Big liver, big spleen: the organomegaly numbers measured in infants, not pooled from children
also called hepatomegaly in infant ALL · splenomegaly in infant ALL · hepatosplenomegaly infant frequency · organomegaly at diagnosis
In cohorts made only of babies under one year, an enlarged liver or spleen was found in roughly 65-88% of them, which is as common as or commoner than the 64%/61% figure pooled from children of all ages. But a small liver edge is normal in a healthy baby, so the sign is least reliable in exactly the age group where it is most often recorded.
The same thing, for a scientist
The graph's existing organomegaly entry takes its frequencies from Clarke et al.'s meta-analysis of 33 childhood cohorts aged 0-18 (n = 3084): hepatomegaly 64%, splenomegaly 61%, hepatosplenomegaly 42% [E1]. Infants are a small minority of that denominator. Three cohorts restricted to the first year of life give the infant-specific figures: the Tokyo Children's Cancer Study Group series of 62 infants with ALL recorded hepatomegaly in 40 (reported as 70.0%) and splenomegaly in 40 (reported as 68.3%) [E2]; a 144-case Chinese single-centre infant leukaemia series recorded hepatosplenomegaly in 87.5% [E3]; and a 20-year single-hospital US series of 9 infants recorded hepatosplenomegaly in 6 of 8 assessed [E4]. Neonatal disease specifically is described as presenting with hepatosplenomegaly, leukaemia cutis and/or hyperleucocytosis as its characteristic triad [E5]. The direction of the correction is therefore upward, not downward - but the sign's specificity falls as the frequency rises, because a liver edge palpable 1-2 cm below the costal margin and a spleen tip palpable in the first months of life are normal findings in healthy infants, and none of these cohorts states the threshold in centimetres at which it called an organ enlarged.
What the infant-only cohorts actually report1 deeper
| Population | n | Hepatomegaly | Splenomegaly | Hepatosplenomegaly |
|---|---|---|---|---|
| Children 0-18, 33 pooled cohorts 1 | 3084 | 64% | 61% | 42% |
| Infants <1 y with ALL, TCCSG Japan 2 | 62 | 40 patients, reported 70.0% | 40 patients, reported 68.3% | not stated separately |
| Infant leukaemia (ALL + AML + MPAL), single centre China 3 | 144 | not stated separately | not stated separately | 87.5% |
| Infant ALL, single US hospital, all <6 months 4 | 9 | not stated separately | not stated separately | 6 of 8 |
The three infant cohorts are small, retrospective and span 1977-2014 across three countries, so they are not a pooled estimate and should not be read as one. What they establish is a direction: organomegaly is at least as frequent in infants as in the pooled childhood figure, and probably more so.
An arithmetic discrepancy worth flagging
In the TCCSG report the same count - 40 patients - is given as 70.0% for hepatomegaly and 68.3% for splenomegaly 2. Those percentages imply denominators of about 57 and 59 respectively, not the 62 infants in the cohort. The abstract does not state how many infants were examined for each sign. The honest reading is 40/62 = 64.5% as a lower bound for each, with the published percentages reproduced as printed and the discrepancy noted. This graph reports both rather than silently picking one.
Why a frequency this high is less useful than it looks
Three things degrade the sign in this age group, and none of them is captured by a percentage:
- The normal baseline is not zero. A liver edge 1-2 cm below the right costal margin is a normal finding through infancy, and a spleen tip is palpable in a substantial minority of healthy neonates. A cohort that codes "hepatomegaly: yes/no" without a centimetre threshold is measuring the examiner as much as the child.
- Nobody states the threshold. None of 2, 3 or 4 gives the size criterion used. Comparing 70% to 64% across studies that may have used different cut-offs is not a comparison.
- The abdomen is easier to feel in an infant than in a teenager, which biases the infant frequency upward independently of tumour burden.
The sign that carries real information in an infant is not the presence of a palpable liver but a hard, several-centimetre, rapidly enlarging one, and no cohort in this beat reports that stratified by centimetres.
Correction to the foundation of this beat
concept-hepatosplenomegaly-infant-leukemia sources its headline frequencies from 1, an age 0-18 meta-analysis, and applies them to infants. That entry is not wrong about the pooled number; it is applying it outside the population it estimates. The infant-specific figures above should be its headline, with the pooled number retained as the comparator. This entry does not edit that file - the coordinator owns that correction - but records what the replacement numbers are and where they come from.
| metric | presenting sign frequency |
|---|---|
| population | infants <1 year with acute leukaemia |
| comparator population | children 0-18 with leukaemia |
pooled childhood
{
"hepatomegaly_pct": 64,
"splenomegaly_pct": 61,
"hepatosplenomegaly_pct": 42,
"n": 3084,
"source_pmid": "27647842"
}infant cohorts — the numbers behind the chart
[
{
"cohort": "TCCSG Japan infant ALL",
"n": 62,
"hepatomegaly_pct_reported": 70,
"splenomegaly_pct_reported": 68.3,
"pmid": "10530057",
"title": "Clinical and biological aspects of acute lymphoblastic leukemia in 62 infants: retrospective analysis of the Tokyo Children's Cancer Study Group.",
"authors": "Isoyama K, Okawa H, Hayashi Y, et al.",
"venue": "Pediatr Int",
"year": 1999,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"cohort": "single centre China infant leukaemia",
"n": 144,
"hepatosplenomegaly_pct": 87.5,
"pmid": "26714909",
"title": "[Clinical analysis of 144 cases of infant leukaemia].",
"authors": "Wen X, Guan X, Xian Y, et al.",
"venue": "Nan Fang Yi Ke Da Xue Xue Bao",
"year": 2015,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"cohort": "single US hospital infant ALL",
"n": 9,
"hepatosplenomegaly": "6 of 8",
"pmid": "18800262",
"title": "Infant acute lymphoblastic leukemia: a 20-year children's hospital experience.",
"authors": "Murray RA, Thom G, Gardner RV, et al.",
"venue": "Fetal Pediatr Pathol",
"year": 2008,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]- Has any infant ALL cohort recorded liver and spleen size in centimetres below the costal margin rather than as a binary, so that the sign can be separated from normal infant anatomy?
- Does organomegaly at diagnosis carry independent prognostic information in infants once white cell count is in the model, or is it simply a proxy for tumour burden?
- Is the 70% versus 87.5% gap between the Japanese and Chinese infant cohorts definitional, lineage-driven (the Chinese series includes AML) or real?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review
In a meta-analysis of 33 childhood leukaemia cohorts (n = 3084, ages 0-18), hepatomegaly was present in 64%, splenomegaly in 61% and hepatosplenomegaly in 42%.
Clinical presentation of childhood leukaemia: a systematic review and meta-analysis. · Clarke RT, Van den Bruel A, Bankhead C, Mitchell CD, Phillips B, Thompson MJ · Arch Dis Child 2016What this evidence cannot show
All 33 included studies were cohort studies without control groups, so these are frequencies among diagnosed cases, not diagnostic performance. The age range is 0-18 and infants are a small minority of the denominator; the paper does not present an infant-stratified analysis. Frequencies from case series are inflated by ascertainment: a sign is recorded when someone looked for it.
Evidence that argues against this (1)
- The same paper's own design is the counter-argument to using it here: it pools across 0-18 with no age stratum, so it cannot speak to infants, and the infant-only cohorts [E2][E3] give higher figures. Wen X et al. 2015
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In 62 infants under 1 year with ALL treated at 30 Tokyo Children's Cancer Study Group institutions between 1977 and 1995, hepatomegaly was found in 40 patients (reported as 70.0%) and splenomegaly in 40 patients (reported as 68.3%).
Clinical and biological aspects of acute lymphoblastic leukemia in 62 infants: retrospective analysis of the Tokyo Children's Cancer Study Group. · Isoyama K, Okawa H, Hayashi Y, Hanada R, Okimoto Y, Maeda M, et al. · Pediatr Int 1999What this evidence cannot show
Retrospective, 1977-1995, pre-molecular era (MLL rearrangement was tested in only 13 of 62). The two printed percentages are inconsistent with a denominator of 62 and the paper does not state how many infants were examined for each sign. No size threshold for 'enlarged' is given. Abstract read in full; full text not retrieved.
Evidence that argues against this (1)
- A later and larger infant series from a different country reports a considerably higher hepatosplenomegaly frequency (87.5% of 144), so the TCCSG figure is not a stable estimate. Wen X et al. 2015
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Among 144 infants with acute leukaemia at a single Chinese centre (1993-2014), 87.5% had hepatosplenomegaly, and six patients under 6 months of age had skin infiltration.
[Clinical analysis of 144 cases of infant leukaemia]. · Wen X, Guan X, Xian Y, Dou Y, Guo Y, Xiao J, et al. · Nan Fang Yi Ke Da Xue Xue Bao 2015What this evidence cannot show
Single centre, retrospective over 21 years, mixed lineage (83 ALL, 55 AML, 1 hybrid, 5 unclassifiable) so not an ALL-specific figure. Most patients abandoned treatment after diagnosis, which selects the cohort towards those who presented floridly. Chinese-language paper; the English abstract was read, the full text was not. No size threshold given.
Evidence that argues against this (1)
- The TCCSG infant ALL cohort reports about 70%, roughly 18 points lower, on a purely lymphoid population - the gap is as likely to be definitional as biological. Isoyama K et al. 1999
- E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
In a 20-year single-hospital series of 9 infants with ALL (all under 6 months), 6 of 8 assessed had hepatosplenomegaly and white cell counts ranged from 42,000 to 1.6 million per microlitre.
Infant acute lymphoblastic leukemia: a 20-year children's hospital experience. · Murray RA, Thom G, Gardner RV, Craver RD · Fetal Pediatr Pathol 2008What this evidence cannot show
n = 9. The denominator for the organomegaly figure is 8, not 9, and the paper does not say why one infant is missing. Single institution, pathology-department review.
Evidence that argues against this (1)
- At n = 9 the 95% confidence interval on 6/8 runs roughly from 35% to 97%, so this series is compatible with almost any of the other reported frequencies and discriminates between none of them. Isoyama K et al. 1999
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review
Neonatal leukaemia (first month of life) characteristically presents with hepatosplenomegaly, leukaemia cutis and/or hyperleucocytosis, and congenital infection must be excluded before the diagnosis is made.
Diagnosis and management of neonatal leukaemia. · van der Linden MH, Creemers S, Pieters R · Semin Fetal Neonatal Med 2012What this evidence cannot show
A narrative review rather than a cohort, so 'characteristically' is not a measured frequency. It covers neonatal AML and ALL together and neonatal AML is the commoner of the two, so the triad is weighted towards myeloid disease.
Evidence that argues against this (1)
- A 2026 neonatal case with the full clinical triad had no KMT2A rearrangement on molecular testing, so the triad does not identify the KMT2A-rearranged biology this graph is about. Bustamante-Ordonez MA et al. 2026
infant-specific organomegaly frequency replaces the age 0-18 pooled figure
the pooled 64%/61% figures are estimated outside the infant population they are applied to
- ← related toHypothesis: the spread in infant presenting-feature frequencies is mostly measurement, not biology
the organomegaly spread with no stated threshold