Main

Oxygen toxicity is thought to play an important role in the pathogenesis of a number of diseases in the newborn infant(1), including BPD(2). Oxygen free radical-induced lipid peroxidation can lead to loss of membrane fluidity and integrity, and protein oxidation can cause structural and functional alterations(3). These events are thought to contribute to cell dysfunction and death. In one study of preterm infants with respiratory distress syndrome, high concentrations of ethane and pentane, both volatile products of lipid peroxidation, were correlated with poor respiratory outcome and death(4). In another study of infants with hyaline membrane disease, high tracheal aspirate concentrations of protein-carbonyls, a protein oxidation product, were associated with the development of BPD(5).

In preterm infants, naturally occurring antioxidant defense systems, may be insufficiently developed. These systems include antioxidant enzymes(6), such as superoxide dismutases, glutathione peroxidases, and catalase, as well as small molecular weight antioxidants, such as glutathione and the antioxidant vitamins E (α-TP) and C (AA).

In preterm infants, cord plasma concentrations of lipid soluble vitamins, including α-TP, are low(7). In contrast, these infants are born with high plasma concentrations of AA; however, these levels fall precipitously after birth(8, 9). In addition, AA concentrations in tracheal aspirates of preterm infants who require mechanical ventilatory support for hyaline membrane disease are even lower than corresponding plasma levels(8).

Therefore, in this pilot study, we used a baboon model of BPD closely resembling the disease in human premature infants(10) to test the hypothesis that early postnatal administration of a high dose combination of AA and α-TP modifies the pulmonary response to prolonged hyperoxia. Specifically, we hypothesized that this intervention will reduce concentrations of biochemical markers of lipid peroxidation in tracheal aspirates, improve lung function and result in less dramatic histopathologic lung changes.

METHODS

Animals. All animal care procedures were performed in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals. The protocol was reviewed and approved by the animal care committees of the Southwest Foundation for Biomedical Research in San Antonio and the University of Texas Health Science Center at San Antonio.

Five premature baboons were delivered by hysterotomy at 140 ± 2 d of gestation (normal gestation 180 d). Gestational ages were determined by matings that were timed by observation of female perineal sex skin changes and confirmed by ultrasound examination at intervals during pregnancy.

Vitamin supplementation. Four animals (hyperoxic antioxidant group) were treated with 100% oxygen and were assigned to early high dose AA (100 mg·kg·-1·d-1) and high dose α-TP (20 mg·kg·-1·d-1) supplementation (in addition to standard multivitamin supplementation) for the 10-d study period. The vitamins were added to a dextrose solution during the first 24 h of life and to TPN solutions from d 2-10 and infused continuously over 24 h each day. One animal received only standard multivitamin supplementation with its TPN (no AA or α-TP during the first 24 h of life, then 16 mg of AA/150 mL of TPN and 1.4 mg of α-TP/150 mL of TPN) to provide baseline data for plasma and tracheal aspirate concentrations of AA, α-TP, and CD. An additional 10 historic animals treated comparably with 100% oxygen were combined with this animal (hyperoxic control group,n = 11). Historical data from 11 O2-treated as required animals were used as a normoxic control group (Fig. 1).

Figure 1
figure 1

Assignment of antioxidant vitamin supplementation and oxygen exposure in pilot study animals and historic control animals. Biochemical measurements of plasma and tracheal aspirate vitamin concentrations and markers of lipid peroxidation were only obtained in the pilot study animals (hyperoxic antioxidant group, n = 4; hyperoxic control group, n = 1). Ventilatory parameters and histopathologic data were available for all animals.

Full size image

Ventilatory support and routine care. All animals were resuscitated with endotracheal intubation, anesthetized with ketamine hydrochloride, and instrumented with umbilical arterial and venous lines. Catheter and endotracheal tube positions were confirmed by radiography.

Animals were ventilated with a time-cycled pressure-limited infant ventilator (either Bear Cub, Bournes, Riverside, CA, or Infant Star, Infrasonics, Inc., San Diego, CA) from the time of intubation. In all groups, arterial Pco2 was maintained between 40 and 55 mm Hg by varying ventilator rate (intial rate 30 breaths/min, maximum 60 breaths/min) and tidal volume with attempts made to avoid both over- and underinflation of the lungs. On the basis of previous experience where higher ventilatory pressures were associated with an extremely high incidence of air leak, peak inspiratory pressure was initially set to maintain a Paw of 12-13 cm H2O.

In the hyperoxic animals (hyperoxic antioxidant group and hyperoxic control group), inspiratory oxygen concentration (Fio2) was maintained at 1.0 throughout the entire 10-d study period with adjustments of Paw made to maintain arterial Po2 between 200 and 400 mm Hg. This was achieved primarily through manipulations of Paw and attempts to optimize lung inflation. When ventilation was adequate, changes in Paw were accomplished by adjustment of positive end-expiratory pressure; if both oxygenation and ventilation were inadequate, PIP was increased (to a limit of 45 cm H2O) to achieve oxygenation goals. In the normoxic control animals, the goals for lung inflation was the same, but Fio2 was adjusted to keep arterial Po2 between 60 and 80 mm Hg.

If animals could not be adequately oxygenated and/or ventilated with rates up to 60 breaths/min and PIP of 45 cm H2O, high frequency oscillatory ventilation would be used to reach these goals, changing back to conventional ventilation usually at 48 h of life.

Weaning of ventilator rate and tidal volume was accomplished as tolerated. All animals were kept on positive end-expiratory pressure at a rate of≥8-10 breaths/min for the duration of the 10-d experimental period.

All premature baboons received sedation with ketamine and diazepam as needed. Ketamine and local infiltration with 1% xylocaine were used for invasive procedures. All animals received ampicillin (100 mg·kg·-1·d-1 i.v.) and gentamicin (5 mg·kg·-1·d-1 i.v.). Routine nursing care, administration of i.v. fluids, and blood replacement therapy was done as previously described(11).

Biochemical methods. Plasma concentrations of AA andα-TP were measured in cord blood samples and daily in infant blood samples. Tracheal aspirate concentrations of AA and α-TP, as well as CD(a marker of lipid peroxidation) were also measured daily. Urea levels in paired plasma and tracheal aspirate samples were compared to calculate dilution factors for the tracheal aspirate samples(12).

Measurements of AA. One hundred microliters of heparinized plasma were mixed with 100 μL of ice-cold 40% (wt/vol) metaphosphoric acid in distilled water containing 50 mM DTT, vortexed for 30 s, and centrifuged at 3000 rpm for 10 min. The clear supernates were stored at-74 °C until assay. At the time of AA analysis, extracts were thawed at room temperature, vortexed for 30 s, and centrifuged at 4 °C and 3000 rpm for 5 min. AA analysis was performed by paired-ion reversed-phase HPLC coupled with electrochemical detection as described by us(13).

Tracheal aspirate samples were collected by a standardized procedure. After instillation of 0.5 mL of normal saline, a suction catheter was inserted slightly distal to the tip of the endotracheal tube, and 80 cm H2O of negative pressure was applied. The secretions were collected into a sterile glass trap, and the catheter was rinsed with sterile normal saline to obtain a total sample volume of 0.5 mL. The samples were centrifuged at 3000 rpm for 10 min. One hundred microliters of supernate were mixed with 100 μL of ice-cold 40% (wt/vol) metaphosphoric acid in distilled water containing 50 mM DTT, vortexed for 30 s, and centrifuged at 3000 rpm for 10 min. The clear supernates were stored at -74 °C until assay. At the time of AA analysis, tracheal aspirate samples were thawed and processed as described for the plasma samples.

Measurements of α-TP. Blood and tracheal aspirate sampling was as outlined above for AA measurements. Immediately after collection, the samples were protected from light exposure by wrapping with aluminum foil. The samples were stored at -74 °C until assay. At the time of α-TP analysis, extracts were thawed at room temperature, vortexed for 30 s, and centrifuged at 4 °C and 3000 rpm for 15 min. α-TP was measured by the HPLC method with detection at 288 nm after extraction with hexane. The detection limit of this assay is 0.7 μmol/L(14).

Measurement of CD. Tracheal aspirates were obtained as outlined above for AA measurements, and stored at -74 °C until analysis. Within 21 d of storage, the samples were thawed in room air, vortexed for 30 s, and centrifuged at 4 °C and 3000 rpm for 15 min. The lipid portion of the tracheal aspirates was extracted according to a modified Bligh and Dyer(15) method. The lipid extracts were dried under N2 and dissolved in spectrophotometric grade cyclohexane, and the absorbance was measured at 234 nm. The concentration of the CD was calculated using an extinction coefficient of 28 000 M-1 cm-1(16).

Physiologic measures of lung function. Ventilator settings (PIP, Paw, ventilator rate) and arterial blood gases were recorded every 12 h for the 10-d study period. OI were calculated for each animal at 0.5, 1, 2, 3, 4, 5, and 10 d of life (OI = (100 × Paw× Fio2)/arterial Po2).

At the time of autopsy, PV curves were done. The trachea was cannulated and connected to a PV apparatus as previously described(17). After completion of two standard inflations and deflations from 0 to 35 cm H2O, the lung was inflated in stepwise 5-cm H2O increments with a 2-min pause at each pressure to allow for volume equilibration. The system volume was recorded and the process repeated. After stepwise inflation to 35 cm H2O, the lung was deflated in 5-cm H2O decrements. Measured system volumes were adjusted for gas compression values of the apparatus to find the actual lung volumes. Adjusted lung volumes were plotted against pressure to develop the PV curve.

Pathologic methods. At necropsy, the right lower lobe was removed, weighed, and intrabronchially fixed with phosphate-buffered 4% paraformaldehyde and 0.1% glutaraldehyde at 20 cm H2O constant airway pressure for 48 h. Historical controls were inflated comparably and fixed with buffered paraformal-dehyde-glutaraldehyde fixative or Carnoy's. After fixation, the tissue was sectioned into three transverse levels, embedded in Paraplast, sectioned at 4 μm, and stained with hematoxylin and eosin. The slides were reviewed for evidence of lung injury without knowledge of treatment group, and then compared with the historical controls.

Statistical methods. Linear regression analysis was used to evaluate correlations between plasma and tracheal aspirate AA concentrations, tracheal aspirate AA and CD concentrations, as well as plasmaα-TP and tracheal aspirate CD concentrations. The two-sample t test was used to compare mean plasma and tracheal aspirate concentrations of AA, mean plasma concentrations of α-TP, and mean tracheal aspirate concentrations of CD between animals receiving high dose and standard antioxidant vitamin supplementation. For physiologic data (PIP, Paw, OI), means were compared using one-way ANOVA. If a significant difference was identified, the post hoc Student-Newman-Keuls test was applied for determination of which groups were significantly different. If a wide variance was noted, a Kruskal-Wallis one-way ANOVA test was also applied. Comparisons of the same variables at different time points were analyzed by ANOVA for repeated measures. Statistical significance was accepted if the null hypothesis was rejected at the p < 0.05 level.

RESULTS

The four animals that received early high dose AA and α-TP supplements had an actual mean daily AA and α-TP intake of 106 and 23 mg·kg·-1·d-1, respectively. The 100% O2-treated animal that received only standard multivitamin supplementation had an actual daily intake of AA and α-TP of 11 and 1 mg·kg·-1·d-1, respectively. All historical controls were also managed with a similar total parenteral nutrition regimen including early introduction of amino acids, as well as standard multivitamin and trace mineral supplementation.

One animal in the treatment group died of severe respiratory failure on day of life 7. Data obtained during this time period have been included in the results. The remaining four animals survived the 10-d study period.

High frequency oscillatory ventilation was never used in the hyperoxic antioxidant group, but six and five animals in the hyperoxic control group and the normoxic control group, respectively, received this form of ventilatory support.

Biochemical results. Early high dose antioxidant vitamin administration effectively raised plasma and tracheal aspirate concentrations of AA. Mean plasma AA concentration in all samples obtained from treated animals was 395 ± 216 μmol/L compared with 37 ± 22 μmol/L in the control animal (p < 0.001), and mean tracheal aspirate AA concentration was 286 ± 205 μmol/L in the treated animals compared with 62 ± 35 μmol/L in the control animal(p < 0.002). There was a strong positive correlation (r = 0.97) between the AA concentrations in plasma and tracheal aspirates (Fig. 2).

Figure 2
figure 2

Correlation between plasma AA ([AA]p) and tracheal aspirate AA ([AA]ta) concentrations in four animals treated with high dose antioxidant vitamin supplementation (hyperoxic antioxidant group) and one animal receiving standard antioxidant vitamin supplementation(hyperoxic control animal) (r = 0.97). All five animals were exposed to an Fio2 of 1.0 over the entire 10-d study period.

Full size image

Mean plasma concentrations of α-TP were also increased in the treated animals (24.6 ± 17.5 μmol/L compared with 10.1 ± 2.5μmol/L in the control animal, p < 0.001), but to a lesser extent than AA concentrations. In all animals, tracheal aspirate concentrations of α-TP were always below the detection limit of the assay (0.7 μmol/L).

The increase in plasma and tracheal aspirate AA concentrations in the treated animals did not have any apparent effect on tracheal aspirate concentrations of CD (mean concentrations 1050 ± 1111 μmol/L versus 482 ± 333 μmol/L, in treated animals and control, respectively, p = 0.09). No statistically significant correlation was observed between tracheal aspirate CD concentrations and plasma AA concentrations (r = 0.56, Fig. 3A), tracheal aspirate AA concentrations (r = 0.65, Fig. 3B) or plasma α-TP concentrations (r = 0.19, Fig. 3C).

Figure 3
figure 3

(A) Relationship between plasma AA ([AA]p and tracheal aspirate CD ([CD]ta) concentrations (r = 0.56),(B) between tracheal aspirate AA ([AA]ta) and tracheal aspirate CD ([CD]ta) concentrations (r = 0.65), and(C) between plasma α-TP ([α-TP]p) and tracheal aspirate CD ([CD]ta) concentrations (r = 0.19) in four animals treated with high dose antioxidant vitamin supplementation (hyperoxic antioxidant group) and one animal receiving standard antioxidant vitamin supplementation (hyperoxic control animal). All five animals were exposed to an Fio2 of 1.0 over the entire 10-d study period.

Full size image

Respiratory parameters. The respiratory data are presented in Fig. 4,A-C. The 72-h study time proved to be the time period when many of the measurements in the hyperoxic antioxidant group diverged from the normoxic and/or hyperoxic control groups. Importantly, the hyperoxic antioxidant group required higher PIP and Paw and had worse OI. Starting at 72 h, PIP values were significantly higher in the hyperoxic antioxidant group when compared with those of the normoxic animals (Fig. 4A). Paw in the hyperoxic antioxidant animals were significantly higher than those of the normoxic group at 72 h(p < 0.05), and remained elevated for the duration of the study when compared with both normoxic and hyperoxic control groups (Fig. 4B). Likewise, the OI values of the hyperoxic antioxidant group were worse at 72 h than those in the two other study groups and significantly so at 72-h, 96-h (compared with normoxic controls), and 240-h (compared with both normoxic and hyperoxic controls) study times (Fig. 4C). There were significant reductions over time for PIP, Paw, and OI in the hyperoxic and normoxic groups, whereas this was not demonstrated in the hyperoxic antioxidant group (ANOVA by repeated measures). Consistent with these observations, PV curves from three animals of the hyperoxic antioxidant group showed no improved compliance when compared with PV curves from three hyperoxic control animals (Fig. 5).

Figure 4
figure 4

Time course of (A) PIP (*significant vs normoxic control group by one-way ANOVA with post hoc Student-Newman-Keuls test), and (C) OI (#sigificant vs others by Kruskal-Willis; *significant vs each other by one-way ANOVA with post hoc Student-Newman-Keuls test) in hyperoxic antioxidant group (dark gray bars, n = 4), hyperoxic control group(light gray bars, n = 11), and normoxic control group (white bars, n = 11). The graphs represent means ± SEM.

Full size image
Figure 5
figure 5

PV curves in three animals of the hyperoxic antioxidant group (dark gray symbols) and three animals of the hyperoxic control group (light gray symbols).

Full size image

Histopathology. Finally, early supplementation of high dose antioxidant vitamins had no effect on lung disease at the histopathologic level. Lungs from normoxic animals showed normal and well expanded saccules (primitive alveoli) after the 10-d study period (Fig. 6A). The hyperoxic and hyperoxic antioxidant groups had comparable pathologic findings. An inflation pattern of alternating overexpansion and atelectasis was evident. Within the saccular/alveolar spaces, focal hemorrhages, edema and increased numbers of alveolar macrophages and scattered polymorphonuclear cells were present. Peribronchiolar fibrosis extended into the surrounding saccular walls (Figs. 6,B and C).

Figure 6
figure 6

(A) Normoxic control animal. The lung shows the characteristic saccules with scattered alveoli (arrows) branching from their walls. Only a few alveolar macrophages are present. br, bronchiole. Hematoxylin and eosin, ×125. (B) Hyperoxic control animal. The saccular/alveolar walls are thickened and show increased cellularity when compared with the normoxic control. Alveolar macrophages and some protein debris and scattered red blood cells are present in the alveolar/saccular spaces. br, bronchiole. Hematoxylin and eosin, ×125. (C) Hyperoxic antioxidant animal. In this micrograph, there are more red blood cells and protein in the alveolar/succular spaces than depicted in the hyperoxic control animal, but this finding was present in both hyperoxic groups. Likewise, the alveolar/saccular walls show increased cellularity and early fibroproliferative changes comparable to the hyperoxic control group. br, bronchiole. Hematoxylin and eosin, ×125.

Full size image

DISCUSSION

Lung immaturity, oxygen toxicity, and ventilator-induced trauma are thought to be major factors in the pathogenesis of BPD in preterm infants with respiratory insufficiency. Coalson et al.(10) have shown that prolonged exposure of preterm baboons to positive pressure ventilation and an Fio2 of 100% reporducibly leads to histopathologic changes that closely mimic BPD seen in human preterm infants. In this model, baboons are delivered by cesarean section at 140 d of gestation (approximately 75% of normal gestation). These premature baboons require mechanical ventilatory support for respiratory failure due to hyaline membrane disease. The histopathologic changes seen after 10 d of hyperoxia are thought to be mainly due to oxygen free radical-mediated injury because lung changes are minimal in normoxic control animals. The model thus seems suitable to explore effects of antioxidant strategies aimed at the prevention of BPD.

AA is the most powerful water-soluble inhibitor of lipid peroxidation in human plasma(13). AA effectively scavenges a variety of reactive oxygen species, including superoxide (O2-) and hydroxyl(HO·) radicals(18), and suppresses the inactivation of antiproteases(19) and surfactant(20) by oxidants released from activated neutrophils. A pulmonary antioxidant role of AA is further supported by decreased systemic AA levels in cigarette smokers which appears to be mainly caused by AA oxidation by oxidants contained in the gas phase of cigarette smoke(21). Similarly, plasma AA concentrations in patients with adult respiratory distress syndrome are extremely low(22). Most interestingly, AA supplementation has been shown to attenuate ozone-induced(23) and NO2-induced(24) airway hyperresponsiveness in humans. Mohsenin(25) has shown that supplementation with a combination of AA and α-TP leads to inhibition of lipid peroxidation in lungs of healthy volunteers exposed to oxidant stress in the form of inhaled NO2.

α-TP plays an important role in protecting biologic membranes from peroxidative injury, mainly by its ability to reduce peroxyl radicals(26, 27), thus acting as a chain-breaking antioxidant. Interestingly, α-TP is secreted by lower respiratory tract epithelial cells together with surfactant(28). AA is thought to contribute to the regeneration of membrane-bound α-TP(29). This interaction may be of particular importance at the interface of the respiratory tract lining fluid and the lipid bilayer of the respiratory tract epithelial cells(30).

AA supplementation has been reported to effectively improve survival in a variety of laboratory animals exposed to hyperoxia. Tokunaga et al.(31) reported a 100% survival rate after 72 h of hyperoxia (Fio2 > 95%) in rats pretreated with AA compared with only 50% in control animals. This was accompanied by significantly lower concentrations of malondialdehyde (a marker of lipid peroxidation) in serum and lungs(31).

In the present study, early high dose supplementation with AA andα-TP in preterm baboons effectively increased plasma concentrations of AA and α-TP. These changes were accompanied by significant increases in tracheal aspirate concentrations of AA; respiratory tract lining fluid concentrations of α-TP are low(30), and the sensitivity of our α-TP assay (0.7 mmol/L) was too low to detect any potential changes in α-TP concentrations in the dilute tracheal aspirate or bronchoalveolar lavage fluid samples.

However, high dose antioxidant vitamin supplementation had no beneficial effect on the pulmonary response to hyperoxia. At the biochemical level, concentrations of lipid peroxidation products (CD) in tracheal aspirates were unchanged. Clinically, when compared with normoxic control animals, the antioxidant treated animals had worse respiratory parameters, with significant differences in mean airway pressures and oxygenation indices beyond 72 h of life. Finally, at the histopathologic level, lungs of treated animals were indistinguishable from the hyperoxic control animals. The consistency of the respiratory abnormalities during the clinical course of the antioxidant treated group precluded the use of additional animals to test this antioxidant regimen.

The lung is protected from free radical-mediated injury by a complex system of intra- and extracellular antioxidants, including superoxide dismutase, glutathione peroxidase, and catalase, low molecular weight antioxidants(including antioxidant vitamins), metal-binding proteins, as well as sacrificial proteins and unsaturated lipids(30). Modification of only two components of this complex system may be insufficient to modify the pulmonary response to the high degree of oxidant stress used in this baboon model of BPD. A successful antioxidant prevention strategy may have to be more comprehensive, and may have to include other known respiratory tract lining fluid antioxidants, such as glutathione and metal-binding proteins, as well as antioxienzymes to enhance intracellular antioxidant protection.

By measuring total AA concentrations only, we could not distinguish between reduced (antioxidatively active) and oxidized (antioxidatively inactive) AA. Interestingly, Buffinton and Doe(32) have described decreased activity of GSH/NADPH-dependent dehydroascorbic acid reductase in the mucosa of patients with inflammatory bowel disease. If the activity of this enzyme is also decreased in the inflamed lungs of premature baboons, this could perhaps in part explain why our intervention failed to protect the lungs from hyperoxic injury. Related to this, it has been reported that premature infants have low levels of antioxidant enzymes, including those necessary to maintain cellular glutathione levels and redox status. Therefore GSH levels in these infants may not be sufficient to keep AA in the reduced form(33). Because reduced AA may be important in regeneratingα-TP from its oxidized form, low levels of GSH could result in limited activity of both antioxidants. On the other hand, both AA(34) and α-TP(35) have been shown to affect the activity of the nuclear transcription factor, NFκB activation by high levels of AA and/or α-TP may prevent the up-regulation of antioxidant enzymes under oxidative stress conditions.

It is also possible that antioxidants alone are insufficient to alter the pulmonary inflammatory response seen in this model because injury mechanisms other than those mediated by oxygen free radicals are involved in the pathogenesis of BPD. For example, an imbalance between neutrophil-derived proteases (e.g. elastase) and antiproteases, such asα1-antitrypsin, has been shown to put preterm infants at risk to develop BPD(36, 37). Furthermore, our animals were not only exposed to hyperoxia, but also required mechanical ventilatory support. Undoubtedly, the use of positive pressure ventilation contributes to the pathogenesis of BPD(38). This may also in part explain the discrepancy between our findings and the reported benefit of antioxidant strategies in other animal models of hyperoxic lung injury(31).

Finally, because the animals in the hyperoxic antioxidant group apparently had a worse clinical course than the animals in the hyperoxic control group[more blood pressure instabilities (data not shown), higher Paw beyond 72 h of life (Fig. 4B), worse OI at 240 h of life (Fig. 4C).] the possibility of a potential negative effect of the high dose antioxidant regimen in this model of BPD has to be considered. It has been suggested that AA can have prooxidant effects in the presence of free transition metal ions (Fenton chemistry)(39), and potentially redox active nonprotein-bound iron has been detected in both preterm and term infants(40). Other investigators, however, although confirming increased AA oxidation in the presence of nonprotein-bound iron, found no evidence of an associated increased formation of hydroxyl radicals(41). Our own investigations(42) also suggest that even in the presence of nonprotein-bound iron AA acts as an inhibitor of lipid peroxidation. Prophylaxis of BPD with vitamin E alone has been unsuccessful(43). Increased incidences of both sepsis and late onset necrotizing enterocolitis have been reported with the use of pharmacologic doses of vitamin E in retinopathy of prematurity prevention trials(44, 45). In addition, an unusual syndrome consisting of hepatomegaly, jaundice, and ascites has been associated with the i.v. use of a vitamin E product, E-Ferol (vitamin E solubilized with polysorbate-80 and polysorbate-20)(46). A possible explanation for these side effects of α-TP was put forth by Hale et al.(47) who observed massive accumulation of vitamin E emulsion in phagocytic cells of the spleen, and to a lesser extent liver and lung after rapid i.v. injection. No toxicity was associated with the polysorbate emulsifier alone(47). Our animals treated with high dose antioxidant vitamins, however, showed no liver or spleen pathology at necropsy.

In summary, early high dose supplementation with the antioxidants AA andα-TP did not prevent BPD in ventilated preterm baboons exposed to hyperoxia over a period of 10 d.