Transferring a subject to 100 % oxygen following a period of hypoxia, paradoxically may worsen the symptoms of hypoxia (Latham, 1951). The author suggested that this ‘oxygen paradox’ was due to a fall in arterial pressure occurring during oxygen breathing. In this study Latham’s findings have been extended by measuring cerebral blood flow velocity, to investigate changes in cerebral blood flow in this situation and their relationship to changes in arterial pressure and arterial PCO2.
The subjects were seven males aged 24 ± 6 years with no history of cardiovascular or respiratory disorders and taking no medication. The study was approved by the local ethics committee. Each subject performed two protocols. For 10 min they breathed room air whilst lying supine. Hypoxia was then induced by breathing 8 % oxygen for 5 min. In one protocol they were suddenly switched to breathing 100 % oxygen for 5 min and finally to room air. In the other protocol, hypoxia was followed by breathing air for 5 min and then finally 100 % oxygen for 5 min. Throughout testing continuous measurements were made of arterial blood pressure, heart rate, pulmonary ventilation, end-tidal PCO2 and PO2, mean blood flow velocity in the right middle cerebral artery (CBFV) using transcranial Doppler ultrasound (EME TC2-64), forearm blood flow and, as a safety measure, arterial oxygen saturation.
During hypoxia ventilation increased causing a 19.4 ± 1.4 % (mean ± S.E.M.) fall in end-tidal PCO2 below the control level of 43 ± 2.7 mmHg. Despite this fall in PCO2, low PO2 increased CBFV by 8 ± 1.5 % (2-way ANOVA; P < 0.05), but when the subjects were switched to 100 % oxygen, CBFV fell to 11 ± 1.5 % below the control level. During this period of oxygen breathing end-tidal PCO2 remained low at 13 ± 1.4 mmHg below the control level. However, in subjects who breathed room air following hypoxia, PCO2 recovered to control levels and there was no fall in CBFV. Subsequent administration of 100 % oxygen did not depress CBFV. The reduction of CBFV was significantly correlated (r = 0.81) with the fall in end-tidal PCO2. There were no changes in arterial pressure, heart rate or forearm blood flow suggestive of cardiovascular syncope.
Since no subject in this study showed the full signs and symptoms of oxygen paradox, the possibility of cardiovascular syncope as an important contributor cannot be ruled out. The results of this study do suggest that persistence of low arterial PCO2 during oxygen breathing, which induces cerebral vasoconstriction, may be important in causing the oxygen paradox.