Evaluation of Equivalent Keratometry Readings Obtained by Pentacam HR (High Resolution)

Purpose To assess the repeatability of Equivalent Keratometry Readings (EKRs) obtained by the Pentacam HR (high resolution) in untreated and post-LASIK eyes, and to compare them with the keratometry (K) values obtained by other algorithms. Methods In this prospective study, 100 untreated eyes and 71 post-LASIK eyes were included. In the untreated group, each eye received 3 consecutive scans using the Pentacam HR, and EKR values in all central corneal zone, the true net power (Knet) and the simulated K (SimK) were obtained for each scan. In the post-LASIK group, each eye received subjective refraction and 3 consecutive scans with the Pentacam HR preoperatively. During the 3-month post-surgery exam, the same examinations and the use of an IOLMaster were conducted for each eye. The EKRs in all zone, the Knet, the mean K (Km) by IOLMaster and the K values by clinical history method (KCHM) were obtained. The repeatability of the EKRs was assessed by the within-subject standard deviation (Sw), 2.77Sw, coefficient of variation (CVw) and intraclass correlation coefficient (ICC). The bonferroni corrected multiple comparisons were performed to analyze the differences among the EKRs and K values calculated by other algorithms within the 2 groups. The 95% limits of agreement (LoA) were calculated. Results The EKR values in all central corneal zone were repeatable in both the untreated group (Sw≦0.19 D, 2.77Sw≦0.52 D, CVw≦1%, ICC≧0.978) and the post-LASIK group (Sw≦0.22 D, 2.77Sw≦0.62 D, CVw≦1%, ICC≧0.980). In the untreated group, the EKR in 4mm zone was close to SimK (P = 1.000), and the 95% LoA was (-0.13 to 0.15 D). The difference between Knet and SimK was -1.30±0.13 D (95% LoA -1.55 to -1.55 D, P<0.001). In the post-LASIK group, all the EKRs were significantly higher than KCHM (all P<0.001). The differences between the EKR in 4mm zone and KCHM, the EKR in 7mm zone and KCHM, Knet and KCHM, Km and KCHM, SimK and Knet were 0.64±0.50 D (95% LoA, -0.33 to 1.62 D), 1.77±0.88 D (95% LoA, 0.04 to 3.51 D), -0.98±0.48 D (95% LoA, -1.92 to -0.04 D), 0.64±0.53 D (95% LoA, -0.40 to 1.68 D), and 1.73±0.20 D (95% LoA, 1.33 to 2.13 D), respectively. Conclusions The EKRs obtained by the Pentacam HR were repeatable in both untreated eyes and post-LASIK eyes. Compared to the total corneal power obtained by the clinical history method, the EKR values generally overestimated the total corneal power in post-LASIK eyes. So, further calibrations for the EKR values should be conducted, before they were used for the total corneal power assessment in post-LASIK eyes.


Introduction
The accurate assessment of the total corneal power in eyes after corneal refractive surgery is essential for the prediction of intraocular lens (IOL) power [1][2][3]. Currently, there are several methods to assess total corneal power after corneal refractive surgery, which can be divided into two categories. One requires preoperative data, including the clinical history method (K CHM ) [4] and double-K method [5,6]. The other category does not require preoperative data and includes the rigid contact lens method [7], the Maloney method [8][9][10], Shammas formula [11,12] and the BESSt formula [13].
The Holladay's Equivalent Keratometry Readings (EKRs) obtained by the Pentacam HR belongs to the latter [14]. However, the precision of EKRs for the assessment of total corneal power after corneal refractive surgery is still controversial. The repeatability of the EKR (4.5 mm central corneal zone) from untreated eyes has been shown to be suboptimal [15]. Additionally, it was reported that the EKRs were not accurate for the prediction of IOL power in both untreated eyes and postoperative eyes [16].
The purpose of the present study was to assess the repeatability of the EKR in all central corneal zones obtained by the Pentacam HR. In untreated eyes, we compared the EKRs in all zones with the simulated keratometry (SimK) obtained by the Pentacam HR, and in post-LASIK eyes, we mainly compared the EKRs in all zones with K CHM .

Subjects and Methods
The present study was conducted at the Eye Hospital of Wenzhou Medical University (Wenzhou, China) between August and December 2011. The study was performed in accordance with the principles stated in the Declaration of Helsinki and was approved by the Office of Research Ethical Committee, Eye Hospital of Wenzhou Medical University. All subjects were informed of the purpose of the research and provided signed informed consent. Two groups were included in this study, the untreated group and the post-LASIK group. For the untreated group, inclusion criteria were healthy subjects aged 18 to 40 years old, without communication or cooperation disorders. All the subjects had the following qualities: 1) a preoperative best corrected visual acuity of 1.0 or more, 2) a corneal astigmatism of less than 2.0 diopter (D), and 3) an intraocular pressure range of 10 mmHg to 21 mmHg. The exclusion criteria included the following: 1) history of ocular pathology, 2) history of corneal or intraocular trauma, 3) previous ocular surgery, 4) the wearing of hard contact lenses within 4 weeks or soft contact lenses within 2 weeks, and 5) severe dry eye (tear film break-up time shorter than 5 seconds). For the post-LASIK group, besides the inclusion and exclusion criteria stated above, the postoperative uncorrected visual acuity was no less than 1.0 without corneal opacities.
In the untreated group, subjects received 3 consecutive scans using the Pentacam HR (Oculus, Germany) after a routine ophthalmic examination (including subjective refraction, slit lamp examination, fundus examination and intraocular pressure). In addition to the preoperative exams of the untreated group, the post-LASIK group received exams more than three months after the surgery, including the following: subjective refraction, the Pentacam HR (with 3 consecutive scans) and IOLMaster (Carl Zeiss). The Pentacam HR was performed in a dark room. Subjects were told to place their chin on the chinrest, focus on the indicator, and widely open their eyes after blinking. The Pentacam HR would take pictures automatically within 2 seconds. When the quality specification screen displayed "OK", the result was considered valid. Three valid scans for each eye were obtained. The mean keratometry (K m ) from post-LASIK eyes was obtained by the IOLMaster. In the post-LASIK group, the surgery was performed using the Allegretto laser system (Lumenis, Inc., USA). The diameters of optical zone ranged from 6.0 mm to 7.0 mm during the surgery.

Parameters for assessment of corneal power in the Pentacam HR
The anterior and posterior central corneal curvature, within a certain diameter (in meter), were defined as r anterior and r posterior , respectively. The refractive index of air (n 0 ) is 1.000. The standard corneal refractive index (n) is 1.3375. The real corneal refractive index (n 1 ) is 1.376. The aqueous refractive index (n 2 ) is 1.336. The ratio of posterior and anterior curvature from an untreated cornea (R 1 ) is 0.822 [17][18][19]. The ratio of simulated keratometry and anterior corneal power (R 2 ) is (1.3375-1.000)/(1.376-1.000) = 0.8976.
The SimK was calculated based on the anterior central corneal curvature within certain range and the standard corneal refractive index as follows: SimK ¼ ðn À n 0 Þ=r anterior : The true net power (K net ) [20] represents the sum of the anterior and posterior corneal power and is calculated using the following formula: K net ¼ ðn 1 À n 0 Þ=r anterior þ ðn 2 À n 1 Þ=r posterior : The EKR [14] was advanced by Holladay and his colleagues specifically for the evaluation of total corneal power after corneal refractive surgery and has been obtained in Pentacam HR since 2006. In the Pentacam HR's software interface "Holladay EKR Detail Report" values are given within 1 mm, 2 mm, 3 mm, 4 mm, 4.5 mm, 5 mm, 6 mm and 7 mm of the central corneal diameter, respectively. The values were calculated using the following formula: EKR ðDÞ ¼ ðn 1 À 1Þ=r anterior þ ðn À 1Þð1 À 1=R 2 Þ Â R 1 =r posterior This formula above, however, can be simplified to: EKR ðDÞ ¼ 0:376=r anterior À 0:03165=r posterior : In this study, the mean EKR within 1 mm, 2 mm, 3 mm, 4 mm, 4.5 mm, 5 mm, 6 mm and 7 mm zones were calculated and were abbreviated as EKR1, EKR2, EKR3, EKR4, EKR4.5, EKR5, EKR6 and EKR7, respectively.
The following three parameters are needed for the calculation of K CHM [4]: 1) the preoperative spherical equivalent (SEQ pre ) at corneal plane, 2) the postoperative spherical equivalent (SEQ post ) at the corneal plane and 3) the preoperative simulated keratometry (preK). First, the amount of refraction change in diopters before and after the surgery is calculated using this formula: Second, the K CHM is calculated based on the following formula: In this study, preK was obtained by the Pentacam HR, and we regarded the K CHM as the benchmark to assess the EKR, SimK and K net obtained by the Pentacam HR in the post-LASIK group.

Statistical analysis
All data were entered into an Excel spreadsheet (Microsoft Excel 2010 Crop.), and MedCalc Software (Vision11.4.2.0, MedCalc, Inc.) was used for statistical analysis. The results were expressed as the mean ± standard deviation (SD). A P value less than 0.05 was considered statistically significant. The normality of all data distributions was confirmed by the Kolmogorov-Smirnov test (all P>0.05), and parametric statistical tests were used for data analysis. Repeated measures ANOVA was applied to analyze the repeatability of EKR in different zones and the SimK and K net in both the untreated and post-LASIK groups . The within-subject SD (S w ) of three consecutive measurements were calculated. The repeatability limit was defined as 2.77S w , which means an interval within which 95% of the differences between measurements are expected to lie. The within-subject coefficient of variance (CV w ) was calculated as the ratio of the S w to the overall mean. Using the CV w , data with different means can be compared with each other. A lower CV w is associated with higher repeatability. The intraclass correlation coefficients (ICCs) were based on the analysis of variance for a two-way mixed-effects model with an absolute agreement for consistency of individual measurements and the 95% confidence interval (CI) was calculated. The ICC (ranging from 0 to 1) assesses the consistency for data sets of repeated measurements. As the ICC nears a value of 1, the measurement consistency increases, and a value more than 0.9 indicates acceptable clinical reliability [21]. The repeated measures analysis of variance with the Bonferroni corrected multiple comparisons was applied to compare the EKRs, SimK and K net in the untreated group, and the EKRs, SimK, K net and K CHM in post-LASIK group. The 95% limits of agreement (LoA) were calculated.

Results
In the untreated group, 100 eyes (47 men and 53 women) were included, and the mean manifest spherical equivalent refraction was -3.23±1.76 D. In the post-LASIK group, 71 eyes (24 men and 19 women) were included. The mean manifest spherical equivalent refraction before the surgery and at least three months after surgery were -5.01±1.91 D (range -9.50 D to -1.50 D) and 0.03±0.38 D (range -1.50 D to 0.88 D), respectively. The mean amount of refraction change in diopters before and after the surgery was 5.05±1.81 D (range 1.25 to 8.88 D).
The repeatability of the EKRs in different zones, SimK and K net (Table 1) In both the untreated group and the post-LASIK group, all the S w , 2.77S w and CV w values were less than 0.22 D, 0.62 D and 1%, respectively, and all ICCs were higher than 0.978. The Pentacam HR performed high repeatability of all the EKRs, SimK and K net values from both untreated eyes and post-LASIK eyes.
In the post-LASIK group, it indicated that the trend from EKR1 to EKR7 gradually decreased to EKR4 (39.04±1.67 D) as the minimum, then increased to EKR7 (Fig 2). There was no statistical significance between EKR4 and EKR4.5 (P = 0.396), EKR3 and EKR4.5 (P = 0.559). K CHM was 38.40±1.86 D. The differences between EKR4 and K CHM , EKR7 and K CHM , K net and K CHM , K m and K CHM , SimK and K net were 0.

Discussion
The EKR assessment, advanced by Holladay and co-workers, was used to assess the total corneal power after corneal refractive surgery. In the study by Holladay et al [14], the Pentacam   [15] used the Pentacam HR to measure 100 untreated eyes and found that the repeatability of EKR maps and the Holladay report Equivalent K1, K2 and Km (EKR4.5) were poor. In contrast, in the present study, EKRs in all zones from untreated eyes were repeatable. According to McAlinden et al, only the EKR in 4.5 mm was evaluated, and we evaluated the EKRs in all central corneal zones. Moreover, this study was the first to report the repeatability of EKRs in all zones in post-LASIK eyes, and the results for the EKRs were repeated in all central corneal zones. Interestingly, in the post-LASIK group, it indicated that the trend from EKR1 to EKR7 gradually decreased to EKR4 as the minimum, then increased to EKR7 (Fig 2). This result was different from the studies by Falavarlani et al [22] (35 eyes after PRK) and Savini et al [23]  Currently, there is no generally accepted device that can exactly measure and calculate the total corneal power after corneal refractive surgery. The clinical history method, which was presented in 1989 by Holladay et al, has been regarded as the gold standard in several published studies [13,14,[22][23][24]. In our study, K CHM was smaller than all the EKRs, and compared to EKR4 (the smallest among all the EKRs), the difference was 0.64±0.50 D (P<0.001). The K m obtained by the IOLMaster was equal to EKR4 and similar to SimK. We compared our study with Falavarjani et al [22] and Savini et al [23]. The K CHM , SimK and EKR in our study were higher than in Savini et al, but lower than in Falavarjani et al. One possibility may be due to the different preoperative refraction of the subjects. In our study, the mean preoperative spherical equivalent was -5.01±1.91 D (range -9.50 D to -1.50 D). According to Savini et al and In the event of similar mean corneal power, a higher refractive error will lead to a larger amount of surgical correction. This leads to a flatter cornea with smaller postoperative corneal power. Based on all three studies, we speculate that the EKR overestimates the corneal power in post-LASIK eyes. Two observations support this notion. First, in our study the EKR4 (the smallest among all the EKRs) was 0.6 to 0.7 D higher than the gold standard K CHM . Second, the SimK, K m , EKR4 and EKR4.5 in our study were very close to each other (the same results were shown by Savini et al and Falavajani et al). As SimK and K m were calculated by the same formula (SimK = 1.3375/r anterior ), which explains why no statistical significance was observed between SimK and K m . For untreated eyes, we could obtain accurate results using this formula; however, for eyes after corneal refractive surgery, because of the change of ratio between the posterior and anterior corneal curvature, the corneal power was overestimated [25,26]. And in our study, EKR4 (the minimum of EKRs) was equal to K m , and similar to SimK. Therefore, we conclude that EKR4 also overestimates the corneal power in post-LASIK eyes.
In our study, SimK was 1.30±0.13 D higher than K net in the untreated group, and in the post-LASIK group the difference increased to 1.68±0.42 D. The difference was attributed to two factors. First, the two parameters were calculated using a different corneal refractive index based on different assumptions. For SimK, the standard corneal refractive index (1.3375) was adopted considering the cornea as a thin lens; for K net , the real corneal refractive index (1.376) and aqueous refractive index (1.336) were adopted. Savini et al [27] used the Pentacam to measure 71 preoperative eyes, and found that SimK was 1.25 D higher than K net , which was very close to our study (1.30±0.13 D). Thus, we believe that this factor might cause the deviation of 1.2 to 1.3 D. It should be noted that K net , which is the sum of anterior and posterior corneal power (without consideration of CCT), was different from the total corneal power based on the Gaussian thick lens formula (with consideration of CCT). The difference, which was between 0.12 to 0.25 D [20], could be regarded as acceptable in the clinic. Second, for post-LASIK eyes the ratio of the posterior and anterior corneal curvature is considerably different from of the Gullstrand model eye value of 0.883. Therefore, the standard corneal refractive index based on the Gullstrand model eye caused the wrong calculation, explaining the deviation of approximately 0.4 D.
There were two limitations to our study. First, we only studied untreated eyes and myopic post-LASIK eyes, the eyes after RK and hyperopia corneal refractive surgery were not involved. Second, the repeatability of the clinical history method was not confirmed. The clinical history method has been suggested not to be the most accurate for the prediction of IOL power after corneal refractive surgery [28].
In conclusion, the EKR obtained by the Pentacam HR were repeatable in both untreated eyes and post-LASIK eyes. Compared to the clinical history method, EKRs generally overestimated the corneal power in post-LASIK eyes. EKR4 and EKR4.5 were closest to the clinical history method, but were still overestimated by 0.6 to 0.7 D.