Minimizing blood sampling in preterm infants

Iatrogenic blood loss contributes substantially to anemia and transfusion exposure in preterm infants. The effectiveness of interventions aimed at minimizing blood sampling remains unclear. Most prior trials have focused on placental transfusion strategies rather than blood conservation during postnatal care.

To evaluate the benefits and harms of strategies to minimize blood sampling, compared with standard blood sampling, in preterm infants.

We used CENTRAL, PubMed, two other databases, and two trial registers, together with reference checking and handsearching conference abstracts, to identify the studies included in the review. The latest search date was September 2025.

We included parallel-group RCTs, quasi-randomized trials, and cluster-randomized trials. We excluded cross-over randomized trials and non-randomized cohort studies. We included preterm infants born before 37 weeks' gestation and admitted to a neonatal unit. We included interventions that reduced blood sampling or blood loss, with or without erythropoiesis-stimulating agents. Our comparisons were: • cord blood or placental blood sampling versus sampling solely from the infant; • sampling of newborn non-blood fluids (e.g. urine, saliva) versus blood sampling alone; • devices for micro-sampling versus no use of micro-sampling; • devices that reintroduce the blood after analysis versus not using these devices; • use of strict protocols to limit neonatal blood sampling versus not using these protocols; • combinations of the above strategies.

Our outcomes were: all-cause mortality (until hospital discharge); cerebral palsy, blindness, and sensorineural deafness requiring amplification (at 18 to 24 months' corrected age); severe intraventricular hemorrhage (sIVH) (during the first week of life); bronchopulmonary dysplasia (at 36 weeks' postmenstrual age [PMA]); and retinopathy of prematurity (ROP) requiring treatment (until hospital discharge).

We used the Cochrane risk of bias 2 (RoB 2) tool to assess bias.

Where possible, we synthesized results using fixed-effect meta-analysis, calculating risk ratios (RRs) for dichotomous outcomes and mean differences (MDs) for continuous outcomes, with 95% confidence intervals (CIs). We used the Mantel-Haenszel method for dichotomous outcomes and the inverse-variance method for continuous outcomes. We used GRADE to assess the certainty of the evidence.

We included five RCTs with 365 preterm infants, sample sizes ranging from 20 to 102 (100 analyzed). In all studies, the infants' mean gestational age was less than 29 weeks. Two studies were conducted in India, two in the USA, and one in Israel. Two studies compared cord blood or placental blood sampling versus sampling solely from the infant; two studies compared devices that reintroduced the blood after analysis versus not using these devices; and one study compared the use of strict protocols to limit neonatal blood sampling versus not using these protocols.

Cord blood or placental blood sampling versus sampling solely from the infant. The evidence is very uncertain about the effect of cord or placental blood sampling on all-cause mortality by hospital discharge (RR 0.65, 95% CI 0.22 to 1.89; I² = 0%; 2 studies, 160 participants; very low-certainty evidence). Cord blood or placental blood sampling may reduce sIVH during the first week of life (RR 0.35, 95% CI 0.14 to 0.89; I² = 0%; 2 studies, 160 participants; low-certainty evidence) and bronchopulmonary dysplasia at 36 weeks' PMA (RR 0.78, 95% CI 0.58 to 1.06; I² = 0%; 2 studies, 152 participants; low-certainty evidence), and likely reduces ROP requiring treatment compared to sampling solely from the infant (RR 0.52, 95% CI 0.32 to 0.85; I² = 0%; 2 studies, 152 participants; moderate-certainty evidence). No studies reported cerebral palsy, blindness, or sensorineural deafness requiring amplification. Devices that reintroduce the blood after analysis versus not using these devices. The evidence is very uncertain about the effect of devices that reintroduce the blood after analysis on: sIVH during the first week of life (RR 1.17, 95% CI 0.46 to 2.96; I² not applicable; 1 study, 93 participants; very low-certainty evidence); bronchopulmonary dysplasia at 36 weeks' PMA (RR 1.14, 95% CI 0.81 to 1.60; I² not applicable; 1 study, 93 participants; very low-certainty evidence); and ROP requiring treatment by hospital discharge (RR 1.49, 95% CI 0.36 to 6.18; I² not applicable; 1 study, 74 participants; very low-certainty evidence). No studies reported all-cause mortality, cerebral palsy, blindness, or sensorineural deafness requiring amplification. Use of strict protocols to limit neonatal blood sampling versus not using these protocols. The evidence is very uncertain about the effect of using strict protocols to limit blood sampling on: all-cause mortality by hospital discharge (RR 0.96, 95% CI 0.25 to 3.63; I² not applicable; 1 study, 100 participants; very low-certainty evidence); sIVH during the first week of life (RR 0.72, 95% CI 0.17 to 3.06; I² not applicable; 1 study, 100 participants; very low-certainty evidence); bronchopulmonary dysplasia at 36 weeks' PMA (RR 0.86, 95% CI 0.61 to 1.22; I² not applicable; 1 study, 100 participants; very low-certainty evidence); and ROP requiring treatment by hospital discharge (RR 1.26, 95% CI 0.75 to 2.12; I² not applicable; 1 study, 100 participants; very low-certainty evidence). Cerebral palsy, blindness, and sensorineural deafness requiring amplification were not reported.

Cord or placental blood sampling for admission laboratory tests likely reduces ROP requiring treatment and may reduce sIVH; effects on bronchopulmonary dysplasia may favor the intervention, while effects on mortality are very uncertain. The evidence is very uncertain regarding the effects on mortality and major morbidities of (1) devices that reintroduce blood after analysis and (2) strict sampling protocols. Cord or placental sampling may be considered a feasible blood-conservation strategy. The benefits of other approaches remain unproven in randomized trials. None of the five included studies reported long-term outcomes such as cerebral palsy, blindness, or sensorineural deafness requiring amplification. Future trials should be large, high quality, and powered to detect effects on short-term morbidity and long-term neurodevelopment. They should also evaluate innovative approaches to reduce blood sampling, including micro-sampling technologies and the use of non-blood body fluids.

This review received no dedicated funding.

Protocol available via DOI 10.1002/14651858.CD016077.