ORIGINAL ARTICLES THE PREMATURE INFANTS IN NEED OF TRANSFUSION (PINT) STUDY: A RANDOMIZED, CONTROLLED TRIAL OF A RESTRICTIVE (LOW) VERSUS LIBERAL (HIGH) TRANSFUSION THRESHOLD FOR EXTREMELY LOW BIRTH WEIGHT INFANTS HARESH KIRPALANI, MSC, FRCP(UK),1 ROBIN K. WHYTE, MB, FRCP(C),1 CHAD ANDERSEN, MBBS, FRACP, ELIZABETH V. ASZTALOS, MSC, FRCP(C), NANCY HEDDLE, MSC, MORRIS A. BLAJCHMAN, MD, FRCP(C), ABRAHAM PELIOWSKI, MD, FRCP(C), ANGEL RIOS, MD, MEENA LACORTE, MD, ROBERT CONNELLY, MD, FRCP(C), KEITH BARRINGTON, MB, FRCP(C), ROBIN S. ROBERTS, M.TECH, FOR THE PINT INVESTIGATORS* Objective To determine whether extremely low birth weight infants (ELBW) trans- fused at lower hemoglobin thresholds versus higher thresholds have different rates of survival or morbidity at discharge. Study design Infants weighing 1000 g birth weight were randomly assigned within 48 hours of birth to a transfusion algorithm of either low or high hemoglobin transfusion thresholds. The composite primary outcome was death before home discharge or survival with any of either severe retinopathy, bronchopulmonary dysplasia, or brain injury on cranial ultrasound. Morbidity outcomes were assessed, blinded to allocation. Results Four hundred fifty-one infants were randomly assigned to low (n 223) or high (n 228) hemoglobin thresholds. Groups were similar, with mean birth weight of 770 g and gestational age of 26 weeks. Fewer infants received one or more transfusions in the low threshold group (89% low versus 95% high, P .037). Rates of the primary outcome were 74.0% in the low threshold group and 69.7% in the high (P .25 risk difference, 2.7% 95% CI ���3.7% to 9.2%). There were no statistically significant differ- ences between groups in any secondary outcome. Conclusions In extremely low birth weight infants, maintaining a higher hemoglobin level results in more infants receiving transfusions but confers little evidence of benefit. (J Pediatr 2006 149:301-7) Eadmissions.1 xtremely low birth weight infants (ELBW 1000 g) accounted for 0.4% of all Canadian births in 1996 to 1997 and 9% of all neonatal intensive care unit These newborn infants rapidly become anemic from a combination of frequent laboratory blood sampling and an immature hematopoietic system,2-4 leading to the transfusion of allogeneic red blood cells (RBCs) in at least 94%.5 RBC transfusion guidelines available at the start of this trial6,7 recommended the maintenance of the hemoglobin of ELBW infants at ���physiologic��� levels, but the justification for such an intervention is poorly supported. The risks and benefits of RBC transfusions to ELBW infants are unclear: Limiting RBC transfusions may reduce transfusion-associated infec- tion and iron overload, but the resulting low hemoglobin levels may result in the morbidities associated with chronic anemic hypoxemia. A study in adult patients in critical care8 reported no benefit in survival from liberally transfusing allogeneic RBCs to maintain a hemoglobin concentration between 100 and 120 g/L. A restrictive RBC BPD Bronchopulmonary dysplasia ELBW Extremely low birth weight PINT Premature infants in need of transfusion RBC Red blood cell ROP Retinopathy of prematurity See editorial, p 287 From Pediatrics and Clinical Epidemiol- ogy and Biostatistics, McMaster Univer- sity, Hamilton, Ontario, Canada Depart- ment of Pediatrics, Dalhousie University and IWK Health Centre, Halifax, Nova Scotia, Canada Mercy Hospital for Women, Melbourne, Victoria, Australia Sunnybrook and Women���s Health Sci- ence Center, University of Toronto, To- ronto, Ontario, Canada Canadian Blood Services and Transfusion Medicine, Mc- Master University, Hamilton, Ontario, Canada Royal Alexandra Hospital, Edm- onton, Alberta, Canada Albany Medical Center, Albany, New York Brooklyn Hospital Center, Brooklyn, New York Kingston General Hospital, Kingston, On- tario, Canada McGill University, Mon- treal, Quebec, Canada and Clinical Trials Methodology Group, Clinical Epidemiol- ogy and Biostatistics, McMaster Univer- sity, Hamilton, Ontario, Canada. *A complete list of the PINT Investigators is available at www.jpeds.com. This study was supported by the Canadian Institutes Health Research (FR No. 41549) 2000-2004. Registration Number: www.clinicaltrials.gov NCT 00182390. Submitted for publication Nov 18, 2005 last revision received Feb 23, 2006 ac- cepted May 8, 2006. Reprint requests: Dr Haresh Kirpalani, Mc- Master University, HSC 3N11F, 1200 Main Street West, Hamilton, Ontario, Canada L8N 3Z5. E-mail: kirpalan@mcmaster.ca. 1Drs Kirpalani and Whyte are co-principals. 0022-3476/$ - see front matter Copyright �� 2006 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2006.05.011 301

transfusion strategy (maintaining the hemoglobin concentra- tion between 70 and 90 g/L) was at least as good as, and, for some subgroups, better than the liberal strategy. In this randomized, controlled trial, we randomly as- signed ELBW newborn infants to RBC transfusion algo- rithms, incorporating restrictive or liberal (low or high) he- moglobin ���thresholds.��� The primary outcome was a composite of either death before home discharge or survival with severe morbidity, where severe morbidity was defined as one or more of (1) retinopathy of prematurity (ROP), (2) bronchopulmonary dysplasia (BPD), or (3) brain injury on cranial ultrasound. This composite outcome captured the competing outcomes of death, or survival with severe mor- bidity, in one measure. The purpose of this trial was to determine whether maintaining ELBW infants at a restrictive hemoglobin level, by adopting a restrictive (lower) RBC transfusion threshold, led to a difference in clinical outcome compared with a liberal RBC transfusion threshold, in the first weeks of post-natal life. METHODS Study Population Newborn infants were enrolled from 10 neonatal inten- sive care units in Canada, the United States, and Australia (Appendix available at www.jpeds.com). Eligible babies were required to be of birth weight 1000 g, gestational age 31 weeks, and to be 48 hours old at the time of enrollment. Infants deemed non-viable by the attending physician were excluded, as were those with cyanotic heart disease, congenital anemia, acute shock, transfusion after 6 hours of age, or known parental opposition to transfusions, a family history of anemia and hemolytic disease, or where the attending physi- cian anticipated using erythropoietin. Informed consent was requested from the parents or guardians. After consent, in- fants were individually randomly assigned to either the re- strictive or the liberal threshold for RBC transfusion. Treat- ment allocation was by telephone to an interactive voice system at the coordinating center. The concealed computer- generated randomization sequence was stratified by center and birth weight ( 750 g, 751 to 999 g) and incorporated random block sizes. The study was approved by each local research ethics board. Intervention Infants were allocated to a transfusion algorithm of low or high hemoglobin thresholds for transfusion (Table I). The algorithm was developed by consensus among participating sites, against the background of then current guidelines.6,7 The appropriate hemoglobin threshold was adjusted for cap- illary or arterial/venous blood sampling according to prior work.9 The thresholds developed were based on whether or not the infant was receiving respiratory support (assisted ventilation, continuous positive airway pressure, or supple- mental oxygen) and on post-natal age. The protocol did not dictate how or how often the hemoglobin value was to be determined, but all hemoglobin values were recorded. Phle- botomy losses were not recorded. Transfusion was indicated whenever the hemoglobin equaled or fell below the threshold value. In addition, attending neonatologists were allowed to give non-algorithm dictated RBC transfusions in the event of shock, severe sepsis, coagulation defects, surgery, or for un- anticipated emergencies. RBC transfusions were given within 6 hours of the hemoglobin determination, but as soon as possible after the attainment of a threshold value. All trans- fusions consisted of washed, packed RBCs (homologous or directed donation) given at 15 mL/kg at rates determined by local policy. No attempt was made to blind caregivers with respect to treatment group, as the concealment of hemoglobin values was considered both unethical and impractical. The transfusion algorithm was distributed widely and placed prominently at the site of care of each study patient. The allocated transfusion algorithm was maintained after dis- charge from tertiary care to the level 2 nursery, but data collection was limited to transfusions, hemoglobin levels prompting transfusion, and hemoglobin at discharge. All cen- ters used iron supplementation according to local guidelines. Data Collection All data, including Score for Acute Physiology (SNAP) scores, were collected from entry to discharge by using spe- cifically designed case report forms. Routine care of ELBW infants at all participating institutions included first-week and pre-discharge cranial ultrasounds and pre-discharge retinal examination. Donor exposure to RBCs, platelets, and fresh Table I. Hemoglobin threshold levels (g/L) triggering RBC transfusion Age in days Blood sampling Low threshold High threshold Respiratory support No respiratory support Respiratory support No respiratory support 1-7 Capillary 115 100 135 120 Central 104 90 122 109 8-14 Capillary 100 85 120 100 Central 90 77 109 90 15 Capillary 85 75 100 85 Central 77 68 90 77 302 Kirpalani et al The Journal of Pediatrics ��� September 2006