Neonatal Thrombocytopenia Interpreter
Neonatal Thrombocytopenia Interpreter
A low platelet count in a newborn is read by when it appeared at least as much as by how low it is. Under 72 hours of age it usually points at the pregnancy — placental insufficiency, growth restriction, pre-eclampsia, perinatal asphyxia. After 72 hours it usually points at sepsis or necrotising enterocolitis. And a severe, isolated thrombocytopenia in an otherwise well term infant is the one pattern that is neither, and the one that needs a specific treatment urgently. The transfusion threshold is lower than most people expect, and there is a randomised trial behind that.
Count, timing, and what it points at
Count + timing → likely causeA term infant, birth weight on the 50th centile, feeding well and clinically normal at 14 hours of age. A full blood count taken because of scattered petechiae noticed by the midwife shows a platelet count of 42 ×10⁹/L. No bleeding. The film confirms it and shows no clumps. The mother’s platelet count is 268 ×10⁹/L.
The bands, the timing rule, and the threshold that a trial moved
Timing — early onset is a count that first falls within 72 hours of birth; late onset is after 72 hours. Use the age at which it first fell, not today’s age.
Early onset points at placental insufficiency, intrauterine growth restriction, maternal pre-eclampsia, perinatal asphyxia, alloimmune or maternal immune disease, and early-onset sepsis.
Late onset points at sepsis and necrotising enterocolitis, which between them account for most of it, then drugs, thrombosis and the rarities.
Transfusion threshold — below 25 ×10⁹/L in a non-bleeding neonate. Above that, platelets are not routinely given. Major bleeding overrides the threshold entirely.
- 72 hours
- the discriminator, and it works because it separates causes that operated before birth from causes that arrived after it. It is not a biological boundary and a count that falls at 60 hours in an infant developing sepsis is a late-onset picture in everything but the clock. Use it as the first question, not the last
- 25 ×10⁹/L
- the transfusion threshold, and the number most likely to be wrong in a reader’s memory. A randomised trial of 660 preterm infants born before 34 weeks with severe thrombocytopenia compared thresholds of 50 and 25 ×10⁹/L. Death or major bleeding within 28 days occurred in 26 per cent (85 of 324) at the higher threshold and 19 per cent (61 of 329) at the lower one — odds ratio 1.57, 95% confidence interval 1.06 to 2.32. The more liberal threshold did harm, by about 8 percentage points, or roughly one extra death or major bleed for every 13 infants managed liberally
- isolated, severe, well
- the three words that should make you think alloimmune. A well term infant with a severe thrombocytopenia and nothing else wrong is not explained by placental insufficiency (which causes growth restriction), by asphyxia (which causes encephalopathy), or by sepsis (which causes illness). What is left is immune, and the alloimmune form is the one with a specific treatment and a specific risk
- the mother’s platelet count
- the single question that separates the two immune causes and it costs nothing. In neonatal alloimmune thrombocytopenia the mother’s count is normal, because the antibody is directed at an antigen she does not have. In maternal immune thrombocytopenia her count is low, and the infant’s is usually milder
Worked example
A term infant, birth weight on the 50th centile, feeding well and clinically normal at 14 hours of age. A full blood count taken because of scattered petechiae noticed by the midwife shows a platelet count of 42 ×10⁹/L. No bleeding. The film confirms it and shows no clumps. The mother's platelet count is 268 ×10⁹/L.
No major bleeding, so the overriding rule does not apply
The count first fell under 72 hours of age, so this is early onset
Term, appropriately grown, well, not septic — none of the usual early-onset explanations fits
42 is below 50, so this is severe by the usual classification, and it is isolated: no anaemia, no neutropenia, no coagulopathy
A well term infant with a severe isolated thrombocytopenia and a normal maternal count is the alloimmune presentation — the normal maternal count is what distinguishes it from maternal immune thrombocytopenia
So: cranial ultrasound now, maternal and paternal samples for HPA typing and maternal antibody screening, and a discussion about HPA-1a and HPA-5b negative platelets
Note what the answer is not: 42 is above 25, so this infant does not meet the routine transfusion threshold — but the threshold is not the question here, the diagnosis is, and a falling count in confirmed alloimmune disease is treated ahead of it
What the timing points at
| Early onset — under 72 hours | Late onset — 72 hours or more | |
|---|---|---|
| Commonest causes | Placental insufficiency: growth restriction, pre-eclampsia, maternal diabetes | Sepsis and necrotising enterocolitis, which between them account for most cases |
| Next | Perinatal asphyxia | Drugs — indometacin, ibuprofen, vancomycin, linezolid, heparin and others |
| Then | Alloimmune, or maternal immune thrombocytopenia | Thrombosis around a central line; congenital infection presenting late |
| Also | Early-onset bacterial sepsis; congenital infection | Inherited thrombocytopenia declaring itself |
| Typical severity | Mild to moderate, unless immune | Variable; a falling count tracks severity of the underlying illness |
| Typical course | Nadir around day 4, recovered by day 7 to 10 | Follows the underlying illness |
PlaNeT-2 — why the threshold is 25 and not 50
| Threshold 50 ×10⁹/L | Threshold 25 ×10⁹/L | |
|---|---|---|
| Infants randomised | 324 | 329 |
| Death or major bleeding within 28 days | 85 (26%) | 61 (19%) |
| Odds ratio (95% CI) | 1.57 (1.06 to 2.32), P = 0.02 | Reference |
| Absolute difference | About 8 percentage points of harm | — |
| Number needed to harm | About 1 extra death or major bleed for every 13 infants managed at the higher threshold | — |
Three questions, in order: is it real, when did it start, and is the baby well
Thrombocytopenia is the commonest haematological abnormality in a neonatal unit, and most of it needs no treatment at all. The first question is whether the count is real: a clotted or underfilled EDTA sample, platelet clumping and the rare EDTA-dependent pseudothrombocytopenia together account for more unexpected severe thrombocytopenias than any diagnosis, and the film settles all three in minutes. Only once the count is confirmed does the differential open.
The second question is timing, and it is the one that does most of the work. A count that first falls within 72 hours of birth reflects something that operated before delivery. The commonest cause by a distance is chronic fetal hypoxia from placental insufficiency — growth restriction, maternal pre-eclampsia, maternal diabetes — which suppresses fetal megakaryocytopoiesis and produces a mild to moderate count that reaches its nadir around day 4 and recovers by day 7 to 10 without any intervention. Perinatal asphyxia does something similar over a shorter course. A count that first falls after 72 hours reflects something that arrived after delivery, and the two dominant causes are late-onset sepsis and necrotising enterocolitis, with drugs, line-associated thrombosis and the rarities behind them.
The third question is whether the baby is well, and it exists to catch the one pattern that neither of the first two explains. A term infant, appropriately grown, feeding normally, with no sepsis and no asphyxia, who has a severe isolated thrombocytopenia in the first days of life, is not described by the placental story or the sepsis story. That infant has an immune cause, and the one that needs urgent specific treatment is neonatal alloimmune thrombocytopenia, in which maternal IgG against a paternally inherited platelet antigen — HPA-1a or HPA-5b in over 90 per cent of UK cases — crosses the placenta and destroys fetal platelets. The feared complication is intracranial haemorrhage, frequently antenatal, so the cranial ultrasound is urgent rather than routine. The single cheapest discriminator is the mother’s own platelet count: normal in alloimmune disease, low in maternal immune thrombocytopenia.
The transfusion threshold is lower than most clinicians expect and the reason is a randomised trial rather than a consensus. PlaNeT-2 randomised 660 preterm infants born before 34 weeks with severe thrombocytopenia to transfusion thresholds of 50 or 25 ×10⁹/L and found death or major bleeding within 28 days in 26 per cent at the higher threshold against 19 per cent at the lower — odds ratio 1.57, 95% confidence interval 1.06 to 2.32. The liberal arm did harm. Current guidance follows: transfuse a non-bleeding neonate below 25 ×10⁹/L, and do not transfuse routinely at or above it. Major bleeding, and a confirmed or suspected alloimmune case, are the situations in which the threshold stops being the question. The platelet count unit converter handles the units and the neonatal direct antiglobulin test interpreter the coincident haemolysis.
Frequently asked questions
What platelet count is normal in a newborn?
At or above 150 ×10⁹/L, at every gestation. The lower limit of normal is not lowered for prematurity, which surprises people — a count of 120 ×10⁹/L in a 27-weeker is thrombocytopenia in exactly the sense it would be in a term infant, even though it is common enough on a neonatal unit to look unremarkable. The usual severity bands are mild 100 to 150, moderate 50 to 99 and severe below 50 ×10⁹/L; a second convention splits the severe band again at 30 into severe and very severe. Nothing in management turns on which convention is used, because the transfusion threshold of 25 ×10⁹/L sits below both boundaries.
Why does it matter whether neonatal thrombocytopenia started before or after 72 hours?
Because the two groups have almost entirely different causes. Before 72 hours the explanation is usually something that operated in utero — placental insufficiency with intrauterine growth restriction or maternal pre-eclampsia, perinatal asphyxia, an immune cause, or congenital or early-onset infection. After 72 hours the dominant causes are late-onset sepsis and necrotising enterocolitis, followed by drugs and line-associated thrombosis. Use the age at which the count first fell, not the age today: a count that was already low on day one and is still low on day five is an early-onset problem, not a late-onset one.
When should a neonate with a low platelet count be transfused?
Below 25 ×10⁹/L if the infant is not bleeding. The British Society for Haematology’s 2020 addendum states that preterm neonates with a count below 25 ×10⁹/L should receive platelet transfusions alongside treatment of the underlying cause, and that non-bleeding neonates should not routinely be transfused at a count of 25 ×10⁹/L or above. Major bleeding overrides the threshold entirely — you transfuse on the bleeding, not the number. Suspected or confirmed alloimmune disease is the other situation in which the threshold stops being the main question, because there the specific product matters more than the trigger.
Is a restrictive platelet transfusion threshold really safe in preterm babies?
The evidence says the liberal threshold is the unsafe one, which is the opposite of what intuition suggests. PlaNeT-2 randomised 660 preterm infants born before 34 weeks with severe thrombocytopenia to transfusion at a platelet count of 50 or 25 ×10⁹/L. Death or major bleeding within 28 days occurred in 26 per cent (85 of 324) in the higher-threshold group and 19 per cent (61 of 329) in the lower-threshold group, with an odds ratio of 1.57 and a 95% confidence interval of 1.06 to 2.32. That is about 8 percentage points of absolute harm from transfusing more, or roughly one extra death or major bleed for every 13 infants managed liberally. The restrictive threshold is not a compromise made to save platelets; it is the arm that did better.
How do I recognise neonatal alloimmune thrombocytopenia?
The pattern is a severe, isolated thrombocytopenia in an otherwise well term infant in the first days of life, with a normal maternal platelet count. Nothing else fits: placental insufficiency causes growth restriction, asphyxia causes encephalopathy, sepsis causes an ill baby. The mother’s count is the cheapest discriminator — it is normal in alloimmune disease, because the antibody targets an antigen she does not carry, and low in maternal immune thrombocytopenia. Three actions follow: an urgent cranial ultrasound, because intracranial haemorrhage is the feared complication and is often antenatal; maternal and paternal samples to the reference laboratory for HPA typing and maternal antibody screening; and treatment with HPA-1a and HPA-5b negative platelets, which cover over 90 per cent of UK cases. Do not delay transfusion waiting for matched units in a bleeding or severely affected infant — random-donor platelets are given immediately and intravenous immunoglobulin is added where the response is poor.
Does a low platelet count in a baby with necrotising enterocolitis need correcting?
Not for the number’s sake. In late-onset thrombocytopenia with sepsis or necrotising enterocolitis the count is doing two things: it is a consequence of consumption, marrow suppression and often disseminated intravascular coagulation, and it is a severity marker whose rate of fall tracks the extent of bowel involvement and is one of the things that prompts a surgical opinion. Transfusing it does not change the trajectory of the underlying illness, and the threshold stays at 25 ×10⁹/L in a non-bleeding infant. What is worth doing is checking the coagulation screen and fibrinogen, because the thrombocytopenia of established necrotising enterocolitis rarely travels alone, and repeating the count often enough to see the trend rather than a single value.
Related calculators
References
- Curley A, Stanworth SJ, Willoughby K, et al; PlaNeT2 MATISSE Collaborators. Randomized trial of platelet-transfusion thresholds in neonates. N Engl J Med. 2019;380(3):242–251. doi:10.1056/NEJMoa1807320. 660 preterm infants born before 34 weeks with severe thrombocytopenia, randomised to a platelet-count transfusion threshold of 50,000 or 25,000 per cubic millimetre. Death or major bleeding within 28 days: 26% (85 of 324) at the higher threshold against 19% (61 of 329) at the lower; odds ratio 1.57, 95% CI 1.06 to 2.32, P = 0.02.
- New HV, Berryman J, Bolton-Maggs PHB, et al; British Society for Haematology. Guidelines on transfusion for fetuses, neonates and older children. Br J Haematol. 2016;175(5):784–828. doi:10.1111/bjh.14233. The source for the neonatal alloimmune thrombocytopenia recommendations quoted here: that UK Blood Services stock platelets negative for HPA-1a and HPA-5b, "antibodies to which are responsible for over 90% of cases"; that intravenous immunoglobulin is considered where the response to those platelets is poor or matched units are unavailable; and that the suggested threshold of 25 ×10⁹/L in the absence of bleeding is the same as for neonates without NAIT.
- British Society for Haematology. Guidelines on transfusion for fetuses, neonates and older children (Br J Haematol. 2016;175:784–828): addendum August 2020. Br J Haematol. 2020;191(5):725–727. doi:10.1111/bjh.17109. Verbatim: "For preterm neonates with very severe thrombocytopenia (platelet count below 25 × 10⁹/L) platelet transfusions should be administered in addition to treating the underlying cause of the thrombocytopenia (2C). For non-bleeding neonates platelet transfusions should not be routinely administered if platelet count is ≥ 25 × 10⁹/L (1B)."
- Roberts I, Murray NA. Neonatal thrombocytopenia: causes and management. Arch Dis Child Fetal Neonatal Ed. 2003;88(5):F359–F364. doi:10.1136/fn.88.5.f359. The source of the severity bands used here (mild 100–150, moderate 50–99, severe below 50 ×10⁹/L) and of the early-onset/late-onset framework at 72 hours.
- Chakravorty S, Roberts I. How I manage neonatal thrombocytopenia. Br J Haematol. 2012;156(2):155–162. doi:10.1111/j.1365-2141.2011.08892.x. The practical approach this page follows, including the maternal platelet count as the discriminator between alloimmune and maternal immune disease.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.