When Your Premature Baby's Clotting Test Looks "Off": What a New Study Says About Plasma and Brain Bleeds

Understanding a 2026 Study on Fresh-Frozen Plasma Transfusion and Brain Bleeding in Premature Infants

A new study of nearly 750 premature babies found that giving plasma to correct an abnormal blood-clotting test does not, on its own, protect against brain bleeding — and in the most premature infants, it may actually be linked to a higher chance of it. Using careful statistical methods designed to make an observational study behave more like a controlled trial, researchers concluded that a baby's age at birth, not the clotting number by itself, should guide this decision.

Why This Question Matters to Families in the NICU

If your baby was born early and spent time in the neonatal intensive care unit (NICU), you may already be familiar with two of the most worrying words a parent can hear: brain bleed. Doctors call this intraventricular hemorrhage, or IVH — bleeding into the fluid-filled spaces (ventricles) deep inside the brain. It happens because premature babies' brains contain a delicate, richly supplied strip of tissue called the germinal matrix, which is still actively producing new brain cells before birth. In a full-term baby, this tissue has mostly matured and toughened up by the time of delivery; in a very premature baby, it is still fragile, and its blood vessels can be thin and prone to rupture, especially when blood flow to the brain rises and falls unpredictably ([1]). Doctors monitor for this risk with routine brain ultrasounds and grade any bleeding found on a scale of I to IV, a system devised by Dr. Lu-Ann Papile and colleagues back in 1978 that neonatal teams around the world still use today ([2]).

At the same time, blood tests that measure how quickly a baby's blood clots — one common one is called the APTT — are often abnormal in premature babies. This is expected: a newborn's clotting system is still maturing, and a baby born many weeks early simply has not yet built up the same levels of clotting proteins that an older infant or adult has. Scientists first mapped out this normal developmental pattern decades ago ([3]), and it means that a "high" clotting-time result in a preemie often reflects normal prematurity rather than a dangerous bleeding problem.

How Families and Doctors Used to Face This Problem

For a long time, it was reasonable to assume that fixing an abnormal clotting test would help prevent brain bleeding, and some early evidence pointed that way. A study published in 2009 followed a group of extremely premature babies (born at 23 to 26 weeks) and reported that a program of routine clotting checks followed by early plasma transfusion — plasma is the straw-colored liquid part of blood, rich in clotting proteins — appeared to go along with less severe brain bleeding than had been seen before ([4]). Understandably, many NICUs took this to heart, and giving plasma for an abnormal clotting test, even without any visible bleeding, became routine practice in a lot of units for years afterward.

Over time, though, scientists studying newborn blood more closely found that the "normal" range for clotting tests in premature babies is wide and shifts a great deal depending on how early a baby was born, meaning a great many "abnormal" results may not be abnormal at all for that baby's stage of development ([5]). By the mid-2010s, expert transfusion guidelines had shifted to specifically recommend against giving plasma to a baby who was not actively bleeding just because of an abnormal test result ([6]). Yet old habits are hard to shake in busy hospital units: a 2023 review at one NICU found that well under two-thirds of plasma transfusions given could actually be justified by the medical evidence, with many of the unsupported orders being exactly this kind of precautionary, non-bleeding use ([7]). This tug-of-war between an intuitive old practice and mounting doubts about it is exactly the gap the new study set out to address.

What the Researchers Did

The team, based at Daping Hospital in Chongqing, China, could not ethically run a trial that randomly assigned some babies to get plasma and others not to — that would mean deliberately withholding a treatment some doctors believed in, or deliberately giving one others doubted, without real evidence of benefit. Instead, they used ten years of hospital records (2015 to 2025) and applied a technique called "target trial emulation," which means designing the analysis, on paper, exactly as if it were a real controlled trial, and then following that design as closely as possible using the real-world data that already existed — a rigorous approach specifically developed to avoid the kinds of misleading comparisons that can creep into ordinary observational studies ([8]).

Out of more than 12,000 premature babies whose records were reviewed, the researchers found 748 who had a markedly prolonged clotting test in their first two days of life and had not yet developed a brain bleed. Of these, 434 went on to receive a single plasma transfusion, and 314 did not. The team then used statistical modeling — including a kind of artificial-intelligence tool called a "causal forest," which can sift through many babies' individual characteristics to check whether a treatment helps some infants more than others, rather than assuming it works the same way for everyone — to compare outcomes between the two groups as fairly as possible ([9]); the study itself is the source paper for this article ([10]).

What They Found, in Concrete Terms

Overall, about 1 in 5 transfused babies (95 of 434, or 21.9%) developed some degree of brain bleeding, compared with about 1 in 6 babies who were not transfused (51 of 314, or 16.2%). At first glance that might look like plasma made things worse across the board. But once the researchers accounted for the fact that the babies who received plasma were, on average, smaller, sicker, and more premature to begin with — exactly the babies doctors are most inclined to treat — the overall difference was no longer statistically meaningful. In plain terms: on average, across the whole group, plasma neither clearly helped nor clearly hurt.

The more important finding was what happened when the researchers looked separately at the most and least premature babies in the study. Below about 32 weeks of gestation, babies who received plasma had a meaningfully higher chance of brain bleeding than similar babies who did not. At or above about 32 weeks, there was a hint that plasma might modestly lower the risk, though this difference was small enough that it could have occurred by chance. Put simply: the very smallest, most fragile babies — the ones doctors often feel most compelled to treat — appeared to be the ones for whom this treatment might do more harm than good.

What This Means for Your Baby's Care

If your baby is extremely premature and has a clotting test come back abnormal, this research suggests it is reasonable to ask your care team why a plasma transfusion is, or is not, being considered, and whether there is a specific concern beyond the number on the lab report — such as active bleeding or a very low platelet count — driving that decision. This study does not mean plasma is never appropriate; babies who are actively bleeding or who have other clear indications still stand to benefit from it, just as they always have. What it challenges is the older habit of transfusing automatically, for the test result alone, especially in the smallest babies. Because this was one hospital's experience looked at after the fact, rather than a multi-hospital trial designed and run in advance, it should be read as a strong signal for further research and thoughtful practice review, not as a final rule. Your NICU team can help translate what this means for your baby's individual situation.

What Researchers Are Working On Next

The study's own authors describe their findings, especially the pattern of who might be helped versus harmed, as an important lead rather than a closed case, and they are calling for larger studies across multiple hospitals to confirm it. Future research is also likely to explore whether different clotting-test thresholds, or additional factors like platelet counts, could help identify, ahead of time, which individual babies are genuinely likely to benefit from plasma — moving the field away from a one-size-fits-all number and toward care tailored to each baby. This mirrors a broader trend in newborn intensive care toward giving blood products more selectively and only when the evidence supports it, a shift also visible in recent research on platelet and red-blood-cell transfusion practices for premature infants.

It is worth remembering that studies like this one exist precisely because so much of neonatal medicine has, historically, had to be learned by careful observation of babies who were already being treated, rather than from large upfront trials — brain bleeding in preemies was recognized as a distinct problem worth systematically tracking only in the late 1970s, and the science of how a newborn's blood clots was mapped out only in the decades since. Every generation of NICU research narrows the gap between habit and evidence a little further, and this study is one more step in narrowing it specifically for plasma transfusion.

References

  1. Ballabh P. Pathogenesis and prevention of intraventricular hemorrhage. Clin Perinatol. 2014;41(1):47-67. doi:10.1016/j.clp.2013.09.007 ↩
  2. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr. 1978;92(4):529-534. doi:10.1016/S0022-3476(78)80282-080282-0) ↩
  3. Andrew M, Paes B, Milner R, Johnston M, Mitchell L, Tollefsen DM, Powers P. Development of the human coagulation system in the full-term infant. Blood. 1987;70(1):165-172. doi:10.1182/blood.V70.1.165.165 ↩
  4. Dani C, Poggi C, Ceciarini F, Bertini G, Pratesi S, Rubaltelli FF. Coagulopathy screening and early plasma treatment for the prevention of intraventricular hemorrhage in preterm infants. Transfusion. 2009;49(12):2637-2644. doi:10.1111/j.1537-2995.2009.02328.x ↩
  5. Christensen RD, Baer VL, Lambert DK, Henry E, Ilstrup SJ, Bennett ST. Reference intervals for common coagulation tests of preterm infants. Transfusion. 2014;54(3):627-632. doi:10.1111/trf.12322 ↩
  6. Keir AK, Stanworth SJ. Fresh-frozen plasma should not be given to nonbleeding premature infants with "abnormal" coagulation tests. Transfusion. 2014;54(6):1681. doi:10.1111/trf.12589 ↩
  7. Sarkar S, Brimacombe M, Herson V. Fresh frozen plasma and cryoprecipitate: Can we safely reduce their use in the NICU? J Perinatol. 2023;43(2):226-230. doi:10.1038/s41372-022-01438-x ↩
  8. Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial is not available. Am J Epidemiol. 2016;183(8):758-764. doi:10.1093/aje/kwv254 ↩
  9. Wager S, Athey S. Estimation and inference of heterogeneous treatment effects using random forests. J Am Stat Assoc. 2018;113(523):1228-1242. doi:10.1080/01621459.2017.1319839 ↩
  10. Tang B, Yu Q, Zhang J, Duan X, Meng T, Chen Z, Ran Q, Zhang Y, Tian S, Hu Z, Peng W. Effect of fresh-frozen plasma transfusion on intraventricular hemorrhage risk in preterm infants with prolonged APTT: a target trial emulation with causal forest analysis. Front Neurosci. 2026;20:1950034. doi:10.3389/fnins.2026.1950034 ↩