Newborn jaundice — what's normal, when to worry

The question

My baby looks yellow — is this normal, and when does jaundice need treatment?

Short answer

Jaundice is extremely common — about 60 in 100 term babies and 80 in 100 premature babies develop some yellowing, usually peaking around days 3–5 [2]. In most cases it is harmless and fades on its own within a week or two. But bilirubin, the pigment causing the yellow colour, can damage the brain at very high levels — so jaundice in the first 24 hours, or deepening jaundice, needs prompt bilirubin measurement, and treatment (usually light therapy) starts when levels cross thresholds based on the baby's age in hours [1].

What the strongest evidence says

How common it is

NHS parent guidance reports jaundice in about 60% of term babies and 80% of babies born before 37 weeks [2] — guideline epidemiology rather than a single cohort study, but consistent with the clinical picture that most newborns get some jaundice - usually appearing on days 2-4, peaking around days 3-5, and resolving on its own after 1-2 weeks [5].

Why it happens

Jaundice is bilirubin deposited in the skin. Most newborn jaundice comes from two things happening at once: increased red blood cell breakdown (newborns turn over red cells fast) and decreased bilirubin excretion (the newborn liver is still learning to clear it) [5]. Breastfeeding, haemolysis (e.g. blood-group incompatibility), and some metabolic or genetic conditions raise the risk further [5].

When to act is decided by numbers, not colour

NICE guidance (CG98) sets treatment thresholds using bilirubin level, age in hours, gestational age, and risk factors — not how yellow the baby looks [1]. For a well baby born at ≥38 weeks, phototherapy starts above roughly [1]:

• 200 µmol/L at 24 hours
• 250 µmol/L at 48 hours
• 300 µmol/L at 72 hours
• 350 µmol/L at 96 hours or later

For babies born before 38 weeks, the treatment chart is different (thresholds are lower) [1]. Two other risk factors — a sibling who needed phototherapy, and the mother intending to breastfeed exclusively — don't change the thresholds themselves; they mean bilirubin gets rechecked sooner (within 18 hours rather than 24) when levels are approaching the threshold, because they predict significant jaundice [1].

These numbers are guides for your clinician — the NICE age-in-hours chart is the tool, not this page. Never try to interpret a bilirubin result against the list above on your own: the thresholds interact with age in hours and risk factors [1].

The first-24-hours rule

Jaundice visible within the first day of life is different: NICE says measure bilirubin within 2 hours and assess urgently, because early jaundice is more likely to have an underlying cause [1].

Two kinds of breastfeeding-related jaundice

The 2022 American Academy of Pediatrics guideline carefully separates them [4]: early suboptimal-intake jaundice — not enough milk getting in, often peaking days 3–5 — from later breast-milk jaundice, which persists for weeks despite good feeding and weight gain and is usually benign. Confusing the two causes unnecessary guilt in one direction and missed feeding problems in the other. The AAP also recommends against giving water or sugar water for jaundice [4].

What treatment does — and what it hasn't been proven to do

Phototherapy (blue light) lowers bilirubin — shown against no treatment in randomized trials [5]. How much it prevents brain injury is less directly proven: trials measured bilirubin levels, not brain outcomes, and most didn't report kernicterus [5]. Exchange transfusion (replacing the baby's blood) is the rescue treatment; remarkably, no randomized trial has ever tested it against no treatment — its use rests on consensus and the fact that it reliably removes bilirubin [5].

How rare the worst outcome is

Severe high bilirubin (>510 µmol/L) occurred in 7.1 per 100,000 UK live births in prospective national surveillance; of 108 such cases, 14 showed signs of bilirubin brain injury [3]. Estimates of actual kernicterus (lasting brain damage) in high-income countries range from about 1–2 per 100,000 births [6] — but definitions vary between sources, so treat any single number with caution.

What it means for the parents

Jaundice turns the first week into a sequence of checks — midwife visits, heel-prick bilirubin tests, maybe a readmission — at exactly the moment you're learning to feed a newborn. A few evidence-grounded realities:

• The checks are the burden, and that's normal. Frequent bilirubin rechecks (within 18 hours if risk factors, 24 hours otherwise [1]) mean disrupted days and broken sleep for you. NICE explicitly tells clinicians to offer reassurance that jaundice is usually transient and harmless and to avoid causing unnecessary anxiety [1] — if your care team is alarming rather than reassuring, that's a care problem, not a you problem.
• Phototherapy separates you from your baby. Treatment usually means the baby lying under lights, often in hospital, with feeds continuing but cuddles interrupted. No trial has measured what this does to parent–infant bonding [5] — the absence of evidence isn't evidence of absence, and it's okay to find it upsetting.
• Breastfeeding and jaundice collide emotionally. Early jaundice is often tangled up with feeding difficulties, and the AAP's distinction matters here: if the issue is suboptimal intake, the answer is feeding support, not guilt; if it's breast-milk jaundice with a thriving baby, it usually needs nothing but time [4]. Either way, NICE says breastfeeding can usually continue through jaundice and its treatment [1].
• Pale stools and dark urine change the picture. Jaundice lasting beyond 2 weeks (3 weeks if your baby was born before 37 weeks), or with pale stools and dark urine, needs assessment for rarer causes (including liver conditions like biliary atresia) [1][7] — this is a different pathway from ordinary newborn jaundice, and worth raising rather than waiting out.
• Ask the number-based question. When the midwife looks at your yellow baby, ask: "Is this within the normal range for her age in hours?" It invites the answer the guidance actually runs on — bilirubin, age in hours, risk factors — rather than a vague reassurance [1].
• If readmission happens, plan for it. Phototherapy usually means an unexpected hospital stay in the first week. Keep a small bag packed, decide in advance which parent stays overnight, and line up cover for anything else (pets, other children, work) — tired decisions at midnight are worse decisions.

What remains uncertain

• Exactly how much phototherapy reduces brain injury (trials measured bilirubin, not brain outcomes) [5].
• The long-term harms of phototherapy: the systematic-review verdict is "insufficient evidence" [5], and the 2022 AAP guideline notes some evidence of a small increased risk of subsequent childhood epilepsy after phototherapy — judging the benefit to outweigh that small potential risk when bilirubin is at treatment thresholds [4]. Do not let anyone tell you it is proven risk-free.
• The true current incidence of kernicterus: estimates vary by definition and source — about 1–2 per 100,000 births in high-income countries in one review [6], but definitions differ, so treat any single number with caution.
• The evidence suggests current thresholds and phototherapy keep brain injury rare — but the word is suggests, because the brain-injury evidence is observational (rating B).
• Home phototherapy (used in some countries) was not covered by this topic's evidence review — the evidence here concerns hospital-based phototherapy as practised under NICE guidance.

Benefits and risks in absolute terms

Benefits of monitoring and treatment:

• Most jaundice resolves without treatment. For the minority whose bilirubin climbs, phototherapy lowers it [5], and the UK's surveillance shows severe high bilirubin at only 7.1 per 100,000 births with 14 of 108 cases showing brain injury signs [3] — in a country where 60 in 100 term babies get some jaundice [2]. The system catches the dangerous tail.

Risks and downsides:

• Readmissions and rechecks: the price of catching rare danger is that many babies get extra tests and some get readmitted for phototherapy they may only marginally have needed — the thresholds are consensus-based, not trial-derived [1].
• Phototherapy's harms are incompletely quantified ("insufficient evidence" overall [5], with the small epilepsy signal noted in the AAP guideline [4]) — ask your clinician what's known and what's not rather than assuming it is risk-free.
• Exchange transfusion is invasive with real risks — an estimated mortality of 3–4 per 1,000 exchanged infants and 5–10% permanent sequelae in survivors (observational estimates) [5] — and is reserved for extreme levels [1][5].

The overall picture: a common, mostly harmless phenomenon with a rare dangerous tail — and a monitoring system designed to catch that tail early [1][3].

Practical considerations

• Look, don't guess. If your baby's skin or the whites of the eyes look yellow, mention it to your midwife — a visual check is the trigger for bilirubin measurement, not the diagnosis [1]. You can press gently on the skin and look for a yellow tinge where it pales as you let go — a technique both NICE and NHS parent guidance describe [1][2]. Also note feeding, wet nappies, and stool colour.
• Keep a note of when you first noticed yellowing and how it has changed. Clinicians will ask about timing and trajectory, and sleep-deprived memory is unreliable [1].
• Know what the bilirubin test involves. It's a small blood sample (heel prick or vein). Ask what the number was and which threshold they're comparing it against — you're entitled to understand the decision being made about your baby [1].
• Feeding is part of the treatment. Frequent effective feeds help clear bilirubin through the stools; if feeding is difficult, ask for feeding support early rather than struggling alone [4].
• What the pathway looks like if treatment is needed: a blood test confirms the bilirubin level → phototherapy in hospital, with the baby under the lights between feeds → repeat blood tests to track the level → home once it falls safely. You can usually keep feeding throughout [1][5]. Exchange transfusion is rare and reserved for extreme levels [1].
• Do not sunbathe your baby as treatment. Sunlight for jaundice has only very-low-certainty evidence from one small trial and no safety data [8] — use proper medical assessment instead.
• Never try to interpret bilirubin numbers yourself against the thresholds above — they interact with age in hours and risk factors, and your clinician uses the full chart [1].

When to talk to your doctor, midwife, or pediatrician

Urgently (same day): jaundice visible in the first 24 hours; rapidly deepening yellowing; baby very sleepy or hard to wake, floppy or unusually stiff, arching their back, a high-pitched cry, or not feeding well — these can be signs of bilirubin affecting the brain; pale stools with dark urine; jaundice persisting beyond 2 weeks (3 weeks if born before 37 weeks) [1][5][7]. Routinely: any yellowing you're unsure about at a midwife or health-visitor visit — checking is always reasonable, and NICE expects clinicians to reassess bilirubin whenever jaundice is a concern [1]. Trust the trajectory: yellowing that is deepening rather than fading after day 5 deserves a call even without other symptoms.

References

1. NICE Clinical Guideline CG98. Jaundice in newborn babies under 28 days. https://www.nice.org.uk/guidance/cg98/chapter/recommendations — A
2. Norfolk and Norwich University Hospital NHS. Jaundice in newborn babies (parent leaflet). https://www.nnuh.nhs.uk/publication/download/jaundice-in-newborn-babies-v4/ — C (guideline epidemiology, not a primary cohort)
3. Manning D et al. Prospective surveillance study of severe hyperbilirubinaemia in the newborn in the UK and Ireland. Arch Dis Child Fetal Neonatal Ed. 2007. https://data.parliament.uk/DepositedPapers/Files/DEP2008-1258/DEP2008-1258.pdf — B
4. Kemper AR et al. Clinical Practice Guideline Revision: Management of Hyperbilirubinemia in the Newborn Infant 35 or More Weeks of Gestation. Pediatrics. 2022;150(3):e2022058859. https://publications.aap.org/pediatrics/article/150/3/e2022058859/188726/Clinical-Practice-Guideline-Revision-Management-of — A
5. BMJ Clinical Evidence. Neonatal jaundice (systematic review). https://pmc.ncbi.nlm.nih.gov/articles/PMC3217664/ — B (trials measured bilirubin, not brain injury)
6. Hamza A. Kernicterus. Autops Case Rep. 2019;9(1):e2018057. PMID: 30863731. https://pmc.ncbi.nlm.nih.gov/articles/PMC6394357/ — C (secondary synthesis; definitions vary)
7. South West London NHS. Jaundice in babies (parent guidance, including prolonged-jaundice assessment). https://www.swlondon-healthiertogether.nhs.uk/download_file/775/3567 — D
8. Horn D et al. Sunlight for the prevention and treatment of hyperbilirubinemia in term and late preterm neonates. Cochrane Database Syst Rev. 2021;(7):CD013277. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD013277.pub2/abstract?cookiesEnabled — C (3 small RCTs at high risk of bias; very-low-certainty evidence)

Changelog

• 2026-09-08: Topic created (draft). Awaiting independent review. (Superseded 2026-09-09: author-review model, no external review before v1.)
• 2026-09-09: Full author review (v1 author-review model; reviewer Guille, executed by AI agent under his explicit delegation of 2026-09-09). Threshold numbers (200/250/300/350 µmol/L at 24/48/72/96h) re-verified against the NICE CG98 threshold table and kept — they demonstrate decisions are number-based, with the "not DIY" framing retained (decision 1). Risk-factor wording corrected: sibling phototherapy and exclusive-breastfeeding intent affect recheck timing (18h vs 24h), not thresholds — only <38 weeks gestation uses a lower threshold chart (verified against NICE 1.4.1). Phototherapy harms reframed: BMJ Clinical Evidence "insufficient evidence" verdict verified, plus the AAP 2022 note of a small possible childhood-epilepsy signal (benefit judged to outweigh it at thresholds). Added exchange-transfusion absolute risks (3–4 per 1,000 mortality; 5–10% permanent sequelae, observational estimates). Urgent red flags expanded with acute-bilirubin-encephalopathy signs (arching, high-pitched cry, unusual stiffness); prolonged-jaundice cutoff tightened to NICE's >14 days term / >21 days preterm. Restored the skin-blanching check technique (wrongly removed at second pass; NICE 1.2.5 + NHS leaflet). Kernicterus incidence kept at 1–2/100,000 with definitions-vary caveat (verified against Hamza 2019). Verdict: approve.