Surfactant Treatment of Neonates With Respiratory Failure and Group B Streptococcal Infection
Objective. Connatal pneumonia caused by group B streptococcal (GBS) infection may be associated with surfactant dysfunction. We investigated the effects of surfactant treatment in term and preterm neonates with GBS infection and respiratory failure, in comparison with corresponding data from a control population of noninfected infants treated with surfactant for respiratory distress syndrome (RDS).
Design/Methods. The study comprised 118 infants with respiratory failure, clinical and/or laboratory signs of acute inflammatory disease, and GBS infection proven by culture results. They were recruited retrospectively from a database of patients treated with surfactant at 28 neonatology units participating in European multicenter trials (1987–1993) and prospectively from the same units in the following years. A nonrandomized control group of 236 noninfected infants was selected from the same database. The primary parameters evaluated were oxygen requirement, ventilator settings, and incidence of complications.
Results. Median birth weight in the GBS study group was 1468 g (25th–75th percentiles: 1015–2170), and median gestational age was 30 (27–33) weeks. Thirty-one percent of the infants weighed >2000 g. Median age at surfactant treatment was 6 hours. The mean initial surfactant dose was 142 mg/kg (standard deviation: 53). Ninety of the infants were treated with Curosurf (Chiesi Farmaceutici, Parma, Italy), 13 with Survanta (Abboth GmbH, Wiesbaden, Germany), 12 with Alveofact (Dr Karl Thomae GmbH, Biberach, Germany), and 3 with Exosurf (Wellcome GmbH, Burgwedel, Germany). Within 1 hour of surfactant treatment, median fraction of inspiratory oxygen was reduced from .84 (25th–75th percentiles: .63–1.0) to .50 (.35–.80). The incidence of complications in the study group (mortality: 30%; pneumothorax: 16%; intracranial hemorrhage: 42%) was high, compared with infants with RDS.
Conclusions. Surfactant therapy improves gas exchange in the majority of patients with GBS pneumonia. The response to surfactant is slower than in infants with RDS, and repeated surfactant doses are often needed. The mortality and morbidity are substantial, considering the relatively high mean birth weight of the treated infants.
- GBS =
- group B streptococcus(al) •
- RDS =
- respiratory distress syndrome •
- CRP =
- C-reactive protein •
- I:T ratio =
- immature to total ratio •
- BW =
- body weight •
- Pao2 =
- partial pressure of arterial oxygen •
- MAP =
- mean airway pressure •
- ICH =
- intracranial hemorrhage •
- OR =
- odds ratio •
- CI =
- confidence interval
Group B streptococcus (GBS) is a common cause of systemic and pulmonary infections in the neonatal period. One to 4 per 1000 newborn infants suffer from early-onset GBS septicemia. Although the prognosis of term neonates with connatal GBS infections has improved considerably over the last decade, mortality is still substantial in affected premature infants.1
Most neonates with systemic or pulmonary GBS infection have respiratory symptoms.1,2 Clinical, laboratory, and radiologic signs cannot differentiate with certainty between idiopathic respiratory distress syndrome (RDS) and GBS pneumonia in the early course of the disease.2 Data from animal experiments indicate that respiratory failure in neonatal GBS pneumonia is in part caused by surfactant deficiency. We could recently demonstrate that surfactant improved lung function and mitigated bacterial growth in immature rabbits with experimentally induced neonatal GBS pneumonia.3 There is also anecdotal evidence from small uncontrolled clinical trials4,5 and a recent multicenter study6 that surfactant treatment can improve gas exchange in infants with congenital pneumonia.
The aim of the present study was to evaluate the effects of surfactant treatment on oxygen requirements, ventilator settings, and outcome parameters in term and preterm neonates with severe respiratory failure triggered by GBS infection.
The prospective study started in May 1993 and continued until April 1998. In addition, we asked for retrospective evaluation of surfactant-treated, GBS-infected infants from 1987 onward. Data from 28 European centers (see “Appendix”) were collected using a standardized case record form requesting information on pregnancy complications, laboratory and radiologic findings, gas exchange, and ventilatory parameters before and after surfactant replacement as well as outcome parameters. For definition of outcome variables and disease severity, we used identical diagnostic criteria as previously applied in trials organized by the Collaborative European Multicenter Study Group.7–11
Infants eligible for the study had to fulfill the following criteria:
GBS infection verified by bacterial cultures (from blood, cerebrospinal fluid, tracheal aspirate fluid, gastric aspirate fluid, and/or skin swabs). In addition, clinical (eg, lethargy, greyish skin color, apnea) and/or laboratory signs (C-reactive protein [CRP]: >1 mg/dL; leukocytopenia: <10 000/μL; immature to total ratio of neutrophilic granulocytes [I:T ratio]: >.16; thrombocytopenia: <150 000/μL) of acute inflammatory disease had to be present to exclude infants with pure colonization.
Respiratory failure requiring mechanical ventilation.
Infants with severe malformations or evidence of GBS infection only by immunologic methods (latex agglutination tests) were not included.
From the database of previous clinical trials of surfactant replacement therapy,8–10 a control group of noninfected neonates was recruited. The selected infants were treated with surfactant for severe RDS but had no evidence of bacterial infection in blood cultures or tracheal aspirate fluid. Radiologic signs of pneumonia (patchy infiltrates) had to be absent in the first week of life.
The documentation of study variables (ie, characterization of infants, ventilatory parameters, and outcome) in the previous studies was identical to that in the present investigation, which enabled us to perform a pooled analysis of the data. Imbalances between the GBS and the noninfected control group with respect to the distribution of some variables—resulting from the eligibility criteria for recruiting infants in the clinical trials—were accounted for in the statistical analysis by the propensity score technique (see below).
Descriptive statistical methods were used to characterize the distribution of relevant variables in the GBS or the RDS group. The statistical comparisons performed between groups of patients are based on Fisher's exact test, the Wilcoxon rank-sum test (for comparisons between the RDS and the GBS groups and comparisons between term and preterm infants), the Kruskal-Wallis test (for comparisons between >2 groups of patients), the Wilcoxon signed-rank test (for comparions of oxygen demand and mean airway pressure between time points), and the log-rank test. Details are given in Tables 1 and 5 and Figs 1 and 2. These direct comparisons, however, must be interpreted with caution, because the RDS infants cannot be considered equivalent to controls from a randomized, controlled trial.
The relative risk of an adverse outcome (death or intracranial hemorrhage [ICH] or oxygen dependency on day 28 in GBS-infected vs noninfected RDS infants) was calculated by means of the recently described propensity score adjustment technique12,13 to account for the effects of different distribution of sex, gestational age, birth weight, Fio 2 before therapy, and Apgar score at 5 minutes. All of these variables have been shown to influence the outcome of infants in previous surfactant trials.11 All statistical analyses were performed with the statistical program SAS, Version 6.12 (SAS Institute Inc, Cary, NC).
Characterization of Study Infants
One hundred twenty-four completed case record forms were obtained from the 28 centers listed in the “Appendix.” Fifty of these cases (42%) were collected retrospectively until May 1993. Six patients were excluded because of violation of the entry criteria. In 2 of these cases, GBS colonization rather than GBS infection had to be assumed, because the infants did not show clinical or laboratory signs of infection. In the other 4 excluded patients, GBS colonization was diagnosed prenatally or in placental swabs, but there was no direct evidence of GBS infection in cultures taken from the infants (eg, attributable to prenatal antibiotic treatment of the mother). Median birth weight was close to 1500 g, 36 infants (31%) weighed >2000 g and 28 (24%) had a birth weight <1000 g. The majority of infants were preterm. However, in 23 infants (19%) gestational age was >35 weeks, indicating that severe respiratory failure can also develop in term infants with presumably mature surfactant system.
Other characterization variables are listed in Table 1. The control group consisted of 236 noninfected infants. These patients tended to be more immature. However, differences in birth weight or gestational age were accounted for in the outcome analysis by the propensity score technique.
Clinical and Laboratory Signs of Infection
Prolonged rupture of membranes and clinical evidence of maternal infection (fever, elevated level of CRP) were frequent complications in the case series (Table 2). However, culture results from rectal/vaginal swabs were available in only 49 pregnancies (42%). Fifty-five percent of these cultures were GBS-positive. Prenatal antibiotic treatment was initiated in 25% of cases (Table 2).
Clinical signs of infection (eg, lethargy, greyish skin color, apnea) were present in 72% of the infants (Table 3). Leukocytopenia was the most common laboratory finding indicating infection. The median age at diagnosis was 4 hours (25th– 75th percentiles: 1–17), indicating that the onset of infection was probably intrauterine in many cases. I:T ratio of neutrophilic granulocytes increased >.16 in 60% of patients at a median age of 8 (2–22) hours; however, differential blood counts were not available on a 24-hour basis in all units. As expected from previous studies, serum CRP reacted more slowly and turned positive (>1 mg/dL) at a median age of 19 (6–30) hours.
Septicemia/meningitis with GBS-positive blood or cerebrospinal fluid cultures was found in 40% of infants. Sixty percent of the GBS-positive blood cultures were obtained within the first 24 hours of life. Only 13% of GBS-positive blood cultures were obtained after 72 hours of life. Lumbar puncture was only performed in 23 neonates. Four had meningitis with GBS present in cultures from cerebrospinal fluid (Table 3). One infant was included in the trial in the septicemia group without available culture results on admission. The infant died within 24 hours, and GBS was isolated in postmortem cultures from the spleen and the lung.
In 27 infants (23%) blood cultures were negative but GBS was isolated from tracheal and/or gastric aspirate fluid. However, such cultures were not routinely performed in all centers. In 44 neonates (37%), GBS could only be isolated from skin swabs.
Antibiotic treatment was initiated in 50% of infants within 2 hours of birth. The most commonly used combination was ampicillin/amoxycillin plus an aminoglycoside. Immunoglobulins were administered to 27 infants (23%).
Median age at surfactant treatment was 6 hours and a relatively high mean initial phospholipid dose of 142 mg/kg body weight (BW) was given (Table 4). Most infants were treated with multiple surfactant doses, and 1 infant received no less than 8 doses (cumulative dose: 890 mg/kg BW). In 2 infants the synthetic surfactant Exosurf (Wellcome GmbH, Burgwedel, Germany) was given initially without success, and a modified natural surfactant preparation was used for retreatment. In 1 infant 2 natural surfactants of different origin were used. Because of the small numbers and because the infants were not randomized to various treatment groups, no further attempts were made to compare the efficacy of the different surfactants used in this study.
Radiologic findings on chest films before surfactant replacement, graded according to Giedion et al,14 were compatible with RDS grade III or IV in 56% of cases in GBS-infected neonates. Patchy infiltrates indicating pneumonia were found in 36%. In the noninfected control group, 83% of infants had radiologic findings corresponding to RDS grade III or IV (see Fig 1).
In GBS infected patients, the median Fio 2 could be reduced within 1 hour from .84 (25th–75th percentiles: .63–1.0) to .50 (.35–.80;P < .01). However, there was a considerable number of nonresponders. Twenty-five percent of the GBS-infected neonates still needed >80% oxygen 1 hour after surfactant instillation (Fig 2A). The improvement in gas exchange (partial pressure of arterial oxygen [Pao 2]/Fio 2) was clearly slower than in the control group of infants with RDS without infection (Fig 2B). From 1 to 72 hours after surfactant treatment, Fio 2 values were significantly lower (P < .001, Wilcoxon rank-sum test) and Pao 2/Fio 2values were significantly higher (P < .01) in the RDS group. Mean airway pressure (MAP) could only be reduced very slowly after surfactant replacement (Fig 2C). Significant differences (P < .05) in MAP between the groups could only be observed from 1 to 48 hours after surfactant treatment.
In infants (n = 14) receiving an initial surfactant dose below 100 mg/kg BW, the median reduction in Fio 2 within 1 hours of surfactant instillation was only .10 (25th–75th percentiles: .00–.23). After a dose of 100 to 199 mg/kg BW (n = 62), Fio 2 could be reduced by .24 (.05–.40) within 1 hours of surfactant therapy. A phospholipid dose of ≥200 mg/kg BW (n = 42) decreased Fio 2 by .30 (.02–.40) within 60 minutes. Although this indicates a clear trend toward better acute improvement in gas exchange with a surfactant dose ≥200 mg/kg BW, these differences did not reach the limit level for statistical significance (P = .07) because of relatively small numbers in the subgroups and a considerable variability in the individual response to surfactant treatment.
Neonates with GBS septicemia and/or meningitis demonstrated a slower response to surfactant treatment (Fio 2 1 hours after surfactant treatment: .60 [.40–.95]; P < .05, Kruskal-Wallis test vs all other groups) than did infants with cultural proof of GBS in only gastric/tracheal aspirate fluid or skin swabs (Fig 3).
Twenty-three neonates (19%) had a gestational age of >35 weeks, which is usually associated with mature lung function. However, surfactant improved gas exchange in these infants as well, indicating that secondary surfactant deficiency had developed during the course of the GBS infection. These infants demonstrated a good initial response to surfactant but needed significantly more oxygen throughout the first 24 hours than did preterm GBS-infected neonates (P < .05 at 12 hours and P < .01 at 24 hours, Wilcoxon rank-sum test, after surfactant treatment), in whom some degree of primary surfactant deficiency attributable to immaturity might have contributed to the disease (Fig 4).
The complication rate in the GBS-infected infants was high: overall 30% died, 16% developed pneumothorax, and 43% ICH (Table 5). Severe intraventricular hemorrhage (ICH grade III or IV) was found in 19% of the cases. In the noninfected GBS group mortality was 19% and the incidence of ICH or pneumothorax was 35% or 13%, respectively. Prognosis was worst in the GBS septicemia subgroup, in which 49% of the patients died. In contrast, mortality in the groups with proof of GBS infection only in skin swabs or tracheal/gastric aspirate fluid was 19% and 14%, respectively. Interestingly, a low incidence of patent ductus arteriosus requiring indomethacin treatment or surgical ligation was observed in GBS-infected patients. This might be the consequence of increased pulmonary vascular resistance in the infected neonates.
The median duration of hospital stay in the GBS group was 32 days (25th–75th percentiles: 5–68). The infants were ventilated for 5 (3–11) days and needed supplemental oxygen for 7 (3–21) days. In 5 infants high-frequency oscillatory ventilation was applied before surfactant treatment. In 13 other infants, high-frequency oscillatory ventilation was initiated within 72 hours of surfactant instillation. Five infants were treated with inhaled nitric oxide for pulmonary hypertension. Two patients underwent extracorporeal membrane oxygenation and both survived.
Because differences in variables influencing outcome, such as birth weight or oxygen demand before surfactant treatment, were observed between the GBS and the RDS groups, the propensity score technique was used to calculate odds ratios (ORs) and asymptotic 95% confidence intervals (CIs) to describe the relative risks of different complications for a surfactant-treated, GBS-infected patient, compared with noninfected controls with similar disease severity receiving surfactant treatment for RDS. At 28 days the combined relative risk for death, ICH, and/or chronic lung disease was more than doubled (OR: 2.25; 95% CI: 1.3–3.97) for GBS-infected infants.
GBS colonization is a common finding in pregnant women with incidences ranging from 10% to 40%. In approximately one half of the infants born to colonized mothers, GBS can be detected on the skin of the neonate. However, only 1% of infants born to colonized mothers develop severe systemic and/or pulmonary infection.15–17Infection with GBS can cause a disease similar to acute (adult) RDS in term and preterm neonates. Attack rate, morbidity, and mortality of GBS-induced respiratory failure is high in premature infants, especially after prolonged rupture of the membranes.1However, severe respiratory failure can also develop in GBS-infected term infants with basically mature lung function. More than one third of the surfactant-treated, GBS-infected infants in the present study had a BW above 2000 g or a gestational age above 35 weeks. This category of infants was excluded from treatment in most of the previous surfactant trials.
The early onset of respiratory symptoms and the fact that leukocytopenia is often present soon after birth (at a median age of 4 hours in this study) indicate that infection often begins in utero. In our study 50% of all infants had received antibiotic treatment before 2 hours of age, which is in keeping with the clinical strategy of liberal administration of antimicrobial chemotherapy used in most centers. This strategy implies that all infants with severe RDS should immediately be treated with antibiotics for at least 48 hours until the results from laboratory investigations and bacterial cultures are available.
Recent data indicate that intrauterine infection and the resulting increased cytokine levels might have detrimental effects on cerebral and systemic perfusion.18 These findings might at least in part explain the high incidence of ICH in the relatively mature infants treated in our investigation and, furthermore, suggest that even immediate postnatal antibiotic treatment may only reduce morbidity and mortality to a certain extent. For this reason different strategies for antenatal screening and clinical management of perinatal GBS infections have been suggested.16,17 In our study ∼38% of mothers suffered from prolonged rupture of membranes and 36% had signs of maternal infection. However, only 25% of the mothers received antenatal antibiotic treatment. A recent controlled, randomized trial confirmed the advantages of intravenous antibiotic treatment in all pregnant women with rupture of membranes and imminent premature birth.19 Screening and prophylaxis, thus, seem the most promising ways to reduce perinatal mortality and morbidity related to GBS infections. A recent population-based study estimated that effective vaccination of pregnant women might prevent as many as 4100 cases of severe neonatal GBS septicemia per year in the United States, resulting in an estimated savings of 131 million dollars.20
We were able to demonstrate a significant improvement in gas exchange in GBS infected neonates after surfactant treatment. The design of the study did not allow the conclusion that this improvement was associated with a decrease in mortality or complication rate. Unfortunately, it is not possible to study the effects of surfactant for treatment of GBS pneumonia in a prospective, randomized, controlled trial. Although leukocytopenia, increased I:T ratio, and elevated serum CRP develop in the majority of cases within the first 24 hours, there is no way to distinguish with certainty between idiopathic RDS and pneumonia in a premature infant during the first hours after birth, ie, at the time when surfactant therapy should be considered. Because we and others have demonstrated that surfactant improves gas exchange and survival in infants with immature lungs, it would seem unethical to perform a controlled, randomized study withholding surfactant treatment from the control group—especially because in infants with RDS, good response to surfactant treatment corresponds to a low complication rate.11,21 Moreover, in a premature infant both lung immaturity and pulmonary or systemic infection might contribute to the severity of respiratory failure. In our study this seems to be true particularly for infants with proof of GBS infection only from skin swabs or gastric/tracheal aspirates. However, it has been shown that premature infants with GBS isolated from blood cultures nearly uniformly demonstrate severe clinical signs of infection, with RDS being the one of most frequently diagnosed presenting symptoms.1
Term infants and neonates with GBS septicemia demonstrated a slower response to surfactant treatment. This might be because of a larger amount of surfactant inhibitors in the bronchoalveolar space. Probably, circulatory problems and a certain degree of pulmonary hypertension contributed to the disease severity in this subgroup of patients. In our study, mortality in the group of infants with GBS septicemia was as high as 49%, indicating that problems other than surfactant deficiency play a decisive role in the pathophysiology of the disease.
When we compared the response to surfactant treatment with a group of noninfected neonates receiving surfactant treatment for severe RDS, a higher percentage of nonresponders, a slower reduction in oxygen demand, and an increased incidence of complications were observed in the GBS-infected patients. Even after adjustment for confounding variables by means of the propensity score technique, the relative risk for GBS-infected infants to die or develop ICH or chronic lung diseases was more than doubled. Possibly a higher dose of surfactant and earlier treatment might have improved these results. In a recently published study on surfactant treatment of meconium aspiration syndrome administered before 6 hours of age, a good effect on oxygenation was only observed after a cumulative dose of 300 mg/kg BW had been instilled into the airways.22 An initial surfactant dose of 300 mg/kg was also used successfully in adult patients with acute (adult) RDS caused by sepsis.23 Early treatment with large doses of surfactant may be required to overcome the presence of intraalveolar surfactant inhibitors in this category of patients and prevent ventilator-induced lung injury.
Apart from its well-characterized biophysical properties, surfactant probably also plays an important role in the antibacterial defense system of the lung.24 In agreement with data presented by other investigators,25 we found that surfactant reduced bacterial proliferation in an animal model of neonatal GBS pneumonia.3,26 Recently, the significance of the hydrophilic surfactant proteins SP-A and SP-D for host defense has been demonstrated in experiments with GBS-infected, SP-A-deficient mice.27 SP-A and SP-D are not present in any of the surfactants used in this study. In the future new synthetic surfactant preparations will probably be designed to enhance bacteriostatic effects and maximize resistance to inhibition by plasma proteins leaking into the airways.28
We conclude that surfactant replacement, if necessary with high and repeated doses, improves gas exchange in term and preterm neonates with respiratory failure attributable to GBS infection. However, in view of the high mortality and morbidity of the disease, prevention strategies, hopefully including vaccination of pregnant women against GBS,15–17 remain of prime importance. The possibility to use surfactant as a carrier for antibiotics29,30 or specific immunoglobulins31 against the polysaccharide capsule of GBS is under current investigation in animal models.
Trial collaborators (alphabetical order of places): J. Koppe, Academic Medical Center, Amsterdam, The Netherlands; C. Papagaroufalis, Aghia Sophia Children's Hospital, Athens, Greece; F. Porz, U. Bernsau, Kinderklinik, Klinikum Augsburg, Germany; X. Kravel, A. Riverola, Hospital San Juan De Deu, Barcelona, Spain; D. Sweet, H. L. Halliday, Royal Maternity Hospital, Belfast, Northern Ireland, United Kingdom; A. Valls i Soler, Hospital De Cruces, Barakaldo-Bizkaia, Spain; M. Wagner, H. Segerer, M. Obladen, Kaisevin Auguste Victoria Krankenhaus, Virchow Klinikum, Berlin, Germany; H. Boenisch, Kinderklinik, Städtische Kliniken, Braunschweig, Germany; Anne Clercx, University Hospital, Bruxelles, Belgium; J. Dinger, Universitäts-Kinderklinik, Dresden, Germany; G. Rinaldi, R. Magaldi, Neonatology Service, Hospital of Foggia, Foggia, Italy; P. Tuo, G. Silvestri, G. Gaslini Sick Children Hospital, Genua, Italy; C. Roll, L. Hanssler, Universitäts-Kinderklinik, Essen, Germany; H. Schiffmann, R. Noth, Universitäts-Kinderklinik, Göttingen, Germany; L. Pralle, S. Domhof, Department of Medical Statitistics, Göttingen, Germany; G. Pichler, W. Müller, Universitäts-Kinderklinik, Graz, Austria; C. Altfeld, J. Natzschka, Kinderkrankenhaus auf der Bult, Hannover, Germany; P. Groneck, Kinderklinik, Amsterdamer Str, Köln, Germany; F. Morandi, Neonatal Pathology Service, Lecco Hospital, Lecco (Como), Italy; J. Babnik, Department of Obstetrics, Ljubljana, Slovenia; G. Compagnoni, Neonatal Pathology Service, Mantua Hospital, Mantua, Italy; B. Foresti, G. Marraro, Department of Anesthesiology, L. Mandic Hospital, Merate, Italy; F. Porcelli, G. Motta, Pediatric Hospital V. Buzzi, Milano, Italy; N. W. Svenningsen, University Hospital, Lund, Sweden; V. Zanardo, Neonatal Intensive Care Unit, Pediatric Hospital, Padua, Italy; M. Kassis, M. Voyer, H. Walti, J. P. Relier, Institute de Puericulture de Paris and Hôpital Port Royal, Paris, France; G. Bevilacqua, S. Parmigiani, Department of Neonatology, Pediatric University Hospital, Parma, Italy; G. Noack, Karolinska and St Göran's Hospital, Stockholm, Sweden; M. de Kleine, St Joseph Ziekenhus, Veldhoven, The Netherlands.
The trial is dedicated to the memory of our friend and collaborator Nils W. Svenningsen from the University of Lund in Sweden who died on January 30, 1998.
Some of the studies from which the patients were recruited were supported by Chiesi Farmaceutici and/or Ares Serono. However, the studies were designed, conducted, and analyzed independently of the pharmaceutical industry.
The study was supported by the German (DFG He 2072/2-1/2-2) and the Swedish Medical Research Councils (Project 3351), Konung Oscar II:s Jubileumsfond and a collaborative project (Project 313/S-PPP) of the German Academic Exchange Service (Deutscher Akademischer Austauschdienst) and the Swedish Institute.
- Received July 6, 1999.
- Accepted February 16, 2000.
Reprint requests to (E.H.) Department of Pediatrics, University of Göttingen, Robert-Koch Str 40, D-37075, Göttingen, Germany. E-mail:
- Auten RL,
- Notter RH,
- Kendig JW,
- et al.
- Khammash H,
- Perlman M,
- Wojtulewicz J,
- et al.
- Collaborative European Multicenter Study Group
- Speer CP,
- Harms K,
- Herting E,
- et al.
- Speer CP,
- Robertson B,
- Curstedt T,
- et al.
- American Academy of Pediatrics, Committee on Infectious Diseases and Committee on Fetus and Newborn
- Findlay RD,
- Taeusch HW,
- Walther FJ
- Wright JR. Host defense functions of surfactant. In: Rooney SA, ed. Lung Surfactant: Cellular and Molecular Processing. Austin, TX: R. G. Landes; 1998:191–214
- LeVine AM,
- Bruno MD,
- Huelsman KM,
- Ross GF,
- Whitsett JA,
- Korfhagen T
- Sherman MP, Hall SL, Johnson JT, Aeberhard EE. Formulating a surfactant therapy for preterm infants who have bacterial pneumonia and respiratory distress syndrome. In: Bevilacqua G, Parmigiani S, Robertson B, eds. Surfactant in Clinical Practice. London, United Kingdom: Harwood Academic Publishers; 1993:103–111
- Copyright © 2000 American Academy of Pediatrics