The etiology of chronic lung disease in preterm infants is thought to be multifactorial. Factors implicated in the pathogenesis of CLD include barotrauma associated with mechanical ventilation, fluid overload, patent ductus arteriosus, genetic predisposition, infection, and damage to the lung from free oxygen radicals (oxygen toxicity). The relative contribution of these factors is poorly understood.
In adults, free radicals have been implicated in aging, atherosclerosis, cataracts, neurodegenerative disorders, neoplasia and other disorders (Knight 1998). In healthy humans a balance exists between oxygen derived free radical production and their inactivation by antioxidant defences. Disturbances in this balance may contribute to the pathogenesis of certain disease processes seen in the preterm infant such as chronic lung disease (Saugstad 1990, Kelly 1993, Fardy 1995) retinopathy of prematurity (Saugstad 1990, Kelly 1993), intraventricular hemorrhage (Kelly 1993) and periventricular leukomalacia (Volpe 1997). Preterm infants are often exposed to excessive oxidative stress because they are exposed to high oxygen concentrations due to surfactant deficiency and immature lungs. In addition, preterm infants have inadequate antioxidant defences and are not able to induce antioxyenzymes in response to oxidative stress (Saugstad 1998, Davis 1998). Also, inflammation and infection, which are closely linked to oxidative stress, are more common in preterm infants (Saugstad 1998).
A free radical is an atom, molecule or compound that contains an unpaired electron. Radicals produced endogenously in the body include the superoxide and hydroxyl radicals, hydrogen peroxide, hypochlorous acid, peroxynitrite, and nitric oxide. Free radicals are produced in abundance in all cells. However numerous natural defences exist either to prevent their formation or to neutralize them after they are produced. One of these defences is the intracellular enzyme superoxide dismutase, which dismutates the extremely toxic superoxide radical into potentially less toxic hydrogen peroxide. Superoxide dismutase appears in two forms: one in the cytoplasm of the cell or in the extracellular spaces with two subunits, each with one equivalent of Cu2+ and Zn2+; the other in the mitochondrion with Mn2+ as its subunit.
This systematic review will review all randomized trials of exogenously administered superoxide dismutase for the prevention of chronic lung disease in preterm infants receiving mechanical ventilation.
1. Published manuscripts: We searched Medline (1966 - 2000) and the Cochrane Controlled Trials Register (CCTR) from the Cochrane Library (2000, issue 4). We did not limit the search to any language. We used the following keywords for the search: {bronchopulmonary dysplasia OR chronic lung disease} AND superoxide dismutase, limited to human studies in newborn infants (infant, newborn). From the resulting studies we manually extracted studies on preterm infants undergoing mechanical ventilation. We also hand searched the reference lists of articles obtained.
2. Published abstracts: We hand searched the abstracts of the Society for Pediatric Research (USA, published in Pediatric Research) from 1983 - 2000. We identified abstracts that had the key word superoxide dismutase in the title.
3. Selection process: We selected only randomized controlled trials fulfilling the selection criteria described in the previous section. Selection was done separately by two investigators. There were no disagreements about which studies to include or exclude.
We originally intended to divide the review into two sections: Studies in which superoxide dismutase was used in preterm infants at risk of developing respiratory distress syndrome and those where it was used in preterm infants who had already developed respiratory distress syndrome. However, the subjects in the two studies selected for inclusion in this review were preterm infants who had already developed respiratory distress syndrome and were receiving mechanical ventilation. Therefore we did not divide the review into the sections originally planned.
1. Data extraction and entry: Two investigators extracted, assessed and coded separately all data for each study, using a form that was designed specifically for this review. There was no disagreement between the two reviewers. We did not have to obtain additional information from the original authors. For each study, final data was entered into RevMan by one reviewer (GKS) and then checked by a second reviewer (RFS).
2. Planned subgroup analyses: We originally planned subgroup analyses based on gestational age at birth, type of superoxide dismutase used and the dose of superoxide dismutase used. However we did not perform such analysis in the review because only two studies were selected for inclusion.
3. Criteria for assessing the methodological quality of the studies: We used the standard method of the Cochrane Neonatal Review Group. We assessed the methodological quality of the studies by assessing the risk for four types of bias (selection, performance, attrition and detection). Each study was assessed separately by two reviewers. Disagreements were resolved by discussion between the reviewers. Information about the method of randomization used in Davis 1997 was provided by JMD, one of the authors of this review.
4. Obtaining combined outcomes: From the data reported by Davis 1997, we added the number of infants who had bronchopulmonary dysplasia at 28 days to the number of infants who were dead at 28 days to obtain the combined outcome of bronchopulmonary dysplasia or death at 28 days. Similarly, from the data reported by Rosenfeld 1984, we added the number of survivors who had respiratory problems after discharge to the number of infants who had died before discharge to obtain the combined outcome of respiratory problems after discharge or death.
5. Statistical analysis: We used the standard method of the Cochrane Neonatal Review Group. In assessing the treatment effects we used the relative risk and risk difference for categorical outcomes. For outcomes measured on a continuous scale we used the mean difference. Whenever possible, 95% confidence intervals were used in addition to point estimates of treatment effects. We used a fixed effect model for meta-analysis.
Rosenfeld 1984 was a single center study and enrolled infants with respiratory distress syndrome who were on mechanical ventilation. Infants with major congenital anomalies were excluded, as were infants with other causes of respiratory distress, including aspiration, sepsis, pneumonia and drug withdrawal. Infants had a mean gestation of 28.7 weeks (range 24 - 35 weeks) and a mean birth weight of 1154 grams ( range 560 - 2260 grams). Surfactant was not used in this study. Rosenfeld 1984 used bovine superoxide dismutase 0.25 mg/kg or an equal volume of saline placebo administered subcutaneously every 12 hours until ventilation or continuous positive airway pressure were no longer needed and the infants were in room air. The time of initiation of therapy is not mentioned in the article. The outcomes reported in this study included bronchopulmonary dysplasia, findings on chest radiograph and plasma superoxide dismutase levels. Bronchopulmonary dysplasia was not defined in this study and appears to have been diagnosed on radiologic criteria without any clinical criteria such as oxygen requirement.
Davis 1997 was a multicenter study from six
participating hospitals. This study enrolled infants with respiratory distress
syndrome who were on mechanical ventilation. Surfactant was used as part
of therapy. Infants in this study had a birth weight between 700 grams
to 1300 grams, with a mean gestation of approximately 27 weeks. Infants
with major congenital anomalies were excluded, as were infants with congenital
infection and perinatal asphyxia. Davis 1997 used recombinant human superoxide
dismutase or saline placebo administered intratracheally in an initial
volume of 1 ml/kg. The initial drug or placebo was administered separately
from surfactant, in two aliquots over a 1-minute period within 30 to 120
minutes after surfactant administration. Repeat doses (suspended in 2 ml/kg
of saline) were administered every 48 hours for up to seven doses, as long
as the infant continued to require intubation and mechanical ventilation.
Two dosage regimens were used for superoxide dismutase - 2.5 mg/kg and
5 mg/kg. In this review we combined the data from patients from both dosage
regimens as no dose-response relationship was evident in the results. Davis
1997 reported the following outcomes: bronchopulmonary dysplasia (defined
as oxygen dependency at 28 days of life with an abnormal chest radiograph),
oxygen dependency at 36 weeks postmenstrual age, apnea, patent ductus arteriosus,
intraventricular hemorrhage all grades, intraventricular hemorrhage grades
III and IV, necrotizing enterocolitis, renal failure, sepsis, retinopathy
of prematurity, mortality at 28 days, overall mortality, plasma superoxide
dismutase levels, urine superoxide dismutase levels, neutrophil chemotactic
activity in tracheal aspirates, albumin levels in tracheal aspirates and
antibodies to recombinant human superoxide dismutase.
Randomization: Both studies allocated treatment by randomization. Rosenfeld 1984 used "random selection charts" for randomization. Davis 1997 used computer generated randomization and concealed allocation by using sealed envelopes.
Blinding of treatment: In both studies clinicians were masked to the nature of the intervention. Saline placebo was used in the control group in each study.
Blinding of outcome assessment: In Rosenfeld 1984 the radiologists reading chest radiographs at three and 12 months during outpatient follow up were masked to the nature of the therapy. In Davis 1997 chest radiographs and cranial ultrasounds were interpreted by a single pediatric radiologist who was also blinded to treatment assignment.
Exclusion after randomization: Rosenfeld 1984 excluded 14 patients who died in their reporting of outcomes other than mortality and reported clinical and radiologic outcomes only in survivors. There were no exclusions by Davis 1997.
Chronic lung disease (oxygen dependency at 36 weeks postconceptual age): Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Chronic lung disease was present in two of 22 infants who received superoxide dismutase and in one of 11 infants who received placebo (relative risk 1.00, 95% confidence limits 0.10, 9.86).
Bronchopulmonary dysplasia (oxygen dependency with an abnormal chest radiograph at 28 days): Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Three of the 22 infants who received superoxide dismutase and none of the 11 who received placebo developed this outcome (relative risk 3.65, 95% confidence limits 0.21, 65.05).
Death at 28 days: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. One of 22 infants who received superoxide dismutase and one of the 11 infants in the placebo group died by 28 days of life (relative risk 0.50, 95% confidence limits 0.03, 7.26).
Bronchopulmonary dysplasia (oxygen dependency with an abnormal chest radiograph) or death at 28 days: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. This outcome occurred in four of 22 infants who received superoxide dismutase and one of the 11 infants who received placebo (relative risk 2.00, 95% confidence limits 0.25, 15.82).
Death prior to discharge: Neither Davis 1997 nor Rosenfeld 1984 detected a difference in this outcome between infants who received superoxide dismutase and those who received placebo. In Davis 1997 death prior to hospital discharge occurred in three of 22 infants who received superoxide dismutase and in one of the 11 infants in the placebo group. In Rosenfeld 1984 death prior to discharge occurred in seven of the 21 infants who received superoxide dismutase and in seven of the 24 infants the placebo group. The typical estimate from the meta-analysis of these two trials suggests no difference in the risk of death prior to discharge (relative risk 1.20, 95% confidence limits 0.53, 2.71).
Chest radiograph abnormalities in survivors: Rosenfeld 1984 reported that chest radiograph abnormalities in survivors were less frequent in surviving infants who received superoxide dismutase than in those survivors who received placebo. This outcome occurred in three of 14 surviving infants who received superoxide dismutase and in 12 of 17 surviving infants who received placebo (relative risk 0.30, 95% confidence limits 0.11, 0.87).
Respiratory problems after discharge: Rosenfeld 1984 reported respiratory problems after discharge from the neonatal intensive care unit to occur less frequently in survivors who had received superoxide dismutase than in survivors who had received placebo. This outcome occurred in three of 14 surviving infants who had received superoxide dismutase and in 11 of 17 survivors in the placebo group (relative risk 0.33, 95% confidence limits 0.11, 0.96).
Respiratory problems after discharge or death: We derived this outcome from the data of Rosenfeld 1984 by combining the numbers of the infants who died with those survivors who subsequently developed clinical findings of bronchopulmonary dysplasia after discharge from the neonatal intensive care unit. There was no difference in this outcome between infants who received superoxide dismutase and those who received placebo. This outcome occurred in 10 of the 21 infants who received superoxide dismutase and in 18 of 24 infants who received placebo (relative risk 0.63, 95% confidence limits 0.38, 1.05).
Apnea: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Apnea occurred in 10 of 22 infants who received superoxide dismutase and in six of the 11 infants in the placebo group (relative risk 0.83, 95% confidence limits 0.41, 1.69).
Patent ductus arteriosus: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Patent ductus arteriosus occurred in 10 of 22 infants who received superoxide dismutase and in five of the 11 infants in the placebo group (relative risk 1.00, 95% confidence limits 0.45, 2.21). Rosenfeld 1984 reported that the incidence of patent ductus arteriosus was similar in the two groups, though no numbers were reported.
Intraventricular hemorrhage (all grades): Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Intraventricular hemorrhage (all grades) occurred in two of 22 infants who received superoxide dismutase and in two of the 11 infants in the placebo group (relative risk 0.50, 95% confidence limits 0.08, 3.09). Rosenfeld 1984 reported a similar incidence of intraventricular hemorrhage in the two groups, though no numbers were reported.
Intraventricular hemorrhage (grades III/IV): Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Intraventricular hemorrhage (grades III/IV ) occurred in one of 22 infants who received superoxide dismutase and in none of the 11 infants in the placebo group (relative risk 1.57, 95% confidence limits 0.07, 35.58 ).
Necrotizing enterocolitis: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. Necrotizing enterocolitis occurred in two of 22 infants who received superoxide dismutase and in one of the 11 infants in the placebo group (relative risk 1.00, 95% confidence limits 0.10, 9.86).
Sepsis: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. It occurred in seven of 22 infants who received superoxide dismutase and one of the 11 infants in the placebo group (relative risk 3.50, 95% confidence limits 0.49, 25.01).
Retinopathy of prematurity: Davis 1997 did not detect a difference in this outcome between infants who received superoxide dismutase and those who received placebo. It occurred in nine of 22 infants who received superoxide dismutase and in three of the 11 infants in the placebo group (relative risk 1.50, 95% confidence limits 0.51, 4.45).
Respiratory support: Rosenfeld 1984 did not detect a difference between the treatment and placebo groups in total days of oxygen therapy, days of mechanical ventilation at various rates, mean peak FiO2 and distribution of days at various oxygen concentrations, and mean peak inspiratory pressures during the first week. However, the total duration of continuous positive airway pressure was significantly less in the superoxide dismutase group than the placebo group (4.9 vs 9.7 days, p<0.03) because of fewer days of continuous positive airway pressure > 5 cm H2O (2.3 vs 6.1 days, p < 0.02). No standard deviations were provided in the publication, therefore the data have not been entered into the data tables.
Neurodevelopmental outcome: Neither Davis 1997 nor Rosenfeld 1984 reported neurodevelopmental outcome.
Based on the small number of patients exposed to recombinant human superoxide dismutase in the trial by Davis 1997 and in one other non-randomized trial by Rosenfeld 1996, it appears to be well tolerated in the short term, both by the subcutaneous route and by the intratracheal route.
One recent study has not been included in this review as it is currently reported only in the form of abstracts (Davis 1999). This was a multicenter randomized controlled trial comparing recombinant human superoxide dismutase to placebo in preterm infants 600 - 1200 grams. When this study is published, it will be assessed for inclusion in this systematic review in a future update.
In summary, based on currently available published trials, there is insufficient evidence to draw firm conclusions about the efficacy of superoxide dismutase in preventing chronic lung disease of prematurity. Data from a small number of treated infants suggest that it is well tolerated and has no serious adverse effects.
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Davis 1997 | Blinding of randomization: Yes
Blinding of intervention: Yes Complete follow-up: Yes Blinding of outcome measurement: Yes only for assessment of chest radiographs and cranial ultrasounds. |
Preterm infants with birth weight 700 to 1300 grams, less than 24 hours old, intubated and on mechanical ventilation for treatment of respiratory distress syndrome. Had received surfactant within first 24 hours of life. | Experimental group 1:
Superoxide dismutase 2.5 mg/kg (n=11) Experimental group 2: Superoxide dismutase 5 mg/kg (n=11) In both these groups the drug was given intratracheally, repeated every 48 hours until extubated or until 7 doses completed, whichever was earlier. Control group: Equal volume of saline intratracheally, repeated every 48 hours until extubated or until 7 doses completed, whichever was earlier (n=11). |
Oxygen dependency with abnormal chest radiograph at 28 days, oxygen dependency at 36 weeks postconceptual age, intraventricular hemorrhage all grades, intraventricular hemorrhage, grades 3 and 4, patent ductus arteriosus, retinopathy of prematurity, necrotizing enterocolitis, sepsis, mortality at 28 days, overall mortality. | Small sample size. In this review we combined the data from patients from both dosage regimens as no dose-response relationship was evident in the results. | A |
| Rosenfeld 1984 | Blinding of randomization: Can't tell
Blinding of intervention: Yes Complete follow-up: No Blinding of outcome measurement: Yes for radiologic evaluation. No for other outcomes. |
Preterm infants admitted to the neonatal intensive care unit with respiratory distress syndrome, ventilator dependent with fraction of inspired oxygen > 0.70 at 24 hours of age to maintain PaO2 > or = 50 torr. | Experimental: Bovine superoxide dismutase 0.25 mg/kg subcutaneously
every 12 hours until no longer needing ventilator or continuous positive
airway pressure and maintained on room air (n=21).
Control (Placebo): Equal volume of saline subcutaneously (n=24). |
Radiologic changes of bronchopulmonary dysplasia. | Study done in pre-surfactant era. Definition of bronchopulmonay dysplasia different from current definition. Small sample size. | B |
| Study | Reason for exclusion |
| Davis 2000 | Neurodevelopmental abnormalities were compared between patients who received superoxide dismutase and those who received placebo. This was done by combining patients who received superoxide dismutase in two different studies and by combining patients in the placebo groups from those studies. One of the two studies had random allocation of patients to treatment / placebo and the other had non-random allocation of treatment and placebo (groups were studied sequentially). |
| Rosenfeld 1996 | Non-random allocation of treatment and placebo (groups studied sequentially). |
Davis JM, Rosenfeld WN, Richter SE, Parad R, Gewolb IH, Spitzer AR, Carlo WA, Couser RJ, Price A, Flaster E, Kassem N, Edwards L, Tierney J and Horowitz S. Safety and pharmacokinetics of multiple doses of recombinant human CuZn superoxide dismutase administered intratracheally to premature neonates with respiratory distress syndrome. Pediatrics 1997;100:24-30.
Rosenfeld 1984 {published data only}
Rosenfeld W, Evans H, Concepcion L, Jhaveri R, Schaeffer H, Friedman A. Prevention of bronchopulmonary dysplasia by administration of bovine superoxide dismutase in preterm infants with respiratory distress syndrome. J Pediatr 1984;105:781-785.
Davis JM, Richter SE, Biswas S, Rosenfeld WN, Parton L, Gewolb IH Parad R, Carlo W, Couser RJ, Baumgart S, Atluru V, Salerno L, Kassem N. Long-term follow-up of premature infants treated with prophylactic, intratracheal recombinant human CuZn superoxide dismutase. J Perinatol 2000;4:213-216.
Rosenfeld 1996 {published data only}
Rosenfeld WN, Davis JM, Parton L, Richter SE, Price A, Flaster E, Kassem N. Safety and pharmacokinetics of recombinant human superoxide dismutase administered intratracheally to premature neonates with respiratory distress syndrome. Pediatrics 1996;97:811-817.
Davis JM, Rosenfeld WN, Parad R, Richter S, Gewolb I, Couser R, Parton L, Carlo W, Hudak M, Mammel M, Davidson D, Gertsmann D, Ramanathan R, Kinsella J, Baumgart S, Donn S, Raju T, Salerno LM, Huang W, Barton N. Improved pulmonary outcome at one year corrected age in premature neonataes treated with recombinant human superoxide dismutase. Pediatr Res 2000;47:395A (Abstract no. 2333).
* Davis JM, Rosenfeld WN, Richter SE, Parad R, Gewolb IH, Couser R, Price A, Kinsella JP, Donn SM, Gertsmann D, Ramanathan R, Raju T,Hudak M, Baumgart S, Carlo W, Davidson D, Mammel MC, Parton L, Salerno L, Huang W, Barton N. The effects of multiple doses of recombinant human CuZn superoxide dismutase (rhSOD) in premature infants with respiratory distress syndrome (RDS). Pediatr Res 1999;45:193A (Abstract no.1129).
* indicates the primary reference for the study
Bancalari E, Abdenour GE, Feller R et al. Bronchopulmonary dysplasia: clinical presentation. J Pediatr 1979;95:819-823.
Davis JM. Superoxide dismutase: a role in the prevention of chronic lung disease. Biol Neonate 1998;74 (S1):29 - 34.
Fardy C, Silverman M. Antioxidants in neonatal lung disease. Arch Dis Child 1995;73:F112-117.
Farrell PA, Fiascone JM. Bronchopulmonary dysplasia in the 1990s: a review for the pediatrician. Curr Probl Pediatr 1997;27:129-163.
Kelly FJ. Free radical disorders of preterm infants. Br Med Bull 1993;49:668-678.
Knight JA. Free radicals: their history and current status in aging and disease. Ann Clin Lab Sci 1998;28:331-346.
O'Brodovich HM, Mellins RB. Bronchopulmonary dysplasia. Unresolved neonatal acute lung injury. Am Rev Respir Dis 1985;132:694-709.
Saugstad O. Oxygen toxicity in the neonatal period. Acta Paediatr Scand 1990;79:881-892.
Saugstad OD. Chronic lung disease: the role of oxidative stress. Biol Neonate 1998;74(S1):21-28.
Shennan AT, Dunn MS, Ohlsson A et al. Abnormal pulmonary outcomes in premature infants: prediction from oxygen requirements in the neonatal period. Pediatrics 1988;82:527-532.
Volpe JJ. Brain injury in the premature infant. Neuropathology, clinical aspects, pathogenesis and prevention. Clin Perinatol 1997;24:567-587.
01.02 Bronchopulmonary dysplasia (28 days)
01.04 Bronchopulmonary dysplasia or death at 28 days
01.05 Death prior to discharge
01.06 Chest radiograph abnormalities in survivors
01.07 Respiratory problems after discharge in survivors
01.08 Respiratory problems after discharge or death before discharge
01.10 Patent ductus arteriosus
01.11 Intraventricular hemorrhage
01.12 Intraventricular hemorrhage, Grades III / IV
01.13 Necrotizing enterocolitis
01.15 Retinopathy of prematurity