Scleral intraocular lens fixation is gaining more importance over the last years. Subluxated IOLs, pseudoexfoliation syndrome and trauma are the main reasons for scleral IOL fixation. Besides well-established iris fixation and IOL suturing, two relatively new techniques have been developed in recent years. Little is known about the effective lens position, and thus IOL power calculation in these cases is challenging.

The Carlevale and the Yamane methods are two innovative procedures which allow for scleral lens fixation regardless of ACD or iris status. Both techniques are based on the implantation of foldable lenses, which are inserted through a relative astigmatism neutral 2.4–2.8 mm posterior limbal incision, leading to functionally satisfying postoperative results. However, it has to be evaluated if these techniques also fulfill the patient’s increasing expectation of visual function and refractive outcome [9, 10, 13].

Only few studies have reported the refractive outcome of either of the two methods, but so far, none have compared them directly in a prospective manner [11, 14].

Here, we evaluated and compared the PE of 53 consecutive patients who underwent secondary IOL implantation performing the Yamane three-piece IOL technique or secondary Carlevale IOL implantation. Additionally, we analyzed the influence of biometric parameters on the PE.

There was no significant difference in age, postoperative SE, postoperative visual acuity, AL, Kmean, ELP and WTW between the YC and CG.

Surprisingly, there was no statistical difference in IOL decentration (YG: 0.57 ± 0.37 mm, CG: 0.38 ± 0.21 mm) since we expected a higher degree of decentration within the YG as IOL centration is based on the manual adjustment of the IOL position by pulling on the haptics and different flange lengths which appears to be less precise then using the Carlevale IOL.

IOL tilt can cause higher order aberrations and refractive (spherical and torical) errors [15, 16]. Haptic design is very likely responsible for the lower amount of tilt in the CG (6.45 ± 2.03°) as compared to the YG (7.67 ± 3.70°). With the Carlevale IOL, the haptics are placed at the exact opposite positions decreasing the possibility of tilt and decentration. The scleral fixation of a three-piece IOL in the Yamane technique induces torque and stress on the haptic-optic junctions, subsequently leading to IOL tilt. Another source of tilt is that the haptics are passed through an intrascleral tunnel whereby the length and angle of the tunnel in the eye are prone to be asymmetrical despite the external placement of the 30G needle 180° apart. A recent study including 39 patients who underwent Yamane technique reported a mean tilt of 2.4° which was far below our results. But the group measured tilt by calculating the mean of the tilt along the vertical and horizontal axes using an AS-OCT whereas we reported the highest tilt independently of its axis measured over six sectorial B-scans in an AS-OCT [17].

Regarding the calculation of the PE in the YG, we found that the SRK/T formula led to a more myopic result, whereas the Holladay 1 and Hoffer Q to a more hyperopic result than predicted. Our findings considering the Hoffer Q and Holladay 1 formulae are comparable to what was recently published by McMillin et al. [14]. In their study, 40 patients underwent IOL implantation using the Yamane technique with the following formulae being used: Holladay 1, SRK/T Hoffer Q and Barrett formula. Their PEs were between + 0.46 D (Barrett2) and + 0.67 D (Hoffer Q). The difference with the SRK/T results between their hyperopic shift and our myopic shift is somehow surprising, as they stated that their haptic externalization points were placed 2 mm behind the limbus as originally described by Yamane [8], whereas we chose 2.5 mm so we would expect a hyperopic shift. This incongruence might be due to forward or backward shift of the IOL optic compared to the haptic fixation point or high variability between the surgeons’ personal definition of the surgical limbus as a landmark. Due to some cases of optic-iris-capture within our pilot series, which we performed prior to this study, we chose to externalize the haptics 0.5 mm posterior to the recommended location. The lower rates of optic-iris-capture in Yamane’s patients, in the above-mentioned study, might be due to the use of an IOL with a 7 mm optic diameter, compared to the IOL we implanted with a common optic diameter of 6 mm [8].

The evaluation of the Barrett formula is of subordinate importance as it is based on ACD and lens thickness, of which both are not available in pseudophakic or aphakic patients. Considering the MAE, they reported values between 0.73 D (Holladay 1) and 0.86 D (Hoffer Q), which are comparable to ours [14].

Compared to the Yamane method, the IOL haptics in the Carlevale method are externalized 1.5 mm posterior to the limbus. However, this IOL is specifically designed for scleral fixation with a particular optic and haptic design. The optic diameter is 6.5 mm and each haptic is attached at two points on the optic converging into one anchor haptic creating a kind of plate haptic with a single scleral attachment on each side of the optic. We did not observe any optic-iris-capture in our study patients. The IOL power constants are adjusted to this scleral fixation point, and thus in the CG, we found that SRK/T and Holladay 1 formula resulted in a more myopic outcome than predicted. On the contrary, the Hoffer Q resulted in a more hyperopic outcome, similar to that in the YG.

Our PE results in the CG agree with the ones published by Vaiano et al. [11]. They retrospectively analyzed the refractive PE of 25 patients and adapted the IOL power calculation constants to subsequently level the mean PE to zero. They reported slightly higher standard deviations for the PEs which were between ± 0.89 D (SRK/T) and ± 0.95 D (Hoffer Q) depending on the formula used, compared to ours which were between ± 0.74 D (Holladay 1) and ± 0.80 D (SRK/T).

They reported MAEs between 0.62 D (SRK/T) and 0.67 D (Hoffer Q) which were similar to our results 0.61 D (Holladay 1) and 0.68 D (SRK/T).

The proportion of eyes within ± 0.5 D was higher in Vaiano’s analysis (56% SRK/T to 64% Holladay 1) compared to ours (45% SRK/T to 52% Holladay 1), though in our study, more eyes were within ± 1.0 D (68% Holladay 1, 72% Hoffer Q & SRK/T in Vaiano et al.’s study; 72% SRK/T, 76% Hoffer Q & Holladay 1 in our study) and none were beyond 2.0 D [11].

Across our groups, no difference in the PE between corresponding formulae could be found. Concerning the AE, no difference between the formulae could be observed within and across groups.

Multiple studies reporting the PE after Iris claw implantation showed a PE between 0.99 ± 0.57 D and 1.1 ± 0.94 D. The proportion of eyes within ± 1.0 D AE was around 60%. The PE and AE using the Yamane or Carlevale method were lower as compared to the above results [18,19,20]. Besides inferior refractive outcome, iris claw lenses need larger incisions which result in higher postoperative astigmatism and possible wound leakage in the early postoperative follow-up [18, 21, 22].

Modern formulae lead to a highly predictable refractive outcome after cataract surgery in which 66%–71% of eyes are within an error range of 0.5 D and between 90% and 97% within 1.0 D depending on eye length, formula and publication [23,24,25]. Scleral fixated IOLs are not as precise as compared to in-the-bag IOL implantation. Possible reasons are effects of biometric parameters on the power calculation and high variabilities in the ELP. Additionally, patient numbers are much lower and subsequently, extreme eye lengths are rarer and therefore statistical analyses and IOL constant optimization become less precise.

Regarding the influence of presurgical biometric data on the PE, we did not find any statistically significant nor clinically relevant correlations between AL, Kmean and the PE in the CG, so no recommendation about which formula to use in certain eye lengths and corneal radii can be made.

We did not find any significant correlation between AL and the PE in the YG which is in line with what was published by McMillin [14]. Although we did not find any correlation between WTW and the PE in either group, Mularoni et al. reported that two of their patients with an WTW of over 12.6 mm led to extensive haptic stretching which shifted the FIL-SFF IOL optic anteriorly [26]. Three of our patients showed a WTW above 12.5 mm, however we only observed a slight myopic shift in one of the patients (Holladay 1: −0.5 D). Furthermore, we did not find any significant correlations between AL and ELP nor WTW and ELP.

Our final goal was to evaluate the influence of the ELP on the PE. In the CG, increasing ELP led to a more hyperopic result in both uni- and multivariable analyses, reaching R2 values from 0.28 up to 0.44 whereas there was only very little correlation (max R2 = 0.2) between ELP and PE for the YG. This discrepancy might be explained by differences in tilt or the fewer number of patients.

For the CG however, ELP was the parameter with highest influence on refractive outcome in our analysis, highlighting the importance of reliable predictability of the IOL position and reliable intraoperative lens positioning. Similarly, a clinically relevant correlation between ELP and PE was found for sutured secondary IOL fixation after pars plana vitrectomy (ppVE), where a 1.0 mm more posterior externalization point led to a + 1.0 D hyperopic PE [27]. Hence, for scleral lens fixation techniques, precise and reliable placement of haptics is an important factor for ELP and the final refractive outcome, especially when compared to in-the-bag IOL placement, where the ELP is less likely to be influenced directly by the surgeon.

The results of our study must be interpreted with certain limitations as the number of patients in both groups is small and there are no eyes with very long or very short AL or large WTW values. Although small astigmatism-neutral posterior limbal incisions were made, minimal amounts of corneal astigmatism could be induced and subsequently influence the refractive outcome. Additionally, while all surgeries were performed by a single, highly experienced surgeon, scleral IOL fixation may be affected to a higher degree by individual operating style and technique in comparison to primary in-the-bag placement.