Highlights
– In a 4‑week point‑prevalence survey across 44 PICUs in Spain, the United Kingdom, and Italy (2023), 12.8% (348/2713) of children received ≥1 RBC transfusion.
– Hemoglobin level was documented as the primary indication in 54.6% of transfusion events; in 45.1% no additional physiologic trigger was recorded.
– Median pre‑transfusion hemoglobin was 8.3 g/dL (IQR 7.2–9.9); values varied by country and exceeded conservative thresholds in many cases.
– Number of transfusions was independently associated with 28‑day PICU mortality after adjustment for admission reason and PIM score.
Background: Why this matters
Red blood cell (RBC) transfusion is one of the most common and potentially lifesaving interventions in pediatric intensive care, but it carries risks including transfusion‑associated circulatory overload, infection, and immunomodulation. Over the last two decades the field has shifted toward restrictive transfusion strategies where supported by evidence; however, the extent to which clinical practice aligns with evidence and guideline recommendations across European PICUs has been incompletely characterized. The European PEdiatric TRAnsfusion Practices in PICU (E‑PETRA) investigators conducted a contemporary point‑prevalence study to quantify RBC use, triggers, thresholds, and early outcomes in real‑world PICU settings.
Study design and methods
This international observational point‑prevalence study sampled four prespecified 7‑day blocks between March and July 2023 in 44 PICUs across Spain, the United Kingdom, and Italy. Eligible patients were aged 1 month to 17 years who received ≥1 RBC transfusion during their PICU stay in those blocks. No interventions were applied. Data collected included indication for transfusion (clinician‑documented), physiologic triggers (e.g., bleeding, cardiovascular instability, extracorporeal support), pre‑transfusion hemoglobin concentration, number of transfusions, and outcomes including 28‑day PICU mortality. The dataset included 2713 PICU admissions during the study windows; 527 transfusion events were recorded in 348 patients.
Key results
Prevalence and distribution
– Overall, 348/2713 patients (12.8%) received at least one RBC transfusion during the sampled intervals, accounting for 527 transfusions.
– Country‑specific proportions of patients transfused were: Italy 17.3% (66/382), United Kingdom 13.9% (166/1195), and Spain 10.2% (116/1136).
Indications and physiologic triggers
– The primary indication recorded was low hemoglobin in 54.6% of transfusion events. Other primary indications were bleeding (10.6%), cardiovascular instability (10.5%), and extracorporeal support (10.1%).
– Importantly, 45.1% of transfusions had no physiologic trigger documented beyond hemoglobin level; this suggests many decisions were driven largely by a numeric hemoglobin threshold rather than integrated physiologic assessment.
Pre‑transfusion hemoglobin thresholds
– Median pre‑transfusion hemoglobin across all transfusions was 8.3 g/dL (IQR 7.2–9.9 g/dL).
– Median hemoglobin varied significantly by country: Spain 7.8 g/dL, United Kingdom 8.6 g/dL, Italy 8.9 g/dL (p < 0.001).
– When cardiac patients were excluded (a group with frequently different transfusion practice), the overall median hemoglobin threshold was 7.4 g/dL (IQR 6.8–8.6 g/dL) and there were no statistically significant differences between countries (p > 0.05).
Outcomes and associations
– Overall 28‑day PICU mortality among the 348 transfused patients was 10.7%.
– The number of transfusions was associated with increased mortality even after adjustment for admission reason and Pediatric Index of Mortality (PIM) score. The observational design precludes causal inference, but the association raises concerns about potential harms of greater transfusion exposure or that transfusion number is a marker of underlying severity.
Subgroup observations
– Cardiac patients and those on extracorporeal support accounted for a minority of transfusions but tended to have higher pre‑transfusion hemoglobin targets, consistent with clinician concern about oxygen delivery in these groups.
Expert commentary and interpretation
Practice versus evidence: The TRIPICU randomized trial (restrictive threshold ~7 g/dL) and subsequent guideline recommendations have supported a restrictive transfusion approach for hemodynamically stable, non hemorrhaging pediatric ICU patients. The current point‑prevalence data show that hemoglobin still predominates as the trigger for transfusion in European PICUs, and that median pre‑transfusion hemoglobin often exceeds 7 g/dL — particularly in some countries and in cardiac or ECMO patients. This suggests incomplete translation of restrictive strategies into routine practice for many children.
Why might clinicians transfuse above 7 g/dL?
– Perceived risk of tissue hypoxia in vulnerable groups (cardiac disease, ECMO, severe sepsis) leads clinicians to select higher targets, despite limited RCT evidence supporting higher thresholds in many subgroups.
– Lack of easy, validated physiologic bedside markers of tissue oxygen delivery that reliably guide transfusion decisions in children.
– Institutional culture, training, and local protocols that favor percent‑hematocrit or hemoglobin cutoffs over dynamic physiologic assessment.
Risk–benefit balance and observed association with mortality
– The association between transfusion count and mortality is consistent with adult and pediatric observational literature linking transfusion exposure to adverse outcomes, though confounding by severity is likely. Transfusion may be both a marker of severity and a contributor to risk (e.g., volume overload, transfusion‑related immunomodulation).
Implications for practice
– The high proportion of transfusions driven by hemoglobin alone and the variability across countries indicate an opportunity for standardizing care and implementing goal‑directed transfusion strategies that combine a restrictive hemoglobin threshold with physiologic triggers and reassessment after single‑unit transfusion.
Recommended pragmatic approach
– For hemodynamically stable, non‑cardiac PICU patients, default to a restrictive hemoglobin threshold (≈7.0 g/dL) as supported by randomized data, unless clear physiologic indications exist.
– Integrate physiologic markers—ongoing bleeding, hemodynamic instability despite vasoactive support, rising lactate, low mixed/central venous oxygen saturation, or signs of inadequate end‑organ perfusion—into the decision to transfuse.
– Consider single‑unit transfusion with prompt reassessment of clinical and laboratory endpoints rather than routine multi‑unit transfusion.
– For cardiac, ECMO, or complex congenital heart disease patients, individualize thresholds while prioritizing studies that clarify optimal targets in these high‑risk groups.
Limitations of the study
– Point‑prevalence design captures practice in specific 7‑day windows and may not reflect year‑round patterns; however, the chosen blocks spanned spring to mid‑summer to mitigate season bias.
– Observational data cannot establish causality for the association between transfusion exposure and mortality; residual confounding by illness severity or unmeasured factors is likely.
– Documentation of physiologic triggers relied on clinical recording practices which may underreport nuanced physiologic reasoning.
– The study sampled three European countries and 44 PICUs; results may not generalize to other regions or smaller facilities with differing resources.
Research and implementation priorities
– Implementation studies to close the gap between evidence and practice: audit‑and‑feedback, decision support tools, and transfusion bundles that operationalize restrictive thresholds plus physiologic triggers.
– Pragmatic randomized trials in specific high‑risk subgroups (congenital heart disease, ECMO, severe sepsis) to determine safe and effective hemoglobin targets.
– Development and validation of bedside physiologic indices or composite scores that better predict tissue oxygenation and can guide transfusion decisions beyond a single hemoglobin value.
– Longitudinal outcomes research to disentangle whether transfusions contribute causally to worse outcomes or are markers of severity, including propensity‑matched and instrumental variable analyses.
Conclusions
This 2023 European point‑prevalence study demonstrates that while a minority of PICU patients receive RBC transfusions during a month of care, decisions are frequently dominated by hemoglobin concentration alone, often above conservative thresholds. There is meaningful intercountry variability and an observed association between transfusion exposure and higher 28‑day mortality. These findings underscore the need to standardize transfusion practice through implementation of evidence‑based restrictive thresholds for stable patients, paired with physiologic, goal‑directed triggers and reassessment practices, and to target research toward groups where uncertainty remains.
Funding and clinicaltrials.gov
This analysis is based on the published E‑PETRA study. The original article should be consulted for study‑level funding statements and declarations. This point‑prevalence observational study was not a clinical trial and has no ClinicalTrials.gov identifier.
References
Butragueño‑Laiseca L, Ray S, Sarfatti A, Stanworth SJ, Campos Rodríguez R, Gómez‑Zamora A, Hernández Yuste A, Benítez Gómez I, de Lama Caro‑Patón G, Giorni C, Lampugnani E, Daverio M, Chiusolo F; European PEdiatric TRAnsfusion Practices in PICU (E‑PETRA) Investigators and the United Kingdom Pediatric Critical Care Society Study Group (PCCS‑SG). Red Cell Transfusion During Pediatric Intensive Care: A 28‑Day Point Prevalence Study in Three European Countries in 2023. Pediatr Crit Care Med. 2025 Oct 1;26(10):e1231–e1241. doi: 10.1097/PCC.0000000000003805. Epub 2025 Aug 7. PMID: 40772802.
Lacroix J, Hebert PC, Hutchison JS, Hume HA, Tucci M, et al.; TRIPICU Investigators. Transfusion strategies for patients in pediatric intensive care units. N Engl J Med. 2007;356(19):1609–1619. (TRIPICU trial supporting restrictive transfusion threshold ~7 g/dL in stable critically ill children.)
Author note
This summary was prepared by a clinical medical writer to interpret and contextualize the E‑PETRA 2023 point‑prevalence findings for clinicians and policy makers. Clinicians should consult the full original manuscript for detailed methods and local institutional transfusion guidelines when making bedside decisions.
