The voluntary hyperventilation (VHV) test is used in many clinical examinations. However, arterial hypoxemia following a clinical VHV test is not a well-studied phenomenon. We analyzed the arterial blood gases (ABGs) of 61 patients during a VHV test. The ABG were taken prior to (PaO2-Prior), immediately following (PaO2-Immediate), and 5 min after (PaO2-After) the VHV test. The patients' average PaO2 rose significantly (p < 0.0001) from the PaO2-Prior (88 +/- 8 mm Hg; mean +/- SD) to the PaO2-Immediate (118 +/- 13 mm Hg) and then dropped significantly (p < 0.0001) to the PaO2-After (74 +/- 16 mm Hg). Two of the 20 patients who experienced an angina pectoris attack (AP(+)) following the VHV test showed severe arterial hypoxemia (PaO2-After < 60 mm Hg), whereas 9 of the 41 patients who did not experience an angina pectoris attack (AP(-)) showed a PaO2-After < 60 mm Hg. The PaO2-After did not correlate with the PaO2-Prior. The decrease in the PaO2-Prior to After did not correlate significantly with the left ventricular ejection fraction rate (n = 58, r = 0.18, not significant). However, the decrease in the PaO2-Prior to After correlated well with the degree of recovery of the PaCO2 following the VHV test (r = -0.69, p < 0.0001). The age, gender ratio, changes in arterial blood gases, number of patients who experienced PaO2-After < 60 mm Hg, and left-ventricular ejection fraction rate were not significantly different between the AP(-) and AP(+) groups. Posthyperventilation hypoxemia developed frequently following the VHV test during coronary angiography. Although this arterial hypoxemia was not directly correlated with the occurrence of AP attacks following VHV in this study, continuous SaO2 monitoring is recommended whenever a VHV test is used as a diagnostic technique to avoid the potentially deleterious effects of arterial hypoxemia.