Urea cycle disorders and hyperammonaemia — branching viva

Opening framework

My framework has four layers. First, the recognition — a well neonate deteriorating 24 to 72 hours into feeds with a respiratory alkalosis and encephalopathy is a urea cycle disorder until proven otherwise, and the ammonia is the test that decides it. Second, the emergency protocol — treat on suspicion: stop protein, calorie-load, scavenge nitrogen, give arginine, and dialyse at threshold. Third, the biochemistry — the cycle spans the mitochondrion and cytosol, and each block produces a recognisable metabolite pattern. Fourth, the family — an X-linked OTC diagnosis obliges carrier testing of the mother and sisters and reproductive counselling. [1]

The 'treat on suspicion' emergency protocol

The cardinal principle is that treatment begins before the enzyme or molecular diagnosis returns, because outcome tracks peak ammonia multiplied by time to definitive treatment. The protocol is H.A.L.T.: halt protein and give heat (10% glucose + intralipid ± insulin to switch off catabolism, the single most important physiological move), alternative-pathway scavengers (sodium benzoate and sodium phenylbutyrate, which excrete nitrogen as hippurate and phenylacetylglutamine), load the cycle downstream with intravenous arginine or citrulline, and take the ammonia off with dialysis if it exceeds threshold. [1] [5]

The dialysis thresholds follow the consensus guidance: in a neonate I proceed to haemofiltration or haemodialysis when ammonia exceeds roughly 500 micromoles per litre, or at a lower threshold of 200 to 300 with encephalopathy or a clearly rising trend despite medical therapy. Continuous kidney replacement therapy is preferred in the haemodynamically unstable neonate. The objective is to halve the ammonia within hours, because every hour of cerebral oedema adds to permanent injury — so I do not wait for the ammonia to plateau or for the diagnosis. [5]

Localising the defect with the metabolic panel

The first-tier panel is interpreted together: ammonia, plasma amino acids, urinary orotic acid, blood gas, glucose, lactate, ketones, liver function and coagulation, and acylcarnitines. The localising logic is positional. A proximal block produces low citrulline and low arginine; if the block is at OTC, carbamoyl phosphate is shunted into the pyrimidine pathway and urinary orotic acid is high, whereas if it is at CPS1 or NAGS, orotic acid is low or absent. A distal block produces accumulation further down: high citrulline in citrullinaemia type I (ASS1), argininosuccinic acid in ASL deficiency, and high arginine in arginase deficiency. [1]

This pattern converts a single blood test plus urinary orotic acid into a near-diagnostic first-tier screen. I confirm with molecular sequencing of the candidate gene — which also enables cascade carrier testing and prenatal or preimplantation diagnosis for the family — and I reserve liver enzyme assay for cases where sequencing is uninformative. The functional OTC variant data confirm that residual enzyme activity predicts the neonatal-versus-late-onset presentation, which is why partial defects can hide for years. [12]

The ammonia-glutamine mechanism of cerebral oedema

The brain injury is driven by glutamine, not ammonia directly. Ammonia crosses the blood-brain barrier freely and is fixed into glutamine by astrocytic glutamine synthetase. Glutamine is osmotically active: it draws water into the astrocyte, the cell swells, mitochondrial function fails, reactive oxygen species accumulate, and the result is cytotoxic and vasogenic cerebral oedema, raised intracranial pressure, and ultimately neuronal death and herniation. The injury is proportional to both the peak ammonia and its duration — which is precisely why speed of reduction is the dominant determinant of outcome. [3]

Branch: the mother, an OTC heterozygote

For the mother found to carry an OTC pathogenic variant, I apply the X-linked counselling framework. Affected males usually present with severe neonatal hyperammonaemia, while heterozygous females show a variable phenotype driven by skewed X-inactivation — some asymptomatic, some decompensating postpartum or after a protein load. Each pregnancy carries a 50 percent chance of transmission, with affected male fetuses at risk of severe neonatal disease and female fetuses at risk of variable expression. I offer prenatal or preimplantation genetic diagnosis, and I arrange carrier testing of her daughters and sisters. [12]

Closing: liver transplantation and the trap to avoid

For severe neonatal OTC disease, early liver transplantation — performed in a metabolically stable window once the acute crisis is controlled — corrects the hepatic enzyme defect, restores metabolic stability, and liberates the child from dietary restriction and the constant threat of decompensation. The trap is to view transplantation as a rescue for established brain injury; it cannot reverse that, which is why it is complementary to, not a substitute for, excellent emergency care. The closing point is that the speed of the acute response — measured ammonia, early scavengers, dialysis at threshold — is the single most powerful modifier of the lifelong neurodevelopmental outcome. [10] [1]