Influence of optional biometric variables on refractive outcomes predicted through the Kane and Barrett UII formulas
Free Paper Details
First Author: M.Raimundo PORTUGAL
Co Author(s): E. Neves J. Simão C. Lobo J. Murta
Modern intraocular lens (IOL) calculation formulas depend on a set of mandatory biometric parameters – axial length, keratometry, anterior chamber depth – as well as optional biometric parameters. This can be seen in the Barrett Universal II and the Kane formula which contain optional parameters such as lens thickness (LT) and white-to-white distance (WTW) or LT and central corneal thickness (CCT), respectively. Such parameters may not be measurable using older optical biometry devices. Currently there is a lack of evidence regarding their effect on refractive outcomes. We compare outcomes of cataract surgery with and without the use of optional parameters.
Tertiary university hospital.
Retrospective consecutive case study of eyes that underwent uncomplicated cataract surgery with trifocal intraocular lens implantation (Alcon PanOptix®). We compared the predicted spherical equivalent (SE) for the implanted lens as obtained through the Kane and Barrett Universal II (UII) formulas to the postoperative SE obtained by subjective refraction 6-12 post-operatively. To assess the influence of optional variables for the Barrett UII and the Kane formulas back-calculation without those parameters was performed. Primary outcomes included mean arithmetic error (ME), standard error (SD), mean absolute error (MAE) and median absolute errors (MedAE).
Seventy-nine eyes were included (n=79). After optimization, the MAE obtained through the Kane and Barrett UII formulas without optional parameters was 0.207 and 0.213, respectively (p>0.05). MedAE was 0.171, 0.173, respectively (p>0.05). With the inclusion of both optional parameters for the Barrett UII and Kane formulas, the post-operative SE prediction changed by more than 0.1D (absolute values) in only 15.2% (n=12) and 17.7% (n=14) of eyes, respectively. In this subset, a more accurate prediction was seen in 50% of cases (n=6 and n=7), with a worse prediction in the other 50% (n=6 and n=7).
In our sample, the inclusion of optional biometric variables in the Barrett UII and Kane formulas did not provide any significant refractive benefit.