Investigation of corneal epithelial thickness and irregularity by optical coherence tomography after transepithelial photorefractive keratectomy

ABSTRACT

Clinical relevance

Corneal epithelial healing after refractive surgery is a clinically significant issue, especially for surface ablation procedures, and this can be monitored using optical coherence tomography (OCT).

Background

The aim of this work is to investigate the corneal epithelial thickness and irregularity by OCT after transepithelial photorefractive keratectomy (t-PRK) and analyse its correlation with visual and refractive outcomes.

Methods

Patients aged ≥18 years with myopia, with or without astigmatism, who underwent t-PRK between May 2020 and August 2021 were included. All participants were subjected to complete ophthalmic examinations and OCT pachymetry at every follow-up visit. Patients were followed up at 1 week and 1, 3, and 6 months postoperatively.

Results

A total of 67 patients (126 eyes) were enrolled in this study. One month postoperatively, spherical equivalent refraction and visual acuity achieved preliminary stability. However, central corneal epithelial thickness (CCET) and standard deviation of the corneal epithelial thickness (SD_cet) took 3–6 months to progressive recovery. Patients with higher baseline spherical equivalent refraction were associated with slower epithelial recovery. At every follow-up time point, a significant superior-inferior difference in the minimum corneal epithelial thickness area was observed. Higher stromal haze was correlated with higher spherical equivalent refraction (both baseline and residual) but had no relation with visual outcomes. There was a significant correlation between higher CCET with a better uncorrected distance visual acuity and lower corneal epithelial thickness irregularity.

Conclusions

CCET and SD_cet measured by OCT seem to be a good auxiliary indicator for reflecting the status of corneal wound recovery after t-PRK surgery. However, a well-designed randomised control study is needed to confirm the study results.

KEYWORDS:

• Corneal epithelial thickness
• myopia
• optical coherence tomography
• transepithelial photorefractive keratectomy

Introduction

The concept of transepithelial photorefractive keratectomy (t-PRK) surgery was first introduced in the 1990s.¹ The original, classic two-step t-PRK process comprises ablation of the corneal epithelium via laser phototherapeutic keratectomy, followed by stromal ablation via photorefractive keratectomy (PRK).^2,3 A single-step method was launched in 2007, as the Amaris laser platform (SCHWIND eye-tech-solutions GmbH, Kleinostheim, Germany).

Surface ablation procedures, such as conventional PRK or t-PRK, prevent the flap-related complications of laser in situ keratomileusis and provoke a lower rate of postoperative corneal ectasia.^4–6 However, conventional surface ablation procedures are associated with significant postoperative pain, slower visual recovery, and corneal stromal hazing, dissuading patients from undergoing surgery.^7,8 Corneal epithelial healing after surgery is a clinically significant issue, especially for surface ablation procedures. The average time for re-epithelialization after t-PRK surgery is approximately 2–3 days.^9–11 A faster epithelial healing rate in t-PRK compared to conventional PRK methods was noted in previous studies; nevertheless, this did not result in better visual or refractive outcomes.^7,9–12

In contrast to corneal re-epithelialization, the recovery of corneal epithelial thickness takes 3–6 months, with the stable thickness achieved between 3 and 12 months after refractive surgery.^13,14 Postoperative epithelial hyperplasia and stromal wound healing have been suggested to be responsible for postoperative refractive regression.^15–19 A significant correlation has been reported between changes in spherical equivalent and epithelial thickness, from before surgery to 6 months postoperatively, in the paracentral and peripheral zones after PRK for myopia.¹³

Clinical methods of corneal epithelial thickness measurement include in vivo confocal microscopy, very high-frequency digital ultrasound, optical coherence tomography (OCT), and Scheimpflug tomography. Among these technologies, OCT provides a noncontact and wider area of measurement, which is more applicable for investigating the corneal epithelial thickness healing patterns.²⁰ The corneal epithelial thickness mapping by OCT has shown high repeatability and reproducibility and is equivalent to other measuring techniques.^21–24 The aim of this study is to evaluate the corneal epithelial thickness and irregularity by OCT after t-PRK in myopic eyes and analyse its correlation with patients’ visual and refractive outcomes.

Methods

Participants and ophthalmic examinations

Patients aged ≥ 18 years with low, moderate, or high spherical equivalent (defined as ≥ −3.00 dioptres (D), < −3.00 D and ≥ −6.00D, and < −6.00D, respectively) and with or without astigmatism who underwent t-PRK in the Tri-Service General Hospital, Taipei, Taiwan, between May 2020 and August 2021, were included in this retrospective, non-randomised, clinical study. Corneal astigmatism was further classified into three types based on which degree the steepest meridian lies [against-the-rule (ATR): 0°–30° and 150°–180°; with-the-rule (WTR): 60°–120°; oblique astigmatism (OA): 31°–59° and 121°–149°].

This study adhered to the tenets of Declaration of Helsinki and was approved by the Institutional Review Board of Tri-Service General Hospital (TSGHIRB No.: A202105167). The requirement for patient informed consent was waived. All study procedures were performed following the relevant guidelines and regulations.

The exclusion criteria were preoperative corrected distance visual acuity (CDVA) less than 6/6, unstable refraction within a 3-month follow-up period preoperatively, severe ocular surface disease (dry eye disease, keratoconus, corneal opacity, corneal dystrophy, vernal keratoconjunctivitis, herpetic keratitis, and other ocular surface infections or inflammation), eyelid abnormalities or dysfunctions, any posterior segment abnormality, as well as the history of previous eye surgery, pregnancy, and systemic disorders such as diabetes mellitus, hypertension, and autoimmune diseases.

Patients were advised to discontinue contact lenses wearing for at least 4 weeks before the preoperative examinations and surgery. All patients underwent complete preoperative and postoperative ophthalmic examinations, including preoperative CDVA, postoperative uncorrected distance visual acuity (UDVA), manifest and automated refractions, slit-lamp biomicroscopy, non-contact tonometry, corneal topography (Orbscan; Bausch & Lomb, Inc., Laval, Canada), and OCT pachymetry (Avanti RTVue XR; Optovue, Inc., Fremont, CA, USA). Patients were arranged for outpatient follow-up visits at 1 week and at 1, 3, and 6 months postoperatively (each follow-up time point is defined as the time period from the surgical day [day 0]).

Epithelial thickness measured by optical coherence tomography

All eligible participants underwent OCT pachymetry at baseline and every follow-up visit. The measurement at 1 week follow-up was done after removal of the bandage contact lens and rest for at least 30 minutes. The OCT pachymetry report contained both a corneal pachymetry map and an epithelium map. Each image was assessed by the treating physician (CML or THW), and images with a signal strength index of < 30 were excluded from further analysis.

Epithelium maps consist of 25 sectors, including a 2-mm radius central zone and eight octants equally distributed within paracentral (2–5 mm), mid-peripheral (5–7 mm), and peripheral (7–9 mm) annular zones. The mean epithelial thickness was shown within each sector in epithelium maps. The areas of minimum and maximum thickness were indicated with symbols on the map (Figure 1). Epithelium statistics within the central 7 mm, including superior and inferior (2–7 mm) mean thickness, minimum (min) and maximum (max) thickness, standard deviation of the corneal epithelial thickness (SD_cet), and the difference between the minimum and the maximum thickness (min-max), were calculated and exported on the report of OCT pachymetry. The irregularity of corneal epithelial thickness was reflected by the min-max and SD_cet on the epithelium map.

Figure 1. An example of epithelium map from the OCT pachymetry report. The epithelium maps were drawn according to the default colour-coded scale. It consists of 25 sectors, including a 2-mm radius central zone and eight octants equally distributed within paracentral (2–5 mm), mid-peripheral (5–7 mm), and peripheral (7–9 mm) annular zones. Mean epithelial thickness was shown within each sector and the area of minimum and maximum thickness were indicated as * and + on the map, respectively.

Surgical procedure and postoperative regimen

All surgeries were performed by one of the two experienced cornea specialists (CML or THW) using the SCHWIND Amaris 1050RS laser system with SmartPulse technology (SmartSurf^ACE procedure). The ablation was based on algorithms calculated using the ORK-CAM (ORK-Custom Ablation Manager) software planning module. All procedures were performed with at least a 6.50-mm ablation zone. Topical anaesthesia with proparacaine hydrochloride 0.5% eye drops (Alcaine; Alcon, Inc., Geneva, Switzerland) was first applied to the ocular surface.

Corneal epithelial removal and spheroelliptical ablation of stroma were performed in a single step using an excimer laser with one ablation profile. After ablation, mitomycin-C 0.02% was applied to the stromal bed for 12–30 s, depending on the patient’s refractive error (12 s, 20 s, and 30 s for low, moderate, and high spherical equivalent group, respectively), followed by balanced salt solution irrigation. A bandage contact lens (PureVision2; Bausch & Lomb, Inc.) was placed to cover the ocular surface and removed at the 1-week follow-up visit.

Postoperatively, topical levofloxacin 0.5% (Cravit; Santen, Inc., Osaka, Japan) four times a day, and nepafenac 0.1% (Nevanac, Alcon, Inc.) three times a day were applied for 1 week. In addition, fluorometholone 0.1% (Flumetholon, Santen, Inc.) four times a day and preservative-free artificial tears [Optive Fusion® Lubricant Eye Drops (Unit Dose), Allergan Pharmaceuticals (Pty), Ltd.] every 2 h during the daytime were prescribed after surgery and gradually tapered over 3 months.

Statistical analyses

The main outcome measurements of the study included central corneal epithelial thickness (CCET), SD_cet, visual acuity (VA; logMAR), and spherical equivalent refraction at baseline and several postoperative time points (1 week, 1 month, 3 months, and 6 months from day 0) after t-PRK surgery. Unless stated otherwise, all parameters were assumed to be at the specific time point. The Kolmogorov-Smirnov test was used to test the normality of all variables. Categorical variables are reported as numbers and percentages, and continuous variables are reported as mean ± SD. Continuous variables were compared using an independent Student’s t-test; Pearson’s chi-squared and Fisher’s exact tests were used to evaluate the differences in categorical variables.

A generalised estimating equation was used to calculate the parameters. Correlation analyses between CCET, SD_cet, and haze with refraction, UDVA, and OCT pachymetry parameters were expressed as Spearman’s rank correlation coefficient. Statistical significance was defined as a p-value<0.05. All statistical analyses were performed using the IBM Statistical Package for the Social Sciences (SPSS) Statistics for Windows, Version 22.0. (IBM Corp., Armonk, NY, USA).

Results

A total of 67 patients (126 myopic eyes) who underwent t-PRK surgery at the Tri-Service General Hospital between May 2020 and August 2021 were retrospectively reviewed. These 67 patients included 41 men (61.2%) and 26 women (38.8%), with a mean age of 26.79 years (SD, 6.92; range, 18–50). Eight patients received t-PRK only in one eye. There were 40, 62, and 24 eyes with low, moderate, and high spherical equivalent refractions, respectively. All of the included eyes achieved a baseline CDVA of 6/6 or higher.

Table 1 shows the baseline demographic characteristics among the low, moderate, and high spherical equivalent groups. There was no significant difference between the preoperative CCET, type of astigmatism, and age between the groups. However, patient sex, sphere power, cylinder power, and spherical equivalent refraction were statistically different among groups.

Table 1. Baseline characteristics of myopic eyes that underwent transepithelial photorefractive keratectomy, according to spherical equivalent degree (n = 126).

	Overall (n = 126)	Low SE (≥ −3.00 D) (n = 40)	Moderate SE (< −3.00 D and ≥ −6.00D) (n = 62)	High SE (< −6.00D) (n = 24)	p-value
Sex, n (%)					0.006
Male	77(61.1)	31 (77.5)^c	37 (59.7)	9 (37.5)^c
Female	49(38.9)	9 (22.5)^c	25 (40.3)	15 (62.5)^c
Age (years)	26.98 ± 6.94 (18 to 50)	25.70 ± 5.71 (19 to 38)	26.67 ± 6.85 (19 to 45)	29.87 ± 8.37 (18 to 50)	0.068
CCET(μm)	53.66 ± 3.21 (47 to 66)	54.60 ± 3.75 (48 to 66)	53.44 ± 2.50 (47 to 60)	52.67 ± 3.57 (47 to 61)	0.129
Sphere (D)	−3.69 ± 2.01 (−10.50 to 0)	−1.66 ± 0.77^a,c (−3.00 to 0)	−3.77 ± 0.80^a,b (−5.50 to −2.25)	−6.86 ± 1.24^b,c (−10.50 to −5.25)	<0.001
Cylinder (D)	−1.08 ± 0.79 (−3.00 to 0)	−0.77 ± 0.62^c (−2.75 to 0)	−1.05 ± 0.73^b (−3 to 0)	−1.66 ± 0.91^b,c (−3 to 0)	<0.001
Astigmatism type, n (%)					0.279
Without	14(11.11)	5(12.5)	8(12.9)	1(4.17)
ATR	13(10.32)	6(15)	5(8.06)	2(8.33)
WTR	88(69.84)	26(65)	41(66.13)	21(87.5)
OA	11(8.73)	3(7.5)	8(12.9)	0
SE refraction (D)	−4.23 ± 2.14 (−11 to −0.63)	−2.05 ± 0.74^a,c (−3.00 to −0.63)	−4.29 ± 0.76^a,b (−5.88 to −3.13)	−7.69 ± 1.32^b,c (−11 to −6.08)	<0.001

ATR, Against-the-rule; CCET, central corneal epithelial thickness; OA, Oblique astigmatism; SE, spherical equivalent; WTR, With-the-rule.

Continuous variables are expressed as mean±SD (range); categorical variables are displayed as number (percentage).

^ap<0.05 for Low vs Moderate.

^bp<0.05 for Moderate vs High.

^cp<0.05 for Low vs High.

Table 2 summarises the overall changes in VA, refraction, CCET, and other parameters of OCT pachymetry at every follow-up visit, from baseline to 6 months postoperatively. As this study focused on the postoperative corneal epithelial remodelling, the baseline OCT parameters (except CCET) were not recorded. The mean CCET after surgery was significantly lower than that at baseline, up until 6 months postoperatively. The postoperative UDVA was relatively stable at 1 month and significantly superior to the baseline CDVA at 6 months.

Table 2. Changes of visual acuity, refraction, central corneal epithelial thickness, other parameters of OCT pachymetry, and haze from baseline to 6 months postoperatively.

	Baseline (n = 126)	1 week (n = 93)	1 month (n = 109)	3 month (n = 95)	6 month (n = 29)
VA (logMAR)*	0 ± 0	0.09 ± 0.11^a (0.061 to 0.11)	0.01 ± 0.06^b (−0.01 to 0.02)	−0.01 ± 0.08 (−0.03 to 0.003)	−0.06 ± 0.10^a,b,c (−0.09 to −0.03)
Sphere (D)	−3.69 ± 2.01 (−4.04 to −3.33)	−0.55 ± 0.89^a (−0.74 to −0.35)	0.10 ± 0.60^a,b (−0.02 to 0.22)	0.10 ± 0.59^a,b (−0.02 to 0.23)	−0.10 ± 0.69^a,b,c (−0.20 to 0.40)
Cylinder (D)	−1.08 ± 0.79 (−1.22 to −0.94)	−0.72 ± 0.61 (−0.85 to −0.58)	−0.57 ± 0.35^a,b (−0.64 to −0.50)	−0.48 ± 0.36^a,b (−0.56 to −0.41)	−0.34 ± 0.23^a,b (−0.44 to −0.24)
SE refraction (D)	−4.23 ± 2.14 (−4.61 to −3.85)	−0.91 ± 0.95^a (−1.12 to −0.70)	−0.19 ± 0.58^a,b (−0.30 to −0.07)	−0.14 ± 0.57^a,b (−0.25 to −0.02)	−0.07 ± 0.73^a,b (−0.39 to 0.25)
CCET (μm)	53.66 ± 3.2 (53.09 to 54.22)	45.59 ± 7.49^a (43.94 to 47.23)	45.47 ± 4.74^a (44.55 to 46.38)	51.70 ± 4.14^a,b,c (50.81 to 52.56)	55.96 ± 4.64^b,c,d (53.95 to 57.96)
Superior (μm)	-	38.96 ± 7.62 (37.29 to 40.64)	43.92 ± 5.53^b (42.85 to 44.99)	51.77 ± 4.71^b,c (50.83 to 52.81)	55.39 ± 5.17^b,c (53.16 to 57.63)
Inferior (μm)	-	45.71 ± 8.79 (43.78 to 47.64)	48.65 ± 4.63^b (47.75 to 49.54)	54.22 ± 4.91^b,c (53.43 to 55.08)	57.83 ± 3.54^b,c,d (56.30 to 59.36)
Min (μm)	-	25.20 ± 8.01 (23.44 to 26.95)	32.66 ± 8.54^b (31.00 to 34.31)	45.76 ± 5.08^b,c (44.77 to 46.90)	48.90 ± 5.66^b,c,d (47.30 to 52.44)
Max (μm)	-	59.98 ± 9.88 (57.81 to 62.15)	57.21 ± 5.78^b (56.09 to 58.33)	60.04 ± 4.38^c (59.11 to 60.96)	63.70 ± 5.01^c (61.53 to 65.86)
Min-Max (μm)	-	−34.78 ± 11.84 (−37.38 to −32.18)	−24.55 ± 10.05^b (−26.50 to −22.61)	−14.29 ± 5.12^b,c (−15.27 to −13.13)	−13.83 ± 4.94^b,c (−15.96 to −11.69)
SDcet (μm)	-	7.66 ± 3.29 (6.94 to 8.38)	4.79 ± 2.30^b (4.35 to 5.24)	2.88 ± 0.90^b,c (2.67 to 3.05)	2.80 ± 0.92^b,c (2.40 to 3.20)
Haze, n (%)	-	5(5.4; 1 to 11)	24(22; 14 to 30)	44(46.3; 37 to 57.7)^c	18(62.1; 43.3 to 80.9)^d

CCET, central corneal epithelial thickness; OCT, optical coherence tomography; SD_cet, standard deviation of corneal epithelial thickness; SE, spherical equivalent; VA, visual acuity.

Continuous variables are expressed as mean±SD (95% CI); categorical variables are displayed as number (percentage; 95% CI).

*indicate baseline corrected distance visual acuity and postoperative uncorrected distance visual acuity.

^ap<0.05 compared with baseline.

^bp<0.05 compared with 1 week.

^cp<0.05 compared with 1 month.

^dp<0.05 compared with 3 month.

All refractive parameters, including sphere power, cylinder power (except 1st week), and spherical equivalent refraction, showed a significant decrease from the baseline at every postoperative follow-up time point. The spherical equivalent refraction showed no statistical difference at the 1-month, 3-month, and 6-month time points. Considering other parameters of OCT pachymetry, the superior mean, inferior mean, and minimum corneal epithelial thickness gradually increased from 1 week to 6 months after refractive surgery.

The min-max and SD_cet gradually decreased postoperatively; they were significantly lower at the 3-month and 6-month points as compared to the 1-week and 1-month points. The temporal changes of VA, spherical equivalent refraction, CCET, and SD_cet are presented in Figure 2A–D.

Figure 2. The temporal changes of visual acuity, A: spherical equivalent refraction, B: central corneal epithelial thickness, C: and standard deviation of corneal epithelial thickness, D: Solid line, dotted line, short-dashed line, and long-dashed line represent values of the overall, low SE, moderate SE, and high SE group, respectively.

CCET, central corneal epithelial thickness; SDcet, standard deviation of corneal epithelial thickness; SE, spherical equivalent; UDVA, uncorrected distance visual acuity.

*indicate baseline corrected distance visual acuity and postoperative uncorrected distance visual acuity.

A subgroup analysis comparing postoperative parameters between the low, moderate, and high spherical equivalent groups at different follow-up time points (1-week, 1-month, and 3-month) was also performed. The results are presented in Table 3. At 1-week postoperatively, the mean CCET, UDVA, and spherical equivalent refraction showed no significant difference between groups. The min-max and SD_cet were significantly lower in the low spherical equivalent group than in the moderate and high groups. The cylinder power and maximum epithelial thickness were highest in the high spherical equivalent group.

Table 3. Postoperative variables at 1 week, 1 month, and 3 months among different spherical equivalent groups.

	Low SE (≥ −3.00 D) (n = 40)	Moderate SE (< −3.00 D and ≥ −6.00D) (n = 62)	High SE (< −6.00D) (n = 24)	p value
1 week
CCET (μm)	45.10 ± 6.08 (42.87 to 47.26)	45.03 ± 7.62 (42.90 to 47.60)	47.93 ± 9.56 (43.13 to 51.87)	0.410
UDVA (logMAR)	0.069 ± 0.11 (0.04 to 0.11)	0.085 ± 0.10 (0.65 to 0.13)	0.12 ± 0.14 (0.08 to 0.20)	0.358
Cylinder (D)	−0.59 ± 0.41^c (−0.89 to −0.53)	−0.66 ± 0.72 (−0.91 to −0.49)	−1.10 ± 0.52^c (−1.53 to −0.92)	0.022
SE refraction (D)	−0.68 ± 0.87 (−1.17 to −0.56)	−0.92 ± 0.90 (−1.53 to −0.84)	−1.33 ± 1.14 (−2.17 to −0.87)	0.094
Superior (μm)	39.77 ± 6.26 (37.80 to 42.91)	37.84 ± 7.69 (34.67 to 39.88)	40.13 ± 9.83 (36.25 to 45.53)	0.480
Inferior (μm)	44.37 ± 7.82 (41.43 to 47.09)	46.32 ± 9.17 (43.62 to 49.20)	46.87 ± 9.89 (41.01 to 50.44)	0.571
Min (μm)	27.93 ± 6.19 (25.38 to 29.97)	23.92 ± 8.34 (20.05 to 25.90)	22.87 ± 9.31 (18.90 to 27.43)	0.056
Max (μm)	56.17 ± 8.30^c (53.51 to 59.85)	61.54 ± 11.10 (58.13 to 64.78)	63.73 ± 7.26^c (60.41 to 67.37)	0.021
Min-Max (μm)	−29.00 ± 10.97^a,c (−33.02 to −24.98)	−38.48 ± 10.99^a (−41.82 to −35.14)	−40.72 ± 9.72^c (−45.56 to −35.89)	<0.001
SDcet (μm)	6.06 ± 2.68^a,c (5.15 to 7.11)	8.12 ± 2.97^a (7.53 to 9.33)	9.71 ± 3.81^c (7.80 to 11.30)	0.001
Haze, n (%)*	1 (3.2)	4 (9.1)	0	0.284
1 month
CCET (μm)	46.15 ± 4.69 (44.47 to 47.71)	44.95 ± 4.76 (43.67 to 46.13)	46.00 ± 4.84 (43.40 to 48.23)	0.462
UDVA (logMAR)	−0.01 ± 0.06^c (−0.03 to 0.01)	0.01 ± 0.06^b (−0.01 to0.03)	0.04 ± 0.05^b,c (0.02 to 0.06)	0.031
Cylinder (D)	−0.51 ± 0.30 (−0.60 to −0.41)	−0.58 ± 0.33 (−0.66 to −0.49)	−0.68 ± 0.53 (−0.91 to −0.37)	0.301
SE refraction (D)	−0.20 ± 0.59 (−0.28 to 0.12)	−0.23 ± 0.56 (−0.38 to −0.09)	−0.02 ± 0.67 (−0.53 to 0.27)	0.366
Superior (μm)	44.21 ± 5.63 (42.21 to 46.09)	44.41 ± 5.81 (43.01 to 46.04)	41.21 ± 3.09 (40.09 to 44.03)	0.142
Inferior (μm)	50.03 ± 3.39 (48.68 to 51.08)	48.09 ± 4.81 (46.79 to 49.28)	47.71 ± 5.90 (44.48 to 50.40)	0.112
Min (μm)	33.33 ± 7.91 (30.44 to 35.91)	33.52 ± 9.09 (31.22 to 35.93)	27.50 ± 5.83 (25.39 to 32.98)	0.051
Max (μm)	57.42 ± 5.10 (55.48 to 59.05)	57.05 ± 6.18 (55.64 to 59.04)	57.36 ± 5.98 (54.01 to 60.12)	0.953
Min-Max (μm)	−24.09 ± 9.34 (−27.30 to −20.88)	−23.53 ± 10.44 (−26.50 to −21.03)	−29.86 ± 9.93 (−33.19 to −22.56)	0.101
SDcet (μm)	4.59 ± 2.29^c (3.80 to 5.40)	4.49 ± 2.09^b (3.99 to 5.09)	6.54 ± 2.58^b,c (4.59 to 7.52)	0.008
Haze, n (%)*	5 (14.7)	15 (25.4)	4 (25)	0.463
3 month
CCET (μm)	54.18 ± 3.60^a,c (52.49 to 55.40)	51.24 ± 3.39^a (49.98 to 51.98)	48.82 ± 4.65^c (46.76 to 51.35)	<0.001
UDVA (logMAR)	−0.05 ± 0.07^c (−0.06 to −0.015)	−0.02 ± 0.08^b (−0.04 to 0.01)	0.05 ± 0.07^b,c (0.01 to 0.08)	<0.001
Cylinder (D)	−0.30 ± 0.22^c (−0.45 to −0.24)	−0.46 ± 0.27^b (−0.54 to −0.38)	−0.84 ± 0.49^b,c (−1.09 to −0.59)	<0.001
SE refraction (D)	−0.01 ± 0.43 (−0.09 to 0.22)	−0.21 ± 0.59 (−0.49 to −0.06)	−0.18 ± 0.68 (−0.53 to 0.16)	0.311
Superior (μm)	53.79 ± 4.40^c (51.94 to 55.30)	51.78 ± 4.28^b (50.01 to 52.66)	48.41 ± 4.60^b,c (46.39 to 50.95)	0.001
Inferior (μm)	56.50 ± 4.85^a,c (54.71 to 57.78)	53.93 ± 3.27^a,b (52.38 to 54.57)	51.24 ± 3.40^b,c (49.78 to 53.33)	<0.001
Min (μm)	47.54 ± 4.86^a,c (45.39 to 49.17)	46.60 ± 4.53^a,b (44.37 to 47.51)	40.59 ± 3.39^b,c (39.12 to 42.55)	<0.001
Max (μm)	61.25 ± 3.77 (59.80 to 62.62)	59.84 ± 4.52 (58.79 to 61.46)	58.59 ± 4.66 (56.57 to 61.32)	0.129
Min-Max (μm)	−13.71 ± 4.67^c (−15.73 to −12.13)	−13.24 ± 5.10^b (−16.01 to −12.36)	−18.00 ± 4.40^b,c (−20.25 to −15.98)	0.003
SDcet (μm)	2.81 ± 0.85^c (2.50 to 3.15)	2.64 ± 0.80^b (2.50 to 3.11)	3.63 ± 0.88^b,c (3.24 to 4.10)	<0.001
Haze, n (%)*	9 (31) ^c	19 (39.6)	16 (88.9)^c	<0.001

CCET, central corneal epithelial thickness; SD_cet, standard deviation of corneal epithelial thickness; SE, spherical equivalent; UDVA, uncorrected distance visual acuity.

Continuous variables are expressed as mean±SD (95% CI); categorical variables are displayed as number (percentage).

* Fisher’s exact test.

^ap<0.05 for Low vs Moderate.

^bp<0.05 for Moderate vs High.

^cp<0.05 for Low vs High.

At 1-month postoperatively, both UDVA and SD_cet showed significant improvement in the low and moderate spherical equivalent groups than in the high group. At the 3-month follow-up examination, the UDVA, cylinder power, superior mean, min-max, and SD_cet were significantly improved in the low and moderate spherical equivalent groups than in the high group. The CCET, inferior mean, and minimum epithelial thickness were significantly improved in the low spherical equivalent group than in the other groups. Furthermore, the percentage of haze was also significantly lower in the low spherical equivalent group than in the high group.

Table 4 shows the correlation analysis among CCET, SD_cet, and haze with logMAR UDVA, refraction, and OCT pachymetry parameters at the 3-month follow-up examination. LogMAR UDVA and SD_cet were significantly negatively correlated with CCET (ρ = −0.340, p = 0.001, and ρ= −0.206, p  = 0.045, respectively). Spherical equivalent refraction (baseline) and all other OCT pachymetry parameters, including superior mean, inferior mean, minimum, maximum, and min-max epithelium values showed a significant positive correlation with CCET (ρ = 0.372, 0.780, 0.862, 0.696, 0.591, and 0.211, respectively; all p < 0.001 except for min-max with a p-value of 0.040).

Table 4. Correlation between central corneal epithelial thickness, standard deviation of corneal epithelial thickness, and haze with postoperative variables at 3-month follow up.

	CCET		SDcet		Haze
	ρ#	p-value	ρ#	p-value	ρ#	p-value
UDVA (logMAR)	−0.340	0.001	0.046	0.660	0.171	0.082
Sphere (D; residual)	0.021	0.546	0.251	0.017	−0.321	0.001
Cylinder (D; residual)	0.179	0.091	−0.288	0.006	−0.137	0.175
SE refraction (D; residual)	0.100	0.350	0.187	0.078	−0.410	<0.001
SE refraction(D; baseline)	0.372	<0.001	−0.308	0.002	−0.307	0.002
CCET(μm)	-	-	−0.206	0.045	−0.099	0.343
Superior (μm)	0.780	<0.001	−0.269	0.008	−0.034	0.745
Inferior (μm)	0.862	<0.001	−0.080	0.443	−0.224	0.031
Min (μm)	0.696	<0.001	−0.582	<0.001	−0.093	0.374
Max (μm)	0.591	<0.001	0.412	<0.001	−0.135	0.198
Min-Max (μm)	0.211	0.040	−0.934	<0.001	0.030	0.774
SDcet (μm)	−0.206	0.045	-	-	−0.052	0.622

CCET, central corneal epithelial thickness; SD_cet, standard deviation of corneal epithelial thickness; SE, spherical equivalent; UDVA, uncorrected distance visual acuity.

#Spearman’s rank correlation coefficient.

SD_cet had a significant positive correlation with sphere power (residual), and maximum epithelium thickness (ρ = 0.251, p = 0.017; and ρ = 0.412, p < 0.001, respectively). Furthermore, the cylinder power (residual), spherical equivalent refraction (baseline), CCET, superior mean, minimum, and min-max of epithelial thickness were significantly negatively correlated with SD_cet (ρ=-0.288, p=0.006; ρ=-0.308, p = 0.002; ρ=-0.206, p = 0.045; ρ = -0.269, p = 0.008; ρ=-0.582, p < 0.001; and ρ=-0.934, p < 0.001, respectively). As for haze, sphere power (residual), spherical equivalent refraction (residual), spherical equivalent refraction (baseline), and inferior mean epithelial thickness were significantly negatively correlated with haze (ρ=-0.321, p = 0.001; ρ=-0.307, p = 0.002; ρ=-0.410, p < 0.001; and ρ=-0.224, p = 0.031, respectively).

To evaluate the recovery of corneal epithelial thickness over different areas, the OCT epithelium map was further divided into four quadrants (superior-nasal [SN], superior-temporal [ST], inferior-nasal [IN], and inferior-temporal [IT]) and the minimum epithelial thickness area of all studied eyes was recorded. At each follow-up time point, the percentage of the minimum epithelial thickness area in four quadrants was calculated. Figure 3 showed the results obtained at the 1, 3, and 6-month follow-up examination. The percentage, from higher to lower, was SN, ST, IT, and IN with significant differences between any two groups at all follow-up time point (all p < 0.05). The proportion of baseline astigmatism types at 1, 3, and 6-month follow-up within each quadrant of the minimum corneal epithelial thickness was analysed, which showed no statistical difference (see Table S1).

Figure 3. The percentage of minimum corneal epithelial thickness area in four quadrants at 1, 3, and 6-month follow up time points.

SN, superior-nasal; ST, superior-temporal; IN, inferior-nasal; IT, inferior-temporal.

*The percentages referred to the proportions of the four quadrants at each specific follow-up time point.

**. p < 0.05 between any two of the four quadrants at all follow-up time points.

Discussion

The healing process of the corneal epithelium involves the integration and cooperation between epithelial proliferation, differentiation, cell adhesion, migration, and apoptosis, which are orchestrated by a myriad of cytokines.^25–27 Epithelial renewal enables the integrity of the corneal structure and maintains its smooth surface and refractive function. As Figure 2 showed, an immediate decrease in the spherical equivalent refraction was observed after t-PRK. The condition became stable after the 1-month follow-up examination. However, a significantly lower CCET was observed at 1 week, 1 month, and 3 months compared to baseline CCET. The cornea returned to its preoperative corneal epithelial thickness at >3 months postoperatively, which is consistent with previous findings.^13,14

The postoperative UDVA showed no difference with baseline CDVA after 1 month and was significantly better than baseline at the 6-month follow-up examination. The condition may be a result of wavefront refractive procedures to reduce higher‐order aberrations and improve visual quality. However, a Cochrane review concluded that no important difference in the clinical outcomes was observed between wavefront versus conventional refractive surgery at 6 and 12 months postoperatively.²⁸ Further comparison studies of pre-operative CDVA and post-t-PRK UDVA are required to provide more information. Except for VA, all the other OCT pachymetry parameters showed greater improvement at 6 months than at earlier follow-up time points. All of these findings suggest that, as expected, it takes a longer time for the corneal epithelium to return to its preoperative structural status after t-PRK than what it needs to recover from an epithelial wound.

A subgroup analysis based on different baseline spherical equivalent refractions was performed. A higher CCET and maximum epithelial thickness were observed in the high spherical equivalent group at 1 week, with the condition progressively reversing thereafter. This may result from temporary epithelial hyperplasia, which occurs in the early postoperative months after refractive surgery.¹⁷ Another possible mechanism is the tendency of the epithelium to fill depressions and reduce any corneal asymmetry after surgery.²⁹ At the 3-month follow-up examination, the high spherical equivalent group showed a poorer recovery status in CCET, UDVA, cylinder power, superior average, inferior average, minimum, min-max of corneal epithelial thickness, SD_cet, and haze than the other groups, which implies a slower epithelial wound healing in higher baseline spherical equivalent refraction eyes. However, the high spherical equivalent group reached an average of 0.05 logMAR UDVA and −0.18 residual spherical equivalent refraction at 3-months, which is clinically satisfactory.

A correlation analysis was also performed in this study. At 3-months postoperatively, CCET and SD_cet were significantly correlated to each other and they both showed a significant correlation with spherical equivalent refraction (baseline). Patients with a higher refractive error at baseline would have a lower CCET and higher SD_cet at 3 months postoperatively. Both CCET and SD_cet were also correlated with superior average, minimum, maximum, and min-max of corneal epithelial thickness. These findings suggested that CCET and SD_cet were important and valuable parameters for estimating corneal wound recovery and remodelling status after t-PRK.

A higher CCET was correlated to a better visual outcome. Higher haze was correlated with higher sphere power (residual), higher spherical equivalent refraction (residual), higher spherical equivalent refraction (baseline), and lower inferior mean corneal epithelial thickness. Many studies have shown a higher rate of stromal haze in high preoperative spherical equivalent patients who underwent PRK.^30–33 Nevertheless, patients with haze developed only grade 0.5 to grade 1 haze in this study, which may be the reason why the haze showed no correlation with patients’ visual outcomes.

Interestingly, a significant superior-inferior difference in the minimum corneal epithelial thickness area on the OCT pachymetry test was found at every follow-up time point. The percentage of the SN region was significantly higher among the four quadrants. A further analysis of the proportions of baseline astigmatism types in each of the four quadrants failed to find any statistical significance.

Several explanations are possible. First, gravity may cause an inclination of the corneal migration towards the inferior region. The gravitational force over the tear film, which contains a biochemically complex mixture of factors, would cause a greater tear volume to remain over the inferior meniscus for a long time.^34–36 Second, the corneal sub-basal nerve plexus is arranged in a radiating pattern (whorl-like complex), converging towards an area approximately 1–2 mm inferior to the central cornea.³⁷ The distribution of corneal nerves may have an influence on the remodelling of corneal epithelium.³⁸ Third, the mechanical effect of the eyelids over the ocular surface, either the upper eyelid coverage or the shear forces during blinking, may influence corneal epithelial growth.^39,40

This is thought to be the first study to demonstrate the corneal epithelial remodelling after t-PRK by OCT epithelium map. However, this study had some limitations. First, the study included participants from a single centre and had a non-randomised retrospective study design. The sample size was small and the follow-up period was short. Nevertheless, most of the corneal remodelling and epithelial thickening occurred within 3–6 months after the refractive surgery, whereas the stability of epithelial thickness has been observed between 3 and 12 months.^13,14

Another limitation is that all study participants were from the Taiwanese population, were relatively young, and a statistical difference was noted initially in the sex of the three spherical equivalent groups. Finally, it remains unclear how the warpage effect of contact lens wearing can influence the OCT measurement results at the 1 week follow-up examination as the contact lens was removed at that time point.

Conclusions

Spherical equivalent refraction and VA achieved preliminary stability 1 month after t-PRK. In contrast, CCET and SD_cet took 3–6 months to progressive recovery. The high baseline spherical equivalent group was found to have a slower recovery postoperatively, and a correlation between higher spherical equivalent refraction (baseline) with lower CCET, higher SD_cet, and higher haze incidence at 3 months was noted. The SN area was revealed to be the slowest area of corneal epithelial thickness recovery among four quadrants, and an apparent superior-inferior difference was noted in t-PRK.

Higher stromal haze was correlated with higher spherical equivalent refraction (both baseline and residual) but had no relation to visual outcome. This might have been attributed to the fact that all participants developed only low-grade haze in this study. A significant correlation was also observed between higher CCET with better UDVA and lower corneal epithelial irregularity.

Both CCET and SD_cet measured by OCT seem to be a good auxiliary indicator for reflecting the status of corneal surface irregularity and wound recovery after t-PRK surgery. A well-designed randomised control study is needed to confirm the study results.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08164622.2023.2197107

Data availability statement

The original contributions presented in the study are included in the article/supplementary materials. Further inquiries can be directed to the corresponding author.

Additional information

Funding

This study was supported by the Tri-Service General Hospital Research Foundation [grant number: TSGH-D-112173, TSGH-E-112244]. The sponsor has no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.