Distribution of ocular biometric parameters and optimal model of anterior chamber depth regression in 28,709 adult cataract patients in China using swept‐source optical biometry

Background The purpose of this study was to evaluate the ocular biometric parameters in adult cataract patients from China and create an anterior chamber depth (ACD) regression model. Methods The ocular biometric records of 28,709 right eyes of cataract surgery candidates who were treated at Aier Eye Hospitals in nine cities from 2018 to 2019 were retrospectively analyzed. All measurements were taken with IOLMaster 700. We included patients who were at least 40 years old and were diagnosed with cataract. Results The mean age of the patients was 68.6 ± 11.0 years. The mean values recorded were as follows: axial length (AL), 24.17 ± 2.47 mm; mean keratometry (Km) value, 44.26 ± 1.70 D; corneal astigmatism (CA), 1.06 ± 0.96 D; ACD, 3.02 ± 0.45 mm; lens thickness (LT), 4.52 ± 0.45 mm; central corneal thickness (CCT), 0.534 ± 0.04 mm; and white to white (WTW) corneal diameter, 11.64 ± 0.46 mm. ACD correlated positively with AL (Spearman coefficient, 0.544) and WTW (0.300), but negatively with LT (-0.660) and age (-0.285) (all P < 0.01). In the multivariate regression analysis of ACD, which included LT, AL, WTW, sex, Km, CCT, and age, there was a reasonable prediction with adjusted R2 = 0.641. Conclusions Cataract patients with longer AL and wider WTW have deeper ACD. With increasing age and lens thickening ACD becomes shallower. Based on the standardized coefficients of ACD multivariate regression analysis from the study, LT is the main factor that affects ACD, and is followed by AL.


Background
Modern cataract surgery has transformed from a rehabilitative procedure to a refractive operation. This can be attributed to the development of measuring devices, lens calculation formulas, surgical technology, and intraocular lenses. Accurate prediction of refractive outcomes after cataract surgery is crucial. When biometry is based on optical coherence interferometry (Zeiss IOLMaster), the error from anterior chamber depth (ACD) prediction amounts to approximately 42 % of the total refractive prediction error [1]. Olsen [2] found that postoperative ACD is significantly predictable by a five-variable regression method that incorporates preoperative axial length (AL), ACD, keratometry (K), lens thickness (LT), and refraction. Therefore, it is essential to acquire the biometric characteristics of cataractous eyes to enable the utilization of precise ocular biometric parameters for accurate intraocular lens calculation.
Previous research has confirmed that female sex, older age, shorter AL, shallow ACD, and smaller LT were risk factors for angle closure [3]. These risk factors are also closely associated with elderly cataract patients.
To the best of our knowledge, there is no study of a large series of biometry measurements acquired with the IOLMaster 700 device in the existing literature. The purpose of this study was to review and evaluate their biometric parameters and create an ACD regression model for cataract surgery candidates aged 40 years or older. The results of this study may provide a new reference for cataract patients based on a large group of patients.

Study population
The medical records of 28 The ocular parameters were measured with a noncontact swept-source optical biometer (IOLMaster 700, Carl Zeiss Meditec AG, Jena, Germany). The inclusion criteria were as follows: (1) patients diagnosed with cataracts; (2) patients who were 40 years or older; and (3) patients with good quality measurements with the IOLMaster 700. The exclusion criteria were as follows: (1) patients who had complications with other eye diseases including corneal diseases, retinal diseases, glaucoma or inflammatory eye diseases; (2) patients who had a history of trauma or ocular surgery; and (3) patients who had used drugs that affected anterior chamber depth and pupil diameter, such as atropine or pilocarpidine eye drops.

Device and measurement
The IOLMaster 700 has a swept source optical coherence tomography (SS-OCT) scanning capacity with 1, 050 nm laser infrared light and measures AL, ACD, LT, and central corneal thickness (CCT). The device can obtain OCT images of the macula and visualizes the measurement of the AL of the eye (Fig. 1). The IOL Master 700 uses telecentric keratometry for K measurements. Corneal power is measured in two meridians: the greatest and least radii of curvature (K 1 , K 2 ). Corneal astigmatism (CA) is calculated as the absolute difference between K 1 and K 2 values. Mean keratometry (Km) is the average of K 1 and K 2 . It provides keratometry measurements in the central 2.5 mm zone and uses a refractive index of 1.3375 for the biometry parameters. ACD is measured from the corneal epithelium to the anterior lens surface [4]. A single capturing process can provide multiple measurements for each parameter. We controlled the quality of the measurements according to the manufacturer instructions. All patients were tested by experienced examiners. The compound values of AL, ACD and LT are derived from the average of the six measurements. The compound values of K, CCT and WTW are derived from the average of the three measurements. We selected measurements with a standard deviation of less than 20 microns.

Statistical analysis
To avoid correlation between the right and left eyes, only the right eye of each subject was used for statistical analyses. Data were processed with the SPSS software version 19.0 (IBM, USA). We performed a four-step analysis. First, we stratified the parameters according to sex and age groups. The study participants were classified into 5 groups based on age (40-49 years, 50-59 years, 60-69 years, 70-79 years, over 80 years). All eyes were divided into five groups based on AL; they were also categorized into five groups based on ACD. Second, the Mann-Whitney U Test was used to compare the ocular parameters according to sex and AL group. The chisquare test was used to compare the frequency distribution. Third, we evaluated the association between age and other ocular parameters using Spearman's correlation. Finally, we performed a multivariate analysis with ACD as the dependent parameter and all other parameters that were significantly associated with ACD as independent variables.
Variation inflation factors were calculated to investigate multicollinearity. All P-values were two-sided and considered statistically significant when the values were less than 0.05.

Results
A total of 28,709 patients, including 11,753 men (40.9 %) and 16,956 women (59.1 %) diagnosed of cataract, were selected according to our inclusion and exclusion criteria. The mean age of the patients was 68.6 ± 11.0 years (range, 40-101 years). The mean values measured were as follows: AL, 24.17 ± 2.47 mm; mean keratometry (Km) value, 44.26 ± 1.70 D; corneal astigmatism (CA), 1.06 ± 0.96 D; ACD, 3.02 ± 0.45 mm; lens thickness (LT), 4.52 ± 0.45 mm; central corneal thickness (CCT), 0.534 ± 0.04 mm; and white to white (WTW) corneal diameter, 11.64 ± 0.46 mm. The details of the ocular characteristics of the right eyes stratified according to age groups are shown in Table 1. The subgroups and the distribution of AL and ACD are shown in Table 2. Figure 2 illustrates the percentage of patients in the AL groups stratified according to age. The proportion of patients with long AL (AL > 25 mm) decreased with age.
In this study population, the AL was statistically significantly longer in male subjects than in female subjects (24.39 mm versus 24.02 mm, P < 0.001). The ACD was significantly deeper in men than in women (3.12 mm versus 2.96 mm, P < 0.001). Men had a thicker LT than women (4.53 mm versus 4.52 mm, P < 0.05). The Km was flatter in men than in women (43.75 D versus 44.62 D, P < 0.001). The WTW was significantly wider in men than in women (11.78 mm versus 11.54 mm, P < 0.001). The above results were shown in Table 3.
We also compared the ACD of the three AL groups (< 22 mm, 22-25 mm and ≥ 25 mm) and found that the differences between them were statistically significant (P < 0.001). The ACD of the patients in the group with AL < 22 mm was the shallowest (Fig. 3).
Spearman correlation analysis revealed that age, AL, ACD, LT, Km, and WTW were all significantly correlated with each other (P < 0.05). ACD correlated positively with AL (r = 0.544, P < 0.01) and WTW (r = 0.300, P < 0.01), but negatively with LT (r = -0.660, P < 0.01) and age (r = -0.285, P < 0.01). ACD was highly correlated with LT and AL. The complete matrix of the correlation is presented in Table 3.
Variation inflation factors for variables were below 2, indicating that there was no clear evidence of multicollinearity. In the multivariate regression analysis of the final model, which included LT, AL, WTW, Km, sex, CCT, and age, there was a reasonable prediction with adjusted R 2 = 0.641 (Table 4).

Discussion
We reviewed and evaluated the biometry parameters of cataract surgery candidates and created an ACD regression model for predicting the refractive outcomes of cataract surgery. Optical biometry has been well accepted as the gold standard since the introduction of the IOLMaster optical biometer in 1999 [5]. As a newly available swept-source OCT-based optical biometry device, the IOLMaster 700 provides OCT imaging of the macula and visualizes the measurement of the AL of the eye. The repeatability and reproducibility of the sweptsource optical biometer is excellent and its agreement with a standard biometer is very high; better lens penetration ability and more accurate AL measurements are obtained with the swept-source biometer than with the standard biometer [6]. The SS-OCT biometer showed better penetration in dense posterior subcapsular cataracts, measuring AL successfully in 96 % of cases which possibly was a result of the reduced scattering and attenuation from ocular opacities by the 1055 nm light source used [7]. In the present study, ocular biometry was performed with the IOLMaster 700, which can capture all parameters in a single process. It is convenient for evaluating patients, especially elderly patients.
A gradual descending trend in average age was reported in five studies of patients who underwent cataract surgery (Table 5) [8][9][10][11][12]. To minimize selection bias, we compared the proportion of patients under 70 years of age in our study and with those of Cui's [10] studies, which represented the cataract surgery scenario in China 10 years ago (July 2007 and June 2011). Both studies included cataract candidates scheduled for phacoemulsification, most of whom came from the urban area. The result, which indicated that there was a higher ratio of cataract patients under 70 years of age currently than there was 10 years ago, was statistically significant (χ 2 = 199.008, P < 0.001). Over the last 10 years, the development of phacoemulsification and femtosecond laser-  [13,14] and the popularity of functional intraocular lenses [15,16] have advanced cataract surgery in China, leading to better visual outcomes. Urban Chinese cataract patients in the present study tended to choose surgical treatment earlier than cataract patients did 10 years ago, a decision which could improve their quality of life.
In present study, more female patients scheduled cataract surgery than male patients (59.1 % versus 40.9 %). This finding is almost identical to those of previous published studies [9,17] (60 % versus 40 %). Furthermore, female patients presented for surgery at a slightly older age than male patients (mean: 68.8 years versus 68.4 years, P = 0.005). This is similar to the results of Jivrajka's report (mean: 75 years versus 73 years), which considered that women had a longer life expectancy than men [9]. Moreover, some studies showed that women exhibit a higher prevalence of some types of cataract than men [18,19]. Freeman [20] suggested that a potentially modifiable factor in cataractogenesis may be a woman's exposure to postmenopausal estrogen. This partially explains the sex-related differences in the rates of cataract surgery. Moreover, we believe that elderly women are more concerned about their eye health and are more likely to undergo surgery than elderly men. This needs to be confirmed by further research.
In the present study, female patients presented for surgery with a shorter average AL (mean 24.02 mm versus 24.39 mm, P < 0.001) and shallower ACD (mean 2.96 mm versus 3.12 mm, P < 0.001) than male patients. This echoes the findings of other studies of cataract patients in different countries [21][22][23]. Some studies showed that body height is positively correlated with AL and ACD; taller persons tended to have longer AL and deeper ACD than shorter people, and men are generally taller than women [9,24,25]. The sex-related differences in the AL and ACD measured in the present study may be attributed to the association between ocular dimensions and stature. Further, this is consistent with the higher incidence of primary angle closure (PAC) in women than in men [3,26]. Research confirms that the female sex, older age, shorter AL, shallow ACD, and larger LT are risk factors for angle closure [3,26]. These risk factors are closely associated with elderly cataract patients.
In the present study, the mean AL was 24.17 ± 2.47 mm; this result is similar to those of previous reports by Yu [27] (24.38 ± 2.47 mm) and Huang [28] (24.32 ± 2.42 mm). We found that AL was negatively correlated with age (r = -0.180 P < 0.001) ( Table 4). The longest mean AL was in the 40-49 years age group (26.63 ± 3.40 mm); this mean AL is longer than that of Cui's report (25.97 ± 3.53 mm) [10]. This may be because our study population had a higher proportion of eyes with a longer AL (AL > 26 mm) than the population in Cui's study [10] (14.7 % versus 11.9 %). This is especially apparent in the 40-49 years age group in our study, in which the proportion was as high as 52.1 %. The proportion of eyes  with AL > 25 mm decreased with age (Fig. 2). This corroborates Jivrajka's supposition that high myopia with increased AL predisposes to the development of cataract at a younger age [9]. On the contrary, newer lens formulas and adjustments/transformations are now available to reduce postoperative predictive error in high myopes [29,30]. Furthermore, the levels of satisfaction and spectacle independence experienced by high myopes were reasonably high after implantation of multifocal or trifocal intraocular lens [15,31]. Since these outcomes enhance the confidence of high myopes, they are more inclined to schedule cataract surgery at a younger age.
In the present study, the proportion of eyes with short AL (AL < 22 mm) was 6.4 %, which is similar to that of other studies [10,28]; the proportion of eyes with shallow ACD (ACD < 2.5 mm) was 11.8 %. The refractive status of a patient with a shallow ACD and a short AL after cataract surgery would tend toward a myopic shift. Conversely, a patient with deep ACD and long AL would move toward a hyperopic shift, which was related to the postoperative ACD change [32]. The relative change in ACD after phacoemulsification is larger in short eyes than in normal and long eyes [33]. Melles found that the Barrett Universal Formula had the lowest mean absolute prediction error for eyes with short AL and shallow ACD [34]. This guided us toward choosing the suitable formula for these kinds of patients. The mean ACD in the present study was 3.02 ± 0.45 mm. From the matrix of Spearman correlation of biometric parameters (Table 3) cataract patients with longer AL and wider WTW have deeper ACD. With increasing age and lens thickening ACD becomes shallower. The mean LT was 4.52 ± 0.45 mm and tended to be thicker in older patients and in shorter eyes. These findings are consistent with the findings of other studies [8,9,35]. These results confirm Hoffer's [8] mechanism of emmetropization: the lens thins (or decreases in power) as the eye gets longer (myopic) and thickens (or increases in power) as the eye gets shorter (hyperopic). The Km in the present study was 44.26 ± 1.70 D and tended to be greater in shorter eyes, which is also in line with the mechanism of emmetropization.
Based on the standardized coefficients of ACD multivariate regression analysis, LT is the main factor that affects ACD, and is followed by AL (Table 4). Adjusted R 2 = 0.641 means that 64.1 % of the variance could be explained by this model. A study of an elderly Chinese population showed that approximately one in five people aged 50 years and over developed some form of angle   [26]. In the present study, the mean ACD of patients in the 60-69 age group was 2.55 mm when the AL was < 22 mm (Fig. 3), and it tended to be shallower with age. Therefore, cataract patients with AL < 22 mm should be more concerned about their intraocular pressure when they are over 60 years old. This study has certain limitations. First, the data from the nine hospitals do not completely represent the ocular parameters of the overall population in China. Second, we need to collect more information about anthropometric characteristics, education, occupation, income, and the type of intraocular lens, and evaluate the correlations among them. We will conduct more research to improve on the findings of the analyses.
In summary, we collated data on the ocular biometric parameters of cataract surgery candidates in China, compared them, analyzed their correlations, and created an ACD regression model. This study provided reference values for AL, Km, CA, ACD, LT, CCT and WTW by using IOLMaster 700 for adult cataract patients in China. Cataract patients with longer AL and wider WTW have deeper ACD. With increasing age and lens thickening ACD becomes shallower. Based on the standardized coefficients of ACD multivariate regression analysis from the study, LT is the main factor that affects ACD, and is followed by AL.