ABSTRACT.Purpose: To ascertain the difference of intraocular pressure (IOP) measurement between vitrectomized air-filled and fluid-filled eyes. Methods: Thirty-one eyes of 31 consecutive patients who underwent conventional vitrectomy and intraocular gas tamponade were assessed. After vitrectomy, IOP of the fluid-filled eyes was measured by Tono-Pen. Thereafter, fluid-air exchange was performed, and IOP of the air-filled eyes was measured again. The IOP within each fluid-and air-filled eye was varied by selecting settings on the vitrectomy system, from 10 to 50 mmHg with 5-mmHg increments. Postoperatively, IOP was assessed by both Tono-Pen and Goldmann applanation tonometry (GAT). Linear and nonlinear regression analyses were conducted between intraoperatively measured Tono-Pen readings and actual IOPs. Bland-Altman plot was used to assess the agreements between postoperatively measured Tono-Pen readings and GAT readings. Results: The discrepancy between Tono-Pen readings and actual IOP in fluidfilled eyes was not significant, except for the profound high pressures over 45 mmHg. However, Tono-Pen readings in air-filled eyes were significantly lower than actual IOPs in all ranges, and Tono-Pen increasingly underestimates IOP at higher levels. Intraoperative Tono-Pen readings were correlated significantly with actual IOP and a quadratic equation evidenced the best fit (R 2 = 0.996). Postoperatively, difference of the measurements between TonoPen and GAT was not significant. Conclusion: Tono-Pen and GAT significantly underestimate actual IOP in airfilled eyes. It should be considered that actual IOP would be greater than the measured IOP in gas-filled eyes, even though the IOP is measured as normal.