Therapeutic hypercapnia (TH), an intentional inhalation of CO(2), has been shown to improve pulmonary function in certain models of lung injury. We tested the null hypothesis that TH does not improve hyperoxic lung injury in neonatal rats. The prospective, randomized study was set at Research laboratory in Children's Hospital. Forty-five newborn rats were randomly assigned to three groups (n = 15/group), and exposed to 96 h of normoxia (FiO(2) = 0.21), hyperoxia (FiO(2) > 0.98), and TH (FiO(2) = 0.95, FiCO(2) = 0.05). Lung histology, wet-weight to dry-weight ratio, and concentrations of pro- and anti-inflammatory cytokines (IL-1beta, IL-6, TNF-alpha, and IL-10) were used to evaluate pulmonary damage. Using a scale of 0-4, the total scores for lungs hypercellularity, inflammation, and hemorrhage was significantly increased from a median value of 1.5 in normoxia to 2.5 in hyperoxia (P < 0.05) and 3.0 with TH (P < 0.001, nonparametric ANOVA). The interstitial space relative to the alveolar space, as a measure of hypercellularity, was increased by 18% during hyperoxia and by 44% with TH compared with normoxia. TH significantly increased the size of the interstitial space by 22% compared with hyperoxia (P < 0.001). The lung wet-weight to dry-weight ratio was increased by 10% in both hyperoxic groups (P < 0.001). Both hyperoxic groups showed significant reductions in the concentration of IL-1beta compared with normoxia (P < 0.001), whereas the ratio of IL-1beta to IL-10 was significantly decreased, indicating an anti-inflammatory trend. TH does not prevent histological manifestations of hyperoxic lung injury in spontaneously breathing neonatal rats and may worsen the outcome.