CATARACT SURGERY IN PATIENTS WITH PSEUDOEXFOLIATION GLAUCOMA AND IN NONGLAUCOMATOUS PATIENTS

Background/Aim. The cataract surgery in eyes with and without glaucoma results with the sustained intraocular pressure (IOP) reduction but it is still unknown which glaucomatous patients will achieve clinically significant reduction. The preoperative IOP and some ocular biometric parameters have been shown as potential predictors of the postoperative IOP reduction and the aim of our prospective intervention study was to evaluate that relationship in medically controlled patients with the pseudoexfoliation glaucoma (PXG) and in the nonglaucomatous patients. Methods. Thirty-one PXG patients (31 eyes) and 31 (31 eyes) nonglaucomatous patients, all with clinically significant cataract, were enrolled. The preoperative IOP, anterior chamber depth (ACD), axial length (AL), lens thickness (LT), lens position (LP) [LP = ACD + 0.5 LT], relative lens position (RLP) [RLP = LP / AL] and the pressure-to-depth ratio (PD ratio) [PD ratio = preoperative IOP/preoperative ACD] were evaluated as potential predictors of the IOP change in the 6th postoperative month. Conclusions: The cataract surgery leads to the IOP reduction both in the PXG and in a nonglaucomatous eye and this change predictors are widely available and simply calculable parameters and can be potentially used in managing glaucoma.

Ključne riječi: pseudoeksfolijativni glaukom; operacija katarakte; biometrijski parametri; postoperativna redukcija intraokularnog pritiska Introduction: Glaucoma is the leading cause of irreversible blindness and the only treatment is lowering the intraocular pressure (IOP) on a level on which disease doesn't progress 1 . The pseudoexfoliation glaucoma (PXG) is one of the most complicated forms of glaucoma for treatment because of the high IOP at onset, poor response to medical therapy and faster progression. The PXG is characterized by the pathological production and accumulation of an abnormal extracellular fibrillar material mainly visible on the anterior lens capsule, the pupillary margin, corneal endothelium, lens zonules, trabecular meshwork and often correlates with an increased incidence of cataract formation.Cataract is usually a hard nuclear and a loss of zonules support and poor pupillary dilation makes the cataract surgery challenging and with potential complications like vitreous loss, subluxation or luxation of the lens.
Studies have shown that the cataract surgery leads to an IOP reduction in glaucomatous and nonglaucomatous eyes, and the IOP-reduction effects vary depending on the type of glaucoma and monitoring period 2 . However, a clinically significant IOP reduction does not occur in every patient 3,4 . More recent studies are trying to identify factors that can indicate which patients will achieve a clinically significant IOP reduction after a cataract surgery.
The preoperative IOP has been found to be a significant predictor of the IOP reduction after a cataract surgery. Patients with higher levels of preoperative IOP obtain greater postoperative IOP reduction 2, 3, 5 . Also, some ocular biometric parameters such as anterior chamber depth (ACD), axial length (AL), lens thickness (LT), lens position (LP) [defined as: ACD + 0.5 LT], relative lens position (RLP) [defined as: LP/AL] and the pressure-todepth ratio (PD ratio) [defined as: preoperative IOP/preoperative ACD] has been recognised as potential predictors of postoperative IOP reduction 2, 6, 7, 8, 9, 10., 11 .
To our knowledge, there have been no prospective studies which evaluated these clinical variables in the PXG patients and the aim of this study was to determine if the preoperative IOP, ACD, AL, LT, LP, RLP and PD ratios are related with the postoperative IOP changes in the PXG patients and to compare these findings with those of nonglaucomatous patients.

Materials and Methods
Thirty-one PXG patients (31 eyes) and thirty-one ( According to the following formulas the calculation was as follows: We compared the absolute IOP change and the IOP change percent on the 6th postoperative month as the main outcomes for the predictors of interest. Continuous variables were reported as the mean ± standard deviation (SD). The Mann-Whitney U test was used to assess differences between the groups for ordinal or continuous variables, and the chi-square test (χ 2 ) for categorical variables.
We used the linear mixed-effects regression analysis in order to determine the correlation between the outcome variables and the preoperative factors, including the preoperative IOP, the age, gender and biometric parameters (ACD, AL, LT, LP, RLP, PD ratio). The 8 multivariate linear mixed-effects regression models were created in order to to adjust to potential confounders, including sex, age and the preoperative IOP.
In all models assessing the PD ratio and the IOP change, the preoperative IOP was not included because it is a part of the PD ratio calculation.
The regression coefficients (B), the coefficients of determination (r 2 ) and the statistical significance (p value) were reported. p values ≤0.05 were considered significant. All the statistical analysis was performed using the SPSS software V.21 (SPSS, Inc., Chicago, IL, USA).

Results
The observed group of thirty-one patients with the PXG, and thirty-one patients with the senile cataract as Controls group were included in this prospective study. The demographic characteristics, preoperative biometric measurements, preoperative IOP, postoperative IOP (measurements in the 1st, 3rd and 6th month) and the postoperative IOP change of each group are shown in Table 1 (Table 1). Table 1 The linear mixed-effects regression models were used to shows the association between the absolute IOP change in the 6th month and sex, age, preoperative IOP and ocular biometric parameters for both group (Table 2). Table 2 In the univariate mixed-effects models, for both groups, the Pre-op IOP was found to be a significant predictor of the absolute IOP reduction. In the PXG group, in the univariate and multivariate models, the AL and PD ratios were associated with the significant absolute IOP change 6 months after the cataract surgery, and the RLP in multivariate model. In Controls, parameter PD ratio was the significant predictor of the absolute IOP change both in the univariate and the multivariate analysis ( Table 2). Predictability of the potential predictors was shown throught the coefficient of determination and according to the r 2 value, in the PXG group in the univariate model, the Pre-op IOP (r 2 = 47.3%) was the best predictor of the absolute IOP change , followed by the PD ratio and the AL (r 2 = 18.3% and 12.5%, respectively). In multivariate model in PXG, among the significant predictors, the AL, RLP and PD ratio was the best predictor of the absolute IOP change (r 2 = 57.7%, 51.9% and 14.7%, respectively). In Controls, the best predictor was the Pre-op IOP (r 2 = 22.5%), followed by the PD ratio (r 2 = 16.3% in the univariate and r 2 = 13.7% in the multivariate analysis) ( Table 2). Table 3 shows the association between the percentual IOP change in the 6th postoperative month and sex, age, preoperative IOP and ocular biometric parameters for PXG and Controls, using the linear mixed-effects regression models. The coefficient of determination was also explored. Table 3 In the univariate analysis, in PXG the significant predictors of the percentual IOP change, were the Pre-op IOP and in the univariate and multivariate analysis, the AL, RLP and PD ratio. Among the significant predictors in the univariate model, the best predictability of the percentual IOP change had the Pre-op IOP, followed by the AL, PD ratio and RLP (r 2 = 26.7%, 14.7%, 12.6% and 10.5%, respectively). In the multivariate model, the order of the best predictability among the parameters was as follows: AL, RLP, PD ratio (r 2 = 37.9%, 35.2% and 8.8%, respectively).
Neither the univariate nor the multivariate analysis identified statistically significant predictors of the percentual IOP change in the 6th postoperative month in the Control group (Table 3).

Discussion
The majority of studies, mainly the retrospective ones, examining effect of cataract surgery on the IOP reduction, was conducted in patients with POAG, PACG, and nonglaucomatous patients, and only a few in patients with PXG 2, 14 . A possible explanation for this is the lower PXG incidence compared to other glaucoma forms, as well as its wide variations in incidence and prevalence among different countries globally and in different geographical areas within the same country 15 15 . This discrepancy is undoubtedly influenced by the sample size, but certainly also by the geographical areas with the higher PXG incidence and their distance from the medical centers where cataract surgery is performed, which is the case here.
Our research was inspired by the need to advance our day-to-day clinical practice in the PXG patients treatment. Aware of the diurnal IOP fluctuations 16 , and in order to obtain the most recent values for IOP and PD ratio parameters, the IOP daily curve was performed preoperatively and at each measurement time point postoperatively.
Our study results show that cataract surgery led to the IOP decrease in both PXG patients treated with medicaments and in nonglaucomatous patients. In patients with PXG, the reduction was recorded as early as in the first postoperative month and showed a tendency of further reduction in the third and sixth month. In nonglaucomatous patients, the IOP decrease was the biggest in the third month, and the effect of the decrease began to weaken in the sixth month (Table 1; the absolute and the percent change in IOP).
In the 6th postoperative month in the PXG group, the absolute IOP reduction was -3.23 ± Based on the above, we conclude that our results are consistent with previous studies suggesting an IOP decrease after cataract surgery in the PXG patients and nonglaucomatous patients, as well as the extent of its reduction.
In our study, regression analysis in the univariate model for both groups showed the significant negative correlation of the preoperative IOP (PXG: B= -0.73 ± 0.14; p <0.01; Controls: B= -0.42 ± 0.13; p <0.01) and its absolute postoperative reductions in the 6th month in the sense that the preoperatively higher IOP values are associated with the greater postoperative reductions, which is consistent with the results of other studies 2, 3, 5 . In our sample, this would mean that if the IOP is preoperatively increased by 1 mmHg (relative to the average IOP for the observed group), the absolute postoperative reduction will be greater by the additional 0.73 ± 0.14 mmHg in the PXG and in nonglaucomatous patients by the additional 0.42 ± 0.13 mmHg ( Table 2). For the percent IOP change in the 6th month, this was only the case for the PXG group (B = -2.98 ± 0.90; p <0.01), whereas at the control subjects it had no significance (B = -1.76 ± 0.94; p = 0.072) ( Table 3).
Because the preoperative IOP higher values tend to result in the greater absolute reduction compared to the lower base values, we also examined the relative (percent) IOP change in our work because the percent change may be similar in eyes with different initial IOP measurements.
The preoperative IOP in our study proved to be a significant predictor of both the absolute and the percent postoperative IOP changes in the PXG with predictive ability of r 2 = 47.3% and r 2 = 26.7%, respectively, and in the control group only for the absolute change with predictive ability of r 2 = 22.5% (Table 2 and Table 3). The slightly higher percentage of predictability of the preoperative IOP obtained in our study for the absolute IOP change in the PXG group (r 2 = 47.3% ), may reflect the analysis of the three preoperative IOP measurements in the daily IOP curve test and "improve" the result of a prediction of the postoperative IOP change relative to a single measurement. The statistical phenomenon of "regression to the mean value", which is a consequence of an inadequate number of the basic preoperative IOP measurements, was minimized thanks to a daily curve test conducted preoperatively in our study.
The exact mechanisms of the IOP reduction after cataract surgery are still not clear. Whatever the mechanism, the question is if any other way, except for the preoperative IOP measurement, is possible for determining which patient will benefit from cataract surgery in terms of achieving a clinically significant IOP reduction? This is especially important for glaucoma patients who have low IOP (low-teen) by the medical or the laser therapy, but the disease progression is still present. In such patients, it is not easy to make the decision about a filtering operation known to be frequently accompanied by a range of serious intraoperative or postoperative complications and frequent failure.
In this regard, numerous biometric parameters are observed as possible IOP change predictors after cataract surgery. Recently, papers have been published, where in order to obtain biometric measurements the optical coherence tomography for the anterior eye segment (AS-OCT) has been used, but due to the cost of the equipment such diagnostics are unavailable for most public health institutions, especially in the economically underdeveloped countries. In our study, we analyzed the parameters whose values are easily obtained as part of the patient's preoperative preparation for cataract surgery using optical biometry and ultrasound A-scans available at all centers where cataract surgery is performed. Also, most of the research has so far been done using such equipment, so that our results can be compared with those of other authors.
The results of the biometric measurements obtained from our subjects indicate an average shallower anterior chamber in the PXG group relative to the control group (PXG ACD: 2.90 ± 0.34 mm; Controls ACD: 3.07 ± 0.31 mm; p = 0.066), which is in in line with the results of other studies 17,28 . The probable reason for this is the increased zonular laxity in patients with PXG and lens anteposition. The lens thickness also plays a significant role in this, and was on average higher in the PXG group (PXG LT: 4.65 ± 0.50 mm; Controls LT: 4.34 ± 0.56 mm; p˂0.05). The LP parameter, represented by the sum of ACD and half LT, is expected to be uniform between the groups given that the ACD is higher in the control group. As the AL parameter is also uniform across groups, so is the same case for the RLP parameter represented by the LP and AL relationship. The preoperatively higher IOP and the lower ACD resulted in the significantly higher PD ratio in the PXG group compared to the control group (PXG PD ratio: 5.66 ± 1.15; Controls PD ratio: 4.77 ± 0.82; p˂0.01) ( Table 1).
Examining the correlation of the percent postoperative IOP change in the 6th month and the biometric parameters, in the PXG group, both in univariate and multivariate model, it was significant for the parameters AL, RLP and PD ratio. The parameters AL (r2 = 37.9%) and RLP (r2 = 35.2%) had the highest prediction value for the percent IOP change in multivariate model, while for the same parameters it was low in univariate analysis. The predictability of the PD ratio parameter was low in both univariate and multivariate model (Table 3).
For the control group, no parameter in the analysis model was found to be significant to indicate the percent postoperative IOP change (Table 3). The PD ratio parameter was first introduced by Issa et al. 9 who found that in nonglaucomatous patients the higher PD ratio was followed by the greater postoperative absolute reduction in IOP and indicated it as a strong predictor of this reduction (r 2 = 73.0%). The significant predictability of 34.1% for the PD ratio in nonglaucomatous patients was also established by Dooley et al. 10 Hsu et al. 21 confirmed its significance as a predictor of both the absolute (r 2 = 52.9%) and the percent (r 2 = 39.0%) postoperative IOP reductions in the nonglaucomatous patients, and Coh et al. 8 in patients with POAG.
The RLP is a parameter dependent on the thickness of the natural lens, its anteposition, the depth of the anterior chamber, and the total length of the bulb. Its role as a predictor of IOP change is significant in individuals with a narrow angle, and can be potentially used in individuals with an open angle.