OBJECTIVE. Our goal was to identify trends in patient demographics and indications for and complications related to neonatal exchange transfusion over a 21-year period in a single institution using a uniform protocol for performing the procedure.
METHODS. A retrospective chart review of 107 patients who underwent 141 single- or double-volume exchange transfusions from 1986–2006 was performed. Patients were stratified into 2 groups, 1986–1995 and 1996–2006, on the basis of changes in clinical practice influenced by American Academy of Pediatrics management guidelines for hyperbilirubinemia.
RESULTS. There was a marked decline in the frequency of exchange transfusions per 1000 newborn special care unit admissions over the 21-year study period. Patient demographics and indications for exchange transfusion were similar between groups. A significantly higher proportion of patients in the second time period received intravenous immunoglobulin before exchange transfusion. There was a higher proportion of patients in the 1996–2006 group with a serious underlying condition at the time of exchange transfusion. During that same time period, a lower proportion of patients experienced an adverse event related to the exchange transfusion. Although a similar percentage of patients in both groups experienced hypocalcemia and thrombocytopenia after exchange transfusion, patients treated from 1996–2006 were significantly more likely to receive calcium replacement or platelet transfusion. No deaths were related to exchange transfusion in either time period.
CONCLUSIONS. Improvements in prenatal and postnatal care have led to a sharp decline in the number of exchange transfusions performed. This decline has not led to an increase in complications despite relative inexperience with the procedure.
Exchange transfusion (ECT) was introduced in the late 1940s to decrease the mortality of hemolytic disease of the newborn (HDN) and to prevent kernicterus in surviving patients.1 ECT was subsequently applied to neonatal hyperbilirubinemia from a variety of causes and quickly became one of the most commonly performed neonatal procedures.
In 1968 and 1971, Lucey2,3 accurately predicted that prenatal interventions, particularly the development of Rh-immunoglobulin, coupled with advances in postnatal care such as phototherapy, would lead to a dramatic decline in the number of ECTs performed. Maisels4, in a review that combined data from 3 centers over 40 years, observed a decline in the frequency of ECT and predicted that it would lead to increased complications because of inexperience with the procedure. More recent advances, such as use of intrauterine transfusions and improvements in diagnostic ultrasound,5–8 have likely accelerated this decline in the frequency of ECT.9,10
Since the introduction of ECT, the level of bilirubin at which to initiate this procedure has been a controversial issue. Based on experience with HDN,11 a bilirubin level of 20 mg/dL was used by many centers, including Yale, but some questioned whether it was appropriate to apply this cutoff to patients with nonhemolytic hyperbilirubinemia.12 This debate intensified in the late 1980s and early 1990s, when several reports demonstrated that term infants with nonhemolytic jaundice were not as susceptible to kernicterus as infants with HDN.13,14
In 1994, the American Academy of Pediatrics (AAP) published its first guidelines on the treatment of hyperbilirubinemia.15 These guidelines increased the bilirubin threshold for initiating ECT in term infants without hemolysis and allowed for a trial of intensive phototherapy before an ECT was initiated. In addition, these guidelines encouraged prenatal testing of maternal ABO and Rh types and recommended increased monitoring for hyperbilirubinemia in all infants.15 These interventions had the potential to cause a further decline in the number of patients requiring ECT.16
We hypothesized that changes in prenatal and postnatal care have altered the patient population undergoing ECT, the indication for exchange, and the incidence of ECT-related morbidity and mortality. To examine this, we performed a longitudinal, 21-year review of ECT at a single center.
PATIENTS AND METHODS
Infants who required single- or double-volume ECT and had long-term admissions (>24 hours) in the newborn special care unit (NBSCU) at Yale New Haven Hospital (YNHH) from January 1, 1986, through December 31, 2006, were included. Neonates who received partial ECT for polycythemia or anemia were excluded. Data collection included patient demographics, comorbidities, indication for exchange transfusion, treatment with phototherapy and intravenous immunoglobulin (IVIg), and ECT-related complications.
Patients were divided into 2 groups, 1986–1995 and 1996–2006, based on the AAP guidelines for the management of hyperbilirubinemia published in October 1994 and implemented in the NBSCU at Yale in late 1995. Before this time, the threshold for ECT at YNHH, with or without evidence of hemolysis, was a total serum bilirubin of 20 mg/dL for term infants, with threshold levels decreasing based on birth weight.17 Beginning in late 1995 and continuing to the present time, the threshold for ECT at YNHH was raised to 25 mg/dL for term infants >48 hours old without evidence of hemolysis, but remained at 20 mg/dL for those with hemolysis. Asymptomatic term infants were also provided the opportunity to respond to intensive phototherapy before an ECT was initiated, and all infants were strictly monitored for hyperbilirubinemia as per the AAP guidelines.15
A detailed, step-by-step protocol, provided in the YNHH NBSCU procedure manual, was used for ECT. This technique, as described by Edwards and Fletcher,18 did not change over the 21-year study period.
Indications and Comorbidities
The indications for ECT were hyperbilirubinemia or anemia. Hyperbilirubinemia was further classified by etiology (Rh disease, ABO incompatibility, idiopathic hyperbilirubinemia, and other hematologic diagnoses). Patients were considered to have a significant preexisting comorbidity if they were treated with blood pressure support and/or mechanical ventilation, if they had a major congenital anomaly, or if they had any of the following diagnoses: respiratory distress syndrome, intraventricular hemorrhage (all grades as defined by Papile et al19), necrotizing enterocolitis (NEC; modified Bell's criteria at least stage 2a20), or sepsis (defined as a positive blood culture and/or signs and symptoms consistent with sepsis treated with antibiotics for ≥7 days).
ECT-related complications were defined as any complication, not present before the ECT, which occurred within 7 days after the exchange. They were defined as follows: severe thrombocytopenia, platelet count <50000/mm3; hypocalcemia, serum calcium <8.0 mg/dL or plasma ionized calcium <3.5 mg/dL; seizures, clinical evidence of seizure-like activity treated with antiseizure medication; bradycardia, heart rate <100 beats per minute; apnea, cessation of respirations for >20 seconds; catheter malfunction, central venous or arterial catheter thrombosis or rupture; hyperkalemia, serum potassium >6.5 meq/dL associated with electrocardiogram changes; NEC, modified Bell's criteria at least stage 2a20 diagnosed after the ECT; and ECT-related mortality, ECT-related mortality was defined as any death that was directly related to the ECT and occurred within 7 days after the exchange.
SPSS 13.0 (SPSS Inc, Chicago, IL) and GraphPad Prism 3.0 (GraphPad Software, Inc, San Diego, CA) were used for data analyses. Continuous data were compared by using the Student's t comparison of means. Dichotomous data were compared by using a Pearson's χ2 analysis or Fisher's exact test when at least 1 cell contained a value <5. Trends were analyzed by using linear regression analysis. To incorporate both inborn and outborn neonates into this analysis of trends, the number of ECTs was evaluated per 1000 NBSCU admissions. In evaluating inborn neonates separately, the number of ECTs was evaluated per 1000 live births. A P value of <.05 was considered statistically significant.
This study was approved by the institutional review board of the Yale University School of Medicine.
From January 1, 1986, to December 31, 2006, there were 98901 live births at YNHH and 16389 long-term admissions, inborn and outborn, to the NBSCU. One hundred seven infants underwent 141 ECTs from 1986–2006. Two patients in each time period received a single-volume ECT, with the remaining patients receiving a double-volume or near–double-volume exchange. Over the entire study period, there was a statistically significant decline in the number of ECTs performed per 1000 live births in inborn neonates (r2 = 0.30; P = .010; Fig 1A) and per 1000 NBSCU admission in those both inborn and outborn (r2 = 0.49; P < .001; Fig 1B).
Demographic data were similar between the 2 groups, with no statistically significant differences in gestational age, birth weight, race, gender, or age at ECT (Table 1). The rate of phototherapy before exchange did not differ significantly between groups. Neonates in the 1996–2006 group were significantly more likely to receive IVIg before ECT (P = .016; Table 1).
There were no statistically significant differences in the indications for ECT when comparing the 1986–1995 and 1996–2006 groups (Table 2). The most common indication for ECT was hyperbilirubinemia, which was further subdivided into ABO incompatibility, Rh disease, idiopathic hyperbilirubinemia, and other hematologic diagnoses. Other diagnoses included glucose-6-phosphate dehydrogenase deficiency, pyruvate kinase deficiency, fibrosarcoma with large vascular compartment, hemolytic anemia because of Gram-negative sepsis, congenital acute myelogenous leukemia, α-thalassemia, hereditary pyropoikilocytosis, and hereditary spherocytosis. The most common cause of hyperbilirubinemia requiring ECT was Rh disease. Antibodies to non–D Rh antigens were common, occurring in 40% of patients with Rh disease in the first group and in 64% of patients with Rh disease in the second group.
A smaller proportion of patients in the 1996–2006 group experienced an ECT-related complication (Table 3). This result was not statistically significant, possibly because of the small sample size. We observed a high rate of thrombocytopenia and hypocalcemia after ECT in both the 1986–1995 and the 1996–2006 groups, comparable to previous studies.21,22 Despite similar rates of thrombocytopenia and hypocalcemia, patients treated from 1996–2006 were significantly more likely to be transfused platelets or to be given intravenous calcium (Table 4). The retrospective nature of this study and the small sample size make it difficult to determine the causality of these observations. The higher proportion of preexisting comorbidities in the neonates undergoing ECT from 1996 to 2006 may have resulted in more aggressive management or, alternatively, the difference might stem from unidentified changes in our clinical practice over the last 2 decades.
A total of 5 deaths occurred within 7 days of the ECT, none of which were related to the ECT.
Authors of previous reports have hypothesized that premature infants are more susceptible to complications from ECT.21,22 We observed no significant differences in either time period in the frequency of ECT-related complications or their treatment in neonates <1500 g birth weight compared with those >1500 g. Infants <1500 g did not experience increased rates of thrombocytopenia, hypocalcemia, calcium replacement, or platelet transfusion (data not shown). The small sample size of this premature cohort (n = 19) made it difficult to draw any valid conclusions from the analyses.
These data demonstrate a dramatic decline in the frequency of ECT at YNHH over 2 decades, representing the longest single-center, longitudinal documentation of trends in ECT. This decline is likely multifactorial with contributions from advances in both prenatal and postnatal care, such as middle cerebral artery Doppler studies to noninvasively follow fetal anemia,7,8 and IVIg treatment for patients with hemolysis.23 In addition, adoption of the 1994 AAP guidelines may have contributed to this decline.15
The 1994 AAP guidelines recommend that all infants jaundiced in the first 24 hours of life receive a total serum bilirubin and all infants be assessed for jaundice by a health care provider at 2 to 3 days of life. These guidelines also recommend prenatal testing of maternal ABO and Rh types, prenatal screening for unusual maternal antibodies, and screening of the cord blood if the mother was Rh negative or if the mother's ABO type was unknown. The heightened monitoring of all infants for hyperbilirubinemia may have contributed to early detection and treatment of infants with significant jaundice (hemolytic and nonhemolytic) and, therefore, caused a decline in the number of ECT necessary.
The declining rate of ECT has led to speculation that inexperience with the procedure would result in increased rates of ECT-associated morbidity and mortality.4,21–22 Historically, morbidity and mortality associated with ECT steadily declined from the 1950s through the 1970s, a period of time when the procedure was commonly performed,24–27 seeming to reach a nadir in the mid-1970s when Keenan et al27 reported a serious adverse event rate of 5.2% and a mortality rate of 0.5%. As ECT became a less common occurrence in the NICU, reports of ECT-related morbidity and mortality were also less frequent, but described higher rates of ECT-related complications and mortality.21,22 In 1997, Jackson reported data on 106 neonates from 2 NICUs who underwent 140 ECTs from 1980–1995. He described a high rate of serious complications (12%) in the 25 neonates with significant preexisting comorbidities, as well as an overall mortality attributable to ECT of 2%.21 In 2004, Patra et al22 reported data on 55 neonates from 2 high-volume NICUs who underwent 66 ECTs from 1992–2002. Although their rate of ECT was quite low (<3 ECTs in each unit per year), they reported a high rate of ECT-related adverse events (74%) and a mortality rate of 2%.
In contrast to these smaller studies, our overall rate of ECT-related complications did not increase throughout the study period despite a sharp decline in the number of ECTs performed. The majority of complications were transient and, similar to previous studies,21,22 when serious adverse events occurred, they were observed in patients with significant, preexisting comorbidities. In addition, no cases of ECT-related mortality were observed. We speculate that a combination of factors likely contributed to this observation, including the use of a standardized ECT protocol at our institution and increasing attending neonatologist involvement in ECT procedures. The impression of 1 of the authors (Dr Ehrenkranz), who worked in the NBSCU during the entire 21-year study period, is that attending supervision has increased as the number of ECTs declined and has become nearly universal.
In an era when many neonatal care givers have more experience with advanced therapies such as high-frequency ventilation, dialysis, and extracorporeal membrane oxygenation than with ECT, a standardized protocol for performing ECT may be an important tool for decreasing the number of adverse, procedure-related events. Inclusion of ECT in neonatal education will also help minimize ECT-related morbidity and mortality, even as the frequency of ECT continues to decline.
This work was supported, in part, by National Institute of Child Health and Human Development grant T32 HD07094 (to Dr Steiner).
- Accepted February 16, 2007.
- Address correspondence to Patrick G. Gallagher, MD, Department of Pediatrics, Yale University School of Medicine, 333 Cedar St, PO Box 208064, New Haven, CT 06520-8064. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Bowman JM. Nathan and Oski's Hematology of Infancy and Childhood. 5th ed. Philadelphia, PA: WB Saunders; 1993
- ↵Watchko JF, Oski FA. Bilirubin 20 mg/dL = vigintiphobia. Pediatrics.1983;71 :660– 663
- ↵Newman TB, Maisels MJ. Evaluation and treatment of jaundice in the term newborn: a kinder, gentler approach. Pediatrics.1992;89 :809– 818
- ↵American Academy of Pediatrics, Provisional Committee of Quality Improvement and Subcommittee on Hyperbilirubinemia. Practice Parameter: management of hyperbilirubinemia in the healthy term newborn [published correction appears in Pediatrics. 1995;95:458–461]. Pediatrics.1994;94 :558– 562
- ↵Watchko, JF. Vigintiphobia revisited. Pediatrics.2005;115 :1747– 1753
- ↵Edwards MC, Fletcher MA. Exchange transfusions. In: Fletcher MA, MacDonald MG, Avery GB, eds. Procedures in Neonatology. Philadelphia, PA: JB Lippincott Company; 1983:313–326
- ↵Walsh MC, Kliegman RM, Fanaroff AA. Necrotizing enterocolitis: a practitioner's perspective. Pediatr Rev.1988;9 :219– 226
- ↵Jackson, JC. Adverse events associated with exchange transfusion in healthy and ill newborns. Pediatrics.1997:99(5) . Available at: www.pediatrics.org/cgi/content/full/99/5/e7
- ↵Ip S, Chung M, Kulig J, et al. An evidence-based review of important issues concerning neonatal hyperbilirubinemia. Pediatrics.2004;114(1) . Available at www.pediatrics.org/cgi/content/full/114/1/e130
- Boggs TR Jr, Westphal MC Jr. Mortality of exchange transfusion. Pediatrics.1960;26 :744– 745
- Weldon V, Odell G. Mortality risk of exchange transfusion. Pediatrics.1968;41 :797– 801
- ↵Keenan WJ, Novak KK, Sutherland JM, Bryla DA, Fetterly KL. Morbidity and mortality associated with exchange transfusion. Pediatrics.1985;75(suppl) :422– 426
- American Academy of Pediatrics, Subcommittee on Hyperbilirubinemia. Management of hyperbilirubinemia in the newborn infant 35 weeks or more of gestation [published correction appears in Pediatrics. 2004;114:1138]. Pediatrics.2004;114 :297– 316
- Copyright © 2007 by the American Academy of Pediatrics