Hyperoxia and Oxygen Toxicity
On this page
Direct answer
Oxygen poisons by partial pressure, not by dose alone: above roughly 2 atmospheres absolute of inspired oxygen the brain seizes — the Paul Bert effect of central nervous system toxicity, tremor progressing to generalised convulsions — while prolonged exposure to high fractions at ordinary pressure injures the lung, the Lorrain Smith effect of tracheobronchitis, absorption atelectasis and falling vital capacity. The mechanism is reactive oxygen species: hyperoxia floods the electron transport chain with superoxide and hydrogen peroxide faster than superoxide dismutase, catalase and glutathione defences can clear them, oxidising lipids, proteins and DNA. Diving and hyperbaric practice therefore caps working inspired PO2 at about 1.4-1.6 atmospheres. Two clinical corollaries matter more: absorption atelectasis, because nitrogen washout removes the gas splinting poorly ventilated alveoli which then collapse; and retinopathy of prematurity, where the incompletely vascularised preterm retina is injured by oxygen-driven vascular changes — the reason neonatal oxygen is targeted (typically 88-95% saturation in the very preterm) rather than maximised.
What you must remember
- Two toxicity syndromes: central nervous system (Paul Bert) — tremor, visual symptoms, nausea, convulsions above about 2 ATA, seen in divers and chamber patients; pulmonary (Lorrain Smith) — substernal burning, cough, reduced vital capacity and diffuse alveolar damage from prolonged high fractions even at 1 atmosphere.
- Unit dose concept: toxicity tracks both partial pressure and duration; diving practice limits inspired PO2 to roughly 1.4 ATA working and 1.6 ATA during decompression stops.
- Absorption atelectasis: breathing 100% oxygen washes out alveolar nitrogen; in units with poor ventilation the oxygen is absorbed faster than it enters, alveoli collapse, and shunt develops — the physiological reason routine high-flow oxygen is not benign.
- Hypercapnia in chronic lung disease: oxygen can worsen carbon dioxide retention by removing hypoxic pulmonary vasoconstriction and by the Haldane effect unloading CO2 from haemoglobin, beyond any blunting of hypoxic drive — target saturations near 88-92% in exacerbations of COPD.
- Retinopathy of prematurity: hyperoxia suppresses vascular endothelial growth factor, then relative hypoxia after weaning provokes proliferative neovascularisation; screening programmes target the extremely preterm infant and oxygen is titrated, not maximised.
- Cerebral and retinal vasoconstriction: hyperoxia constricts cerebral vessels (a roughly 10-30% flow reduction at high inspired fractions, commonly quoted) — used therapeutically in cluster headache and studied after head injury.
- Antioxidant framing: the same defences that contain everyday mitochondrial leakage (superoxide dismutases, catalase, glutathione) are simply overwhelmed — hyperoxia is oxidative stress with a named pressure threshold.
A neonatal unit walk-through
A 27-week infant receives routine oxygen in the first days; saturations drift into the high nineties on the pulse oximeter. The modern response is titration, not applause: units target roughly 88-95% in this gestation, titrate blender settings, and screen for retinopathy of prematurity — in India under the national programme for very-low-birth-weight and early-gestation infants, typically from 4-6 weeks of age. The physiology justifies both halves: too little oxygen risks death and cerebral palsy (the trials that lowered targets showed mortality rising when saturations were held too low), too much injures the immature retina through the vascular endothelial growth factor seesaw. Now the adult mirror: a COPD patient given 100% oxygen at home by a well-meaning relative becomes drowsy — absorption atelectasis plus the Haldane effect plus lost hypoxic drive raise carbon dioxide; the answer is controlled oxygen, venturi-delivered, aiming for 88-92% and reassessing blood gases. One molecule, two ages, the same lesson: oxygen is a drug with a therapeutic index.
Where students slip
Treating oxygen as uniformly therapeutic is the root error — the exam wants the dose-response of toxicity by organ and pressure, and the specific thresholds (about 2 ATA for the brain, prolonged near-100% for the lung). Second, students quote "oxygen can stop breathing in COPD patients" as pure hypoxic-drive suppression; the Haldane effect and worsening ventilation-perfusion matching contribute substantially, and saying so marks a strong answer. Third, absorption atelectasis is confused with oxygen's effect on the pulmonary vasculature: one collapses alveoli (mechanical, nitrogen washout), the other dilates them (vasodilatation in ventilated units, redistribution away from hypoxic units lost). Finally, retinopathy of prematurity is explained as "oxygen burns the retina"; it is a two-phase vascular disease — hyperoxic vaso-obliteration then hypoxic proliferation — which is why even fluctuating, not merely high, oxygen is dangerous.
Frequently asked questions
What is the Paul Bert effect?
Central nervous system oxygen toxicity — tremor, visual and gastric symptoms progressing to generalised seizures at inspired partial pressures above roughly 2 atmospheres absolute.
What is the Lorrain Smith effect?
Pulmonary oxygen toxicity from prolonged exposure to high inspired fractions — tracheobronchial inflammation, absorption atelectasis and progressive falls in vital capacity.
Why does 100% oxygen cause absorption atelectasis?
Nitrogen washout removes the non-absorbable gas splinting poorly ventilated alveoli; oxygen then diffuses into blood faster than it is delivered and the alveoli collapse, creating shunt.
How does oxygen worsen hypercapnia in COPD?
Loss of hypoxic pulmonary vasoconstriction worsens ventilation-perfusion matching, the Haldane effect unloads carbon dioxide from haemoglobin, and hypoxic drive is reduced — hence 88-92% targets.
Why is oxygen targeted rather than maximised in preterm infants?
Excess oxygen drives retinopathy of prematurity through vaso-obliteration followed by proliferative neovascularisation, while too little raises mortality — a narrow therapeutic window, typically 88-95% saturation.