Neonatal Hyperammonaemia Interpreter
Neonatal Hyperammonaemia Interpreter
A raised ammonia in a newborn is a metabolic emergency in which the time to treatment predicts the outcome — and it is also, most of the time, a squeezed heel-prick. Three things this page does that a general ammonia page cannot. The neonatal reference interval is much higher than the adult one, so the adult upper limit over-calls. The sampling artefacts are the commonest cause of a raised result and are recognisable. And the differential splits cleanly on the blood gas and the ketones: respiratory alkalosis with a normal anion gap points at a urea cycle defect, a raised anion gap acidosis with ketones at an organic acidaemia.
Is it real, and what is it?
Ammonia + gas + ketones → patternA term infant, 40 hours old, fed normally for the first day and now vomiting, tachypnoeic and lethargic. Treated as possible sepsis. A free-flowing venous ammonia, sent on ice, comes back at 220 µmol/L. The blood gas shows a respiratory alkalosis with a normal anion gap; blood ketones are negative.
The neonatal numbers, and the two tests that split the differential
80 to 150 µmol/L — re-assay in 4 hours, withhold protein.
Above 150 µmol/L — repeat urgently within the hour, stop feeds, contact the metabolic consultant, and start treatment without waiting for the diagnosis.
400 µmol/L or more, or encephalopathic, or not falling on drugs — start haemofiltration urgently, within 6 hours of identification. Metabolic infusions within 30 minutes of the decision to treat.
Urea cycle defect — respiratory alkalosis or normal gas, normal anion gap, no ketones.
Organic acidaemia — raised anion gap metabolic acidosis, ketones present, often with neutropenia, thrombocytopenia, hypoglycaemia and a raised lactate.
- 12 to 82 µmol/L
- the neonatal interval, from 127 hospitalised neonates (43 term, 84 preterm) measured by bromophenol blue photometry on one analyser. It is roughly 1.6 times the adult upper limit. It is one study on one platform and should be read as evidence that the neonatal interval is higher, not as a number to adopt — the correct interval is your own laboratory’s
- every artefact pushes it up
- the single most useful fact about ammonia. Capillary or squeezed sampling, a tourniquet, a struggling or convulsing infant, a warm tube, a delay before separation — all of them raise the measured value and none of them lowers it. Two consequences: a raised result on a bad sample is usually the sample, and a normal result on a bad sample is more reassuring than a normal result on a good one
- respiratory alkalosis
- the finding that turns a septic-looking baby into a metabolic emergency. Ammonia stimulates the respiratory centre directly, so an infant with an untreated urea cycle defect hyperventilates and presents alkalotic with a normal anion gap. Any encephalopathic neonate with a respiratory alkalosis needs an ammonia measured before anything else is concluded
- ketones
- abnormal in a newborn at all, and the cleanest split in the differential. Strongly positive ketones with an acidosis and a raised ammonia is an organic acidaemia; absent ketones with an alkalosis is a urea cycle defect. Urine ketones lag and measure acetoacetate, so blood beta-hydroxybutyrate is the better test
- urine orotic acid
- the test that splits the urea cycle defects themselves. High in ornithine transcarbamylase deficiency, the commonest and the only X-linked one; low in carbamoyl phosphate synthetase 1 and N-acetylglutamate synthase deficiency. It is easy to forget on a urine sample that is being sent for organic acids anyway
Worked example
A term infant, 40 hours old, fed normally for the first day and now vomiting, tachypnoeic and lethargic. Treated as possible sepsis. A free-flowing venous ammonia, sent on ice, comes back at 220 µmol/L. The blood gas shows a respiratory alkalosis with a normal anion gap; blood ketones are negative.
220 is below 400, and the infant is lethargic but not yet comatose or fitting, so the filtration branch is not triggered on either limb — though it will be if the number rises
The sample was free-flowing, on ice and prompt, so this is not an artefact
220 is above 150, which is an action threshold in its own right: repeat within the hour, stop feeds, call the metabolic consultant
The gas shows a respiratory alkalosis with a normal anion gap, and the ketones are negative — the organic acidaemia branch needs both a raised anion gap acidosis and ketones, and neither is present
That leaves the urea cycle pattern, which is exactly what the clinical story fits: a well interval of 24 to 72 hours, then vomiting and encephalopathy, with a respiratory alkalosis rather than the acidosis sepsis would give
So: stop all protein, high-concentration glucose to halt catabolism, ammonia scavengers on metabolic advice, and identify now where this infant could be filtered if the ammonia rises
Diagnostic samples alongside, not before: plasma amino acids (citrulline places the block), plasma acylcarnitines, urine organic acids, and urine orotic acid — high in ornithine transcarbamylase deficiency, low in the two proximal defects
Urea cycle defect against organic acidaemia
| Urea cycle defect | Organic acidaemia | |
|---|---|---|
| Blood gas | Respiratory alkalosis or normal | Metabolic acidosis |
| Anion gap | Normal | Raised |
| Ketones | Negative | Strongly positive |
| Lactate | Normal | Often raised |
| Glucose | Usually normal | Often low |
| Full blood count | Normal | Neutropenia and thrombocytopenia common |
| Key diagnostic urine test | Urine orotic acid — high in OTC deficiency, low in CPS1 and NAGS deficiency | Urine organic acids |
| Key diagnostic blood test | Plasma amino acids — citrulline places the block | Plasma acylcarnitine profile |
Why a raised neonatal ammonia is usually the sample
| Pre-analytical factor | Direction of error |
|---|---|
| Heel-prick or other capillary sample | Raises — cell content and sweat contamination |
| Squeezing the sample out | Raises |
| Tourniquet, or a struggling or convulsing infant | Raises |
| Sample kept warm rather than on ice | Raises |
| Delay before the plasma is separated | Raises |
| Any of the above | Never lowers the result |
A well baby, a well day, then an encephalopathy that looks like sepsis
The classic presentation of a urea cycle defect is a term infant who feeds and behaves normally for the first 24 to 72 hours — the interval during which the mother’s circulation is no longer clearing ammonia but the infant’s own protein load has not yet accumulated — and who then becomes irritable, vomits, feeds poorly, and progresses over hours to lethargy, hypotonia, seizures and coma. Nearly every one of these infants is treated for sepsis first, because that is exactly what it looks like, and the ammonia is often not sent until somebody asks why the cultures are negative. The cost of that delay is neurological and permanent: outcome in hyperammonaemia is governed by the peak concentration and the duration of exposure, not by which enzyme turns out to be missing, which is why treatment begins before the diagnosis is known.
The first job on any raised result, though, is to decide whether it is real, because most raised neonatal ammonias are artefactual. Every recognised pre-analytical problem with ammonia pushes the measurement in the same direction — upward. Capillary and heel-prick samples read falsely high because of cell content and sweat contamination. A tourniquet, a struggling infant or a convulsing one raises it. So does a warm tube, or any delay before the plasma is separated. The correct sample is free-flowing venous or arterial blood, taken without a tourniquet, into the right tube, transported on ice, walked to the laboratory rather than sent by tube, with the laboratory told in advance so it is analysed on arrival. The asymmetry of the errors has a useful corollary that is rarely stated: a normal ammonia on a badly taken sample is more reassuring than a normal ammonia on a perfect one, because the measurement was biased against the answer you got.
The second job is to know what normal is, and this is where a general ammonia range misleads. The neonatal interval is substantially higher than the adult one. A published neonatal reference study of 127 hospitalised newborns under a week old — 43 term and 84 preterm — found a central 95 per cent interval of 12 to 82 µmol/L, with a median of 32 and an upper reference limit of 82, against an adult upper limit of about 50. That is a single cohort on a single analytical platform and should not be adopted as a universal number, but it establishes the direction clearly: applying an adult upper limit to a newborn will over-call. The action thresholds used in practice sit above it — the urea cycle disorder guideline starts treatment in a neonate above 150 µmol/L (its lower adult line of 100 is explicitly for patients outside the neonatal period), re-assays in four hours between 80 and 150 only in a newborn known to be at risk from an affected sibling, and extracorporeal removal is considered at 400 or more or in an encephalopathic infant.
The third job is to place it, and two bedside tests do almost all of the work. Ammonia stimulates the respiratory centre directly, so an infant with an untreated urea cycle defect hyperventilates and presents with a respiratory alkalosis and a normal anion gap — the opposite of the acidosis a septic or shocked infant shows, and the finding that should prompt the ammonia in the first place. An organic acidaemia, by contrast, produces a raised anion gap metabolic acidosis with ketones, and ketones in a newborn are abnormal in themselves. Neutropenia, thrombocytopenia, hypoglycaemia and a raised lactate travel with the organic acidaemias and not with the urea cycle defects. The ammonia interpretation calculator grades ammonia across all ages, the anion gap metabolic acidosis identifier reads the anion gap, the arterial blood gas interpreter the gas, and the neonatal glucose infusion rate calculator the glucose infusion that stops catabolism.
Frequently asked questions
What is the normal ammonia level in a newborn?
Higher than in an adult, which is the single most important thing to know before reading the result. A published neonatal reference study of 127 hospitalised newborns under one week old gave a central 95 per cent reference interval of 12 to 82 µmol/L, with a median of 32 µmol/L — 28 in term and 35 in preterm infants — against an adult upper limit of around 50 µmol/L. That is one cohort measured on one analytical platform, so it should be read as evidence that the neonatal interval is substantially higher rather than as a universal number to adopt; your own laboratory’s neonatal interval is the one that applies. What it does establish is that applying the adult upper limit printed on most reports to a newborn will over-call.
At what ammonia level is a newborn a metabolic emergency?
Above 150 µmol/L the infant is treated as an emergency: repeat urgently within the hour on a properly taken sample, stop all protein feeds, contact the metabolic consultant, and begin treatment without waiting for the diagnosis. Between 80 and 150 µmol/L, the urea cycle disorder guideline’s instruction to re-assay in four hours belongs to its protocol for a newborn at risk because a sibling had an early-onset urea cycle disorder; in other infants the value sits at or just above the top of the neonatal reference interval, and the step is a prompt repeat on a properly taken sample while watching the infant. At 400 µmol/L or more, or in an encephalopathic infant, or where the ammonia is not falling on pharmacological treatment, extracorporeal removal is indicated urgently — guidance in wide use specifies haemofiltration within six hours of identification and metabolic infusions within 30 minutes of the decision to treat. The urgency is because outcome is determined by the peak ammonia and how long it lasts, not by the eventual diagnosis.
Can a heel-prick sample give a falsely high ammonia?
Yes, and it is the commonest cause of a raised neonatal ammonia. Capillary samples read falsely high because of cell content and sweat contamination, and squeezing the heel makes it worse. So does a tourniquet, a struggling or convulsing infant, keeping the sample warm rather than on ice, and any delay before the plasma is separated. Every one of these errors raises the result and none lowers it. Two things follow. A raised ammonia on a capillary or squeezed sample should be repeated on free-flowing venous or arterial blood, on ice, walked to a forewarned laboratory — but treatment should start in parallel rather than waiting for the repeat, because two hours of unnecessary glucose in a well infant costs nothing and two hours of untreated hyperammonaemia is permanent. And conversely, a normal ammonia on a badly handled sample is genuinely reassuring.
How do I tell a urea cycle defect from an organic acidaemia in a newborn?
The blood gas and the ketones split them cleanly. A urea cycle defect produces a respiratory alkalosis or a normal gas with a normal anion gap and no ketones, because ammonia stimulates the respiratory centre and there is no acid accumulating. An organic acidaemia produces a raised anion gap metabolic acidosis with strongly positive ketones, often with hypoglycaemia, a raised lactate, and neutropenia and thrombocytopenia on the full blood count. The diagnostic samples then differ: urine orotic acid and plasma amino acids for the urea cycle defects — orotic acid is high in ornithine transcarbamylase deficiency and low in carbamoyl phosphate synthetase 1 and N-acetylglutamate synthase deficiency — and urine organic acids with a plasma acylcarnitine profile for the organic acidaemias. The emergency treatment is the same for both.
Why does a baby with a urea cycle defect look septic?
Because hyperammonaemic encephalopathy and sepsis produce the same picture in a newborn: poor feeding, vomiting, tachypnoea, temperature instability, lethargy progressing to seizures and coma, in an infant who was well a few hours earlier. The discriminating finding is the blood gas. Sepsis and shock produce a metabolic acidosis; an untreated urea cycle defect produces a respiratory alkalosis with a normal anion gap, because ammonia drives the respiratory centre. A tachypnoeic, encephalopathic newborn who is alkalotic rather than acidotic should have an ammonia measured immediately. The other clue is timing: a well interval of 24 to 72 hours followed by rapid deterioration is very characteristic, and corresponds to the point at which the infant’s own protein load overwhelms a partial or absent enzyme.
Does a normal ammonia exclude a urea cycle disorder?
No. Hyperammonaemia in a partial urea cycle defect is episodic and driven by catabolism, so a single normal value in a well, feeding infant says little about what happens during the next intercurrent illness, fast, or surgical episode. If the clinical suspicion is real — an unexplained encephalopathy, a family history, a previous unexplained neonatal death, or recurrent vomiting with lethargy — the test should be repeated during the illness rather than treated as excluded between episodes. A normal ammonia also does not exclude an organic acidaemia, in which the ammonia can be normal while the acidosis and ketosis are not.
Related calculators
References
- Häberle J, Burlina A, Chakrapani A, et al. Suggested guidelines for the diagnosis and management of urea cycle disorders: first revision. J Inherit Metab Dis. 2019;42(6):1192–1230. doi:10.1002/jimd.12100. The source of the action ladder used here — "if ammonia reaches 80 to 150 μmol/L, re-assay in 4 hours; if ammonia >150 μmol/L … stop feeds" — and of the statements that the prognosis "is strongly influenced by the duration of coma at presentation and peak ammonia levels", that spurious elevations occur from "poor sample processing, tourniquet use, crying, struggling, or convulsions", and that "capillary samples may result in falsely elevated values because of cell content and sweat contamination".
- Diaz J, Tsai M, Bertholf RL, et al. Proposed plasma ammonia reference intervals in a reference group of hospitalized term and preterm neonates. J Appl Lab Med. 2020;5(2):363–372. doi:10.1093/jalm/jfz001. 127 neonates (43 term, 84 preterm) measured by bromophenol blue photometry; central 95% reference interval 12–82 µmol/L, median 32 µmol/L (term 28, preterm 35), upper reference limit 82 µmol/L for neonates under one week of age. One cohort on one platform — cited here as evidence that the neonatal interval is substantially higher than the adult one, not as a universal number.
- Yorkshire and Humber Neonatal and Paediatric Hyperammonaemia Guideline. Sheffield Children’s NHS Foundation Trust. The source of the escalation points quoted here: above 150 µmol/L at any age, repeat urgently within 1 hour and contact the metabolic consultant; at 400 µmol/L or more, or encephalopathic, or pharmacologically resistant, start haemofiltration urgently within 6 hours of identification, with metabolic infusions within 30 minutes of the decision to treat. Also the sampling requirement — a "venous free flowing sample taken and sent on ice and walked quickly to the lab" — and the warning that "heel pricks, squeezed samples, samples at room temperature, and delays in processing" cause falsely raised results.
- Baumgartner MR, Hörster F, Dionisi-Vici C, et al. Proposed guidelines for the diagnosis and management of methylmalonic and propionic acidemia. Orphanet J Rare Dis. 2014;9:130. doi:10.1186/s13023-014-0130-8. The source for the organic acidaemia side of the differential, including the haematological findings — neutropenia and thrombocytopenia — that accompany it and do not accompany a urea cycle defect.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.