · RETROSPECTIVE COHORT

Three data sources: the National Poison Data System (56,535 cases, 1985–2009), the National Battery Ingestion Hotline (8,648 cases), and 13 deaths plus 73 major outcomes from the literature and hotline.

Severe burns with sequelae occurred in just 2 to 2.5 hours.
6.7-fold rise in major or fatal outcomes, 1985–2009 · ingestion of 20–25 mm cells rose from 1% to 18%, lithium cells from 1.3% to 24% · outcomes significantly worse for lithium cells ≥20 mm and children under 4 · 92% of fatal and 56% of major-outcome ingestions were not witnessed · ≥27% of major and 54% of fatal cases were misdiagnosed.

Appraisal. Registry data, so denominators are soft and reporting is voluntary — the true incidence is unknown and the case mix is skewed towards the severe. What registry data can establish well is the pattern among bad outcomes, and that is how it is used here: not to estimate risk, but to say that the fatal cases did not look like emergencies at presentation. The 2–2.5 hour figure is the observed interval in injury cases, not a guaranteed safe window.

NHS England Patient Safety Alert — button batteries

Stage One: Warning. Ref NHS/PSA/W/2014/017, 19 December 2014 · UK NATIONAL ALERT

Issued after review of National Reporting and Learning System and STEIS incidents from 2010. Five cases of severe tissue damage following delays in suspecting, diagnosing or treating ingestion in small children; one child died. A further 241 incidents described batteries swallowed as self-harm by inpatients.

Mechanism stated explicitly as sodium hydroxide generated by the discharged current, not leakage · “Even apparently discharged (‘flat’) batteries can still have this effect” · “button batteries pushed into ears or nostrils can also cause serious injuries” · symptoms of tissue damage can manifest up to 28 days after ingestion · “Removal of the battery alone may be insufficient…” with one child dying of late complications after being treated and sent home.

Appraisal. An incident-report-derived warning, not a trial; its value is that it is the national statement of what is expected of NHS organisations, and that it identifies the specific failure mode — staff not recognising ingestion as an emergency. Note it also flags hearing-aid batteries in older people with confusion or poor vision: six such incidents. This is not only a paediatric hazard.

National Patient Safety Alert — ‘super strong’ magnets

NatPSA/2021/002/NHSPS, 19 May 2021. NHS England & NHS Improvement National Patient Safety Team with RCEM · UK NATIONAL ALERT

NRLS review over three and a half years identified 24 incidents of delayed assessment or treatment. Actions were mandated for completion by 19 August 2021, with the note that failure to act may lead to CQC regulatory action.

Compression of bowel between magnets “can cause necrosis and perforation of the intestines and/or blood vessels within hours” · requires AP and lateral abdominal X-ray and surgical review · “The risks of damage apply to children and adults of any size.”
Failure modes found: unwitnessed ingestion missed; ‘watch and wait’ wrongly applied; reluctance to perform the lateral film; delay in surgical review.

Appraisal. Same design limits as any incident-based alert. Its force is regulatory and its content is corroborated by the RCEM guideline developed alongside it with the British Association of Paediatric Surgeons, the Royal College of Radiologists and the Society of Radiographers.

RCEM Best Practice Guideline — ingestion of super strong magnets in children

Royal College of Emergency Medicine, May 2021 (revised June 2021). Chair: Sarah Stibbards · UK GUIDELINE

Seven recommendations, a management algorithm, and a model patient advice leaflet. The discharge criteria and the 6–12 hour repeat film in this tool are taken from it directly.

Perforation had occurred in 50–75% of symptomatic patients at presentation · fistula formation within 2–5 days · intestinal injury can occur within 8–24 hours despite the child remaining well · abdominal symptoms may not appear for weeks · rare earth magnets usually under 6 mm in diameter · do not use metal detectors.

Appraisal, including a flaw worth knowing. This is consensus guidance from a working group, built on a small non-UK evidence base (its clinical references are US and Korean paediatric series). Two internal points deserve flagging rather than smoothing over:

  • The guideline contradicts itself on magnet strength. The body text says rare earth magnets are “between five and ten times stronger than ceramic magnets”; its own patient leaflet in Appendix 3 says “between seven and fourteen times stronger”. Nothing clinical turns on it, but it is a real inconsistency in the source and this tool does not silently pick one.
  • It disagrees with ESGE on the single magnet — see the Teaching tab, chapter 04. Both positions are defensible for their own population.

ESGE Clinical Guideline — removal of foreign bodies in the upper GI tract in adults

Birk M, Bauerfeind P, Deprez PH, Häfner M, Hartmann D, Hassan C, Hucl T, Lesur G, Aabakken L, Meining A. Endoscopy 2016;48:1–8 · EUROPEAN GUIDELINE

Thirteen GRADE-assessed recommendations. This is the source of every timing threshold in the swallowed arm of the tool, and it is labelled European throughout because the UK has no equivalent.

80–90% of ingested foreign bodies pass spontaneously; 10–20% need endoscopy, <1% surgery · food bolus incidence 13 per 100,000 per year · risk of major complication rises 14.1-fold if impacted >24 h in the oesophagus · sharp-pointed object complication rate up to 35% · objects >5–6 cm carry a 15–35% perforation risk · conservative management of body packing fails in only 2–5% · push technique for food bolus succeeds in >90% · endoscopic success overall >95%, complications 0–5%.

Appraisal. Every one of the thirteen recommendations rests on low quality evidence, by the guideline’s own GRADE assessment, even where the recommendation is graded strong. That is not a criticism so much as a description of the field: you cannot randomise a child with a battery in the oesophagus. The timings are expert consensus anchored to observational data, and they should be read as sensible operational thresholds rather than as validated cut-offs. The guideline explicitly defers to NASPGHAN for children.

BSG and BSPGHAN joint consensus guidelines — eosinophilic oesophagitis

Dhar A, Haboubi HN, Attwood SE, et al. Gut 2022;71:1459–1487. doi:10.1136/gutjnl-2022-327326 · UK GUIDELINE

The strongest UK evidence in this tool, and the answer to what happens after a food bolus is relieved. Recommendations carry GRADE, strength and an explicit level of panel agreement.

Urgent gastroenterology referral and endoscopy on the next available list or as an immediate emergency — Low / Strong / 94% · oesophageal biopsies at index endoscopy — Moderate / Strong / 100% · endoscopy and outpatient review booked after spontaneous resolution — Low / Strong / 83% · maintenance topical steroid reduces recurrent food bolus — Moderate / Strong / 94%.
Six biopsies at two levels. EoE found in up to 46% of food bolus presentations; first presenting symptom in 30% of EoE patients; biopsies not taken in 73%; where not taken, 79% lost to follow-up; delay to diagnosis then ≥2 years; biopsy does not increase perforation risk (10/511, 2%, none biopsy-related).

Appraisal. A formal consensus guideline with a declared development group, GRADE assessment and voting — methodologically the most robust document cited here. The biopsy recommendation rests on moderate evidence and unanimous agreement. The supporting numbers come from observational cohorts, so the 46% figure is a case-mix-dependent estimate rather than a fixed prevalence; the direction of the recommendation is not in doubt.

Resuscitation Council UK — choking algorithms

Adult choking algorithm 2025; Paediatric choking algorithm 2025. Resuscitation Council UK · UK GUIDELINE
Severity determined by effectiveness of cough · effective cough → encourage coughing, continue to check for deterioration · ineffective cough, conscious → 5 back blows then 5 abdominal thrusts, alternating; in an infant, 5 back blows then 5 chest thrusts · unconscious → CPR (adult) / open airway, attempt rescue breaths, continue paediatric BLS · no repeated or blind finger sweeps.

Appraisal. Consensus algorithms derived from ILCOR review; the underlying evidence for back blows versus thrusts is observational and weak, which is why the algorithms emphasise escalation and reassessment rather than claiming superiority for any single manoeuvre.

Manthey 1996 — ultrasound versus radiography for soft-tissue foreign bodies

Manthey DE, Storrow AB, Milbourn JM, Wagner BJ. Ann Emerg Med 1996;28(1):7–9. PMID 8669741 · BLINDED ANIMAL MODEL

Randomised, blinded; two radiologists independently assessed 120 chicken thighs with two-view radiography and 7.5-MHz ultrasound. Sixty contained one foreign body (10 each of gravel, metal, glass, cactus spine, wood, plastic).

Ultrasound sensitivity by material: gravel 40%, metal 45%, glass 50%, cactus spine 30%, wood 50%, plastic 40%. Overall ultrasound sensitivity 43%, specificity 70%. Radiography detected radiopaque objects (gravel, glass, metal) 98% of the time and never detected wood, plastic or cactus spine.

Appraisal. This is a chicken-thigh model, not patients — no overlying anatomy, no clinical context, no sequential scanning by a clinician who knows the history. On its own it would be weak evidence. Its weight comes from the fact that an independent study, using a different design, reached almost the same number (below). The conclusion the authors drew is the asymmetric one and it is the one this tool uses: ultrasound should not be relied on to rule out a retained foreign body.

Fleming 2017 — does a bigger splinter show up better? No.

Fleming ME, Heiner JD, Summers S, April MD, Chin EJ. J Spec Oper Med 2017;17(4):72–75. PMID 29256199 · BLINDED PROSPECTIVE MODEL

Blinded, prospective. Wooden toothpicks of 1, 2.5, 5, 7.5 and 10 mm inserted to a depth of 1 cm in a previously validated chicken soft-tissue model, with five foreign-body-free controls. Fifty emergency physicians with prior ultrasound training each scanned all ten models — 500 examinations.

Sensitivity 48.4% (95% CI 42.1–54.8) · specificity 67.6% (95% CI 61.3–73.2) · sensitivity did not change as object size increased, up to 10 mm.
Authors’ conclusion: “Providers should consider alternative diagnostic modalities if there is persistent clinical concern for a retained, radiolucent, soft-tissue foreign body.”

Appraisal, and why it matters here. Also a model rather than patients, so it inherits the same external validity limits as Manthey. But it is the more useful of the two for the bedside, because the operators were fifty ordinary emergency physicians rather than two radiologists, which is much closer to who actually scans these wounds. It also tested, and refuted, the intuition clinicians most rely on — that a larger splinter is easier to find. This study is why this tool does not present the ultrasound evidence as conflicting. Two blinded experiments, different models, different operators, agree within five percentage points.

References

  1. Litovitz T, Whitaker N, Clark L, White NC, Marsolek M. Emerging battery-ingestion hazard: clinical implications. Pediatrics 2010;125(6):1168–77. PMID 20498173.
  2. NHS England. Patient Safety Alert — Stage One: Warning. Risk of death and serious harm from delays in recognising and treating ingestion of button batteries. Ref NHS/PSA/W/2014/017, 19 December 2014.
  3. NHS England and NHS Improvement National Patient Safety Team with the Royal College of Emergency Medicine. National Patient Safety Alert: Urgent assessment/treatment following ingestion of ‘super strong’ magnets. NatPSA/2021/002/NHSPS, 19 May 2021.
  4. Royal College of Emergency Medicine. Best Practice Guideline: Ingestion of Super Strong Magnets in Children. May 2021, revised June 2021. Working consensus group chaired by Sarah Stibbards, with the British Association of Paediatric Surgeons, the Royal College of Radiologists and the Society of Radiographers.
  5. Birk M, Bauerfeind P, Deprez PH, Häfner M, Hartmann D, Hassan C, Hucl T, Lesur G, Aabakken L, Meining A. Removal of foreign bodies in the upper gastrointestinal tract in adults: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline. Endoscopy 2016;48:1–8. doi:10.1055/s-0042-100456.
  6. Dhar A, Haboubi HN, Attwood SE, Auth MKH, Dunn JM, Sweis R, Morris D, Epstein J, Novelli MR, Hunter H, Cordell A, Hall S, Hayat JO, Kapur K, Moore AR, Read C, Sami SS, Turner PJ, Trudgill NJ. British Society of Gastroenterology (BSG) and British Society of Paediatric Gastroenterology, Hepatology and Nutrition (BSPGHAN) joint consensus guidelines on the diagnosis and management of eosinophilic oesophagitis in children and adults. Gut 2022;71:1459–1487. doi:10.1136/gutjnl-2022-327326.
  7. Resuscitation Council UK. Adult choking algorithm and Paediatric choking algorithm, 2025.
  8. Manthey DE, Storrow AB, Milbourn JM, Wagner BJ. Ultrasound versus radiography in the detection of soft-tissue foreign bodies. Ann Emerg Med 1996;28(1):7–9. PMID 8669741.
  9. Fleming ME, Heiner JD, Summers S, April MD, Chin EJ. Diagnostic Accuracy of Emergency Bedside Ultrasonography to Detect Cutaneous Wooden Foreign Bodies: Does Size Matter? J Spec Oper Med 2017;17(4):72–75. PMID 29256199.
  10. Becq A, Camus M, Dray X. Foreign body ingestion: dos and don’ts. Frontline Gastroenterology 2020. doi:10.1136/flgastro-2020-101450. Cited only as context — a French-authored education review hosted by BSG, not UK guidance.

Discharge advice leaflet

Choose what the patient swallowed or had removed. The leaflet below rewrites itself, and prints on its own — the header, tabs and this panel are hidden on the printed page.

About this tool

A decision aid for foreign bodies presenting to a UK emergency department — swallowed, inhaled, in the nose or ear, or retained in soft tissue, in adults and in children. It covers what to investigate, what to do, who can be discharged, what follow-up is owed, and it prints the advice leaflet.

How the sources are handled

The UK has no single guideline on foreign bodies. It has several very good hazard-specific ones and some real gaps. Rather than blend these into an undifferentiated house view, every recommendation in this tool carries a tag:

  • UK — a UK national alert, royal college guideline or UK society consensus guideline.
  • EUROPEAN — ESGE, used where UK guidance does not exist. Read as European practice, not UK policy.
  • COHORT — an observational study or registry analysis, cited for a specific figure.
  • PRACTICE — established practice with no national guideline behind it, and labelled honestly as such.

Where two good sources disagree — and on the single swallowed magnet, ESGE and RCEM do — the tool shows both and tells you which population each was written for, instead of picking one and hiding the other.

What this tool is not

It does not replace TOXBASE, your regional paediatric surgical centre, or ENT. For button batteries, magnets and caustic co-ingestion, local and regional pathways take precedence over anything here. Follow TOXBASE advice; the National Poisons Information Service is on 0344 892 0111.

It is also not a substitute for clinical judgement in the patient in front of you. The RCEM guideline’s own disclaimer applies equally here: guidance cannot cover all possible scenarios, and responsibility for interpretation and for the patient’s overall care rests with the treating clinician.

Review status

Internally reviewed only. The clinical content has been written and checked against the primary source documents listed in the Evidence tab, but this tool has not had external clinical review or governance sign-off. That wording will stay until someone outside the build has read it.

Version

2026-09-30.3 — first build. Author: Dr Nirmalya Hore. Part of ResusDoc.

Disclaimer. ResusDoc is an educational resource for healthcare professionals. It does not constitute medical advice for patients and does not create a clinical relationship. Doses, thresholds and timings must be checked against your local formulary and pathways before use. No patient-identifiable data is collected or stored by this page.