In this retrospective study, we analysed the differences in visual and refractive outcomes, vault, and rotational stability between horizontal and vertical toric ICL placement by strict 1:1 matched comparison. Although there have been several articles on the safety and efficacy of ICLs, as well as some reports on rotational stability, for the first time, our study investigated the differences in clinical outcomes and rotational stability of toric ICLs at different placement positions [8, 22, 23, 26, 27].
According to the recommendation of the STARR company, all ICLs should be placed within a horizontal 22° deviation (Visian ICL Product Information: Visian ICL For Myopia. Available at http://www.accessdata.fda.gov/cdrh_docs/pdf3/p030016c.pdf). On the one hand, it is difficult to accurately measure the vertical corneal diameter. On the other hand, the rotational stability and clinical effect of vertical placement have not been reported.
Matarazzo et al.  first reported a case of successful vault reduction by vertical placement of the ICL in their study. Subsequently, Kojima et al.  and Fernandez et al.  also reported several cases of extremely high vault reduction by rotation to the vertical position in their respective studies. In our study, the vault of the vertical placement group was significantly lower than that of the horizontal placement group (429 ± 173 µm vs. 612 ± 210 µm), although there was no statistically significant difference in white-to-white diameter between the two groups, and even the horizontal and vertical sulcus-to-sulcus diameters were smaller in the vertical placement group than in the horizontal placement group. This might indicate that vertical placement of the ICL can effectively reduce postoperative vault, especially for patients at the critical value of ICL size selection, and a larger size and vertical placement might be adopted to achieve an ideal vault.
The horizontal placement group in our study, similar to that in previous studies, had a good SI and effective index in terms of visual outcomes. In our study, the SI was 1.21 ± 0.20, no eye lost lines of CDVA, and 66.67% of eyes gained lines of CDVA after the surgery. In addition, the efficacy index was 1.02 ± 0.17, and 100% of eyes had a UDVA of 20/20 or better. A recent study by Chen et al.  found that the SI was 1.19 ± 0.23 and the EI was 1.04 ± 0.27 for ICL (model V4c) implantation. Similarly, Liu et al. [2, 7] also reported in their two studies that the safety factors of ICL implantation were approximately 1.11 ± 0.15 and 1.33 ± 0.60, and the efficacy factors were approximately 1.06 ± 0.15 and 1.14 ± 0.54. These reports were highly consistent with our results. On the other hand, in our study, the clinical outcomes of the vertical placement group were as good as those of the horizontal placement group, with a SI of 1.24 ± 0.22 and an EI of 1.07 ± 0.14. A total of 97.2% of eyes had a UDVA of 20/20 or better, and 100% of eyes had a UDVA of 20/25 or better. There was no significant difference in postoperative UDVA between the two groups (-0.046 ± 0.059 vs. -0.055 ± 0.048, P = 0.472).
Rotational stability is crucial for astigmatism correction by toric ICLs. In this study, OPD-Scan III was used to measure the axial direction of the toric ICLs. Previous studies have shown that compared to conventional methods, OPD-Scan III yields a lower standard deviation value and higher accuracy in measuring the axial direction of the toric ICL [23, 26]. In our study, the absolute degree of rotation at 3 months postoperatively was 3.44° ± 2.72° for the horizontal group and 4.83° ± 2.93° for the vertical group compared to that in the preoperative design. In the two studies in which the crystals were placed horizontally, the absolute degree of rotation at 3 months postoperatively was 3.75° ± 2.92° and 3.39° ± 2.36°, which were similar to the results for our horizontal placement group [26, 31]. Rotation degrees appeared to be higher in the vertically placed group than in the horizontally placed group, although this difference was not statistically significant (P = 0.052). However, if the reference was changed to 2 hours after surgery, the absolute rotation degree was 1.81° ± 1.84° for the horizontal placement group and 1.97° ± 1.70° for the vertical placement group. This suggests that most of the rotation measured at 3 months after surgery might have occurred within 2 hours after surgery, and there was no significant difference in rotation degree between horizontal and vertical placement 2 hours postoperatively. There might be several reasons for this phenomenon. First, in the early postoperative period, the haptics of the ICL might not have been located in the right position of the ciliary sulcus, and it was not sufficiently supportive of the ICL. With the narrowing of the pupil, the iris exerted a downward pressure on the ICL, which can cause the haptics to be better fixed in the corresponding position of the ciliary sulcus and in turn provide the ICL better stability; this is also one of the reasons for the change in the vault after lens implantation . Second, there was usually some viscoelastic agent left under the ICLs in the early postoperative period, which made the ICLs easier to rotate, and this part of the viscoelastic agent would be slowly metabolized. Third, because the upper and lower eyelids cover part of the limbus of the cornea, it is more difficult to create the upper and lower marking points, which might cause certain systematic errors. This might be the reason for the greater deviation of the vertical placement group relative to the planned axis direction.
In this study, we also analysed the influencing factors for postoperative residual astigmatism. Interestingly, postoperative residual astigmatism was not significantly associated with any of the three factors, including ICL rotation, except for a weak positive correlation with AL. In addition to rotation, there seem to be other factors that are more important in influencing postoperative residual astigmatism. We speculated that the stability of the ICL, such as its position deviation and tilt, might have a certain influence on postoperative residual astigmatism. We suspect that due to gravity, the two haptics on the bottom might be more conducive to support of the ICL, making it more likely to remain in the right position. This may explain why the residual astigmatism of the vertical placement group was smaller than that of the horizontal placement group. However, further research is needed to confirm the above speculation. Of course, there is an important premise for the above speculation, that is, the rotation degree of the ICL was small, less than 11° in this study. If the rotation degree was further increased, a large probability of postoperative residual astigmatism would be associated with the rotation degree. In addition, the postoperative residual astigmatism of the two groups was small, which was not enough to have a significant impact on postoperative visual acuity. Therefore, there was no significant difference in postoperative visual acuity between the two groups.
In this study, vector analysis by the Alpins method was used to determine the effect of the toric ICL on astigmatic correction in the two groups. The CI of the horizontal placement group was 1.02 ± 0.17 (range: 0.67 to 1.45), indicating slight overcorrection, while the IOS was 0.20 ± 0.16 (range: 0 to 0.5). For the vertically placed group, the CI was 0.99 ± 0.11 (range: 0.74 to 1.22), indicating slight undercorrection, and the IOS was 0.11 ± 0.13 (range: 0 to 0.5). There was no significant difference in CI between the two groups, while the results for IOS were similar to those for postoperative residual astigmatism. Lee et al.  reported a CI of 0.91 ± 0.21 (range: 0.43 to 1.82) and an IOS of 0.17 ± 0.14 (range: 0.00 to 0.86) in their study on the clinical efficacy of ICLs, and there was an IOS of 0.19 ± 0.11 (range: 0.01 to 0.39) in the study of Hyun et al. ; these findings were similar to the results from our horizontal placement group.
There are several limitations in this study. First, the sample size was not large compared with those of similar studies; however, we adopted a 1:1 paired study that controlled for other factors, and consequently, this study was highly reliable. Second, the follow-up duration of this study was short, but by 3 months after surgery, corneal surgery-induced astigmatism is basically stable; thus, this time point is more commonly used for the study of postoperative astigmatism . Finally, the ICL placement position in the participants included in this study was within 22° of horizontal or vertical. The other placement positions were not within the scope of this investigation and will be further analysed in follow-up studies.