About the Pediatric End-Stage Liver Disease (PELD) Score
PELD ranks children under 12 years on the liver transplant waiting list using five items — bilirubin, INR, albumin, age under one year, and growth failure — and, unlike the adult MELD, it contains no creatinine, because kidney function was not retained in the published paediatric model. Scores of 10 or below indicate lower severity, 11–20 moderate severity with increasing waitlist priority, and above 20 high severity warranting urgent transplant assessment. Its central and well-documented weakness is that it underestimates paediatric waitlist mortality: children frequently need PELD exception points to achieve appropriate priority, and a revised version adding sodium and creatinine (PELD-Na-Cr) raised the area under the curve from 0.799 to 0.854.
Formula
PELD = 10 × [0.480 × ln(bilirubin mg/dL) + 1.857 × ln(INR) − 0.687 × ln(albumin g/dL) + 0.436 (if age < 12 months) + 0.667 (if growth failure)]- bilirubin
- Total bilirubin in mg/dL.
- INR
- International normalised ratio. Its coefficient of 1.857 is the largest, so coagulopathy moves PELD most.
- albumin
- Serum albumin in g/dL, entering negatively — the lower the albumin, the higher the score.
- age < 12 months
- A fixed 0.436 added for infants, reflecting their higher waitlist mortality.
- growth failure
- A fixed 0.667 added when height or weight is more than two standard deviations below the mean for age.
- There is no creatinine term. Renal function is not part of the published paediatric model, which is the single biggest structural difference from adult MELD.
- The age term is a step, not a gradient — it applies in full below 12 months and not at all above, so a score can drop on a birthday without any change in the child's condition.
- The whole bracket is multiplied by 10 and the result rounded, which is why small laboratory changes can move the reported integer.
- Growth failure is defined anthropometrically (more than two standard deviations below the mean for height or weight), not by clinical impression.
Interpreting the result
A score of 10 or below indicates lower severity, and management continues as paediatric hepatology follow-up with close attention to growth and nutrition. Scores of 11–20 signal moderate severity and increasing waitlist priority; nutrition is a modifiable determinant of outcome at this stage, so a paediatric dietitian should be involved rather than merely consulted. Above 20 the child is at high severity and transplant assessment should be urgent. The critical interpretive caveat is that a reassuring PELD is less reassuring than the equivalent adult MELD: because the score systematically under-ranks paediatric waitlist mortality, children commonly require exception points, and a clinically deteriorating child with a modest score should trigger an exception application rather than watchful waiting. Analysis found the original PELD would need roughly nine additional points to equalise children's mortality risk against the age-standardised adult rate.
| Score | Band | What it means | Action |
|---|---|---|---|
| ≤ 10 | Lower severity | Lower predicted risk of death or transfer to intensive care | Continue paediatric hepatology follow-up with growth and nutrition monitoring |
| 11–20 | Moderate severity | Increasing waitlist priority | Ensure transplant assessment is under way; involve a paediatric dietitian, as nutrition is modifiable at this stage |
| > 20 | High severity | High predicted risk of death or intensive-care transfer | Urgent transplant assessment; consider exception points, as PELD under-represents severity in several paediatric conditions |
Scroll the table sideways for every column.
What the PELD / CR Score needs (5 inputs)
- Total bilirubin
- Serum total bilirubin, entered in mg/dL or µmol/L.
- INR
- International normalised ratio — the largest coefficient in the model.
- Albumin
- Serum albumin, which enters with a negative coefficient so that a lower albumin raises the score.
- Age (months)
- Age under 12 months adds a fixed 0.436 to the linear predictor. PELD applies to candidates under 12 years.
- Growth failure
- Height or weight more than two standard deviations below the mean for age; adds a fixed 0.667 to the linear predictor.
Units. The published coefficients expect bilirubin in mg/dL and albumin in g/dL. To convert: bilirubin µmol/L ÷ 17.1 = mg/dL; albumin g/L ÷ 10 = g/dL. This calculator accepts either unit and converts internally.
What it returns
- PELD score
- The rounded score used for waitlist priority.
- Severity band
- Lower, moderate, or high severity.
How it is calculated
PELD was derived from the Studies of Pediatric Liver Transplantation database, a consortium of 29 centres across the United States and Canada, restricted to children with chronic liver disease listed for a first transplant. Two endpoints were modelled: death, and the composite of death or transfer to intensive care. Seventeen candidate factors were narrowed to five that predicted those outcomes — bilirubin, INR, albumin, infancy, and growth failure — with the first three entering on a natural-log scale and the last two as fixed increments. The resulting linear predictor is multiplied by ten and rounded, placing PELD on a scale broadly comparable to MELD so that paediatric and adult candidates can be ranked against one another in a shared allocation system, which is precisely where its calibration problems later surfaced.
Facts & figures
| PELD | Adult MELD | |
|---|---|---|
| Age range | Under 12 years | 12 years and over |
| Bilirubin | Yes | Yes |
| INR | Yes | Yes |
| Albumin | Yes | No |
| Creatinine | No | Yes |
| Growth failure | Yes | No |
| Infancy (< 12 months) | Yes (+0.436) | Not applicable |
Scroll the table sideways for every column.
The two are scaled to be comparable so children and adults can be ranked in one system — but PELD's known underestimation of paediatric waitlist mortality means comparable numbers have not meant comparable risk.
| Model | Cross-validated AUC ROC |
|---|---|
| Original PELD | 0.799 |
| PELD-Na-Cr (adds sodium and creatinine) | 0.854 |
Scroll the table sideways for every column.
Derived in 5,111 children under 12 listed 2005–2017 with 339 waitlist mortality events. The revised model required 9.44 additional points to equalise children's mortality risk with the age-standardised adult rate.
Evidence
Derivation — McDiarmid (SPLIT)
2002 · n = 884Data from the Studies of Pediatric Liver Transplantation consortium of 29 US and Canadian centres, restricted to children with chronic liver disease listed for a first transplant. Two pre-transplant endpoints were evaluated: death (n = 884) and death or transfer to intensive care (n = 779). Seventeen candidate factors were reduced to five.
A model using five objective parameters — bilirubin, INR, albumin, age under one year and growth failure — accurately predicted death, or death and transfer to intensive care, in children awaiting liver transplantation.
Recalibration with sodium and creatinine — Braun
2021 · n = 5,1115,111 children under 12 years with chronic liver disease listed for deceased-donor livers between January 2005 and December 2017, followed from listing to waitlist mortality (death, or removal for being too sick to transplant; 339 events) or 180 days. Linear splines were fitted to the existing PELD components and sodium and creatinine were added.
The updated PELD-Na-Cr achieved a cross-validated AUC ROC of 0.854 against 0.799 for the original PELD, and required 9.44 additional points to equalise children's mortality risk with the age-standardised adult rate — quantifying the long-recognised finding that the original score underestimates paediatric waitlist mortality.
How it compares
PELD / CR Score vs MELD-Na (adult)
PELD and adult MELD-Na are scaled to be compared but are not built from the same variables — PELD swaps creatinine for albumin, growth failure and infancy, and has been shown to under-rank paediatric mortality relative to the adult scale.
MELD-Na relies on bilirubin, INR, creatinine and sodium; PELD uses bilirubin, INR, albumin, infancy and growth failure. The rationale for the divergence is sound — creatinine is an insensitive marker of renal function in small children, while growth and albumin carry real paediatric prognostic weight. The consequence is that equal numbers have not meant equal risk: registry analysis found the original PELD needed roughly nine extra points to bring children's mortality risk into line with the age-standardised adult rate, which is why exception points became routine.
PELD / CR Score vs Child-Pugh score
Child-Pugh grades cirrhosis severity in broad classes and was never designed for children or for allocation; PELD is the paediatric waitlist instrument and is built from objective, verifiable variables for exactly that reason.
Child-Pugh's subjective ascites and encephalopathy items make it unsuitable for rationing, and encephalopathy in particular is difficult to grade reliably in infants and young children. PELD deliberately admits only measurable parameters. Child-Pugh retains descriptive value for paediatric chronic liver disease severity in clinical conversation, but it plays no part in determining waitlist priority.
Pearls & pitfalls
- PELD contains no creatinine. Despite the widely used 'PELD-Cr' label, kidney function is not part of the published paediatric model — a creatinine-containing variant is a separate, later recalibration rather than standard PELD.
- The infancy term is a cliff edge: it adds 0.436 to the linear predictor below 12 months and nothing at all above, so a child's score falls at their first birthday without any change in their liver.
- A low PELD in a visibly unwell child is a reason to apply for exception points, not to wait. Underestimation of paediatric waitlist mortality is a documented property of the score, not an occasional anomaly.
- Growth failure is an anthropometric definition — height or weight more than two standard deviations below the mean for age — so it needs a plotted measurement, not an impression.
- Albumin enters negatively, so albumin infusion lowers the score without improving the underlying disease.
- PELD stops at 12 years. Adolescents are scored on the adult MELD family, which changes which variables count and can shift apparent priority abruptly around that boundary.
Critical actions
- Plot height and weight against population standards before scoring, rather than judging growth failure clinically — it carries a fixed 0.667 and is frequently missed.
- Pursue exception points where the clinical picture outruns the score; several paediatric conditions are known to be under-represented by PELD.
- Involve a paediatric dietitian early, particularly in the moderate band — nutritional status is one of the few modifiable determinants of transplant outcome.
- Recalculate at the intervals policy requires; PELD is a dynamic score and stale values misrepresent priority.
- Confirm the child's age in months and whether they are under 12 years before using PELD at all, since both the infancy term and the PELD/MELD boundary are step changes.
Why this score exists
The authors' brief was to build a score from parameters that were easily obtainable, objective and verifiable, because the purpose was allocation rather than bedside assessment — a score used to ration organs has to resist both error and gaming, which rules out anything requiring clinical judgement. That constraint explains what PELD includes and what it leaves out: growth failure made the cut because it is measurable against population standards, while the subjective severity markers a paediatric hepatologist would weigh did not. It also explains the score's enduring problem. Optimising for verifiability produced a model that ranks children reliably against each other but, as later registry analysis showed, not accurately against adults competing for the same organs.
About the creator
First author, 2002 derivation study
Derived PELD from the Studies of Pediatric Liver Transplantation registry, restricted to objective and verifiable parameters because the score was intended for organ allocation.
Limitations
- It underestimates paediatric waitlist mortality relative to adults competing for the same organs, which is why exception points are so frequently required — the central and long-recognised criticism of the score.
- It omits creatinine, so genuine renal dysfunction in an older child is invisible to it; a recalibrated PELD-Na-Cr adding sodium and creatinine performed considerably better (AUC 0.854 versus 0.799).
- The infancy term is binary rather than continuous, producing a discontinuity at 12 months that does not correspond to any biological threshold.
- It was derived on children with chronic liver disease and does not apply to acute liver failure, which is managed under separate urgent criteria.
- It under-represents severity in specific paediatric conditions — metabolic disease, hepatoblastoma, recurrent cholangitis in biliary atresia — where morbidity is not reflected in bilirubin, INR or albumin.
- Albumin is directly manipulable by infusion, and growth failure, once present, does not reverse quickly, so the score responds asymmetrically to intervention.
If you are the patient
PELD is a score used to work out how urgently a child under 12 with long-term liver disease needs a liver transplant, and how they should be prioritised on the waiting list. It uses three blood tests — bilirubin (which reflects jaundice), INR (a clotting test) and albumin (a protein made by the liver) — together with whether the child is under one year old and whether they are growing poorly for their age. Higher scores mean more urgent need. One important thing families often are not told: this score is known to understate how unwell children are compared with adults waiting for the same organs. That is why transplant teams can apply for extra 'exception' points when a child is clearly sicker than their number suggests. If your child seems unwell but has a low score, that is a recognised limitation of the score itself, and it is worth asking the team directly about exception points.
Frequently asked questions
What is the PELD score and who is it for?#
It is the score used to rank children under 12 years with chronic liver disease on the deceased-donor liver transplant waiting list. It combines bilirubin, INR, albumin, being under 12 months old, and growth failure.
Does PELD include creatinine?#
No. Standard PELD contains no creatinine — kidney function was not retained in the published paediatric model, largely because creatinine is an insensitive marker of renal function in small children. A later recalibration, PELD-Na-Cr, does add sodium and creatinine, but it is a separate model.
What is a high PELD score?#
Above 20 indicates high severity and warrants urgent transplant assessment. Scores of 11–20 indicate moderate severity with increasing waitlist priority, and 10 or below lower severity.
Why do children need PELD exception points?#
Because PELD underestimates paediatric waitlist mortality relative to adults on the same list. Registry analysis found the original score would need roughly nine additional points to equalise children's mortality risk with the age-standardised adult rate, so exception points close a gap in the score itself.
What counts as growth failure in PELD?#
Height or weight more than two standard deviations below the mean for the child's age. It must be measured and plotted against population standards, and it adds a fixed 0.667 to the linear predictor.
What happens to the score when a child turns one?#
The infancy term, worth 0.436 in the linear predictor, is removed. The score therefore falls at 12 months of age even though nothing about the child's liver disease has changed — a known discontinuity in the model.