Analysis of axial Length of Marfan Syndrome with Bilateral Ectopia Lentis by Z-scores for Chinese Patients under 20 years Old

Zexu Chen Eye & ENT Hospital of Fudan University Jiahui Chen Eye & ENT Hospital of Fudan University Min Zhang Eye & ENT Hospital of Fudan University Jialei Zheng Eye & ENT Hospital of Fudan University Michael Deng Eye & ENT Hospital of Fudan University Tianhui Chen Eye & ENT Hospital of Fudan University Yinghong Ji Eye & ENT Hospital of Fudan University Yongxiang Jiang (  yongxiang_jiang@163.com ) Eye & ENT Hospital of Fudan University


Introduction
As an important component of the ocular refractive media, a normal crystalline lens is critical to the development of the visual system. Ectopia lentis (EL) refers to the dislocation of the lens from its physiological position, which may cause refractive errors of varying severity, poor visual acuity, and amblyopia in young patients 1 . The major etiology of congenital EL is Marfan syndrome (MFS, OMIM 154700), the prevalence of which ranges from 4.6 to 6.5 per 100 000 [2][3][4] . MFS is an autosomal dominant hereditary disease caused by brillin 1 (FBN1) gene mutations 5 ; brillin 1 is the main constituent of the micro laments, which are the key structural components of the ciliary zonules 6 . Mutations of FBN1 result in the impaired suspension of the crystalline lens in its anatomic position 7 . In addition to EL, the eyes of patients with MFS are characterized by long axial lengths (AL), low corneal curvatures, severe astigmatism, and dense myopia 8 .
Refractive and biometric characteristics of patients with MFS have been reported [9][10][11][12][13] ; however, most studies have used a relatively small sample size without control groups. To our knowledge, few investigations have examined the distribution of AL and the correlations between AL and other biometrics before lens surgery in young MFS patients with bilateral EL. In this retrospective study of a relatively large group of patients, we aimed to investigate the overall status of axial development in MFS patients younger than 20 years old and the relationships among AL development and patient demographics, including age, sex, family history, and comorbidities. We introduced the age-adjusted Z-scores by using external controls to analyze the correlations of AL development with the corneal curvature radius (CCR), AL/CRC ratio, and white-to-white horizontal corneal diameter (WTW). Our results provide a useful reference for the precise planning of surgical interventions for MFS patients with bilateral EL.

Patient demographics
The ocular biometric parameters of 183 MFS patients under 20 years old with bilateral EL were collated.
Both eyes were enrolled. All participants were Chinese, and 55.49% were male. The age distribution of the participants is shown in Fig. 1 . 2a). There was a signi cant difference in the AL among all the age subgroups (P < 0.001), and the trend in AL increase shown in the scatterplot follows a logarithmic curve (R 2 = 0.145, P < 0.001) (Fig. 2b). Considerable variation in AL was observed among patients of the same age.
The Z-AL score distribution The overall mean Z-score was 1.74 ± 2.68 (range, − 2.55 to 10.68). For the age groups, the mean Z-scores were 1.68 ± 2.37 for under 5 years old, 2.02 ± 3.37 for the 6-8 years of age group, 1.91 ± 2.67 for the 9-11 years of age group, 1.10 ± 2.20 for the 12-15 years of age group, and 1.74 ± 2.11 for the 16-20 years of age group (Fig. 3a). The age difference was successfully adjusted by the Z-AL score because there were no statistical differences in the Z-AL score among the different age groups (P = 0.531) (Fig. 3a). The distribution of AL within the same age group was well-represented by the Z-score. The proportion of the Z-AL score shared a similar pattern among the age subgroups. A considerable variation in AL was observed. In all the age groups, about 30% of patients had a Z-AL score < 0, which means that some MFS patients had a shorter AL than the age-matched control. Extremely long AL (Z-AL score ≥ 3) was observed in about 30% of patients (Fig. 3b).

Relationship between Z-AL score with different factors
To further explore the factors in uencing the AL of the participants, we analyzed the relationships among AL or Z-AL score and laterality, sex, and family history ( Table 1). The right eyes had slightly but signi cantly longer AL than the left ones (P = 0.013). Males seemed to have longer AL, but the difference was not signi cant (P = 0.267). There was no difference between patients with or without a family history of MFS (P = 0.907). In the 20 eyes (10.87%) with microspherophakia, the AL appeared to be larger, but the difference was not statistically signi cant (P = 0.145). Of the 145 eyes (39.84%) with quali ed WTW measurements, megalocornea (WTW ≥ 13) was found in 10 (6.90%). MFS eyes complicated by megalocornea had longer AL on average (P = 0.005). Correlation between AL with biometry parameters as Zscores Z-AL scores were plotted against Z-CCR, Z-WTW, AL/CCR ratio, and corneal astigmatism (Fig. 4). There was a positive correlation between Z-AL and Z-CCR (r = 0.265, P < 0.001) (Fig. 4a) and a positive correlation between Z-AL and AL/CCR ratio (r = 0.762, P < 0.001) (Fig. 4b). No correlation was found between Z-AL and Z-WTW (r = 0.052, P = 0.537) (Fig. 4c). A mild but signi cant correlation was observed between Z-AL and AST (r = 0.151, P = 0.004) (Fig. 4d).

Discussion
About 60% of patients of MFS had EL, which is usually bilateral, progressive, and has a superior direction 8 . The implantation of intraocular lens (IOL) plays a crucial part in the postoperative visual rehabilitation of children. The AL is the most important factor affecting the accuracy of IOL power, and a 1-mm error of measurement can induce an approximate 2.5D refractive error 19 . Previous investigations have shown the AL in the eyes of patients with EL was different from those in normal eyes 20 . However, there have been few studies of the characteristics of the AL in MFS eyes before lens surgery. In this study, we recruited 183 MFS patients with bilateral EL to investigate the distribution and in uence factors of AL in the developing age. The age-adjusted Z-score is a powerful tool to study the in uencing factors and the correlation between AL and other biometrics.
Here, we present the distribution of AL in MFS patients with bilateral EL. Similar to the results for the normative cohort 21 , there was a positive correlation between AL and age in the current study. A logarithmic correlation between AL and age was also previously found in congenital cataract patients 22 .
AL increases dramatically in the rst 2 years of life, then grows at a slower rate until adulthood 22 . The AL growth trend for patients in the current study also presented a logarithmic curvature. However, because of the limited number of patients under 3 years old, the high growth rate normally seen in the early stage of the logarithmic curve is absent in this study. Previous studies found that EL can manifest at an early age: EL is recognized at age 0 in 38% of patients with neonatal MFS 23 , and 12.5% of MFS patients show EL before 3 years of age 8 . The prevalence might have been underestimated because of the di culties for young patients expressing themselves and poor cooperation under ophthalmic examination. The distribution of AL among patients under 3 years old is uncertain and warrants further study.
The Z-score has been used by cardiologists to adjust for body surface area when measuring aortic sinus diameters in various individuals 24 , and new Z-scores have been developed to normalize age, sex, height, and weight 25 . A Z-score ≥ 2.0 (≥ 3.0 for patients under 20 years old) is a major criterium in the diagnosis of MFS 14 . However, the age-adjusted Z-score introduced in the current study was not intended to assist in making a diagnosis based on ocular biometrics, which has been excluded in the 2010 Ghent 2 criteria due to uncertain speci city 14 . In this study, the Z-scores successfully normalized the age differences in AL and other parameters by comparing age-matched published normative values. The distribution of AL across the age groups was well-represented by the Z-score.
The mean AL was 1.74 standard deviations from the normative mean measurement. About 30% of eyes had relatively long AL (Z-AL score ≥ 3), and the largest had a Z-score of 10.68. A long AL has been proven to be a risk factor for retinal detachment 26 and open-angle glaucoma 27 in the eyes of MFS patients.
Thus, patients with longer AL should have more intensive ophthalmic exams, especially when they have a visual complaint. Although a long AL has been recognized as an ocular manifestation of MFS, our study showed that about 30% of the eyes had a short AL (Z-AL score < 0), and the proportion was similar among the different age groups. Thus, young patients with bilateral EL but small AL should not be simply diagnosed as having idiopathic EL without considering the systematic features. Since the manifestations of MFS can be atypical at the young age 28 , a genetic test or careful follow-up is strongly recommended.
Considerable variation was observed among those of the same age, indicating that the development of the AL in MFS patients is associated with multiple factors in addition to age. However, the sex and family history of MFS were not associated with differences in AL. The right eye AL was, on average, slightly longer than the left one, with uncertain clinical implications. One possible explanation is that a variety of FBN1 mutations could contribute to this difference. More than 3000 mutations of FBN1 are known of today 29 , but only 14.20% of the mutations are recurrent 30 . The misfolded brillin-1 and the disorganized elastin in the scleral stroma could lead to stretching of the sclera through the two physiological functions of brillin-1 (structure component or TGF-β signaling mediator) and by different underlying genetic mechanisms (haploinsu ciency or dominant-negative effect) 30 . In addition, the heterogeneity of FBN1 mutations may lead to various degrees of EL, which accelerates globe growth via defocus or deprivation.
Thus, the measurement of AL should be individualized to achieve more precise results, even in patients of the same age. Further genotypic-phenotypic analyses are required to understand the variations in AL and severity of EL. The Z-score will be helpful when analyzing FBN1 mutations and ocular phenotypes in young patients of different ages.
A previous study described the ocular biometrics of MFS eyes based on age strati cation 31 . However, although AL is an important indicator of eyeball development, the relationships between AL and other biometric parameters have been scarcely reported in MFS patients. Correlations among AL and other biometrics were analyzed in the form of the age-adjusted Z-scores, and a positive correlation was found between AL and CCR. The decreasing power of the cornea appears to play an emmetropizing role in response to increasing AL; however, this compensatory capacity reportedly reaches its limit at an AL/CCR ratio of 3 32 . In the linear regression model used in the current study, the AL/CCR ratio exceeded 3 when the Z-AL score was above 2.16. Though, the correlation between AL and preoperative spherical equivalent was not studied because the displaced and deformed lens can affect the total refractive power. Nevertheless, we hypothesized that in the postoperative eyes with centered and stabilized IOL positions, a long-term myopic shift is expected in the developing eyes when the Z-AL exceeds 2.16. Therefore, the selection of IOL power deserves more careful consideration in those patients in order to achieve the refractive target. Microspherophakia and megalocornea are sometimes found in eyes of MFS patients 33,34 . To the best of our knowledge, the prevalence of these two comorbidities in MFS patients with bilateral EL have never been reported. Our study indicated that about 10% of eyes of MFS patients were complicated by microspherophakia, the AL of which seemed to be greater than that of normal eyes but not to a statistically signi cant degree. Megalocornea was found in about 7% of MFS eyes. Although WTW measurements did not change proportionally with AL growth, MFS eyes complicated by megalocornea had signi cantly larger AL on average. Our data show there was high corneal astigmatism in the eyes of MFS patients, which was consistent with previous studies 35,36 . Additionally, we found that patients with smaller AL generally had higher corneal astigmatism. Thus, residual astigmatism from the cornea after lens aspiration and IOL implantation, especially in MFS eyes with smaller AL, is a challenge for postoperative visual rehabilitation.
The results of this study must be assessed within the context of its limitations. First, the data are not a complete representation of MFS patients and only relate to surgical patients with bilateral EL, which prevented us from detecting differences in eyes with unilateral EL or without EL. Second, the external controls used in our study were school-based, cross-sectional studies of young Chinese individuals from a large sample size. The conclusions should be viewed with caution for other ethnic populations. Third, adult patients were excluded from the current study, as AL were shown to be stable in adults (> 30 years old) with MFS during a 10-year follow up 37 . The cross-sectional nature of our study also limited our ability to track changes in AL data over time. Despite these limitations, the results of this study describe the overall distribution of AL in the eyes of Chinese MFS patients under 20 years old with bilateral EL.
In conclusion, the mean AL of MFS was found to be longer than the age-matched controls, especially in the eyes complicated by megalocornea. But one-third of patients still had an AL below the normative average. The considerable individual variation indicates that young patients with bilateral EL but small AL should not be excluded from MFS diagnoses without systematic examinations. The Z-scores normalized the age difference in AL and other parameters by comparing age-matched published normative values, and this technique facilitates the further study of the in uence factors of AL in young patients of different ages. A full understanding of the AL will be helpful for surgical planning and complication monitoring.

Ophthalmic Examinations
The participants were examined by board-certi ed ophthalmologists, and their medical and family histories were evaluated. All patients were examined for EL by slit-lamp examination under complete pupillary dilation. Bilateral EL was de ned as the displacement of the lens from its normal position in both eyes. The AL, CCR, WTW, and corneal astigmatism were measured using partial coherence interferometry (IOLMaster 500, Carl Zeiss Meditec AG, Jena, Germany). Five valid AL readings and three readings of both CCR and WTW were taken.

Z-Score Calculation
The AL and other ocular biometrics were compared to age-matched normative data published in the literature that were cross-sectional, epidemiological investigations of the Chinese population. The Z-score of AL (Z-AL) was calculated using the formula below: Z-AL = (measured AL -normative AL)/ normative standard deviation (SD). The Z-score describes the position of a raw value in terms of its distance from the mean when measured in standard deviation units. For example, a Z-score ≥ 2.0 corresponds to a measurement ≥ 2.0 SD above normal. For patients of 4-18 years of age, the AL, CCR, and WTW were compared with normative measurement data in the Shandong Children Eye Study [15][16][17] . For those over 18 years old, the measurements for 18-year-old children in the same study were employed as reference data. For 3-year-old patients, the Shanghai Pudong Eye Study was used as the reference 18 (Supplementary Table S1 online). Because of a lack of normative measurements for 2-year-olds in the Chinese population, the Z-score was not analyzed in one patient (two eyes). The Z-scores for other biometric parameters, such as Z-CCR and Z-WTW, were calculated in a similar manner.

Statistical Analysis
Continuous variables were described in mean ± SD, and categorical variables were described in counts or proportions as appropriate. The distributions for age and ocular biometric parameters were tested for normality with the Kolmogorov-Smirnov test. The one-way analysis of variance or Kruskal-Wallis test was used, as appropriate, to compare the AL among different age groups. For independent samples, the ttest or Wilcoxon test was used to analyze differences in the AL among patients with different factors, including gender, family history, and comorbidities. And the paired t-test or paired Wilcoxon test was used to evaluate the difference in AL between the left eye and the right eye. A logarithmic correlation between AL and age was calculated for MFS patients. Correlations between the Z-score of AL and Z-scores of other biometric variables were analyzed using Spearman's correlation. All analyses were 2-tailed, and the statistical signi cance was set as a P-value of less than 0.05. Statistical analyses were performed using SPSS version 22.0 (IBM Corp., Armonk, NY, USA).
Declarations Figure 1 The age distribution of MFS patients with bilateral EL Subdivided groups were marked by different colors:

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. AnalysisofALinMFSeyeswithBELbyZscoreSupplementary.pdf