Biliary atresia is the most time-critical diagnosis in neonatal jaundice — it is rare (1:10,000–15,000 births) but invariably fatal if untreated, and the outcome is critically time-dependent. The Kasai portoenterostomy (hepatic portoenterostomy) — an anastomosis between the intestine and the porta hepatis, bypassing the obliterated biliary tree — achieves bile drainage in 60–80% of cases if performed before 60 days of age, but success falls to under 20% if performed after 90 days. Beyond 4 years without successful Kasai, essentially all affected children will require liver transplantation. The UK has implemented mandatory stool colour cards (distributed at the 6-week check), and NICE NG98 explicitly mandates that all neonates with jaundice beyond 14 days must have a split bilirubin performed to calculate the conjugated fraction. The most practical immediate test a GP can perform is urine dipstick — dark urine (conjugated bilirubin is water-soluble and stains the urine) in a jaundiced neonate is a medical emergency requiring same-day referral to paediatric hepatology. Kernicterus from untreated severe unconjugated hyperbilirubinaemia was historically common and caused high rates of cerebral palsy and sensorineural deafness — the introduction of phototherapy has made it rare but not extinct. The opisthotonus (arching of the back) and high-pitched cry in acute bilirubin encephalopathy are pathognomonic — any GP seeing these signs in a jaundiced neonate should call 999 immediately.

The 24-hour rule is the most important rule in neonatal jaundice — any visible jaundice within the first 24 hours of life is pathological by definition, as physiological jaundice cannot occur this early (the mechanisms that produce physiological jaundice — normal red cell breakdown after the high fetal haemoglobin production period — take at least 24 hours to generate significant bilirubin load). Visible jaundice before 24 hours indicates pathological haemolysis, which can cause bilirubin to rise at rates of 17 μmol/L per hour — far faster than phototherapy can clear it. Exchange transfusion may be required within hours. The Rhesus incompatibility situation has been dramatically improved by routine anti-D prophylaxis to all Rh-negative pregnant women — but maternal records must be checked, as missed prophylaxis still occurs. ABO incompatibility (group O mother + group A or B baby) is now the most common cause of significant haemolytic disease of the newborn — it is unpredictable (no maternal sensitisation history) and presents with jaundice in the first 24 hours. G6PD deficiency is X-linked recessive (affects males primarily) and is common in populations from tropical and subtropical regions — it causes haemolytic crisis when the baby is exposed to oxidative stressors (infection, drugs like vitamin K if given in excessive doses, or in the neonatal period can be triggered without obvious trigger).

The distinction between unconjugated and conjugated hyperbilirubinaemia is the critical biochemical branch in neonatal jaundice — unconjugated jaundice (the common, physiological/haemolytic type) can cause kernicterus but is treatable with phototherapy; conjugated jaundice is always pathological and indicates liver or biliary tract disease requiring specialist assessment. The practical bedside test is urine examination — conjugated bilirubin is water-soluble and stains the urine dark brown/orange. A urine dipstick positive for bilirubin in a jaundiced neonate is a same-day hospital emergency. Galactosaemia is a critical metabolic cause of conjugated jaundice — it is caused by deficiency of galactose-1-phosphate uridyltransferase (GALT enzyme) and presents in the first few days of life with jaundice, vomiting, hepatomegaly, hypoglycaemia, and E. coli sepsis (galactose accumulation impairs neutrophil function, predisposing to sepsis). It is included in the UK newborn bloodspot screening programme (heel prick test). Treatment is immediate lactose-free formula. If untreated, it causes liver failure, cataracts, intellectual disability, and ovarian failure in girls. Any sick jaundiced neonate with E. coli sepsis must have galactosaemia excluded immediately — withhold breast milk / lactose-containing formula while results are awaited.

NICE NG98 explicitly states that clinical assessment alone is insufficient for managing neonatal jaundice — the Kramer zones have a sensitivity of only 35% for identifying neonates requiring phototherapy. All neonates with visible jaundice should have their bilirubin measured with TcB (transcutaneous bilirubinometer) or SBR (serum bilirubin), not managed on clinical observation alone. Transcutaneous bilirubinometers (Dräger JM-103, Philips BiliChek) are now widely available in NHS community settings — they use spectrophotometry to estimate bilirubin through the skin and are validated for screening in infants ≥35 weeks' gestation. They allow safe, non-invasive, rapid assessment in any setting. GPs and community midwives who have access to a TcB should use it for all jaundiced neonates presenting in the community. The direct antiglobulin test (DAT) — also called the Coombs test — detects antibody-coated red cells. A positive DAT in a jaundiced neonate confirms immune-mediated haemolysis (Rh incompatibility, ABO incompatibility). A negative DAT in a severely jaundiced neonate should prompt consideration of non-immune haemolysis: G6PD deficiency (red cell enzyme assay), hereditary spherocytosis (osmotic fragility test, ektacytometry), or pyruvate kinase deficiency.

The NICE NG98 threshold charts provide gestational-age-specific phototherapy and exchange transfusion thresholds — the treatment threshold is lower (more conservative) for preterm infants because the blood-brain barrier is less mature and kernicterus can occur at lower bilirubin levels. There are separate charts for infants ≥38 weeks, 35–37+6 weeks, and for infants with additional risk factors (haemolysis, previous sibling requiring exchange transfusion, albumin <25 g/L). The NICE NG98 chart is available as a free app and should be on every GP and midwife smartphone. Referral thresholds: if SBR is at or above the phototherapy line → same-day hospital for phototherapy. If SBR is within 50 μmol/L of the phototherapy line → same-day hospital for review and repeat measurement. The Kasai procedure timing is the most critical factor in biliary atresia outcome — the Jaundice Free Liver campaign has shown that implementing stool colour card screening at the 6-week neonatal check reduces diagnostic delay for biliary atresia. Every GP performing a 6-week check should ask specifically about stool colour and inspect the stool colour card. Pale stools at 6 weeks in a jaundiced infant = same-day referral regardless of how well the baby appears.

The blue light of phototherapy (peak 460 nm) penetrates skin and converts unconjugated bilirubin IXα (the brain-toxic isomer) to lumirubin and configurational isomers that are water-soluble and can be excreted in bile without hepatic conjugation. This bypasses the conjugation defect of the immature neonatal liver. The efficiency of phototherapy is directly proportional to the irradiance (light intensity) and the skin surface area exposed — hence the use of double-surface phototherapy (baby between overhead lights and underneath billi blanket) for higher bilirubin levels. Home phototherapy (billi blankets) is used in some NHS trusts for stable, term infants with physiological jaundice, reducing hospitalisation and supporting breastfeeding. NICE NG98 does not endorse or prohibit home phototherapy — it acknowledges it as an option in some settings. The critical point is that phototherapy cannot be started in primary care without bilirubin measurement and threshold plotting — empirical phototherapy without measurement is not appropriate. The feeding message is critically important: stopping breastfeeding for jaundice is the single most common parental error — parents frequently stop breastfeeding because they have been told that breast milk "causes" jaundice. Breast milk jaundice (prolonged unconjugated jaundice in breastfed infants) is benign and does not cause kernicterus at the levels typically seen. Breastfeeding must be actively supported and continued unless there is a medical reason to supplement.

The GP's role in the 6-week check is particularly important for prolonged jaundice — it is the last formal clinical encounter before the baby is seen again at 8 weeks immunisations. NICE NG98 is explicit: if a baby is still visibly jaundiced at the 6-week check, a split bilirubin must be checked and the stool colour card reviewed. A baby who has been yellow since birth and whose parents "think it's breast milk jaundice" must not be dismissed without investigation. The consequences of missing biliary atresia at the 6-week check and reassuring parents without a split bilirubin are catastrophic — by the time the baby is seen at 8 weeks and jaundice is still present, the Kasai window may be closing. Galactosaemia warrants specific mention at the 6-week check — check the newborn bloodspot result in the red book. The bloodspot screens for galactosaemia (GALT deficiency), phenylketonuria (PKU), hypothyroidism, sickle cell disease, cystic fibrosis, MCADD, homocystinuria, isovaleric acidaemia, glutaric aciduria type 1, and MSUD. GPs should confirm at every early neonatal contact that the heel prick was taken (day 5) and the result was normal.

The breastfeeding message in neonatal jaundice is one of the most common sources of parental misinformation — many parents are told by well-meaning relatives, and sometimes by healthcare professionals, to "stop breastfeeding for 24–48 hours to see if the jaundice improves." This "breastfeeding holiday" test has no evidence base and causes measurable harm: it disrupts breastfeeding establishment during the critical first two weeks, causes maternal engorgement and mastitis, reduces milk supply, and increases formula dependence. The only indication to temporarily supplement breastfeeding is objective evidence of inadequate intake (weight loss >10%, insufficient wet nappies) combined with a rising bilirubin — and even then, the supplement should be expressed breast milk first, formula second, and breastfeeding itself should continue simultaneously. G6PD deficiency is globally the most common enzyme deficiency (400 million affected worldwide) and is highly prevalent in populations from Africa, the Mediterranean, Middle East, and South/Southeast Asia. GPs seeing neonates from these backgrounds should have a lower threshold for G6PD testing in unexplained haemolytic jaundice. G6PD-deficient males can have catastrophic haemolytic crises when exposed to fava beans, certain drugs, or infections even in adult life — parental education at the time of diagnosis has lifelong benefit.

The 6-week GP check is the last safety net for biliary atresia in the UK — after this point, the next routine contact is at 8 weeks immunisations, at which point the Kasai surgery window is closing. UK national audit data (BAPS-CASS study) show that 25% of biliary atresia cases in the UK are still diagnosed after 60 days, with correspondingly worse outcomes. This diagnostic delay is almost entirely attributable to missed clinical assessment at the 6-week check. The BAPS (British Association of Paediatric Surgeons) and the Jaundice Free Liver charity have produced specific GP-targeted education on recognising biliary atresia at the 6-week check. GPs should treat a persistently jaundiced baby at the 6-week check as a potential biliary atresia until proven otherwise — the split bilirubin result and stool colour assessment are non-negotiable. The neurodevelopmental sequelae of kernicterus require long-term follow-up — kernicterus causes the classic tetrad: (1) athetoid/dyskinetic cerebral palsy, (2) sensorineural hearing loss (auditory neuropathy), (3) oculomotor palsy (particularly upgaze palsy), and (4) intellectual disability (variable). Early identification and intervention (cochlear implants for deafness, physiotherapy, SENCO) significantly improve outcomes. GPs should ensure neurodevelopmental review appointments are in place for any neonate who had severe hyperbilirubinaemia (>425 μmol/L) or signs of encephalopathy.