Factors affecting central corneal thickness measurement agreement between Scheimpflug imaging and ultrasound pachymetry in keratoconus

Purpose To evaluate the agreement of central corneal thickness (CCT) measurement between Scheimpflug imaging and ultrasound (U/S) pachymetry in keratoconic eyes, and investigate factors that affect the agreement. Methods This post hoc analysis within a prospective, observational non-randomised study preformed at the Kensington Eye Institute, Toronto, Ontario, Canada, included crosslinking candidates with progressive keratoconus (KC). Main outcome measures were the agreement of CCT measurement between Scheimpflug imaging (Pentacam Oculus, Wetzlar, Germany) and U/S pachymetry (PachPen 24-5100, Accutome Inc., Malvern, Pennsylvania, USA), and factors that affect the agreement. Results A total of 794 keratoconic eyes of 456 subjects with a mean age of 27.6±8.0 years (69.7% males and 49.6% right eyes) were included. Agreement between devices was moderate (intraclass correlation coefficient: 74.9%, Bland-Altman limits of agreement: -48.5 μm to +62.5 μm). In a multivariable analysis, cone decentration (p<0.001, coefficient +10.13 [+6.73 to +13.53 95% CIs]) and Kmax (p<0.001, coefficient +0.68 [+0.46 to +0.90 95% CIs]) were significantly associated (both clinically and statistically) with the level of agreement between the devices; the discrepancy in CCT between the devices increased on average by 10.13 μm for every mm of cone decentration, and by 6.8 μm for every 10D of Kmax. Age, corneal astigmatism and spherical equivalent were statistically but not clinically significant factors affecting agreement. Conclusion The agreement of CCT measurement between Scheimpflug imaging and U/S pachymetry in KC was moderate. To ensure the safety of crosslinking in keratoconic corneas, Scheimpflug and U/S CCT measurements should not be used interchangeably, especially in steep corneas and corneas with decentred cones.