In the present study, we explored the performance of BAD parameters in a sample of Chinese patients with forme fruste keratoconus and those with corneal thickness less than 490um. Our results showed that the influence of CD on the BAD parameters, which had been proved in tomographically normal cornea [2–4], may also affect the FFKC and thin cornea. Then, we found that for eyes with CD ≤ 11.9mm, the rates of abnormal cases in the FFKCG were significantly higher than in the NG for seven indices: BE, Dp, Da, BAD-D, PPImax, PPIavg, and ARTmax. Meanwhile, for eyes with CD > 11.9mm, the rates of abnormal cases in the FFKCG were significantly higher than in the NG only with Dp and PPIavg. For eyes with CD ≤ 11.9mm, the rates of abnormal cases in the TCG were significantly higher than in the NG for nine indices: Df, Dp, Dt, Da, BAD-D, PPImin, PPImax, PPIavg, and ARTmax. For eyes with CD > 11.9mm, the rates of abnormal cases in the TCG were significantly higher than in the NG for seven indices: Dp, Dt, Da, BAD-D, PPImax, PPIavg, and ARTmax. The above data suggest that BAD indices might underestimate the risk of ectasia in the large cornea of Chinese patients, especially for eyes with FFKC.
Previous studies have confirmed the higher rate of false-positive in the BAD parameters in the eyes of the small cornea [2–4]. In the current study, we verified the higher rate of false-negative in the BAD indices in the eyes of CD > 11.9mm with FFKC or thin cornea. The effect of CD on the BAD indices in the large cornea could be attributed to the cornea's asphericity. Compared to the small cornea, the large cornea's curvature flattens more gradually towards the periphery. As a result, a BFS may have a decreased curvature radius (steeper power) in a large cornea. A steeper BFS would result in smaller front and back elevations. For the same reason, the indices of PPIs, which represent the rate of change in corneal thickness from the thinnest point to the periphery, would have a lower rate of change as the overall distance between the thinnest point and periphery is more extended in the large cornea.
We also found that for CD > 11.9mm, the rates of abnormal cases in the FFKCG were significantly higher than in the NG only for Dp and PPIavg. It seemed that the Dp and PPIavg were more sensitive than other indices for detecting subclinical (fruste) ectasia among patients of the large cornea. PPIavg is the average pachymetric progression, and Dp (deviation of normality of pachymetric progression) is the standardized counterpart of PPIavg. According to Ruiseñor et al. , PPIavg is the single metric with the highest sensitivity for subclinical keratoconus among all the ten BAD parameters. To date, a limited number of studies have reported on biomechanical characters of the large cornea. Ding et al.  found that corneal thickness was negatively correlated with cornea diameter. In addition, Montard et al.  demonstrated that corneal diameter was negatively correlated with cornea hysteresis (CH) and corneal resistance factor (CRF), indicating that corneal biomechanics tends to be weaker in the large cornea. Moreover, Shah et al.  revealed a negative correlation between the CH/CRF and spur‑to-spur distance of the ciliary body. The underestimation of the BAD indices in large cornea should be taken into consideration during the screening of refractive surgical candidates, and additional testing should be applied to these patients (e.g., corneal biomechanical testing and epithelial thickness analysis) because these cases could be at higher risk for developing iatrogenic ectasia (keratectasia) after corneal laser vision correction [13–15]. Except for BAD parameters, there are several other indices such as the Corneal Biomechanical Index (CBI), the Tomographic and Biomechanical Index (TBI) (Zhang et al. 2020) and the Screening Corneal Objective Risk of Ectasia Analyzer (SCORE, a software linked to the Orbscan topography system) (Saad & Gatinel. 2012; Chan et al. 2015). However, the CD was not incorporated as a variable in all the systems above. Considering the influence of CD on the diagnostic accuracy of BAD indices for FFKC, further studies are needed to test the influence of CD on the performance of these indexes for the large cornea to improve the diagnostic capacity.
The term FFKC was first described by Amsler  as an incomplete, abortive form of corneal disease that have subtle topographic characteristics but do not reach the threshold of keratoconus suspect. However, because of the ambiguity of definition there are no definitive criteria to help discriminate subclinical keratoconus from normal. A universally accepted criterion for FFKC has not yet been developed. Most studies defined FFKC as no clinical and topographical signs of keratoconus [17–19]. In the current study, FFKC was defined as ‘normal ABCD plus normal IHD’ in which ABCD classification system underlies both anterior and the posterior surfaces and the pachymetry at the thinnest point and the IHD index describes the decentration of elevation data in the vertical direction. We used the strict inclusion because we intended to investigate the earliest or mildest form of FFKC. Our study found that the sensitivity for the detection of FFKC was highest with PPImax (41.6%), followed by PPImin (39.6%) and BAD-D (35.7%). The low sensitivity in our study comparing to other studies (52–80% for BAD-D, 76% for PPImin, 78% for PPImax and 46.7% for ARTmax) [5, 20, 21] might be explained by differences in the definition of FFKC and different values of cut-off in different studies. What is added in the current study is that the sensitivity of the KC indices could become even lower in cornea with CD > 11.9mm.
Besides, we also found that the mean CD for the FFKCG, TCG, and NG groups was 11.9mm, 11.8mm, and 11.7mm, respectively, while the proportion of CD > 11.9mm was 43.6%, 37.5%, and 26% for the three groups respectively. According to Edmund , the mean CD was 11.8mm, and the proportion of CD > 11.9mm was 35.6% in KC eyes. Meanwhile, we found no significant difference in CD between the keratoconic eyes and the FFKC eyes. There is currently no evidence that keratoconic corneal alterations involve a sideward expansion of the cornea [23, 24]. Consequently, the larger CD in FFKCG could be just a sampling error, and further studies with a larger sample are needed to clarify this question.
In this study we employed the Pentacam HR to measure the cornea diameter . Pentacam measures the horizontal CD with an iris camera optic. It can recognize iris landmarks and determine the pupil's location and shape. Pentacam HR was in agreement with other devices, e.g., Orbscan, and has been widely used to determine the size of phakic posterior chamber intraocular lenses [26, 27].
Our data revealed that the influence of CD on the BAD parameters, which had been proved in tomographically normal cornea, may also affect the FFKC and thin cornea. The BAD indices might underestimate the risk of ectasia in the large cornea of Chinese patients, especially for eyes with FFKC. The analytical dimensions should be individualized for eyes with individual CD. Further studies are required to set up a rule for choosing an appropriate analytical dimension for each CD.