Comparison of Two Strategies for Surfactant Prophylaxis in Very Premature Infants: A Multicenter Randomized Trial
Introduction. Previous trials of surfactant therapy in premature infants have demonstrated a survival advantage associated with prophylactic therapy as an immediate bolus, compared with the rescue treatment of established respiratory distress syndrome. The optimal strategy for prophylactic therapy, however, remains controversial. When administered as an endotracheal bolus immediately after delivery, surfactant mixes with the absorbing fetal lung fluid and may reach the alveoli before the onset of lung injury. This approach, however, causes a brief delay in the initiation of standard neonatal resuscitation, including positive pressure ventilation, and is associated with a risk for surfactant delivery into the right main stem bronchus or esophagus. As an alternative approach, surfactant prophylaxis may be administered in small aliquots soon after resuscitation and confirmation of endotracheal tube position. Although this strategy has substantial logistical advantages in clinical practice, its efficacy has not been established.
Objective. The purpose of this study was to determine whether the established benefits of the immediate bolus strategy for surfactant prophylaxis could still be achieved using a postventilatory aliquot strategy after initial standard resuscitation and stabilization.
Design. Multicenter randomized clinical trial with patients randomized before delivery to immediate bolus or postventilatory aliquot therapy.
Participants. Inborn premature infants delivered to mothers at an estimated gestational age of 24 to 28 weeks.
Interventions. Those infants who were randomized to the immediate bolus strategy were intubated as rapidly as possible after birth, and a 3-mL intratracheal bolus of calf lung surfactant extract (Infasurf) was administered before the initiation of positive pressure ventilation. Those infants who were randomized to the postventilatory aliquot strategy received standard resuscitation measures with intubation by 5 minutes of age, if not required earlier. At 10 minutes after birth, 3 mL of surfactant was administered in 4 divided aliquots of 0.75 mL each. Patients in both groups were eligible to receive up to three additional doses of surfactant as rescue therapy in the neonatal intensive care unit, if needed.
Outcome Measures. The primary outcome variable was survival to discharge to home. Secondary variables included neonatal complications and requirement for oxygen therapy at 36 weeks' postmenstrual age.
Results. Among three centers, 651 infants were enrolled and randomized before delivery. Survival to discharge to home was similar for the two strategies for surfactant therapy as prophylaxis: 76% for the immediate bolus group and 80% for the postventilatory aliquot group. In a secondary analysis, the rate of supplemental oxygen administration at 36 weeks' postmenstrual age was 18% for the immediate bolus group and 13% for the postventilatory aliquot group.
Conclusions. Survival to discharge to home was similar with immediate bolus and postventilatory aliquot strategies for surfactant prophylaxis. Because of its logistical advantages in the delivery room and its beneficial effects on prolonged oxygen requirements, we recommend the postventilatory aliquot strategy for surfactant prophylaxis of premature infants delivered before 29 weeks' gestation.
Exogenous surfactant replacement is now standard therapy for the prevention and treatment of the neonatal respiratory distress syndrome. The optimal strategy for the first intratracheal instillation, however, remains controversial.1-10 In a previous trial1 comparing surfactant as immediate bolus prophylaxis and as rescue therapy, we demonstrated better survival and a decreased incidence of pneumothorax in the immediate prophylaxis group.
Exogenous surfactant therapy increases lung volume,11improves oxygenation and may attenuate the lung injury associated with assisted ventilation and supplemental oxygen therapy.12-15The addition of exogenous surfactant to the absorbing fetal lung fluid immediately at birth may facilitate surfactant distribution. Immediate bolus prophylaxis, however, briefly delays the initiation of standard neonatal resuscitation including positive pressure ventilation and carries a risk for surfactant delivery to the esophagus or right main-stem bronchus because endotracheal tube position is not yet verified. Merritt et al2 have described a delayed prophylactic approach with surfactant administration immediately after the initiation of positive pressure ventilation and the successful auscultation of bilateral breath sounds to confirm endotracheal tube position. This strategy, however, has not been tested in a controlled fashion compared with initial surfactant therapy immediately after delivery.
We conducted a multicenter randomized trial in premature infants <29 weeks' gestation by dates to test the hypothesis that survival to discharge to home is equivalent with immediate bolus and delayed postventilatory aliquot strategies for surfactant prophylaxis. The goal of this controlled trial was to determine whether the established benefits of the immediate bolus approach could still be achieved using a postventilatory aliquot strategy which is logistically easier to apply in clinical practice.
The trial was conducted in the perinatal and neonatal units of the Children's Hospital, Albany Medical Center, Albany, New York; the University of Rochester Medical Center (Children's Hospital at Strong), Rochester, New York; and the New York Medical College (Westchester Medical Center), Valhalla, New York. The protocol was reviewed and approved by the research subjects review board of each center and a data and safety monitoring committee at the University of Rochester reviewed two interim analyses.
Patient Selection and Randomization
Women who were expected to deliver viable infants between 24 and 28 weeks' gestation were eligible for participation in the study; written informed consent was obtained before delivery. Gestational age was estimated before delivery on the basis of obstetrical history, supplemented by ultrasound evaluations of the fetus, if available. Antenatal corticosteroids were administered at the discretion of the attending obstetricians. Fetuses with known lethal congenital anomalies were excluded.
Infants were randomly assigned before birth to receive surfactant prophylaxis as an immediate bolus or as 4 postventilatory aliquots at 10 minutes after delivery. Randomization was stratified by two estimated gestational age groups: 24 to 26 and 27 to 28 weeks' gestation. Separate sets of randomization cards for each center were prepared by the University of Rochester Department of Biostatistics in blocks of eight and provided in consecutively numbered, sealed opaque envelopes. The block length was unknown to the investigators. The treatment assignment was determined by the opening of the next sealed envelope just before delivery. In the case of multiple births, a separate envelope was opened for each infant. After delivery, investigators were not obligated to administer surfactant as randomized if the infant was obviously previable or was clearly older than 29 weeks' gestation. Blinding of the family and the nurses and physicians caring for the infant was not attempted because of logistical practicalities.
Protocol for Surfactant Administration
The surfactant used in this study was a calf lung surfactant extract (Infasurf) manufactured by ONY Inc, Amherst, NY. A neonatology attending, fellow, or senior pediatric resident attended the delivery of all study patients. Infants assigned to immediate bolus prophylaxis were intubated as rapidly as possible after delivery. The tube was carefully held in place as an assistant administered a 3 mL (105 mg) intratracheal bolus of Infasurf via a small feeding tube which had been precut to 1 cm shorter than the endotracheal tube. Positive pressure ventilation with an anesthesia bag was then initiated and standard resuscitation measures were carried out as indicated. No attempt was made to hold the chest wall or suppress spontaneous respiratory efforts before intubation and infants were not weighed in the delivery room.
Infants who were assigned to prophylaxis as postventilatory aliquots received standard resuscitation measures with bag and mask ventilation or intubation and positive pressure ventilation as needed. Those infants in the postventilatory aliquot group who did not require emergency intubation for their initial resuscitation were intubated electively after assignment of the 5-minute Apgar score. Tube position was confirmed by successful auscultation of bilateral breath sounds during positive pressure ventilation with an anesthesia bag. Beginning 10 minutes after birth, 3 mL (105 mg) of surfactant were administered intratracheally in 4 equal aliquots of 0.75 mL via a precut feeding tube. The postventilatory dose was administered in 4 divided aliquots to minimize any potential for airway occlusion from the administration of a single bolus to an already air-filled lung. Positive pressure ventilation was administered for 2 minutes between each aliquot, and infants remained in the midline supine position during the dosing protocol. The physicians conducting the resuscitation assigned Apgar scores at 1, 5, and 10 minutes of age. Epinephrine was administered for bradycardia based on the clinical judgment of the physicians. On admission to the neonatal intensive care unit, infants were treated with conventional pressure-limited, time-cycled assisted ventilation. Supplemental oxygen and ventilatory support were adjusted to maintain the partial pressure of arterial oxygen in the range of 6.67 to 9.33 kPa (50 to 70 mm Hg) and the partial pressure of arterial carbon dioxide between 5.33 and 6.00 kPa (40 and 45 mm Hg). Continuous pulse oximetry was used for all infants.
Infants in both groups were eligible to receive a second dose of surfactant, as rescue therapy, (100 mg/kg in 4 aliquots with 2 minutes of hand bagging between each aliquot) after 4 hours of age if there was radiographic evidence of respiratory distress syndrome and the fraction of inspired oxygen (Fio2) was 0.4 or greater and the mean airway pressure was 0.686 kPa (7.0 cm of water) or greater. Patients were eligible to receive third and fourth doses (100 mg/kg in 4 aliquots) at 8-hour intervals if the Fio2 was 0.4 or greater and the mean airway pressure was 0.686 kPa (7.0 cm of water) or greater.
Assessments and Outcome Variables
All data were collected and recorded prospectively by research nurses using study data forms with explicit variable definitions. The primary outcome variable was survival to discharge to home. All three study centers had active back transfer policies to return infants to their local community hospitals (Level I and II units) when intensive care was no longer required. The community hospital records were reviewed to confirm survival to discharge to home and length of hospital stay. Secondary outcome variables were: chronic lung disease, defined as a requirement for oxygen at 28 days of age; and the related variable of a requirement for supplemental oxygen at 36 weeks postmenstrual age. Chorioamnionitis was diagnosed by the attending obstetrician on the basis of maternal fever, uterine tenderness, or purulent fluid. Sepsis at birth was defined as a positive initial blood culture. The modified Ballard examination16 was used for the postnatal estimation of gestational age. Inadvertent administration of surfactant into the right lung was determined from a review of the initial chest roentgenograms.17 The pediatric radiologists who read the chest roentgenograms were not aware of the surfactant treatment group assignments. Pneumothorax was identified on chest roentgenograms or by a positive transillumination of the thorax followed by the needle aspiration of air from the pleural space. Pulmonary hemorrhage was defined by the finding of bright red blood in the endotracheal tube associated with an acute respiratory deterioration. Necrotizing enterocolitis was defined as intestinal pneumatosis, portal air, or necrotic bowel on surgical exploration. Intraventricular hemorrhage was identified at autopsy or by cranial ultrasound examinations and graded with a modification of the system of Papile.18 Patent ductus arteriosus was diagnosed if at least two of the following were present: characteristic murmur, bounding pulses, hyperactive precordium, or evidence of pulmonary edema on chest roentgenograms; or if one of these clinical findings was present and cardiac ultrasonography demonstrated patency of the ductus. Eye examinations for retinopathy of prematurity were initiated at 4 to 6 weeks after birth and repeated every 1 to 3 weeks as indicated. Some infants with chronic lung disease were treated with a course of dexamethasone, at the discretion of the attending neonatologist.
In a previous multicenter randomized trial1 of infants <30 weeks' gestation, we demonstrated an eight percentage point improvement in survival with surfactant replacement therapy as immediate prophylaxis (88%) compared with rescue therapy after the development of the respiratory distress syndrome (80%). Because we considered an eight percentage point difference in survival to be clinically important, the present trial was designed to detect at least an 8% difference in survival with a power of 80% and a significance level of 0.05 with a 2-sided test. The projected sample size was 319 patients in each treatment group.
Statistical analysis was based on intention-to-treat and all randomized patients, including stillbirths and infants considered to be previable at birth, were analyzed according to their original treatment group assignments. The primary outcome variable of survival to discharge to home was analyzed using the Mantel-Haenszel test with stratification by center (Albany, Rochester, and Westchester). In a secondary analysis, the primary variable of survival to discharge to home was analyzed with multiple logistic regression. Variables used in this analysis included center, age group (<27 weeks' and 27 weeks' and greater), surfactant treatment strategy (immediate bolus and postventilatory aliquots), gender, race (white and nonwhite), maternal steroids (1 or more doses versus none), duration of ruptured membranes (<24 hours and 24 hours or more), maternal preeclampsia (yes or no), maternal chorioamnionitis (yes or no), neonatal sepsis at birth (yes or no), and mode of delivery (cesarean section versus vaginal). The multiple logistic regression model also included a factor for the interaction between treatment strategy and age group (ie, a differential treatment difference in the two age groups) and a factor for the interaction between treatment strategy and maternal steroid therapy.
Median Apgar scores were compared using the Wilcoxon rank sum test. Rates of neonatal complications were compared using a multiple logistic regression analysis with adjustment for center, treatment group and gestational age group. In addition to testing for an overall treatment difference, we also tested for an interaction between age group and surfactant treatment.
Between February 12, 1991, and November 29, 1994, 651 infants were enrolled at the three study centers; randomization resulted in 323 patients assigned to the immediate bolus group and 328 patients assigned to the postventilatory aliquot group. In the immediate bolus group, there were 2 stillbirths and 9 infants who were considered to be previable after delivery. In the postventilatory aliquot group, there were 3 stillbirths and 4 patients who were considered to be previable after delivery. Ten patients in the immediate bolus group and 5 patients in the postventilatory aliquot group were considered to be too old for the study at the time of birth, and they did not receive surfactant prophylaxis. Twelve (3.7%) patients who had been randomized to the immediate bolus group received their first dose as aliquots and 5 (1.5%) patients who had been assigned to postventilatory aliquot therapy received their first dose as a bolus. The analysis was intention-to-treat and all randomized patients were included and analyzed according to their original randomized assignment. Baseline characteristics (Table 1) were similar for the two treatment strategies, except for gender. In the immediate bolus group, 45% of the patients were female versus 52% in the postventilatory aliquot group.
The primary outcome variable of survival to discharge to home was similar in the two treatment groups. This rate was 76% (244 of 323) for the immediate bolus group and 80% (261 of 328) for the postventilatory aliquot group (P = .22). Thirty-four percent (27 of 79) of the deaths in the immediate bolus group were respiratory related compared with 37% (25 of 67) for the postventilatory aliquot group. Forty-nine percent (39 of 79) of the deaths in the immediate bolus group occurred during the first 7 days, compared with 55% (37 of 67) for the postventilatory aliquot group. In the secondary analysis of survival by multiple logistic regression, gender was not significant (P = .62). Neonatal sepsis at birth (Table 1) was a significant predictor of death (P = .003). Among the other variables in the logistic regression analysis, the only one to approach statistical significance was the term of interaction between antenatal corticosteroids and surfactant treatment group (P = .07). In both surfactant treatment groups, survival to discharge to home was better in the subgroups exposed to one or more doses of antenatal corticosteroids (Table2).
Median Apgar scores at 1 and 5 minutes, respectively, were similar for the two surfactant strategies (Table3). Two infants who had been randomized to the immediate bolus group were resuscitated before they received their bolus infusions of surfactant. Fifteen percent of the patients in the immediate bolus group received epinephrine as part of their resuscitation, versus 20% of the postventilatory aliquot patients. Problems occurred with the administration of the first surfactant dose in 49 of the immediate bolus patients and in 15 of the postventilatory aliquot patients (P < .0001 by Fisher's exact test). Five infants in the immediate bolus group had known or suspected inadvertent administration of surfactant to the right lung, compared with 7 in the postventilatory group. In the immediate bolus group, 17 patients had known or suspected delivery of the surfactant to the esophagus versus 3 patients for the postventilatory aliquot group. After the administration of the first dose in the delivery room, there was leakage of surfactant around the endotracheal tube in 13 of the patients who received immediate bolus therapy and in 3 of the patients receiving postventilatory aliquot therapy. An obstruction of the endotracheal tube which required reintubation occurred in 14 of the immediate bolus group and in 2 of the postventilatory aliquot group.
Doses of Rescue Surfactant
Twenty-seven percent (87 of 323) of the immediate bolus patients and 25% (82 of 328) of the postventilatory aliquot patients received one or more additional doses of rescue surfactant in the neonatal intensive care unit. The distribution of doses required by the two groups was as follows: two doses: 51 (15.8%) for immediate bolus group and 54 (16.5%) for postventilatory aliquot group; three doses: 27 (8.4%) for the immediate bolus group and 17 (5.2%) for postventilatory aliquot group; four doses: 9 (2.8%) for immediate bolus group and 11 (3.4%) for the postventilatory aliquot group.
Oxygen and Mean Airway Pressure Requirements
Fio2 and mean airway pressure requirements were similar for the two treatment groups during the first 72 hours (Figs 1 and2). The need for supplemental oxygen at 28 days of age and at 36 weeks' postmenstrual age were assessed as secondary outcome variables. The immediate bolus and postventilatory aliquot groups did not differ in the percentage of patients requiring supplemental oxygen at 28 days of age, but fewer infants in the postventilatory aliquot group required supplemental oxygen at 36 weeks' postmenstrual age (13% compared with 18% for the immediate bolus group, P < .03 by multiple logistic regression).
The rates of neonatal complications were similar in the two treatment groups (Table 3). Rates of pneumothorax, pulmonary interstitial emphysema, and pulmonary hemorrhage were similar for the two surfactant treatment groups.
In previous comparison trials of surfactant as prophylaxis and as rescue therapy in premature infants, Kendig et al,1Egberts et al,6 Kattwinkel et al,7 Walti et al,8 and Bevilacqua et al9 demonstrated beneficial effects associated with the prophylactic approach, although the timing of the prophylaxis varied from immediate1 to 15 minutes8 in these reports. In a recent meta-analysis of surfactant as prophylaxis versus rescue, Morley10 showed a 39% reduction in neonatal mortality with prophylaxis compared with rescue therapy (odds ratio of 0.55 in favor of prophylaxis).The data from the multicenter randomized trial reported here demonstrate that the benefits associated with prophylaxis can still be achieved when surfactant administration is delayed by 10 minutes for the initial stabilization of the infant in the delivery room. Moreover, there were significantly fewer administration problems associated with the postventilatory aliquot strategy. An additional significant benefit of the poststabilization strategy was a reduction in the number of infants requiring supplemental oxygen therapy at 36 weeks' postmenstrual age. This was an unexpected benefit for the postventilatory aliquot group and we have no specific hypothesis to explain it. Mean arterial pressure and Fio2 requirements during the first 3 days (Figs 1 and 2) were similar for the two treatment groups and the inspired oxygen requirements were similar for the two treatment groups at 28 days of age.
Our trial comparing two strategies for the administration of surfactant as prophylaxis involves more than just a timing issue. With immediate bolus therapy, exogenous surfactant has a greater opportunity to mix with the rapidly resorbing fetal lung fluid, potentially facilitating its delivery and distribution to the alveoli. However, immediate bolus therapy complicates resuscitation and can result in ineffective surfactant delivery if endotracheal tube placement is improper. These drawbacks can be avoided by delaying surfactant delivery for a brief period after birth. Previous work has indicated that a substantial delay in surfactant administration in rescue therapy, with initial doses typically given several hours or more after birth, can lead to reduced efficacy.1,6-10 The results supported our hypothesis that a short 10-minute delay period for surfactant prophylaxis would allow the necessary logistical advantages while not severely compromising efficacy. The decision to administer the 10-minute poststabilization dose in 4 aliquots as opposed to a single bolus was based on the established safety of the aliquot method for the administration of rescue doses of exogenous surfactant to the air-filled lung. An alternative strategy for surfactant delivery, which we did not study, uses an in-line endotracheal tube adapter equipped with a side port which can be connected to a syringe containing surfactant. This device permits a slow continuous administration of exogenous surfactant while the infant remains connected to the ventilator. This method, which maintains lung inflation during the administration of surfactant, represents an alternative method for the administration of postventilatory surfactant prophylaxis.
Strategies for surfactant prophylaxis have been studied in premature lamb models of the neonatal respiratory distress syndrome. Ingimarsson et al19 and Bjorklund et al15 studied lambs delivered at 126 to 128 days gestation and found physiological and histological evidence of lung injury when the lambs were ventilated immediately after cesarean delivery. After a few manual ventilations of only 8 mL/kg, before surfactant therapy, they found a 38% reduction in static compliance at 4 hours of age, compared with lambs receiving a preventilatory bolus dose of a porcine surfactant.
Cummings et al20 compared preventilatory and postventilatory strategies for the administration of a calf lung surfactant extract (Infasurf) with lambs delivered by cesarean section at 127 days gestation. They performed a tracheotomy for surfactant administration before the umbilical cord was ligated. During a 4-hour study period, the lambs receiving the preventilatory surfactant had better measurements of oxygenation and ventilation than those lambs who were ventilated for 5 minutes before receiving a bolus dose of surfactant. There was no histological evidence of lung injury in either group. The surfactant delivery protocol used in the study of Cummings et al20 was quite different from our current trial in infants and long-term outcomes as emphasized here were not assessed. We did not conduct pulmonary function comparisons during the first 4 hours after birth.
Our original experimental design did not stratify patient enrollment by exposure to antenatal corticosteroids and we did not formulate any pretrial hypotheses21 regarding survival in the subgroup of patients exposed to antenatal corticosteroids. Several recent reports22,23 of the additive beneficial effects associated with antenatal corticosteroids and exogenous surfactant therapy prompted us to include a factor for interaction between surfactant therapy and antenatal corticosteroid therapy in our secondary analysis of survival using multiple logistic regression. This factor for interaction showed a trend toward significance (P = .07).
Byar and Corle24 have stressed the importance of testing for treatment-covariate interactions in the analysis of randomized clinical trials. Such an approach permits the comparison of treatments within risk groups and may identify important stratification groups for future randomized trials.25 Our finding of a trend toward a significant interaction between antenatal corticosteroids and exogenous surfactant therapy has important clinical implications and is consistent with the findings reported by Farrell et al22and Jobe et al.23 Unlike the important, but fixed, covariates of gender, race, maternal age, and duration of ruptured membranes, antenatal corticosteroids represent a modifiable prognostic variable worthy of expanded clinical use. The safety and efficacy of antenatal corticosteroids have recently been emphasized in the form of a National Institutes of Health Consensus Conference.26Ryan and Finer27 have emphasized the importance of ensuring high levels of antenatal corticosteroid utilization in conjunction with future randomized perinatal clinical trials. Antenatal corticosteroids not only induce a maturation of the pulmonary surfactant system, but also reduce the incidence of intraventricular hemorrhage,28induce renal maturation,29 and reduce the incidence of patent ductus arteriosus.30
In this multicenter randomized trial in premature infants <29 weeks' gestation, we have shown that a postventilatory aliquot strategy for surfactant prophylaxis is a safe and effective alternative to the immediate bolus approach. Because of the logistical advantages associated with the postventilatory aliquot strategy in clinical practice, we recommend this approach for surfactant prophylaxis in infants born at <29 weeks' gestation.
This work was supported by a Specialized Center of Research (SCOR) Grant HL-36453 and in part by a General Clinical Research Center (GCRC) Grant MOIRR00044, both from the National Institutes of Health.
We wish to thank ONY, Inc, Amherst, NY for supplying the calf lung surfactant extract (Infasurf) used in this study. We are also indebted to Arthur Watts for statistical programming and to Richard W. Hyde, MD; Robert A. Hoekelman, MD; and Ruth A. Lawrence, MD, for serving on the Data Safety and Monitoring Committee for the study.
The following investigators also participated in this multicenter surfactant trial: from the Albany Medical Center: Sue Boynton, RN; Pauline Graziano, RN; Ann Marie Haber, RN; Allan Geis, MD; Joaquim Pinheiro, MD; Judy Aschner MD; Luis Ochoa, MD; and Marilyn Fisher, MD. From the University of Rochester Medical Center: Sanjiv Amin, MD; Kathleen Brown, RN; Lama Charafeddine, MD; Patricia Chess, MD; Carl D'Angio, MD; Andrew Davey, MD; Jonathan Davis, MD; Margaret Donahue, MD; Okyanus Gurel, MD; Rosemary Higgins, MD; Nancy Landfish, MD; Nirupama Laroia, MD; Siraj Masood, MD; Charles Mercier, MD; Gloria Pryhuber, MD; Raymond Sanders, MD; Patrick Spafford, MD; Christopher Stenberg, MD; Timothy Stevens, MD; Robert Swantz, MD; and Carol Wagner, MD. From the Westchester Medical Center: Ana Marie Castillo, MD; Natalie Dweck, RN; Fernando Ginebra, MD; Lilian Gonzalez, MD; Mazen Khalifeh, MD; Sarvesh Nigam, MD; Mario Reale, MD; Maria Sanchez, MD; Jayesh Shah, MD; Glenn Soriano, MD; and Rolando Vilar, MD.
- Received April 3, 1997.
- Accepted October 2, 1997.
- Address correspondence to: James W. Kendig, MD, Division of Neonatology, Box 651, Department of Pediatrics, University of Rochester Medical Center, 601 Elmwood Ave, Rochester, NY 14642.
Presented in part at the Annual Meeting of the Pediatric Academic Societies, May 9, 1996, Washington DC.
Reprints not available.
- Fio2 =
- fraction of inspired oxygen
- Dunn MS,
- Shennan AT,
- Zayack D,
- Possmayer F
- Egberts J,
- deWinter JP,
- Sedin G,
- et al.
- Kattwinkel J,
- Bloom BT,
- Delmore P,
- et al.
- Bevilacqua G,
- Parmigiani S,
- Robertson B
- Matalon S,
- Holm BA,
- Notter RH
- Huang Y-C,
- Sane AC,
- Simonson SG,
- et al.
- Piantadosi CA,
- Fracica PJ,
- Duhaylongsod FG,
- et al.
- Bjorklund LJ,
- Curstedt T,
- Ingimarsson J,
- et al.
- Ingimarsson J,
- Bjorklund LJ,
- Curstedt T,
- et al.
- Copyright © 1998 American Academy of Pediatrics