With progressively improving visual rehabilitation following cataract surgery, patient expectations of favorable refractive outcomes have increased, ultimately requiring accurate preoperative surgical planning.1 The innovation of intraocular lenses (IOLs) that take into account corneal morphology to correct for preoperative astigmatism has become an option in surgical assessments, but clear corneal incisions (CCIs) and their impact on corneal curvature may induce aberrant postoperative refractive results.2
There have been attempts to minimize surgically induced astigmatism (SIA) by a variety of preoperative calculators and intraoperative techniques.3 Femtosecond laser–assisted cataract surgery (FLACS) was introduced to hopefully induce less SIA through self-sealing triplanar incisions. Preliminary studies have demonstrated that FLACS offers more accurate anterior capsulotomy circularity,4,5 less phacoemulsification time,5 and a reduced need for phacoemulsification energy,5,6 notably in patients with dense nuclear sclerotic cataracts.7
The utility of FLACS in regard to minimizing SIA has shown mixed results. Zhu et al.8 and Serrao et al.9 noted a significantly larger magnitude of SIA in cases where CCIs were created with the femtosecond laser compared to manual incisions. These results were postulated to have been attributed to postoperative CCI edema or due to imprecise localization of the femtosecond laser incisions, namely being closer to the central corneal axis than anticipated.8 In contrast, Mastropasqua et al.10 and a follow-up study by Serrao et al.11 found less SIA with CCIs created by a femtosecond laser. Finally, Nagy et al.,12 Diakonis et al.,13 Ferreira et al.,14 and Férnandez et al.15 all found no difference in SIA between their manual and femtosecond laser cohorts.
Literature investigating SIA between manual and femtosecond laser CCIs has studied the LenSx8,10,12–15 or IntraLase iFS9,11 platforms. To our knowledge, there have been no clinical investigations on the effect of femtosecond laser CCIs created with the Catalys Precision Laser System (Johnson & Johnson Vision, Santa Ana, CA). The purpose of this study was to compare the effect of postoperative corneal morphology and SIA between CCIs created manually or with the Catalys femtosecond laser in a single Canadian surgical center.
Patients and Methods
Study Design
In this retrospective cohort analysis, patients undergoing cataract extraction and IOL implantation from June to September 2018 who were 18 years or older were included from a single high-volume cataract center. This study adhered to the tenets of the Declaration of Helsinki and was approved by the William Osler Health System Research Ethics Board (ID No. 18-0025). All patients completed a written consent form to complete a 1-month and 3-month follow-up for the purposes of this study.
Exclusion criteria were: (1) history of ocular trauma or surgery, (2) previous corneal pathology (ie, endothelial dystrophy, symptomatic superficial punctate keratitis, pterygium, or current corneal ulcer), (3) preoperative corneal astigmatism greater than ±3.00 diopters (D), (4) history of systemic disease affecting ocular functioning, (5) concurrent ocular disease (ie, glaucoma, strabismus, retinopathy, or macular degeneration), and (6) poorly dilating pupils contraindicating FLACS. Patients were removed from initial consideration from inclusion if major intraoperative or postoperative complications occurred, corneal incisions were made incorrectly during surgery, or their postoperative Snellen corrected distance visual acuity (CDVA) was worse than 20/40.16 A sample size calculation was implemented for 64 patients and 104 eyes, which met appropriate power for an alpha value of 0.05.
Surgical Technique
Surgical procedures were performed by four experienced ophthalmologists (SS, EST, HC, and RM) using standard surgical equipment. CCIs were created either manually with a 2.5-mm keratome blade or with the Catalys femtosecond laser. The primary manual and side-port incisions were made at the limbus anterior to conjunctival vascular arcades. Traditional phacoemulsification (Whitestar Signature PRO; Johnson & Johnson Vision) completed the cataract extraction. Phacoemulsification parameters were kept consistent between the manual and femtosecond laser cohorts; there were no differences in preoperative and postoperative management techniques between the two groups.
Femtosecond Laser Platform
After docking with the suction ring onto the patient eye, the surgeon manually adjusted the docking mechanism toward the limbus prior to making the primary incision and side-port guided by the real-time imaging system. The following incision parameters were used throughout the study: single intrastromal at a 9-mm optical zone, 20% uncut anterior/posterior portion, 90° side cut angle, 5 µm horizontal spot spacing, 10 µm vertical spot spacing, 5.0 µJ pulse energy, 30% anterior line distance, and a central line density score of 4. The laser-assisted capsulotomy was a standardized template for a 5-mm diameter scanned capsule with a 600-µm incision depth, horizontal spot spacing of 4 µm, vertical spot spacing of 9 µm, and a pulse energy of 4 µJ. The standardized lens fragmentation template was quadrant softening, with a diameter of 5 mm, segmented soft spacing of 200 µm, grid spacing of 500 µm, horizontal spot spacing of 10 µm, vertical spot spacing of 40 µm, anterior pulse energy of 8 µJ, posterior pulse energy of 10 µJ, anterior capsule safety margin of 500 µm, and a posterior capsule safety margin of 650 µm. The standardized lens fragmentation template included two segmentation repetitions with a limited diameter of 5 mm.
Right eye main incisions were made at 200° with a limbus offset of 0.3 mm, width of 2.5 mm, length of 1.5 mm, no uncut region, 20% uncut anterior/posterior portion, 25 µm uncut central length, 30% anterior depth plane, 70% posterior depth plane 100° anterior side cut angle, and a posterior side cut angle of 45°. The same parameters were used for the right eye side-port incision at 135° axis, 0.8 mm width, 1.2 mm depth, and an anterior side cut angle of 60°.
Left eye main incisions were made at 20° with a limbus offset of 0.3 mm, width of 2.5 mm, length of 1.5 mm, no uncut region, 20% uncut anterior/posterior portion, 25 µm uncut central length, 30% anterior depth plane, 70% posterior depth plane, 100° anterior side cut angle, and a posterior side cut angle of 45°. The same parameters were used for the left eye side-port incision at 315° axis, 0.8 mm width, 1.2 mm depth, and an anterior side cut angle of 60°.
Data Collection
Demographic parameters, comorbidities, and optical biometric measurements taken by the IOLMaster 700 (Carl Zeiss Meditec, Jena, Germany) were recorded. CDVA, intraocular pressure (IOP), corneal topography (Pentacam; Oculus Optikgeräte GmbH, Wetzlar, Germany), central corneal thickness (CCT) (Pentacam), total corneal irregular astigmatism index (Pentacam), irregular corneal astigmatism index (Corneal Analyzer OPD-Scan III; Nidek Co. Ltd, Gamagori, Japan), and anterior segment optical coherence tomography (AS-OCT) images (Visante; Carl Zeiss Meditec) were collected from the preoperative and postoperative 1-week (POW1), 1-month (POM1), and 3-month (POM3) time points. AS-OCT images were manually examined by the same technician to determine the thickness of the CCI and the perpendicular distance from the central axis to the external wound opening (EXD). Intraoperative events and effective phacoemulsification times were recorded.
Primary and Secondary Outcomes
The primary outcome was the mean SIA at the POM3 follow-up, determined by comparing the postoperative Pentacam corneal keratometry measurements at outcome to the preoperative Pentacam keratometry measurements. Secondary outcomes included: CDVA, IOP, CCT, Pentacam total corneal irregular astigmatism, OPD irregular astigmatism index, CCI thickness, EXD, and effective phacoemulsification time.
Statistical Analysis
Microsoft Excel and SPSS software (version 23.0; IBM Corporation, Armonk, NY) were used to perform all statistical testing, with a P value of .05 used to indicate statistical significance. A generalized estimating equation model accounting for within-patient correlation was used. The Alpins vector method, first described in 1993 by Alpins17 (VECTrAK program, version 2.4.2), was used to calculate the SIA. Correlations between SIA and CCI and EXD were tested with the Pearson correlation coefficient.
Results
Overall, 104 eyes of 64 patients undergoing cataract extraction and IOL implantation procedures completed by four experienced ophthalmologists were considered for inclusion. All surgeons adhered to the same aforementioned femtosecond laser and manual incision parameters; no statistical differences were noted between surgeons (Table A, available in the online version of this article). No intraoperative events occurred, and no patients noted postoperative complications meeting exclusion criteria. Table 1 summarizes the baseline parameters between the manual and femtosecond laser CCI cohorts. Of note, there was a significantly longer axial length in the FLACS cohort compared to the manual group (P = .04); however, this can be attributed to a longer posterior chamber with a comparable anterior chamber depth, and therefore was not of concern in our analysis (Table 1).
Table 2 summarizes the differences in intraoperative effective phacoemulsification time and postoperative SIA, CDVA, IOP, CCT, total corneal irregular astigmatism, and irregular astigmatism index at POW1, POM1, and POM3 between the manual and FLACS cohort. Notably, there was a significantly larger mean SIA seen following femtosecond laser CCIs compared to manual incisions at POW1 (manual: 0.51 ± 0.44 D, FLACS: 0.72 ± 0.49 D, P = .02) and POM1 (manual: 0.46 ± 0.36 D, FLACS: 0.64 ± 0.53 D, P = .04). However, at POM3, there was no significant difference in SIA between either cohort (Figure 1). No significant differences or trends were seen in axis of astigmatism, CCT, total corneal irregular astigmatism, or irregular astigmatism index at any postoperative visit (P > .05). CCT, total corneal irregular astigmatism, and irregular astigmatism index values returned to baseline levels by POM3 in both cohorts. No significant difference in effective phacoemulsification time was seen (P = .59).
Table 3 summarizes the postoperative CCI morphology as determined by the AS-OCT at POM1 and POM3. The manual cohort trended to have a larger CCI thickness compared to the FLACS group, with a significantly thicker manual incision site noted at POM3 (manual: 744.26 ± 86.03 µm, FLACS: 703.56 ± 56.75 µm, P = .006). No significant difference in the position of the CCI compared to the central axis was seen at POM1 (P = .17) and POM3 (P = .44). Figure 2 demonstrates the association between SIA and EXD. There was a significant correlation seen between SIA and EXD for both cohorts, with more centrally aligned incisions demonstrating larger postoperative SIA values (P = .01, r = −0.434).
Discussion
The results of our study demonstrate a significantly higher mean SIA in the femtosecond laser group at POW1 (P = .02) and POM1 (P = .04), but no significant difference in SIA between the manual and femtosecond laser groups at POM3 (P = .11). Losing significance coincides with the literature investigating the healing properties of femtosecond laser CCIs, which demonstrate less posterior wound retraction and lower endothelial misalignment by the POM3 follow-up visit.6,18 Of note, the POM3 median SIA was noticeably lower in the femtosecond laser cohort compared to the POM3 mean SIA. Figure A (available in the online version of this article) demonstrates how outliers affected the mean SIA more in the femtosecond laser group compared to a lack of outliers in the manual cohort, leading to a right-skewed FLACS mean SIA. The manual group also demonstrated right-skewed deviation; however, this was not to the same degree as the femtosecond laser group. This finding suggests that a comparison of SIA medians would provide a more accurate measure, and one that should be included in future studies for a more representative statistical presentation.
AS-OCT imaging is a useful tool in demonstrating the positive impact of femtosecond laser CCIs in comparison to manual incisions with regard to corneal morphological properties, notably to determine a possible etiology for the significantly larger SIA in the early postoperative period. Figure B (available in the online version of this article) outlines how the CCI thickness and EXD parameters were calculated by the trained technician. The FLACS mean CCI thickness trended to be smaller throughout the postoperative period, notably demonstrating a significantly thinner value at POM3 (P = .006). In conjunction with the trend for a reduced CCT thickness, an edematous etiology for the increased SIA in the femtosecond laser CCI cohort during the early postoperative period is unlikely. Zhu et al.8 postulated that inaccurate positioning of the LenSx femtosecond laser docking mechanism to be more anterior than expected may have contributed to the significantly larger SIA in their 2017 analysis. However, the EXD was not significantly different at POM1 (P = .17) or POM3 (P = .44) in our analysis. Therefore, the localization of the CCI incision with respect to the central corneal axis was similar and not a source of corneal astigmatic change. This conclusion is consistent with the 2017 study by Bala et al.,19 who determined that femtosecond laser primary CCIs were created in close proximity to their expected location. Bala et al. did note that despite reproducibility in terms of incision length and width, secondary incisions were subject to variable eye tilt and displacement of the docking mechanism of the LenSx platform.19 The direct impact of the secondary incision on SIA is an area of interest for further investigation to determine the etiology of early postoperative increased SIA with Catalys FLACS.
Introduced in 1993,17 the Alpins vector method for SIA analysis has demonstrated accurate modeling for determining the magnitude of keratometric change and axis of astigmatism alteration following cataract surgery. The 2004 report by Alpins and Goggin20 graphically outlined the utility of vector analysis and how to achieve a target astigmatism following cataract surgery. By comparing the postoperative difference in flat and steep keratometry to the difference in preoperative keratometry axes, and noting the axis of astigmatism for each time point, the SIA and axis of astigmatic change can be calculated (Figure C, available in the online version of this article).20 By using a monofocal aspheric IOL implant centered in the capsular bag that does not correct for preoperative astigmatism, the overall change from preoperative to postoperative astigmatism was attributed to SIA. A 2001 clinical evaluation of astigmatism correction through laser refractive surgery by Alpins determined that the vector method can determine SIA through corneal keratometry values; however, underestimation of SIA may be seen due to remodeling of the cornea following surgery.21 With the refined vector analysis model, our statistical analysis offers a reliable investigation of SIA following manual and femtosecond laser CCIs.
There are several limitations to this study. Multiple patients had both eyes enrolled in this study, therefore introducing intraobserver correlation. To limit bias because both eyes are not truly independent, a linear regression model taking into account within-patient correlation was used in statistical analysis. The sample size calculation implemented also took this factor into account and met appropriate power with 64 patients and 104 cases. Second, the measurement of AS-OCT images was subjective to the skill of the technician. To minimize any variability between measurements, the same skilled technician manually measured the CCI thickness and EXD values for each image. For quality control, one author (AP) reviewed every AS-OCT image to confirm consistency across all readings. Finally, inclusion of multiple surgeons introduces variability between cases, notably for the effective phacoemulsification time outcome. Each surgeon was experienced with using the Catalys femtosecond laser and adhered to the protocol for each case.
There was no significant difference in SIA between FLACS with the Catalys femtosecond laser and manual cataract surgery at POM3. Femtosecond laser CCIs led to significantly smaller CCI thickness, and trended toward reduced CCT and comparable CCI position compared to manual corneal incisions.
References
- Tipotsch-Maca SM, Varsits RM, Ginzel C, Vecsei-Marlovits PV. Effect of a multimedia-assisted informed consent procedure on the information gain, satisfaction, and anxiety of cataract surgery patients. J Cataract Refract Surg. 2016;42(1):110–116. doi:10.1016/j.jcrs.2015.08.019 [CrossRef]26948785
- Koch DD, Wang L, Abulafia A, Ofir S, Assia EI. Surgically induced astigmatism. J Refract Surg. 2015;31(8):565. doi:10.3928/1081597X-20150728-03 [CrossRef]26248351
- Holladay JT, Pettit G. Improving toric intraocular lens calculations using total surgically induced astigmatism for a 2.5mm temporal incision. J Cataract Refract Surg. 2019;45(3):272–283. doi:10.1016/j.jcrs.2018.09.028 [CrossRef]
- Day AC, Gartry DS, Maurino V, Allan BD, Stevens JD. Efficacy of anterior capsulotomy creation in femtosecond laser-assisted cataract surgery. J Cataract Refract Surg. 2014;40(12):2031–2034. doi:10.1016/j.jcrs.2014.07.027 [CrossRef]25305149
- Reddy KP, Kandulla J, Auffarth GU. Effectiveness and safety of femtosecond laser-assisted lens fragmentation and anterior capsulotomy versus the manual technique in cataract surgery. J Cataract Refract Surg. 2013;39(9):1297–1306. doi:10.1016/j.jcrs.2013.05.035 [CrossRef]23988242
- Wang X, Zhang Z, Li X, et al. Evaluation of femtosecond laser versus manual clear corneal incisions in cataract surgery using spectral-domain optical coherence tomography. J Refract Surg. 2018;34(1):17–22. doi:10.3928/1081597X-20171109-01 [CrossRef]29315437
- Chen X, Yu Y, Song X, Zhu Y, Wang W, Yao K. Clinical outcomes of femtosecond laser-assisted cataract surgery versus conventional phacoemulsification surgery for hard nuclear cataracts. J Cataract Refract Surg. 2017;43(4):486–491. doi:10.1016/j.jcrs.2017.01.010 [CrossRef]28532933
- Zhu S, Qu N, Wang W, et al. Morphologic features and surgically induced astigmatism of femtosecond laser versus manual clear corneal incisions. J Cataract Refract Surg. 2017;43(11):1430–1435. doi:10.1016/j.jcrs.2017.08.011 [CrossRef]29223232
- Serrao S, Lombardo G, Schiano-Lomoriello D, Rosati M, Lombardo M. Preliminary investigation of corneal wavefront aberration following femtosecond laser clear corneal incision for cataract surgery. Eur J Ophthalmol. 2014;24(6):842–849. doi:10.5301/ejo.5000485 [CrossRef]24846626
- Mastropasqua L, Toto L, Mastropasqua A, et al. Femtosecond laser versus manual clear corneal incision in cataract surgery. J Refract Surg. 2014;30(1):27–33. doi:10.3928/1081597X-20131217-03 [CrossRef]24864325
- Serrao S, Giannini D, Schiano-Lomoriello D, Lombardo G, Lombardo M. New technique for femtosecond laser creation of clear corneal incisions for cataract surgery. J Cataract Refract Surg. 2017;43(1):80–86. doi:10.1016/j.jcrs.2016.08.038 [CrossRef]28317683
- Nagy ZZ, Dunai A, Kránitz K, et al. Evaluation of femtosecond laser-assisted and manual clear corneal incisions and their effect on surgically induced astigmatism and higher-order aberrations. J Refract Surg. 2014;30(8):522–525. doi:10.3928/1081597X-20140711-04 [CrossRef]25325892
- Diakonis VF, Yesilirmak N, Cabot F, et al. Comparison of surgically induced astigmatism between femtosecond laser and manual clear corneal incisions for cataract surgery. J Cataract Refract Surg. 2015;41(10):2075–2080. doi:10.1016/j.jcrs.2015.11.004 [CrossRef]26703282
- Ferreira TB, Ribeiro FJ, Pinheiro J, Ribeiro P, O'Neill JG. Comparison of surgically induced astigmatism and morphologic features resulting from femtosecond laser and manual clear corneal incisions for cataract surgery. J Refract Surg. 2018;34(5):322–329. doi:10.3928/1081597X-20180301-01 [CrossRef]29738588
- Fernández J, Rodríguez-Vallejo M, Martínez J, Tauste A, Piñero DP. Prediction of surgically induced astigmatism in manual and femtosecond laser-assisted clear corneal incisions. Eur J Ophthalmol. 2018;28(4):398–405. doi:10.1177/1120672117747017 [CrossRef]29973075
- Haigis W, Lege B, Miller N, Schneider B. Comparison of immersion ultrasound biometry and partial coherence interferometry for intraocular lens calculation according to Haigis. Graefes Arch Clin Exp Ophthalmol. 2000;238(9):765–773. doi:10.1007/s004170000188 [CrossRef]11045345
- Alpins NA. A new method of analyzing vectors for changes in astigmatism. J Cataract Refract Surg. 1993;19:524–553. doi:10.1016/S0886-3350(13)80617-7 [CrossRef]8355160
- Grewal DS, Basti S. Comparison of morphologic features of clear corneal incisions created with a femtosecond laser or a keratome. J Cataract Refract Surg. 2014;40(4):521–530. doi:10.1016/j.jcrs.2013.11.028 [CrossRef]24568722
- Bala C, Chan T, Meades K. Factors affecting corneal incision position during femtosecond laser-assisted cataract surgery. J Cataract Refract Surg. 2017;43(12):1541–1548. doi:10.1016/j.jcrs.2017.09.024 [CrossRef]
- Alpins NA, Goggin M. Practical astigmatism analysis for refractive outcomes in cataract and refractive surgery. Surv Ophthalmol. 2004;49(1):109–122. doi:10.1016/j.survophthal.2003.10.010 [CrossRef]14711444
- Alpins N. Astigmatism analysis by the Alpins method. J Cataract Refract Surg. 2001;27(1):31–49. doi:10.1016/S0886-3350(00)00798-7 [CrossRef]11165856
Patient Demographics and Baseline Characteristicsa
| Parameter | Manual Group (n = 54) | FLACS Group (n = 50) | P |
|---|
| Age (y) | 70.78 ± 8.87 | 69.40 ± 11.56 | .50 |
| CDVA (logMAR) | 0.41 ± 0.24 | 0.43 ± 0.25 | .70 |
| UDVA (logMAR) | 0.50 ± 0.31 | 0.54 ± 0.28 | .44 |
| IOP (mm Hg) | 14.41 ± 3.63 | 13.66 ± 3.72 | .30 |
| AL (mm) | 23.64 ± 1.22 | 24.18 ± 1.38 | .04 |
| ACD (mm) | 3.11 ± 0.34 | 3.24 ± 0.44 | .12 |
| K1 (D) | 43.27 ± 1.84 | 43.18 ± 1.83 | .80 |
| K2 (D) | 44.25 ± 1.94 | 44.11 ± 1.80 | .71 |
| CCT (µm) | 536.26 ± 70.43 | 533.02 ± 37.07 | .78 |
| Total corneal irregular astigmatism (µm) | 0.27 ± 0.21 | 0.24 ± 0.13 | .41 |
| Irregular astigmatism index | 0.46 ± 0.04 | 0.46 ± 0.06 | .96 |
Statistical Analysis of Intraoperative and Postoperative Outcomesa
| Parameter/Time | Manual Group (n = 54) | FLACS Group (n = 50) | P |
|---|
| Effective phacoemulsification time (s) | 39.55 ± 19.84 | 37.40 ± 20.44 | .59 |
| SIA (D) | | | |
| 1 week | 0.51 ± 0.44 | 0.72 ± 0.49 | .02 |
| 1 month | 0.46 ± 0.36 | 0.64 ± 0.53 | .04 |
| 3 months | 0.45 ± 0.28 | 0.57 ± 0.46 | .11 |
| CDVA (logMAR) | | | |
| 1 week | 0.16 ± 0.13 | 0.11 ± 0.16 | .08 |
| 1 month | 0.13 ± 0.12 | 0.08 ± 0.12 | .03 |
| 3 months | 0.10 ± 0.09 | 0.07 ± 0.09 | .19 |
| UDVA (logMAR) | | | |
| 1 week | 0.23 ± 0.16 | 0.20 ± 0.18 | .16 |
| 1 month | 0.16 ± 0.12 | 0.11 ± 0.13 | .02 |
| 3 months | 0.10 ± 0.09 | 0.09 ± 0.10 | .18 |
| IOP (mm Hg) | | | |
| 1 week | 14.04 ± 3.84 | 13.42 ± 2.96 | .36 |
| 1 month | 13.31 ± 3.28 | 12.94 ± 2.58 | .52 |
| 3 months | 14.02 ± 3.21 | 12.48 ± 2.27 | .06 |
| CCT (µm) | | | |
| 1 week | 553.8 ± 42.5 | 548.6 ± 47.4 | .56 |
| 1 month | 546.52 ± 41.56 | 539.06 ± 40.61 | .36 |
| 3 months | 544.22 ± 37.23 | 534.98 ± 34.86 | .20 |
| Total corneal irregular astigmatism (µm) | | | |
| 1 week | 0.32 ± 0.24 | 0.30 ± 0.13 | .52 |
| 1 month | 0.29 ± 0.21 | 0.26 ± 0.12 | .43 |
| 3 months | 0.29 ± 0.20 | 0.26 ± 0.13 | .36 |
| Irregular astigmatism index | | | |
| 1 week | 0.46 ± 0.08 | 0.49 ± 0.10 | .10 |
| 1 month | 0.47 ± 0.04 | 0.47 ± 0.05 | .86 |
| 3 months | 0.47 ± 0.04 | 0.47 ± 0.05 | .73 |
Postoperative CCI Morphologya
| Parameter | Manual Group (n = 54) | FLACS Group (n = 50) | P |
|---|
| CCI thickness (µm) | | | |
| 1 month | 743.87 ± 133.27 | 702.52 ± 124.60 | .11 |
| 3 months | 744.26 ± 86.03 | 703.56 ± 56.75 | .006 |
| CCI distance (mm) | | | |
| 1 month | 5.77 ± 0.47 | 5.65 ± 0.44 | .17 |
| 3 months | 5.64 ± 0.51 | 5.77 ± 0.35 | .44 |
SIA Stratified by Surgeon
| Surgeon POM3 | Manual Group | FLACS Group | P |
|---|
|
|
|---|
| No. of Patients | SIA | No. of Patients | SIA |
|---|
| SS | 5 | 0.36 ± 0.06 | 10 | 0.47 ± 0.37 | .69 |
| EST | 28 | 0.42 ± 0.25 | 24 | 0.57 ± 0.45 | .15 |
| HC | 7 | 0.56 ± 0.35 | 9 | 0.59 ± 0.49 | .92 |
| RM | 14 | 0.44 ± 0.30 | 7 | 0.75 ± 0.62 | .13 |