Every Drop Counts: How Much Blood the Smallest Premature Babies Lose to Tests, and How Hospitals Can Take Less

Explaining a 22-country European study of blood taken for testing (Houben et al.), and two smaller studies on using cord blood for the first tests (Wu et al.) and a one-drop bedside haemoglobin test (Chang et al.)

Babies born very early need frequent blood tests, and because they have so little blood to begin with, the small amounts taken for testing add up quickly. A large study across 22 European countries found that hospitals differ widely in how much blood their laboratory machines need, and two smaller studies looked at practical ways to take less blood from babies. Here we explain what was found and what it means for families.

Why blood tests add up for the smallest babies

If your baby was born very prematurely, you have probably watched nurses take blood many times a day — from a thin tube placed in a blood vessel, from a vein, or with a quick prick of the heel. These tests help the team check breathing, look for infection, measure sugar and salts, and decide on treatment. They are an important part of good care. But a baby born at 24 or 25 weeks of pregnancy may weigh well under a kilogram, and the total amount of blood in such a tiny body is small. Doctors estimate it at roughly 70 millilitres for every kilogram of body weight [1]. So blood removed for testing, even a few drops at a time, can make up a noticeable share of what the baby has.

Losing blood this way contributes to anaemia, which means having too few red blood cells, the cells that carry oxygen around the body. When anaemia becomes significant, babies may need a blood transfusion, which means receiving donated red blood cells. Transfusions can be lifesaving, but most families and doctors would rather avoid them when they safely can.

What doctors already knew before these studies

Concern about blood taken for tests is not new. For more than two decades, individual hospitals measured it and found large amounts. A study in Leiden, in the Netherlands, followed 20 babies born before 28 weeks and found that each baby needed a median of 47 needle pricks, for blood tests and for placing drips, in the first four weeks of life [2]. ("Median" means the middle value: half the babies had more, half had fewer.) Two Swedish university hospitals found that in the first two weeks, the blood taken for tests in babies born before 28 weeks added up to more than half of the blood they had been born with, and blood gas tests, which check oxygen and acid levels, were the biggest single use [3]. A study in the United States showed that the people drawing blood often take slightly more blood than the laboratory actually needs, especially when tubes have no fill-line and in smaller, sicker babies [4].

Researchers also tested ways to take less. At Stanford in California, switching to a bedside machine that needed much less blood was followed by fewer transfusions in the first two weeks for babies weighing under 1 kilogram [5]. More recently, a carefully designed trial at Surya Children's Hospital in Mumbai, India, compared deliberately limited blood sampling with usual sampling in 102 babies born before 28 weeks and weighing under 1 kilogram. Fewer babies in the limited-sampling group needed a transfusion early on — 41 in every 100 compared with 73 in every 100 [6]. Even so, a review of many studies concluded that good-quality evidence on most blood-saving methods was still scarce [7]. Above all, nobody knew how much blood was being taken across many hospitals and countries, or how much hospitals differed.

What a study across 22 European countries found

The new European study, part of a larger project called the International Neonatal Transfusion Point Prevalence (INSPIRE) study, ran in 64 neonatal intensive care units in 22 European countries between September 2022 and August 2023 [1]. It was observational, meaning researchers recorded what happened without changing anyone's care. The question about blood tests was added because a group of parents advising the researchers said they were worried about how often their babies had blood taken. The study included 320 babies born before 28 weeks. Each day, staff recorded how many of eight common blood tests each baby had, and each hospital reported how much blood its laboratory needed for each type of test.

Almost every baby (97%) had blood taken on the first day, and more than half (58%) were still having tests on day 28. The amount taken was highest in the first week. By day 28, babies born at 24 weeks or earlier had lost an estimated 22 millilitres of blood for every kilogram of birth weight, while babies born at 27 weeks had lost about 8 millilitres per kilogram. The researchers estimated that for babies born at 24 weeks this was about half of the blood they were born with. Our Source Transparency Report explains that the paper's two ways of expressing this result do not fully match, so that particular estimate should be treated as approximate; either way, the amount is substantial.

The most striking finding was how much hospitals differed. To run one of each of eight common tests, some hospitals' machines needed as little as 2.4 millilitres of blood, and others as much as 14.7 millilitres. The typical amount was 4.1 millilitres, a little less than a teaspoon. Babies cared for in hospitals with the smallest-volume machines lost about half as much blood by day 28 as babies in other hospitals — around 8 millilitres per kilogram, compared with about 17 to 20.

You might expect those babies to have needed fewer transfusions, but the numbers were similar: by day 28, about 55 in every 100 babies in the small-volume hospitals had received a transfusion, compared with 53 and 61 in every 100 in the others. The study was not designed to answer that question, and decisions about when to transfuse also differ between hospitals [8].

Using the cord's blood for the first tests

When a baby is born, blood remains in the umbilical cord and placenta after the cord is cut. That blood would normally be thrown away, but it can be used for the baby's first tests. A hospital in Changsha, China, looked back at the records of 138 babies weighing less than 1,500 grams at birth [9]. For 68 of them, the first tests used cord blood; for 70, blood was taken from the baby.

The cord-blood babies had much less blood taken from them in their first day. Among the smallest babies, under 1,250 grams, it was about 5.5 millilitres per kilogram compared with 9.6. In this smallest group, fewer cord-blood babies developed low haemoglobin (the oxygen-carrying part of red blood cells) in the first week — 14 of 47 compared with 15 of 27 — and about two in three got through the first week without a transfusion, compared with about four in ten. But over the whole hospital stay, most of these smallest babies were eventually transfused in both groups (87 and 82 in every 100). Because doctors chose which babies had cord blood used, and those babies were smaller and born earlier, the two groups were not directly comparable, and the authors call for larger, planned studies.

Earlier trials in which babies were assigned to cord-blood sampling by chance, in India and the United States, found that the first transfusion came later [10][11]. The American Academy of Pediatrics now says cord blood can be used for many first tests, and that this can be done even when the cord is clamped later to let extra blood flow into the baby [12]. Delaying cord clamping is itself known to help premature babies survive [13].

A bedside test that needs only a drop

A third study, at Stony Brook Children's Hospital in New York, compared two ways of measuring haemoglobin in 187 newborns [14]. One was a bedside blood-gas machine that needs just 65 microlitres, about one drop; the other was the hospital laboratory's machine, which needs at least 500 microlitres, or half a millilitre. Bedside results came back in about 4 minutes instead of about an hour. For blood from an artery or vein, the two machines agreed closely on average. For heel-prick blood, the bedside machine read slightly higher than the laboratory machine testing blood from the same heel prick; the researchers think differences in how the samples were mixed may explain this. (Heel-prick blood is also known to read higher than blood from a vein [15], but that does not explain this result, because both machines tested heel-prick blood.) Individual readings could still differ by more than the researchers had hoped, so a borderline result may need checking. Importantly, these babies were mostly born only a few weeks early (on average at about 36½ weeks), so the findings may not apply to the tiniest babies, and another study found a different pattern in very premature babies [16].

What this means for your baby

None of this means your baby is having too many tests. Every test should have a reason, and doctors balance the need for information against the blood it takes. These studies suggest that hospitals can often take less blood — by choosing machines that need smaller amounts, by thinking carefully about which tests are needed, and sometimes by using cord blood for the first tests. Taking less blood is a sensible goal, but none of these three studies showed that it reduced the total number of transfusions over a baby's stay.

It is entirely reasonable to ask your baby's team how they try to limit blood taken for tests, whether cord blood can be used for first tests, or how a particular result will change your baby's care. Many units welcome these conversations; managing blood carefully is part of what experts call patient blood management [17].

What researchers are working on next

Researchers want to measure blood loss for each baby more precisely, including the small extra amounts that are drawn but not needed. They hope to run larger studies that test whether small-volume machines, careful test planning and cord-blood sampling actually reduce transfusions and improve longer-term health. Bedside tests also need to be checked in the smallest, most premature babies.

Bringing it together

Very premature babies lose a meaningful amount of blood to tests, most of all in their first week and especially when born earliest. How much they lose depends a great deal on the hospital's laboratory equipment. Using cord blood and one-drop bedside tests are promising ways to take less, and asking about them is a good way for families to take part in their baby's care.

References

  1. Houben NAM, Fustolo-Gunnink S, Fijnvandraat K, et al. Blood loss due to diagnostic testing in extremely preterm infants in 22 European countries: a prospective observational study. EClinicalMedicine. 2026;97:104035. doi:10.1016/j.eclinm.2026.104035 ↩
  2. Counsilman CE, Heeger LE, Tan R, et al. Iatrogenic blood loss in extreme preterm infants due to frequent laboratory tests and procedures. J Matern Fetal Neonatal Med. 2021;34(16):2660-2665. doi:10.1080/14767058.2019.1670800 ↩
  3. Hellström W, Forssell L, Morsing E, Sävman K, Ley D. Neonatal clinical blood sampling led to major blood loss and was associated with bronchopulmonary dysplasia. Acta Paediatr. 2020;109(4):679-687. doi:10.1111/apa.15003 ↩
  4. Lin JC, Strauss RG, Kulhavy JC, et al. Phlebotomy overdraw in the neonatal intensive care nursery. Pediatrics. 2000;106(2):E19. doi:10.1542/peds.106.2.e19 ↩
  5. Madan A, Kumar R, Adams MM, Benitz WE, Geaghan SM, Widness JA. Reduction in red blood cell transfusions using a bedside analyzer in extremely low birth weight infants. J Perinatol. 2005;25(1):21-25. doi:10.1038/sj.jp.7211201 ↩
  6. Balasubramanian H, Bhanushali M, Tripathi V, et al. Effect of Minimization of Early Blood Sampling Losses Among Extremely Premature Neonates: A Randomized Clinical Trial. J Pediatr. 2024;269:114002. doi:10.1016/j.jpeds.2024.114002 ↩
  7. Persad E, Sibrecht G, Ringsten M, et al. Interventions to minimize blood loss in very preterm infants-A systematic review and meta-analysis. PLoS One. 2021;16(2):e0246353. doi:10.1371/journal.pone.0246353 ↩
  8. Houben NAM, Fustolo-Gunnink S, Fijnvandraat K, et al. Red Blood Cell Transfusion in European Neonatal Intensive Care Units, 2022 to 2023. JAMA Netw Open. 2024;7(9):e2434077. doi:10.1001/jamanetworkopen.2024.34077 ↩
  9. Wu H, Cheng L, Lei J, Chu J, Bo T. Umbilical cord blood collection at birth reduces early iatrogenic blood loss in preterm infants with birth weights <1,500 g. Front Pediatr. 2026;14:1739158. doi:10.3389/fped.2026.1739158 ↩
  10. Balasubramanian H, Malpani P, Sindhur M, Kabra NS, Ahmed J, Srinivasan L. Effect of Umbilical Cord Blood Sampling versus Admission Blood Sampling on Requirement of Blood Transfusion in Extremely Preterm Infants: A Randomized Controlled Trial. J Pediatr. 2019;211:39-45.e2. doi:10.1016/j.jpeds.2019.04.033 ↩
  11. Mu TS, Prescott AC, Haischer-Rollo GD, Aden JK, Shapiro JB. Umbilical Cord Blood Use for Admission Blood Tests of VLBW Preterm Neonates: A Randomized Control Trial. Am J Perinatol. 2023;40(10):1119-1125. doi:10.1055/s-0041-1733781 ↩
  12. Kaufman DA, Lucke AM, Cummings JJ. Postnatal Cord Blood Sampling: Clinical Report. Pediatrics. 2025;155(6):e2025071811. doi:10.1542/peds.2025-071811 ↩
  13. Seidler AL, Aberoumand M, Hunter KE, et al. Deferred cord clamping, cord milking, and immediate cord clamping at preterm birth: a systematic review and individual participant data meta-analysis. Lancet. 2023;402(10418):2209-2222. doi:10.1016/S0140-6736(23)02468-602468-6) ↩
  14. Chang V, Hou W, Maduekwe ET. Evaluation of point-of-care hemoglobin testing vs. central laboratory measurements in neonates: a prospective study in neonatal intensive care unit. Front Pediatr. 2026;14:1735203. doi:10.3389/fped.2026.1735203 ↩
  15. Kayiran SM, Ozbek N, Turan M, Gürakan B. Significant differences between capillary and venous complete blood counts in the neonatal period. Clin Lab Haematol. 2003;25(1):9-16. doi:10.1046/j.1365-2257.2003.00484.x ↩
  16. Calandrino A, Montobbio C, Bonato I, et al. Optimizing haemoglobin measurements in VLBW newborns: Insights from a comparative retrospective study. Early Hum Dev. 2024;190:105949. doi:10.1016/j.earlhumdev.2024.105949 ↩
  17. Crighton GL, New HV, Liley HG, Stanworth SJ. Patient blood management, what does this actually mean for neonates and infants?. Transfus Med. 2018;28(2):117-131. doi:10.1111/tme.12525 ↩