Automated keratometry for toric IOL surgery planning may be superior to manual measurement

This study conducted by Warren Hill, MD, and colleagues compared simulated clinical calculations based on dual-zone automated keratometry from an integrated optical biometer (Lenstar LS 900, Haag-Streit International) with manual keratometry results in patients with previously implanted toric IOLs. Although the results suggest the two methods provide similar results, automated keratometry had less variability, which suggests an operational advantage. I wish I had one of these devices.

The authors reviewed records of 96 patients (128 eyes) seen at four clinical sites with manual keratometry and biometry data recorded with the Lenstar LS 900 before toric IOL implantation and refractive follow-up data taken after implantation. Of these eyes, surgery for 56 was planned on the basis of manual keratometry readings, 23 on dual-zone automated keratometry readings and 49 on results from a different device or consideration of multiple devices. The authors calculated simulated refractive outcomes based on mathematically removing the actual IOL implanted and then mathematically inserting the IOLs as determined by manual or automated keratometry from the biometry device.

They found that the actual residual astigmatism was comparable between manual keratometry and automated keratometry from the biometry system (P = 0.67). However, results varied by patient. Also, simulated residual refractive astigmatism was lower for the biometer when the standard deviation of the angle of astigmatism was low.

When a difference of more than one IOL power was calculated between the two methods, the automated keratometry data yielded lower simulated residual refractive astigmatism results in seven of 10 cases. The authors say that the difference in magnitude of the simulated residual refractive astigmatism between methods when there was this level of disagreement (approximately 1 D) argues for having a second or even third method of measuring corneal astigmatism to corroborate the primary method.

They note that the calculated residual refractive astigmatism error (the expected amount based on the surgical planning output result) in the simulations was higher than for the actual surgical planning because no subjective adjustment of IOL selection was permitted in the simulations. They conclude that lowering the variability of expected outcomes might allow surgeons to adopt relatively more aggressive treatment strategies because the likelihood of significantly flipping the axis would be reduced, and this would likely improve residual refractive astigmatism results for all patients.