How Does Cataract Surgery Rate Affect Angle-closure Prevalence

Supplemental Digital Content is available in the text. Précis: A data simulation study suggests that prevalence of occludable angle will decrease when the cataract surgical rate increases in particularly when the surgery focuses on 70+ years old. Purpose: The purpose of this study was to estimate the effects of cataract surgical rates (CSR) on the prevalence of primary angle-closure glaucoma in the Chinese population. Methods: Participants aged 50 years and older from the Liwan Eye Study were included as the study sample. Occludable angle (OA) as a surrogate of primary angle-closure glaucoma was evaluated using static gonioscopy and anterior chamber depth was measured before dilation using A-mode ultrasound. Random sampling was used to generate 50 cohorts with a sample size of 200 for each predefined CSR at 2000, 4000, 6000, 8000, 10,000, 12,000, according to the multinomial distribution. The mean anterior chamber depth and OA rates of each cohort were calculated. Logistic function models of nonlinear least-squares estimation were used to predict the prevalence of OA. Results: Data of the right eye from 1280 participants were included. The prevalence of cataract surgery and OA was 2.27% and 11.3%, respectively. The projected prevalence of OA in the cohorts with CSR of 2000, 4000, 6000, 8000, 10,000 and 12,000 was 11.4% [95% confidence interval (CI), 10.8%-12.0%], 11.2% (95% CI, 10.6%-11.9%), 10.9% (95% CI, 10.3%-11.6%), 11.4% (95% CI, 10.8%-12.1%), 10.8% (95% CI, 10.2-11.4%), and 10.1% (95% CI, 9.46-10.7%), respectively. The OA rates decreased remarkably as CSR increased for those aged 70 years and older. Conclusions: Our study indicated that with CSR increased, the OA prevalence could decrease remarkably especially in the older population. It is advisable to perform cataract surgery at an appropriate time for patients in their late 60s to 70s with significant cataracts.

G laucoma is the leading cause of irreversible blindness worldwide, and the number of people diagnosed with glaucoma is estimated to reach 76 million by 2020 and 111.8 million by 2040. 1 Although primary open-angle glaucoma is the most common subtype of glaucoma, primary angleclosure glaucoma (PACG), which accounts for 25% of all glaucoma globally, is more likely to result in irreversible visual impairment if not properly treated. 2 The prevalence of primary angle-closure (PAC) varies by race. Previous studies suggest that 76.7% of individuals with PACG reside in Asia, among which half of them were Chinese. 3,4 And according to previous reports, 3.1 million Chinese people are blind in at least one eye from PACG. 1,2 Many studies have indicated that eyes with PAC share similar ocular characteristics. [5][6][7][8][9][10][11] Shallow anterior chamber depth (ACD) and lens factors are regarded as cardinal risk factors for angle closure. 6 It is believed that shallow ACD is due to increased lens thickness (LT) and forward positioning of the lens which can cause potential pupillary block and angle crowding. [12][13][14] Considering the lens plays an important role in the development of PAC, lens extraction with intraocular lens (IOL) implantation, has been considered an alternative treatment for PACG with many studies reporting positive results. [15][16][17][18][19] Surgery is the most effective treatment for visually significant cataract, and the number of people requiring cataract surgery worldwide is expected to reach 70.5 million by 2020. 20 Globally, cataract surgery rates (CSR) have increased over recent decades. 21,22 In China, there are ∼30 million people with visually significant cataract and this number will continue to rise in line with the aging population. 23 Significant efforts have been made to increase the CSR in China, with rates increasing from 480 cases per million in 2004 to 2205 in 2017. 24,25 With the increasing rates of cataract surgery, one would expect a commensurate reduction in the prevalence of PACG. Studies from Scotland and Taiwan supported the hypothesis that increased CSR would reduce the occurrence of PACG. [26][27][28] However, no studies have been performed to evaluate the potential impact of CSR on PACG, especially in a Chinese population where there is a low CSR but a large number of cataract and glaucoma patients. Therefore, we investigated the effect of increased CSR on the prevalence of occludable angles (OA) using a simulation study based on data from the Liwan Eye Study. Details of field examinations have been described in detail elsewhere. 29 In brief, 1864 participants aged 50 years and older were recruited from the Liwan District, Guangzhou, China using a cluster random sampling technique. Among these, 1405 took part in the study, and all examinations were carried out between September 2003 to February 2004.

METHODS
Gonioscopy was performed using a Goldmann-type, onemirror lens (Haag Streit Diagnostics, Bern, Switzerland) to observe the structures of the anterior chamber angle before other ocular examinations were performed. Gonioscopy was performed using a narrow beam 1 mm in length by an experienced ophthalmologist (M.H.) at ×25 magnification. Gonioscopy examination was completed on all participants. Dynamic gonioscopy examination of the 4 quadrants was performed after static gonioscopy.
OA, which is synonymous with narrow-angle and suspect PAC, was defined as anterior angles with less of 90 degrees of the posterior trabecular meshwork visible during static gonioscopy. 30,31 LT and ACD were measured with A-scan ultrasonography (Echoscan US1800; Nidek Corp.). LT was defined as the distance from the vertex positions of the anterior to posterior surfaces of the crystalline lens or IOL. ACD was defined as the distance from the anterior corneal surface to the anterior lens capsule or anterior surface of IOL. We found that increase in CSR resulted in a decrease of the mean LT and an increase in the mean ACD which may reduce the prevalence of OA. Ten individual measurements for each parameter were taken and outlier values were deleted. The measurements were repeated when the SD was > 0.13 mm. All ocular parameters are presented for the right eye only.
In the current study, we recruited 1280 subjects aged older than 50 years, among which 390 were 70 years and older, from a total sample population of 12,563 of all ages living in Liwan District, Guangzhou, China. All these 1280 elderly participants were assumed as the main target population for cataract surgery. The number of subjects who had received cataract surgery at baseline was 29, with 62.1% (18/29) of them aged older than 70 years. Therefore, the cataract surgery coverage was 2.27% (29/1280) in the whole study population and 4.62% (18/390) in subjects older than 70 years. The simulated samples under different assumptions of CSR were constructed to estimate the effect of CSR on OA prevalence.
To achieve a CSR of 2000 among the total population (of all ages), 25 (2000/1,000,000×12,563) cataract operations would be required in 1 year in the study population (N = 12,563) according to the definition of CSR (the number of cataract operations performed in 1 year/million population). On the basis of this assumption, the cumulative cataract surgery coverage in the study sample would be about 4.22% [(29+25)/1280] among the recruited 1280 subjects. On the basis of the cataract surgery coverage of 4.22%, 50 simulated samples with an equal size of 200 each sample were constructed using bootstrap resampling methods. 32,33 The same estimation method was applied under different assumptions of CSR ranging from 4000 to 12,000 with an interval of 2000. The simulation was performed using R software (The R Foundation for Statistical Computing, Vienna, Austria).
Normality was checked using the skewness test and analysis of variance was used to determine the variance among different groups. The Mann-Whitney U test was used to analyze the variability of age-specific cataract surgery coverage and age-specific prevalence of OA. A logistic function model of nonlinear least-squares estimation was fitted to investigate the relationship between OA prevalence and ACD and to estimate the age-specific prevalence of OA with different CSRs. Particularly, OA prevalence in subjects older than 70 years with different presumed proportions of cataract surgeries in simulated CSR cohorts was also evaluated. All data analyses were performed using Stata 14.0 (StataCorp, College Station, TX).

RESULTS
Of the 1405 participants who successfully completed examination procedures, 125 were excluded. This consisted of participants with missing data for gonioscopy (4), or ocular biometry (121). This left 1280 participants aged 50 to 91 years with complete examination data for analysis.
The mean age of included participants was 65.0 ± 9.66 years and males accounted for 43.1% (551/1280). There was no significant difference in mean age between males and females (t test, P = 0.874). The age-specific proportion of participants with cataract surgery, measurements of ACD, LT, and prevalence of OA is shown in Table 1. As a baseline, the total proportion of participants who had undergone cataract surgery in this population was 2.27%. The Kruskal-Wallis test suggested that the proportion of participants with surgery shows a positive trend with increasing age (P < 0.001). Participants aged 70 to 74 years had the largest proportion of cataract surgery. The mean age-specific ACD showed a decreasing trend with increasing age, however, in the 70 to 74 age group, the mean ACD was deeper than other age groups.
For the baseline population, the overall prevalence of OA was 11.3% (95% confidence interval, 9.57%-13.1%), and an increasing trend was observed with age. The proportion of subjects with cataract surgery, mean ACD, and mean LT in people with OA and those without OA are summarized in Supplemental Table 2 (Supplemental Digital Content 2, http://links.lww.com/IJG/A478). In total, participants with OA had lower CSR, shallower ACD, and greater LT. The characteristics of different CSR cohorts acquired by sampling are summarized in Table 2. ACD tended to be deeper as CSR increased and OA prevalence tended to be lower as CSR increased. The OA prevalence decreased from 11.4% to 10.1% with CSR increasing from 2000 to 12,000.
The relationship between different CSR and mean ACD is shown in Figure 1. Mean ACD increased with CSR increased, particularly when the CSR increased from 2000 to 6000 and from 8000 to 12,000. The mean ACD in the simulated CSR cohorts were 2.73, 2.75, 2.77, 2.78, 2.80, and 2.82 mm, respectively.
The changing trends of age-specific ACD in different CSR cohorts is illustrated in Figure 2. The mean ACD showed a decreasing trend for those aged of 50 to 70 years but showed an increasing trend for those aged over 70 years.
The relationship between ACD and the prevalence of OA was shown in Figure 3. ACD was divided into 6 groups from 2.00 to 3.50 mm by an interval of 0.30 mm. Figure 3 shows that OA proportions decreased rapidly from ∼35% to 3% when ACD increased from 2.00 to 2.90 mm. OA prevalence was reduced when ACD was > 2.90 mm.
The predictive age-specific rates of OA in different CSR cohorts are illustrated in Figure 4. In each CSR cohort, the rates of OA increased with age in an approximately linear fashion. The rates of OA decreased slightly in each age group with increasing CSR in those less than 65 years. In the population over 75 years, the rate of OA decreased significantly from ∼20.0% to 15.0% as CSR increased from 2000 to 12,000.
Particularly, as shown in Supplemental

DISCUSSION
A better understanding of the effect that CSR has on OA prevalence is of great importance for the management of cataract and health policymakers. Our study attempted to evaluate the changes of the age-specific prevalence of OA with different predefined CSR by using a prediction model from a population-based study of elderly Chinese. The results showed an inverse association between CSR and the prevalence of OA. The prevalence of OA decreased from 11.4% to 10.1% with CSR increased from 2000 to 12,000. This decreasing trend was more remarkable among those aged over 70 years.
Changes in ACD were an important factor by which OA prevalence was affected in different CSR cohorts. A shallower ACD has been identified as the key risk factors for PACG and was traditionally recognized as an effective way of detecting PACG. 6,7,[9][10][11]34 It is assumed that age-related thickening of the lens plays a fundamental role in the shallowing of the ACD. Many studies suggest that cataract extraction can widen the ACD and the anterior chamber angle, 12,15,35 especially for eyes with PACG. 12,16,19,[35][36][37][38] A comparative, nonrandomized, interventional study from Japan showed that after cataract surgery, the chamber angle and ACD in eyes with PACG increased substantially, becoming similar to primary open-angle glaucoma patients or normal eyes over a 12-month follow-up period. 12 Many studies have pointed out that long-term intraocular pressure control in PACG was due to the anatomic changes caused by lens extraction. 18,39,40 The most significant decreasing trends were observed among those aged over 75 years, with the rate of OA declining from ∼20.0% to 15.0%. The possible reason for this may be that cataract surgery coverage was higher among the older population and mean ACD was deeper than groups with lower cataract surgery coverage such as the 60 to 64 and 65 to 69 age groups. 41 Considering that OA is a progressive stage of PACG which will cause blindness without timely treatment, earlier cataract surgery is recommended in glaucoma patients to preserve visual acuity. 16,42 Findings from this investigation suggest that cataract surgery should be performed at a younger age to reduce the incidence of PACG and to improve cataract related vision impairment. 26 Our study indicated that the rate of OA decreased with increasing CSR, especially among people over 70 years of age. A similar inverse relationship between CSR and PAC prevalence has been reported in different ethnicities in previous studies. [26][27][28]43 One prediction study using data from the Meiktila Eye Study 43 estimated that 38.5% of the burden of angle closure could be reduced in the adult population if cataract surgery were performed in 8.8% of eyes with   26 We believe that increasing CSR will change the proportion of PAC disease and reduce the rate of PACG, which is the leading cause of glaucoma-related blindness worldwide. 3 It is reasonable to suggest that Governments promote strategies that focus on increasing CSR, especially in countries such as China who have the largest number of cataract and PACG patients, but low CSR. 23,41 The current study has several strengths, including a large representative sample that was drawn from a population-based study. Given that it was not possible to obtain different predefined CSR cohorts in the same population, we performed repeated sampling according to the multinomial distribution from the eligible database and acquired the necessary data that would simulate cohorts with different CSR. The limitation of this is that results may not fully predict real-world situations. Another potential limitation of this study is that the predictive prevalence of OA among different CSR cohorts may be population-specific, the Liwan Eye Study may not be representative of other populations worldwide, thus, the conclusion may not be generalizable. Further research with larger sample sizes and better designs is required.
To the best of our knowledge, this is the first simulation study on the age-specific prevalence of OA by different CSR using a population-based sample of Chinese adults. Our study showed an inverse relationship between CSR and prevalence of OA, especially among those aged 70 years and older. Besides, the higher proportion of surgeries presumed to perform in subjects older than 70 years, the more significant decreasing trend of OA prevalence. This is a very FIGURE 3. Simulated curve of the association between the prevalence of occludable angle (%) and mean anterior chamber depth (ACD, mm). Figure 3 can be viewed in color online at www. glaucomajournal.com.