Keywords

Tilt and Visual Quality

A tilt of an intraocular lens (IOL) reduces optical quality due to an increase of lower [1, 2] and higher order aberrations [3]. The impact of tilt on positive and negative dysphotopsia as well as chromatic aberrations appears to be uncertain [4, 5]. Aberrations are a problem for any kind of IOL, but especially for aspheric [6,7,8,9], toric [10, 11], extended depth of focus [12, 13], and multifocal IOLs [3].

In the case of an aspherical IOL, tilt leads to a reduction of the aspherical effect up to a worse performance compared to a spherical IOL [6,7,8,9]. A “common” amount of tilt (up to 5°) was not shown to have a relevant influence on the performance of the Strehl ratio in a randomized trial (spherical versus aspherical IOL) [14, 15]. Higher amounts of tilt increase coma (which can mimic astigmatism) [16] and reduce the effect of asphericity [10]. Comparing aspherical, aberration neutral and spherical IOLs in the presence of tilt showed that an aberration neutral IOL outperforms an aspherical IOL [17, 18].

For toric IOLs, tilt has a direct and indirect impact on post-operative astigmatism [10, 11] and it explains approximately 11% of the residual astigmatism error or up to 20% if angle kappa is also taken into account [19].

Multifocal IOLs show a reduced optical quality if tilted. Although this accounts for any type of multifocal IOL, especially the performance of rotationally asymmetric multifocal IOLs decreases with tilt [20, 21].

Measurement of Tilt

Severe tilt may be detected at the slit lamp, although it does not allow any quantification of tilt and the measurement is not reliable.

In general, there are two principal methods to quantify tilt:

  1. 1.

    Cross-sectional scans of the anterior segment

    • Scheimpflug imaging or rotating slit lamp images

    • Optical coherence tomography (OCT)

    • Ultrasound biomicroscopy (UBM)

  2. 2.

    Assessing the Purkinje reflexes

Cross Section-Based Imaging

Tilt quantification with cross-sectional images was introduced in the 1980s [22]. Due to the fact that conventional imaging techniques (except ultrasound) use a light source, imaging behind the iris is not possible. Therefore, this type of tilt quantification uses a fitting concept, where the visible parts of the anterior and posterior lens surfaces are fitted using curved lines (Fig. 61.1). The point of contact is then the estimated equator of the lens. This kind of measurement needs to be performed with a well-dilated pupil in order to assess as much surface of the IOL as possible. In some cases, it is also difficult to identify the anatomical structures of the eye that are necessary to align the points of the reference axis [23].

Fig. 61.1
2 s s O C Ts of phakic and pseudophakic eyes. The anterior side of the cornea appears thickened in the phakic eye compared to the pseudophakic eye. The 3-D results of the eyes are provided as an inset image.

ssOCT images of the phakic (above) and pseudophakic eye (below). The anterior/posterior surfaces of the cornea and the lens are automatically detected [27]

Full size image

More recently, OCT devices have been used to quantify tilt. This concept was shown to be successful for older concepts, such as the time domain OCT [24], but also for more modern devices, such as longitudinal B scans using a swept source OCT [25], or anterior segment swept source OCT devices [26]. Fig. 61.1 shows the large imaging range of up to 13 mm width that allows to measure the region between the epithelium of the cornea and the posterior lens capsule in a single scan [27]. Additionally, tilt was also quantified using a 3-dimensional approach [28,29,30] and a deep learning approach was introduced that allowed to automatically quantify tilt using the scleral spur as a reference [31].

Another possibility is to use a high resolution ultrasound device, often referred to as ultrasound biomicroscopy (UBM), which allows measurements behind the iris [32]. A disadvantage of UBM is that a probe is needed and while the eye is in contact with the probe, the patient cannot fixate on a target. However, it is a good approach for cases where low compliance levels are expected [33]. Although it is more difficult to define the reference axis for UBM scans, it was shown to be beneficial for quantification of out of the bag IOL implantation [34,35,36,37].

Purkinje Reflexes

Purkinje reflexes are another possibility to assess tilt. This concept was already used in the early 1980s [38,39,40]. Since light is reflected at all interfaces of media with a difference in refractive index, these reflections, called Purkinje reflexes, may be used to assess tilt of IOLs.

Two different clinically applicable Purkinje meter systems have been used for the measurement of IOL decentration and tilt [3, 41]. These Purkinje meters use a different algorithm for the analysis. A video camera-based photograph of the reflections from the cornea and the IOL is performed in both devices and with the help of a dedicated software, tilt is calculated [3]. The technique is a non-contact technique which is quick and easy to perform. The improvement and advancement of both systems have been shown to be accurate to measure IOL alignment and to evaluate the effect of IOL misalignment on optical performance [42].

Tabernero et al. [41] improved the measureability of tilt by using a semicircular ring of light emitting diodes. These semicircles are captured and analysed according to their size and distance to each other as well as their position within the pupil (Fig. 61.2).

Fig. 61.2
3 Purkinje structures of an eye. It consists of 3 layers. The outermost to innermost layers are P 4, P 3, and P 1.

Purkinje imaging of a perfectly aligned ophthalmic system—the outer circle represents the pupillary margin, the inner complete dotted circle the first and second (lower half) and the fourth (upper half) Purkinje reflex. The third Purkinje reflex representing the anterior surface of the lens is reflected as a thick dotted half circle [43]

Full size image

As shown in Fig. 61.2, only three semicircles are visible, because the first and second Purkinje reflex (anterior and posterior surface of the cornea) overlap. The distances between the reflexes and the position within the pupil are then plotted as an angular fixation function, where the fixation angle correspondences with the overlapping point of the third (anterior surface of the lens) and the fourth (posterior surface of the lens) Purkinje reflexes. Due to the fact that the patient fixates a central target, IOL tilt and decentration can be measured. This idea was previously described by Guyton et al. [44] in a more manual fashion that was also confirmed in a later study [45].

Another Purkinje meter was developed by Schaeffel [43] and differs from Tabernero’s Purkinje meter in terms of the light source (single LED instead of a semicircle) and the patient has to fixate on an LED target at different positions instead of one central fixation target (Fig. 61.3).

Fig. 61.3
A Purkinje meter structure of an eye. The fields of the meter are focused on the pupil of the eye. The fields are formed using dots and circles. The results and dimensions are displayed on the left side of the structure.

Purkinje meter using dots instead of half circles and an audio system to evaluate the quality of the image [43]

Full size image

In a direct comparison between both Purkinje meters including 30 eyes and inviting both inventors to assist with the measurements, a higher feasibility for the Purkinje meter developed by Tabernero and Artal was found [46]. Comparing only the successfully measured cases, both devices should not be used interchangeably.

In a direct comparison of Scheimpflug imaging and Purkinje meter measurements, both were shown to be reproducible, but the accuracy was higher for the Purkinje meter measurements [47].

Physiological Tilt

A certain amount of tilt is beneficial, as it compensates for horizontal coma [48]. The mean amount of tilt of the crystalline lens was shown to range between 4.3° [49], 4.6° [43], 4.9° [27], and 5.2° [50]. Furthermore, there is a correlation between axial eye length and tilt with shorter eyes having a higher amount of tilt [25, 51]. This should be kept in mind and the term “physiological tilt” should be introduced. There is evidence that the physiological tilt is inferotemporal (the fovea is slightly temporal to the pupillary axis) [49,50,51] and that there is a mirror-symmetry between the eyes [49]. Furthermore, tilt slightly increases (on average less than 0.5°) in the presence of mydriasis [50].

For IOLs, there is a variety of studies assessing the amount of tilt ranging from 2.7° [52], 2.9° [53], 2.9° [28], 3.9° [54], 4.1° [55], 4.8° [51] to 6.2° [49]. Although this list is not complete, it shows the range of tilt. The amount of tilt depends on several factors, such as axial eye length, different measurement and analysis systems, and differences in reference axes. Unfortunately, there is no standardization and different authors have used different definitions and different reference axes so that they cannot be used interchangeably (Fig. 61.4) [56]. This is relevant as some reference axes include angle kappa, whereas others do not [56].

Fig. 61.4
A schematic of the pupillary physiological tilt. The line of sight is a horizontal straight line from the fovea to the fixation target. The pupillary axis line is aligned obliquely at an angle kappa from the line of sight.

Graphical definition of pupillary axis and line of sight and the angle kappa [41]

Full size image

Prediction of Post-operative Tilt

As mentioned above, tilt accounts for more than 10° of the error in toric IOL power calculation and this value increases to almost 20%, if combined with angle kappa [57]. Therefore, predicting tilt and taking it into account would significantly improve toric IOL power calculation [49, 57, 58].

Although prediction of the post-operative amount of tilt is more difficult (correlation of r = 0.4) [49], it was shown that the orientation of tilt can be predicted quite well with a correlation of r = 0.7 (Fig. 61.5) [49]. The average orientation before and after cataract surgery is approximately 16–17° and the predictive power is high [49]. The correlation for the pre- to post-operative amount of tilt was found to be higher (r = 0.5–0.7) in two other studies (Table 61.1) [25, 27]. Axial eye length was not found to be a good predictor of post-operative tilt (r = 0.2) [25].

Fig. 61.5
3 O C Ts of an eye. The O C Ts are analyzed at 3 meridians. The corneal thickness is greater in a phakic eye compared to the pseudophakic eye.

Swept source OCT imaging at three different meridians of the same eye in the phakic state (left) and the pseudophakic state (right) [49]

Full size image
Table 61.1 Data of pre-operative and post-operative amount of tilt in three different studies. * data not in the paper, but calculated from the associated online .xls file
Full size table

Factors Influencing Tilt

Although there is currently no good prediction algorithm on which eye will have severe IOL tilt after cataract surgery, several risk factors were discussed.

Capsulorrhexis

Different aspects of the capsulorrhexis were evaluated concerning their impact on IOL tilt. There is good evidence that the size of the capsulorrhexis has no influence on IOL tilt [55, 59]. Shape and centration of the rhexis were also not found to be clinically relevant in the same two studies. However, an incomplete capsulorrhexis overlap (probably less than 50% overlap—estimation) was found to be a risk factor for tilt [55, 59, 60]. Older techniques used before the introduction of the continuous curvilinear capsulorrhexis, such as the envelop technique, resulted in significantly higher tilt values and should be avoided [61].

As the size and shape of the capsulorrhexis were not shown to have a relevant impact on tilt with modern single-piece IOLs (except for a severe missing overlap), it is likely that femtosecond laser-assisted cataract surgery (FLACS) does not reduce post-operative tilt either. However, it should be mentioned that this was not confirmed in all studies [62].

The bag-in-the-lens IOLs, where the IOL is connected to the anterior and posterior capsulorrhexis edges, were shown to have small amounts of tilt [63, 64]. Although this type of IOL may be used with a meticulously made manual capsulorrhexis, it may be easier to be used with a FLACS made capsulotomy, as the shape and the size of the capsular opening are crucial for the position of the IOL. Incision size was not found to be a relevant factor for predicting tilt [65, 66].

Pseudoexfoliation

Pseudoexfoliation was found to be a relevant risk factor for a post-operative forward tilting of the superior haptic [24, 67, 68] as well as a long-term risk factor for IOL dislocation [69, 70]. Furthermore, pseudoexfoliation is associated with anterior capsule contraction syndrome, which may also result in a tilted IOL [71].

In another study, there was a tendency that a capsular tension ring prevents tilt to a certain degree [72]. This could be beneficial in eyes with pseudoexfoliation, but evidence is scarce and further studies would be necessary for confirmation.

IOL Material and Design

There is general agreement that the influence of IOL material has no or only a minor impact on IOL tilt [73,74,75]. Walkow et al. [76] observed similar results, when assessing the reason for IOL explantation due to decentration or subluxation.

On the other side, the design of the haptics was found to be relevant [73]. This leads to the question, if there is a difference between 1-piece and 3-piece IOLs. A large randomized bilateral comparison found significant differences with the 3-piece IOL showing a significantly higher amount of tilt [53]. This was also confirmed by another randomized trial [53]. Two other studies did not confirm this finding [52, 77]. Although the design of the haptics potentially has an effect on the amount of tilt, the orientation of the haptic position was not found to be relevant [53]. Possibly, the higher tilt in 3-piece designs is a consequence of a slight kinking or bending of the haptic during the implantation process since the haptics have a limited memory compared to the thicker single-piece haptics used.

After-Cataract

Although only mild in extent, posterior capsule opacification, or after-cataract, potentially increases tilt and may be relieved with a posterior Nd:YAG capsulotomy which was shown to decrease tilt back to normal levels [78, 79].

IOL Implantation Outside the Capsular Bag

Three piece IOLs in the sulcus tend to have higher tilt levels (horizontal tilt on average 7.7°) compared to those in the bag IOLs [80]. If this is due to the position in the sulcus itself, or due to the typically compromised posterior capsule has not been identified. Another explanation could be that in the case of sulcus IOLs sometimes one of the haptics unintentionally is positioned in the bag instead of being in the sulcus [37].

For scleral fixated IOLs, slightly higher tilt values were observed compared to those in the bag IOL implantation. For scleral fixated IOLs with a Z-suture, relevant tilt was found in 72% of all cases [81], whereas intrascleral fixation showed lower tilt values of little more than 3°, even though 8% of all cases had an iris capture [82]. Low tilt values were also confirmed for self-sealing scleral pockets measured with UBM [34,35,36] and OCT technology [35] and for long-term results using glue [83].

Furthermore, scleral fixated IOLs showed less tilt, if the sclerectomy was performed with 24 gauge compared to 30 gauge [84]. In the case of relevant post-operative tilt shortening, the length of the haptics was found to be useful to reduce tilt in some cases [85]. There is little information comparing tilt data of scleral fixated IOLs versus iris claw IOLs. One study performed in children showed higher tilt values for scleral fixated IOLs [86].

Combined Surgery

For phacotrabeculectomy, there is no evidence for an increased risk of clinically relevant IOL tilt [87]. Phacovitrectomy potentially increases the risk of IOL tilt, depending on the vitreous tamponade [51]. If air or gas is used, there is evidence for an increased tilt compared to no tamponade [26, 88, 89]. However, this difference was not found to have a significant influence on lower or higher order aberrations and the clinical effect is questionable [88]. It should also be mentioned that a randomized study directly comparing combined phacovitrectomy including endotamponade versus cataract surgery as a stand-alone procedure did not confirm these findings and no difference in tilt was observed [90].

Effect of Tilt on Refraction

The effect of tilt on the induced astigmatism in an aspherical toric IOL depends on several variables:

  • Power of the IOL (spherical equivalent and astigmatism if the lens is toric)

  • Amount and orientation of tilt

As shown by Weikert et al. [10], a non-toric aspheric IOL tilted horizontally (nasal border more anterior, like physiological tilt) will induce against the rule astigmatism. A horizontal tilt of 10° of a 16D and a 28D IOL would result in an induced against the rule astigmatism of 0.33D and 0.56D, respectively (Fig. 61.6).

Fig. 61.6
A line graph plots astigmatism versus I O L tilt. The plots of 16 D, 22 D, and 28 D begin at the origin and form increasing J-shaped curves till (10, 0.35), (10, 0.45), and (10, 0.55), respectively. The values are approximated.

Simulated against the rule astigmatism induced by tilt in an eye with an aspheric IOL for three different IOL powers [10]

Full size image

In the case of a toric IOL oriented at 90°, the horizontal tilt resulted in increased against the rule astigmatism, resulting in overcorrection. If the IOL was oriented at 180°, the consequence would be an undercorrection. It’s curious to observe that this against the rule trend is similar to the effect of the posterior corneal surface astigmatism.

Marcos presented a method to estimate the effect of tilt on astigmatism (in air) using a thin lens formula (Eq. 61.1).

$$ A=P\left\{1+\frac{{\left(\sin \alpha \right)}^2}{3}\right\}\ast {\left(\tan \alpha \right)}^2 $$
(61.1)

Estimating the effect of tilt on astigmatism (A = astigmatism in D, P = power of the IOL in D, α = amount of tilt [91].

A more complex approach would be to use a model with a thin spherical lens. Simplifying the model by neglecting all effects above the second order of aberrations, the Coddington formula may be used [92]. The effect of tilt has to be explained for each order of aberration. Atchison published a thin lens calculation for the effect of tilt on first- and second-order aberrations [92]. According to the Coddington formula, a finite principal ray is sent from an object through a spherical lens and another neighboured ray is sent from the same object through the same lens, where these two rays intersect after refraction [93]. This intersection point consists of focal lines. The two main focal lines are usually called tangential (VT) and sagittal (VS) (Eq. 61.2).

$$ {\displaystyle \begin{array}{l}{V}_S:\frac{n^{\prime }}{s^{\prime }}=\frac{n}{s}+\left({n}^{\prime}\cos {I}^{\prime }-n\cos I\right){c}_S\\ {}{V}_T:\frac{n^{\prime }{\cos}^2{I}^{\prime }}{t^{\prime }}=\frac{n\cos^2I}{t}+\frac{n^{\prime}\cos {I}^{\prime }-n\cos I}{r}\end{array}} $$
(61.2)

Vergence for tangential (VT) and sagittal (VS) focal lines [93].

s and t = distance from the incident point of the ray to the sagittal and tangential point of the image.

cS = curvature of the anterior lens surface.

I and I′ = angles of incidence and refraction.

n and n’ = refractive indices of the object and image spaces (n represents the refractive index of the object side medium and n’ represents the refractive index of the image side medium).

In a very similar fashion, VT’ can be calculated using the Coddington formula for VT, as shown in Eq. (61.3).

$$ {V}_T^{\prime }=V+\left(1+{\delta}^2+\frac{n{\delta}^2}{2\delta}\right)F $$
(61.3)

Vergence for the transversally misaligned focal line [92] (modified).

As for the longitudinal displacement, the formula for the effective lens power can be used (Eq. 61.4).

$$ {V}_{\mathrm{CS}}=\frac{V_S^{\prime }}{1+\frac{d{V}_S^{\prime }}{n}}\;\mathrm{and}\;{V}_{\mathrm{CT}}=\frac{V_T^{\prime }}{1+\frac{d{V}_T^{\prime }}{n}} $$
(61.4)

Converted using the effective lens power [92] (modified) VCS=Vergence of the sagittally misaligned lens on the corneal plane.

VCT = Vergence of the transversally misaligned lens on the corneal plane.

The refractive error is, similar to the longitudinal displacement, the difference between the correct position of the lens and the displaced image of the lens (VC − VCS; VC − VCT).

In a next step, these estimations of the refractive error can be combined to explain the spherical equivalent of the refractive error due to lens displacement (Eq. 61.5).

$$ \varDelta {F}_{\mathrm{SE}}=\varDelta F+\varDelta {F}_S+\frac{\varDelta {F}_T}{2} $$
(61.5)

Effect of longitudinal misalignment and tilt on the spherical equivalent.

This concept [92] was evaluated using tilt, pseudophakic ACD and refraction data of 100 eyes. The correlation between the theoretically predicted refractive error and the actually measured refractive error using subjective and objective refraction was found to be only moderate (r2 = 0.42 (not published)). The most likely reason is the low accuracy of the post-operative manifest refraction.

Summary

Physiological tilt shows a mirror symmetry between both eyes, depends on the axial eye length, is orientated inferotemporally, and does not exceed 5°.Tilt above this physiological level has a significant impact on visual quality, especially for aspheric, toric, and multifocal IOLs. Predicting post-operative tilt was shown to be successful and to improve toric IOL power calculation. There are two concepts for tilt measurements, cross-sectional-based scans (Scheimpflug, OCT, UBM) and imaging of the Purkinje reflexes of the eye. Risk factors for tilt are pseudoexfoliation syndrome, 3-piece IOLs, after cataract, and potentially phacovitrectomy with endotamponade. The capsulorrhexis was found to have a minor influence on tilt.