**4.6 Lens calculation formulas in vitrectomized eyes**

In recent years, the development of new biometric formulas to calculate the power of IOL to be implanted to our patients has allowed the minimization of post-surgical refractive surprises. New biometric calculation formulas use a variety of strategies, such as the inclusion of more predictive ELP values, the use of ray tracing, or artificial intelligence to achieve optimal post-surgical results.

There is currently no consensus on the most accurate method for biometric calculation in vitrectomized patients.

Lamson et al. [26] observed in a retrospective study that refractive outcomes using eight biometric formulas (Holladay 1, SRK/T, Barrett, Hill-radial basis function, Ladas and Holladay 2) were more variable and more hyperopic than in non-vitrectomized populations. The Holladay 2 formula obtained the highest percentages of postoperative refraction with predicted errors between ±0.50 D and ± 1 D. However, we should point out that the study was retrospective and analyzed a reduced sample of patients. In addition, there were important uncontrolled variables such as the implanted IOL model, which was not the same in all participants of the study. This hyperopic shift in vitrectomized eyes also was reported by Lee et al. [68]

Recently, another retrospective study published by Tan et al. [69] evaluated the refractive results obtained in cataract surgery in vitrectomized eyes by applying next-generation formulas (Barett Universal II, EVO, Kane, and Ladas super formula) against traditional formulas (Haigis, Hoffer Q, Holladay 1, and SRK/T) with Wang-Koch axial length adjustment if required. Before the lens constants were optimized, hyperopic outcomes were noted for all formulas, except for the Kane formula, which revealed no statistically significant bias. However, lens constant optimization enabled optimal and comparable results for all formulas.

As general recommendations to calculate IOL in vitrectomized eyes we suggest.

The optimization of the constant in clinical practice or, if not possible, choose a slightly myopic refractive target for the IOL to be implanted (−0.5 D).

Traditional formulas in miopic patients with axial length more than 26 mm should be used with Wang-Koch's correction.

The presence of silicone oil in the vitreous cavity does not change the choice of the biometric formula.

#### **4.7 Considerations in combinated phaco-vitrectomy**

Phaco-vitrectomy is mandatory in cases of retinal surgery with prior cataract. In addition, a large proportion of patients undergoing vitrectomy will develop cataract in the following years. Therefore, phaco-vitrectomy is a common procedure even without prior cataract as it saves costs and risks of a second intervention.

Unlike surgery in previously vitrectomized patients, where the tendency was to a hypocorrection after phacoemulsification (see "LENS CALCULATION FORMULAS IN VITRECTOMIZED EYES"). Phacoemulsification performed concurrently with vitrectomy seems to be associated with myopic shift in the refractive outcome [26, 34, 61, 70] Tranos et al. [70] found that postoperative refractive deviation greater than 0.5 D was associated with shallower ACD and increased macular thickness. Shiraki and Schweitzer [71, 72] related the myopic shift in combined phaco-vitrectomy with the gas tamponade commonly used in cases of retinal detachment. On the other hand, Vandergeest et al. [73] found no tendency toward a myopic shift and they got an elevated percentage of refractive accuracy in combined procedures.

Different from phacoemulsification in previously vitrectomized cases and faced with the variability of published results, our recommendation in cases of phacovitrectomy combined surgery would be to calculate the intraocular lens with a refractive target of zero.
