BACKGROUND: Hip replacement arthroplasty (HRA) is highly traumatic and performed in a lateral position for several hours and dead-space ventilation may increase. So, the difference between arterial and end-tidal PCO2 was investigated depending on the changes in the patient's posture during HRA in elderly patients. METHODS: Forty-three patients scheduled for a HRA were divided into two groups; Adult Group (n = 21, A-Group) and Elderly Group (n = 22, E-Group). Mean arterial pressure (MAP), heart rate (HR), PaO2, PaCO2, and end-tidal carbon dioxide tension (P(ET)CO2) were simultaneously measured at 10 min after anesthesia in a supine position (S[10]), at 30 min intervals from 30 min (L[30]) to 180 min (L[180]) in a lateral position and at 10 min in a supine position after the end of surgery (ES[10]) in both groups. The PaCO2-P(ET)CO2 gradient (P[a-ET]CO2) and dead space ventilation (Vd/ Vt) were calculated. RESULTS: At S(10), P(a-ET)CO2 in the A- and E-Groups was 7.0 +/- 5.0 and 7.2 +/- 3.3 mmHg respectively. From L(30) to ES(10), the P(a-ET)CO2 and the ratio of Vd/Vt in both groups increased significantly and progressively (P < 0.05 vs the control value) and the slope in the E-Group rose two times as compared to that in the A-Group. The correlation coefficient between P(a)CO2 and P(ET)CO2 was very significant from S(10) to L(90) and ES(10) in the A-Group, and from S(10) to L (60) in the E-Group (P < 0.01). CONCLUSIONS: For maintaining adequate ventilation of the elderly patient during HRA, PaCO2 should be measured intermittently along with the position changes after anesthesia in addition to the constant monitoring of PETCO2.