Abstract

In experimental settings, carbon dioxide (CO₂) blood levels affect the perfusion of various tissues. Hypocapnia-induced vasoconstriction impairs blood supply and severe hypocapnia may cause DNA fragmentation in cerebral neurons. In contrast, hypercapnia, acting as a vasodilator, increases mesenteric, renal and cerebral blood flow. Hypercapnia antagonises hyperoxia-induced retinal vasoconstriction but also augments oxygen-induced retinal microvascular degeneration that precedes proliferative retinopathy. In animal models of endotoxin-induced acute lung injury, hypercapnia improves compliance and attenuates alveolar neutrophil infiltration. Peripheral microvessel diameter, blood-flow velocity, and global cardiac output appear to increase with graded arterial hypercapnia up to 80–100 mm Hg, whereas all of these variables start to decrease at higher CO₂ values. These findings are not apt to establish a safe range for target CO₂ levels in ventilated preterm infants. Randomised trials of permissive hypercapnia in preterm infants cover an overall range from 35 to 65 mm Hg. A meta-analysis of all three trials published to date demonstrates the lack of any significant benefit or harm with respect to long-term outcomes such as death, bronchopulmonary dysplasia, retinopathy of prematurity, intravascular haemorrhage (IVH), periventricular leucomalacia (PVL), or neurodevelopmental impairment. In contrast, circumstantial but solid evidence suggests that moderately low CO₂ levels (<30–35 mm Hg) in ventilated preterm infants are strongly associated with the development of PVL. Moreover, both CO₂ extremes and the magnitude of CO₂ fluctuations are associated with severe IVH. This suggests that avoiding rapid CO₂ changes by meticulous and gentle ventilator adjustments may be even more important than absolute CO₂ levels to prevent brain damage of preterm infants.