Background: In order to perform cataract surgery, adequate dilatation of the pupil is essential. This is traditionally achieved by preoperative topical mydriatic eye-drops, commonly cyclopentolate and phenylephrine. This routine has several disadvantages. First, the slow penetration through the cornea delays the onset of mydriasis. Second, the limited bioavailability of topically administered substances with significant systemic absorption may increase the risk for systemic side effects. Third, even if good mydriasis is achieved initially with topical mydriatics (TM), the effect tends to wear off during surgery. In relation to cataract surgery a transient postoperative corneal oedema is sometimes noted, indicating effects on the corneal endothelial pump function. These effects have been ascribed to ultrasonic or mechanical trauma from the phacoemulsification procedure. Corneal endothelial cell loss (ECL) is a commonly studied variable, not least because it is associated with the long-term risk for corneal decompensation. But, there has been a debate whether postoperative corneal swelling after phacoemulsification cataract surgery correlates to ECL. Aims: To evaluate an alternative mydriatic regimen for phacoemulsification cataract surgery: intracameral injection of mydriatics mixed with lidocaine (ICM). Additionally, to determine the correlation between early transient postoperative corneal oedema and permanent ECL after phacoemulsification cataract surgery. Methods: Pupil dilatation with ICM (150 µl of lidocaine 1%, phenylephrine 1.5%, and cyclopentolate 0.1%) was compared to TM (phenylephrine 10% and cyclopentolate 1%) prior to cataract surgery. Additionally, two ICM-groups were randomized to receive either 0.6 µg/ml epinephrine added to the irrigating balanced salt solution or no epinephrine in the irrigation solution. Furthermore, two randomized ICM-groups, with or without cyclopentolate, were analyzed. The patients planned for cataract surgery were examined with ultrasonic pachymetry, specular microscope endothelial photography and Orbscan II slit-scan tomography pre- and postoperatively. Results: With ICM, mydriasis reached 95 ± 3% of its final value within 20 seconds. In the ICM-group, the pupils were smaller than in the TM-group (mean 6.7 ± 1.0 mm versus 7.7 ± 1.0 mm, P<.001), but did not contract intraoperatively as the TM pupils did. Conversely, with ICM the pupil sizes generally increased during the cataract procedures. This increase was significantly greater without epinephrine in the irrigating solution (13 ± 19% versus 4 ± 14%; p = 0.02). No significant differences in pupil sizes were observed between the patients who were given ICM with or without cyclopentolate. The central corneal swelling at the first postoperative day was strongly correlated to the central ECL at 3 months, R2 = 0.785, P < 0.001. Conclusions: ICM is a rapid and safe alternative to TM in phacoemulsification cataract surgery. An irrigating solution without epinephrine can safely be used with ICM. Cyclopentolate, administrated intracamerally, has no immediate additive mydriatic effect to intracameral lidocaine combined with phenylephrine. The degree of permanent corneal endothelial damage in cataract surgery is reflected in the degree of early postoperative corneal swelling.