In the present multicenter study, we demonstrated the correlations between the anterior and posterior corneal surfaces according to different refractive status and anterior segment dimension in a large number of Chinese myopic patients from different parts of mainland China (GZ from the south, SY from the northeast, CD from the southwest, WH and HK from the central). Using the pooled data of the five cohorts with diversity in age, sex and SE, negative correlations were observed between the anterior and posterior Km while positive correlations were observed between the ACA and PCA, between anterior and posterior eccentricity, asl well as between anterior and posterior asphericity. The results were consistent with some previous studies.5,6 We also found that R values of the anterior-posterior corneal correlations varied in different quantile groups of myopia and some corneal parameters (ie, SimKm, PA and CV), but not affected by the quantile levels of WTW and anterior chamber parameters (ie, ACH and ACV).
Assessment of the corneal curvature plays an important role in refractive surgery planning, intraocular lens power calculation and early diagnosis of keratoconus 6,11-13. In traditional keratometry, the cornea is assumed to have a fixed ratio between the anterior and poster corneal curvature, and the anterior corneal curvature is used to calculate the total corneal curvature. With recognition on the changes of anterior and posterior corneal curvature correlations, the ratio between the anterior and posterior curvature radii were adopted to evaluate the risk of hyperopia shift after corneal surgery.14 In the present study, the correlations between the anterior and posterior Km were similar in different myopic groups and quantile groups of PA, CV, ACH, ACV and WTW (Figures 1A, 2A and 3B-F). However, greater R values between the anterior and posterior Km were observed in group Q2 and Q3 than in group Q1 and Q4 of SimKm (Figures 2A and 3A). This finding suggests that the correlations between the anterior and posterior Km are different in corneal with different refractive power, and that measuring both the anterior and posterior corneal curvatures is needed to obtain an accurate total corneal curvature. In the present study, the strength of the correlations between the anterior and posterior Km was strong, with R values of at least -0.85 in all the myopia groups. This result was consistent with a previous study in which the R value was -0.94 for normal eyes and -0.85 for subclinical keratoconus, but the R value was decreased in eyes with confirmed keratoconus5. Our result also indicated that the increase of myopia severity didn’t result in reduced correlation strength between the anterior and posterior corneal curvatures.
Management of PCA during astigmatism correction has become a hot topic in cataract and refractive surgery, and many efforts have been made to predict the PCA using ACA.2,15 However, PCA prediction based on ACA alone is not accurate enough.9 In the present study, the R values for the correlation between ACA and PCA were increased from 0.65 in the LM group to 0.75 in the EHM group, indicating a stronger ACA-PCA correlation in eyes with higher severity of myopia. The R values of ACA-PCA correlation were also slightly increased with quantile levels of SimKm (0.66 to 0.74). These findings indicate that the ACA-PCA correlation is affected by multiple factors and the integration of various anterior segment parameters is needed to make a more accurate PCA prediction.1 Moreover, the moderate ACA-PCA correlations may be due to the greater hereditability of the posterior corneal surface compared to the anterior corneal surface 16,17. Surprisingly, the strength of ACA-PCA correlation seems to be higher in keratoconus eyes compared to normal eyes, although controversy still remains.5-7 This may be due to the unsynchronized changes of the ACA and PCA during the development of keratoconus, and the tendency towards a better alignment of the PCA to the ACA in keratoconus eyes.16-18
In the present study, there were weak correlations (R: 0.29 to 0.45) between the anterior and posterior corneal asphericity in all the quantile groups. These findings were different from previous studies where poor anterior-posterior corneal asphericity correlations in normal eyes1,5 and good correlations in keratoconus eyes were observed.1,2 It was shown that the R values for the anterior-posterior corneal asphericity correlations in normal eyes were 0.03-0.04 and 0.17 in the previous studies.1,5 On the contrary, the R values of the correlations were 0.62-0.89 and 0.86-0.94 in keratoconus eyes.1,2 These findings together suggest that in terms of the anterior-posterior corneal asphericity correlation, myopic eyes are in between the normal eyes and keratoconus eyes. The roles of anterior-posterior corneal asphericity correlations in corneal ectasia warrants further investigation. Like the ACA-PCA correlations, the anterior-posterior corneal asphericity correlations are also affected by multiple factors such as myopia severity, SimKm, PA, and CV, suggesting complex interactions between the anterior-posterior corneal shapes and other anterior segment parameters.
In the present study, the R values for anterior-posterior correlations of corneal astigmatism, eccentricity and asphericity were increased with higher quantile levels of SimKm, but the R values for anterior-posterior correlations of corneal eccentricity and asphericity were decreased with higher quantile levels of PA and CV. These findings suggest that different the corneal parameters have different effects on the anterior-posterior corneal correlations. Further studies are required to investigate the mechanisms of each corneal parameter affecting the anterior-posterior correlations of corneal shapes.
Our study has some limitations. Firstly, the conclusions of our study can only be applied to myopic patients from the same age group. In older patients, the anterior-posterior corneal correlations may be different. Factors affecting these correlations may also be different in older patients. Secondly, the exact mechanisms underlying the anterior-posterior corneal correlations remain a mystery to us, although some affecting factors are identified in the present study. Thirdly, we didn’t compare the anterior-posterior corneal correlations using data obtained from different imaging devices. Thus, we are unable to know whether our results can be used interchangeably with those obtained from other devices. Lastly, due to the cross-sectional design of the study we did not investigate the changes of anterior-posterior corneal correlations after corneal refractive surgery. Further studies are needed to address this issue.
In conclusion, the present study has demonstrated that the anterior-posterior correlations of corneal curvature, astigmatism, eccentricity and asphericity can be affected by the severity of myopia and some other corneal parameters.