Screening of positive dysphotopsia before multifocal intraocular lens implantation

Abstract

Purpose:

To assess the photic phenomena (PP) and positive dysphotopsia in candidates for presbyopia or cataract surgery and to evaluate their relationship with cataract grading systems.

Design:

Retrospective observational.

Methods:

Monocular data for 82 subjects measured during the preoperative screening were retrospectively retrieved from our database. The evaluated variables consisted of two methods for PP measurement: light distortion index (LDI) and parameters obtained from a simulator, both of which were combined with subjective bother related to PP. The cutoff for LDI that better predicted patients passing from slightly to moderately bothersome was estimated. The relationships between LDI and the following objective cataract grading methods were also assessed: objective scatter index (OSI), dysfunctional lens index (DLI), and Pentacam Nucleus Staging (PNS).

Results:

LDI was the best method for measuring PP, which showed a significant correlation with the bothersome question (rho = 0.34, P = 0.002) and also with OSI (rho = 0.67, P < 0.0005), DLI (rho = −0.29, P = 0.007), and PNS (rho = 0.48, P < 0.0005). The number/percentage of patients who found it bothersome was as follows: “Not at all” (18/22%), “Slightly” (41/50%), “Moderately” (15/18.3%), and “Very” (8/9.7%). The cutoff value that predicted the transition from slightly to moderately bothersome was ≥15.20% according to LDI, which could be estimated with the following values for grading: ≥2.8 for OSI, ≤7.6 for DLI, and ≥2 for PNS.

Conclusions:

Patients reporting moderately or higher bothersome levels in the preoperative period and with LDI <15.20%, <2.8 for OSI, >7.6 for DLI, and <2 for PNS might deserve special attention in the multifocal intraocular lens selection.

Positive dysphotopsia (PD) is a topic of great interest in presbyopia and cataract surgery with the implantation of a multifocal intraocular lens (MIOL). This is mainly because some patients implanted with these intraocular lenses (IOLs) can be dissatisfied after surgery due to this phenomenon.[1,2] PD could be defined as any bothersome bright artifacts perceived by patients, such as those experienced around direct bright lights or reflected off objects located in the visual field. PD should be differentiated from the photic phenomena (PP), which are the perceptions of these artifacts without considering how much they bother the patient.[3]

Although PD can be directly attributed to MIOL because PP generated by its design is physically higher than that generated by a monofocal IOL,[4] it is important to note that this PP is similarly generated in all patients. However, only two out of 10 patients have described themselves as being bothered in the long term by a particular trifocal MIOL.[5] Therefore, other factors play an important role, such as the patient’s personality,[6,7] which could influence the level of bothersomeness associated with PP. Unfortunately, there is no standardized questionnaire that allows the surgeon to detect, through assessing personality, patients who could potentially complain about PD in the postoperative period. An alternative to these questionnaires might be to assess PP and PD preoperatively, since patients with normal PP in the preoperative period who also complain about PD at this stage might be affected in the postoperative period with an aggravation of the mentioned symptoms.

PP can be measured directly, with the light distortion analyzer (LDA; CEORLab) and with the use of simulators,[3] or indirectly through measurement of the optical media quality using devices such as double-pass systems, aberrometers, or Scheimpflug technology.[8] The main aim of this study was to assess PP and PD in candidates for presbyopia or cataract surgery and to evaluate their relationship with cataract grading systems used in clinical practice. The information provided by this study could help the surgeon identify patients bothered by PD in the preoperative screening, and such bothersomeness cannot be justified either by a PP or by light scattering related to crystalline lens sclerosis or cataract.

Methods

Subjects and procedures

This study was approved by the Ethics Committee of Research, Almería Center, Torrecardenas Hospital Complex, and conducted in adherence with the tenets of the Declaration of Helsinki.

Data from 82 patients, consecutively measured from October 2020 to May 2021 during the preoperative screening for cataract or presbyopia surgery at Qvision, Ophthalmology Department, VITHAS Almería, Spain, were retrieved from our historical database. Inclusion criteria were patients ranging in age from 45 to 80 years old, whose measurements were taken with the following devices and in the following order in at least one eye: (1) corrected distance visual acuity (CDVA) with an early treatment diabetic retinopathy study (ETDRS) chart (VisionC App, Qvision Academy, Almería, Spain), (2) HD Analyzer (Keeler, Malvern, PA, USA), (3) iTrace (Tracey Technologies Corp., Houston, TX, USA), (4) Pentacam (Oculus, Wetzlar, Germany), (5) Light Distortion Analyzer (CEORLab, University of Minho, Portugal), and (6) DysphotopsiApp (beta version, Qvision Academy, Almería, Spain). The exclusion criteria were any eye disease beyond cataracts that might affect visual performance, such as corneal ectasias, amblyopia, retinal diseases, dry eye, active or recurrent anterior segment pathology, systemic or ocular medications that may affect vision, and previous intraoperative or corneal surgery. Eyes with spherical refraction outside the range from −8 to +5 D were also excluded due to the limitations of the HD Analyzer for automatically correcting the sphere outside this range. All procedures were conducted in the same room in low-mesopic environmental light conditions for one eye and randomized using an iOS App (RandomIZE: Randomization Tool). For eyes with mesopic pupil diameters below 4 mm, one drop of tropicamide (1%) and one drop of phenylephrine (10%) were instilled to dilate the patient’s pupil, thereby enabling all the devices to report results uniformly at 4 mm. In these cases, the measurements of CDVA and PP were conducted before instilling the eye drops to avoid their influence on the results.[9]

Grading cataract

Objective scatter index

HD Analyzer is a double-pass system for evaluating the optical quality using near-infrared light (780 nm), which passes through the optical surfaces of the eye and is reflected off the retina. The system has been described in depth in a previous study,[10] and the procedure for its use in the evaluation with MIOLs has also been discussed.[11] Measurements were taken with external correction of any astigmatism above 0.50 D, using a cylindrical trial lens located in the front holder of the device and allowing the system to automatically correct the spherical refractive error through its internal Badal system.

Dysfunctional lens index

iTrace is a ray-tracing aberrometer combined with a placido disk corneal topographer, which allows for the computation of the internal aberrations of the eye by the difference between total and anterior corneal wavefronts.[8] The dysfunctional lens index (DLI) is employed for grading cataracts. Although the algorithm has not been published, the manufacturer claims that this system uses several factors taken during the measurement beyond internal aberrations. The iTrace was conducted in the conditions previously described, setting up the scan to a diameter of 4 mm and previously performing three corneal topographies to ensure reliability of the capture.

Pentacam Nucleus Staging

Pentacam Wave was used to measure densitometry of the crystalline lens, which represents the amount of backscattered light reported in gray value units.[12] Specifically, the Pentacam Nucleus Staging (PNS) was used as an end point to grade the cataract from 1 to 5.[13] Although the algorithm with the exact cutoff values for the ordinal scale has not been published by the manufacturer, this calculation method is based on the gray value units from a cylindrical, three-dimensional template over the nucleus location, which is automatically recognized by the device.

PP evaluation

Light distortion index

LDA features a central light-emitting diode (LED), which is encircled by smaller LEDs distributed radially along 24 semi-meridians, totaling 240 smaller LEDs [Fig. 1]. This system has been comprehensively described in a previous study,[14] as has the procedure for evaluation with MIOLs.[11] LDA was positioned 2 m away from the patient. The maximum possible field extension was measured with the patient wearing his/her best distance spectacle correction set to infinity, further adjusted by adding +0.50 D for vergence distance correction. The field was constrained only by an angular limit of 4.6°, imposed by the device itself.

Figure 1.

(a) Traffic light simulator using red and yellow LED lights, (b) Light Distortion Analyzer, (c) “DysphotopsiApp” application screen for grading the characteristics of the LED light perception from the traffic light simulator, (d) “DysphotopsiApp” application screen for grading the bothersome caused by the LED lights of the traffic light simulator

PP simulator

The DysphotopsiApp was employed concurrently, requiring the patient to evaluate four characteristics of PP generated by two vertically placed LEDs (red and yellow), simulating traffic lights: starburst/glare size, starburst/glare ratio, light intensity on the left side, and both the intensity and width of the halo ring on the right side [Fig. 1]. Positioned 2 m away from LEDs, the patient was guided to manipulate the sliders to match his/her perception of the observed LED, thereby providing a physical point of reference. Subsequently, two questions appeared consecutively on the screen: (1) How SIMILAR is the simulation to your actual perception of this visual phenomenon? and (2) How BOTHERSOME would you find this phenomenon if it occurred around a bright spotlight? For each question, five possible responses were available: “Not at all…,” “Slightly…,” “Moderately…,” “Very…,” and “Extremely…,” followed by “similar” or “bothersome” for the first and second questions, respectively.

Statistical analysis

One eye per patient was included in the analysis. Mean ± standard deviation (SD) was employed to report centrality and dispersion for normally distributed variables, while median (interquartile range [IQR]) was used for non-normally distributed variables. Distribution assessment was conducted using histograms and the Kolmogorov–Smirnov test. Correlations were evaluated with the Spearman’s rho test, and differences between levels of bothersomeness for the analyzed variables were examined using the Kruskal–Wallis H test. Pairwise comparisons were performed with Bonferroni correction for multiple comparisons.

A logistic regression analysis was performed to calculate the probability of transitioning from “Slightly” to “Moderate” bothersomeness, considering LDI as an independent variable. Bothersomeness responses were dichotomized into ≤”Slightly bothersome” (0) and >“Slightly bothersome” (1). The area under the receiver operating characteristic curve (ROC-AUC) was calculated, and the Youden index was used for determining the diagnostic cutoff value for LDI.[15] Finally, a simple linear regression was conducted to estimate LDI from the grading level of each device that best fitted the previously calculated cut-off. Data analysis was carried out using the IBM Statistical Package for the Social Sciences (SPSS) for Windows statistical software (version 24.0; SPSS, Inc., Chicago, IL, USA).

Results

A total of 52 men and 30 women, with a mean age of 60 ± 9.8 years, were included in the analysis. The median refractive error was 0.75 (2.88) D for the sphere, −0.75 (1.00) D for the cylinder, and 0.50 (3.00) D for the spherical equivalent. Refractive error ranged from −7.50 D to +5.00 D for the sphere, from 0 D to −4.50 D for the cylinder, and from −7.50 D to 4.75 D for the spherical equivalent. Table 1 presents the descriptive results for the measurements of the optical media variables and PP evaluation.

Table 1.

Descriptive statistics of the variables used to grade cataracts and measure the photic phenomena

Device/variable	Mean±SD	Median (IQR)
HD Analyzer
OSI	2.73±3.60	1.4 (1.83)
iTrace
DLI	7.37±2.66	7.98 (4.39)
Pentacam Wave
PNS	1.86±1.31	1 (1.25)
Light distortion analyzer
Light distortion index (%)	15.03±13.86	10.39 (11.29)
Best fit circle radius (mm)	29.49±12.02	26.35±13.55
Irregularity (mm)	0.44±0.43	0.36 (0.43)
DysphotopsiApp
Size (°)	0.28±0.10	0.27 (0.11)
Glare/starburst ratio	0.13±0.30	0 (0)
Glare/starburst intensity	0.89±0.21	1 (0)
Halo width (°)	0.14±0.06	0.13 (0.06)
Halo intensity	0.88±0.22	1 (0.2)

DLI=dysfunctional lens index, IQR=interquartile range, OSI=ocular scatter index, PNS=Pentacam Nucleus Staging, SD=standard deviation

A moderate correlation was found between objective scatter index (OSI) and DLI (rho = −0.43, P < 0.0005). PNS was also correlated with OSI (rho = 0.501, P < 0.0005) and DLI (rho = −0.405, P < 0.0005). The three grading methods were significantly correlated with CDVA in the following order: OSI (rho = 0.66, P < 0.0005), DLI (rho = −0.57, P < 0.0005), and PNS (rho = 0.55, P < 0.0005). Table 2 shows the correlations between the PP evaluation variables and each of the cataract grading methods. LDI was generally correlated with all grading methods, especially with OSI, which also showed significant correlation with the irregularity variable. However, no correlations were found between the variables of DysphotopsiApp and the cataract grading methods.

Table 2.

Correlations between methods for grading cataract and photic phenomena evaluation

	Ocular scatter index	Dysfunctional lens index	Pentacam Nucleus Staging
Light distortion analyzer
Light distortion index (%)	0.67, <0.0005*	−0.29, 0.007*	0.48, <0.0005*
Best fit circle radius (mm)	0.67, <0.0005*	−0.29, 0.009*	0.48, <0.0005*
Irregularity (mm)	0.31, 0.005*	−0.16, 0.15	0.18, 0.12
DysphotopsiaApp
Size (°)	0.09, 0.41	−0.04, −0.76	0.01, 0.9
Glare/starburst ratio	0.19, 0.08	−0.19, 0.08	0.15, 0.19
Intensity glare/starburst	0.02, 0.86	−0.12, 0.30	0.15, 0.19
Halo width (°)	0.03, 0.76	−0.11, 0.34	0.02, 0.86
Halo intensity	−0.05, 0.67	−0.05, 0.64	0.01, 0.89

Correlations assessed with Spearman rho. *P<0.05

The percentage of subjects who rated the level of bothersomeness associated with PP produced by LEDs was as follows: “Not at all” by 18/22%, “Slightly” by 41/50%, “Moderately” by 15/18.3%, “Very” by 8/9.7%, and “Extremely” bothersome by 0%. The similarity of the real perception of LEDs and the one that was configured by the patient as measured through the DysphotopsiApp was rated as “Somehow Similar” by 8/9.8%, “Moderately Similar” by 59/72%, and “Very Similar” by 15/18.3%. The strongest correlation between PD and PP measurement was observed with LDI (rho = 0.34, P = 0.002), followed by the ratio between starburst/glare (rho = 0.35, P = 0.001) measured via the DysphotopsiApp. The size of the phenomena was nonsignificant (rho = 0.21, P = 0.07). No significant correlations were found for the remaining variables of PP measurements listed in Table 2.

Fig. 2 presents the box plot for LDI and various cataract grading methods compared to the responses to the question about bothersomeness to PP. A general median increase was noted across all methods up to the “Moderately” bothersome level, but this was only significant for LDI [as depicted in Fig. 1]. Variability in distribution increased notably at the “Moderately” level and decreased again at the “Very” level. The highest dispersion of results across all levels was found with DLI [Fig. 2c].

Figure 2.

Box plot for the distribution of variables along the possible answers of bothersome to photic phenomena directly measured with the light distortion index (a) or the optical quality variables from which this parameter could be estimated, such as (b) ocular scatter index, (c) dysfunctional lens index, and (d) Pentacam Nucleus Staging

LDI demonstrated a discriminative capacity to detect subjects reporting “Moderately” or higher bothersomeness (ROC-AUC = 0.64, P = 0.05, CI 95%: 0.49 to 0.79). The Youden index was maximized at a value of 0.31, where the cutoff of LDI was 15.20%, achieving a sensitivity/specificity of 0.75/0.57. Fig. 3 illustrates the relationship between the three grading methods and LDI. The best prediction for LDI was achieved with OSI (R² = 0.5, P < 0.0005), followed by DLI (R² = 0.09, P = 0.006) and PNS (R² = 0.06, P = 0.04). The values predicting a cutoff of LDI ≥15.2% were OSI ≥2.8, DLI ≤7.6, and PNS ≥2 [Fig. 3].

Figure 3.

Relationship assessment between the light distortion index and the methods of grading the cataract: (a) ocular scatter index, (b) dysfunctional lens index, and (c) Pentacam Nucleus Staging

Discussion

Despite advances in the evaluation of postoperative PP and PD in clinical studies involving MIOLs,[3,16] there remains a gap in knowledge about identifying patients in the preoperative period who are likely to be bothered in the postoperative period. In this study, we assessed PP and PD in patients with varying degrees of crystalline lens sclerosis and cataracts. The aim was to determine the cutoff point for PP that may result in a moderate or worse PD and to evaluate the relationship between PP and several objective methods for grading cataracts, such as OSI, PNS, and DLI.

Two methods were used in our study to assess PP.[3] The first method was LDA, which has been previously employed in studies involving MIOLs,[17,18,19] and the second was a prototype developed by our research group (DysphotopsiApp). This prototype consists of a simulator that grades PP while the patient looks at two LED lights simulating traffic lights. The aim of this new prototype was to correlate the patient’s responses with his/her direct perception of PP, combining the most commonly reported phenomena: glare, starburst, and halo. Although the majority of patients (72%) found the simulated image moderately similar to their real-life perception, the main limitation was the intensity of LEDs, which exceeded the brightness of the display.

Despite efforts to identify the best simulator, and although some variables from DysphotopsiApp correlated with questions assessing PD, the most effective method for measuring PP was LDA. This method not only showed the strongest correlation with the question for assessing PD, but also consistently correlated with cataract grading methods, particularly with OSI, a correlation not observed with DysphotopsiApp. The cutoff value that best predicted a shift from slightly to moderate bothersomeness was 15.20% for LDI. This value was lower than those reported monocularly by Brito et al.[18] for monofocal IOLs (23.94%), AT Lisa Tri 839M (46.97%), and AT Lisa Bi 909MP (46.97%). However, the results of Brito et al.[18] for AT Lisa Tri 839M (46.97%) differed from those reported in the long term by our research group (18.82%).[5] Other authors have reported monocular LDI for other MIOLs. For instance, Panoptix has an LDI of 36.8%,[17] while M-Plus lenses with powers of +1.50 D and +3.00 D have LDIs of 20.6% and 26.5%, respectively.[19] Similarly, our own research group reported an LDI of 15.21% for Liberty lenses.[11] Unfortunately, previous publications have not specified in their methods section whether measurements were taken with the best distance correction, including a +0.50 D lens to shift the best focus from infinity to a 2-m location.[17,18,19] This omission could partly explain why previous research yielded considerably higher LDI values compared to our experience,[5] as defocus could elevate LDI.[20]

A question was used to assess how bothersome was the perceived PP for the patient, and 28.1% of the subjects reported “moderately” or greater. Our results agree with Ison et al.,[21] who reported 25% of patients finding quite or very bothersome by glare, a value that decreased to 15% after surgery with MIOL implantation. Interestingly, the 9.8% of subjects who reported “very bothersome” by PP in our study had a cataract degree generally equal to or lower than the patients reporting moderately bothersome, especially for OSI and PNS, with median values similar to the patients reporting “not at all” bothersome. These findings support the evidence that other factors such as personality[6,7] could have an influence on PD as the optical quality could not explain that PP is high enough to produce these symptoms.

Our findings could have a very interesting translation to clinical practice by means of combining a question with an objective measure in preoperative screening. First, the following question should be asked: “During the past 7 days, how bothersome have you found it when you experienced visual artifacts around lights at night, such as streetlights, car headlights, or car brake lights? Not at all, Slightly, Moderately, Very, or Extremely bothersome.”[22] Second, one of the following measurements should be taken: LDI, OSI, DLI, or PNS. Finally, patients reporting “Moderately,” “Very,” or “Extremely” bothersome in the preoperative screening and having LDI <15.2%, OSI <2.8, DLI >7.6, or PNS <2 may experience greater bothersomeness after MIOL implantation because there is no optical reason in the preoperative period that could justify this response.

Our research has some limitations due to which the results have to be interpreted with caution. We assessed PD at the monocular level with the best correction. This was decided because cataract grading methods were measured at the monocular level and the question of bothersome was asked when the patient was monocularly looking at the LED light and in reference to that light. These measurement conditions do not match the real-world patient’s experience since patients rarely use spectacle correction after MIOL implantation and the lights are seen in binocular conditions. However, the purpose of this study was not to assess patient-reported outcomes, but to define a method that allows finding an optical explanation to PD. In our opinion, patient-reported outcomes for assessing PD should consider real-world experience data and to seek correlations at the postoperative period, LDI should be measured binocularly without the best correction. This measurement would better represent real-world conditions after surgery with MIOL implantation.

It is also important to note that even though LDI predicted the pass from slightly to moderately bothersome, nearly half of the sample were false positives (0.57), meaning that some patients can have an LDI higher than 15.20% without being bothered by the perceived PP. For this reason, we consider that none of these measurements should be used alone to predict bothersomeness, understanding that the method should combine a bothersome question and a measurement of PP either directly as LDI or indirectly as OSI, DLI, or PNS. Clinically useful data would result when a high bothersome response is obtained in the preoperative screening and one of the previous measurements is below the cutoff values proposed above. On the other hand, it should be considered that the variability in the measurement is a relevant factor since some metrics, such as OSI, can vary by up to 0.6 due to noise factors such as tear film stability.[23]

Conclusion

This is the first study to evaluate LDI in the preoperative period in patients with varying degrees of crystalline lens sclerosis or cataracts. Patients who were very bothered by PP had a median LDI similar to those not bothered at all. This suggests that some patients are already bothered during the preoperative period, despite having good optical quality that does not justify this response. Combining LDI measurement or estimating it with cataract grading systems, along with a single question about light bothersomeness, may be a useful tool for identifying patients who complain about PD in the postoperative period. Future research should include patients assessed preoperatively using this methodology, from diverse sites and geographic locations, and should collect postoperative patient-reported outcomes to confirm the reliability of this new approach for preventing complaints of dysphotopsia.

Financial support and sponsorship

Nil.

Conflicts of interest

The DysphotopsiApp was developed by the Research and Evidence Department of Qvision, headed by Dr. Fernández.