Underestimate of Tonometric Readings after Photorefractive Keratectomy Increases at Higher Intraocular Pressure Levels

Abstract

purpose. To determine whether tonometric readings of increases in intraocular pressure (IOP) during the water-drinking test (WDT) are affected by variations in central corneal thickness (CCT) induced by photorefractive keratectomy (PRK).

methods. Data from 30 randomly selected eyes of 30 patients (18 men and 12 women; mean age, ± SD: 33.9 ± 7.6 years) undergoing bilateral PRK for myopia (−6.57 ± 2.39 D) were obtained. Objective refraction, anterior radius of corneal curvature (R), CCT, and IOP measurements at baseline and at different time intervals after ingestion of 1 L of water within 5 minutes, were performed before and 6 months after PRK. All measured IOPs were recalculated by a correction factor for R and CCT and expressed as corrected intraocular pressure (IOPC) measurements.

results. The mean R ± SD was 7.84 ± 0.20 and 8.76 ± 0.34 mm, and the mean CCT was 544.83 ± 19.69 and 453.97 ± 29.95 μm, before and after PRK, respectively. The mean IOP at baseline was 15.05 ± 2.78 and 9.83 ± 2.56 mm Hg, and during WDT was 18.32 ± 3.42 and 11.42 ± 3.10 mm Hg at 10 minutes, 18.59 ± 2.99 and 11.54 ± 2.54 mm Hg at 20 minutes, 17.80 ± 2.85 and 10.87 ± 2.22 mm Hg at 30 minutes, 16.35 ± 3.02 and 10.26 ± 2.21 mm Hg at 45 minutes, and 14.90 ± 2.52 and 9.81 ± 2.32 mm Hg at 60 minutes, before and after PRK, respectively. The mean IOPC at baseline was 13.64 ± 2.33 and 13.05 ± 2.98 mm Hg, and during WDT was 16.61 ± 2.77 and 15.08 ± 3.59 mm Hg at 10 minutes, 16.96 ± 2.69 and 15.33 ± 2.96 mm Hg at 20 minutes, 16.10 ± 2.50 and 14.42 ± 2.60 mm Hg at 30 minutes, 14.92 ± 2.72 and 13.62 ± 2.65 mm Hg at 45 minutes, 13.82 ± 2.27 and 13.05 ± 2.55 mm Hg at 60 minutes, before and after excimer laser treatment, respectively. Pre- and postoperative IOPs and percentages of IOP increase differed significantly (P < 0.05), in particular at the peak, as did IOPCs but not the percentages of increase in IOPC, apart from the highest values.

conclusions. Corneal changes after PRK for myopia may induce an uneven underestimate of the IOP increases. The inadequacy of a correction factor to compensate for CCT and R at high IOP levels indicates that other biomechanical factors may play a role when the cornea is subjected to dynamic actual IOP variation. Such increase of the well-known underestimate of IOP after PRK at higher actual IOPs may have significant clinical implications in tonometric assessment of subjects at risk of glaucomatous damage.

Recent evidence has stressed the clinical role of the assessment of central corneal thickness (CCT) for correct management of glaucoma, ^{1 2} and pachymetry has been inserted in the current ophthalmologic practice for patients at risk of glaucoma. Indeed, Ehlers et al. ³ and later Whitacre et al. ⁴ reported that applanation tonometry provides accurate readings compared with actual intraocular pressure (IOP) only when CCT is 520 μm, whereas thinner and thicker corneas produce underestimations and overestimations of IOP, respectively. ^{5 6 7 8 9 10}  

Reduction of tonometric readings has been reported in eyes after photorefractive keratectomy (PRK) ^{11 12 13 14} and laser in situ keratomileusis (LASIK). ^{15 16 17} These procedures flatten the anterior corneal surface and reduce CCT proportionally to the extent of myopia, ^{14 18} providing lower IOPs compared with preoperative ones. 

In a recent study, ¹⁹ refractive surgery provided an excellent model for the in vivo study, in the same patient, of the effect of variations in CCT on the IOP readings after the administration of an ocular hypotensive drug. In their sample of myopic patients undergoing PRK, Tamburrelli et al. ¹⁹ demonstrated that CCT reduction provides an erroneous impression of reduced pharmacologic efficacy, that may be avoided after correction of IOP by a proper nomogram. 

The mentioned influence of CCT on IOP reduction readings suggested the present study, in which we evaluated whether similar findings occur in increased IOP as well. Indeed, the misdetection of increased IOP may have negative effects in the clinical management of subjects at risk of glaucomatous damage. To this end, we used the water-drinking test (WDT) ^{20 21} as an in vivo experimental model to induce increased IOP, and the PRK as a surgical means to change corneal parameters, in a study population undergoing refractive surgery for myopia. 

Materials and Methods

Subjects

In this prospective study, data from 30 consecutive randomly selected eyes of 30 patients (18 men and 12 women; mean age, 33.87 ± 7.63 years; range, 22–51) who were undergoing bilateral PRK treatment for myopia were obtained. All patients were healthy and had open angles on gonioscopy and no evidence of external corneal disease or glaucoma. They were evaluated before and 6 months after the surgical procedure. In both conditions, the baseline examinations included objective refraction evaluation by an autorefractometer (model AR-600; Nidek, Aichi, Japan), mean anterior radius of corneal curvature (R) assessment (Keratron Scout; Optikon, Rome, Italy), central pachymetry readings by an ultrasonic pachymeter (Altair 606 AN; Optikon), and IOP measurements by a noncontact tonometer (model TX-10; Canon Ltd., Tokyo, Japan) according to a previous study. ¹⁹ IOP measurements were performed on the same schedule between 9 and 11 AM. The mean of three consecutive readings was considered for statistical analysis, and it was rejected if the standard deviation among measurements exceeded 0.4 mm Hg. 

The design and performance and experimental procedures were clearly formulated in an experimental protocol, which was approved by the institutional ethics review board and adhered to the tenets of the Declaration of Helsinki. Written, informed consent was obtained from each patient before his or her inclusion in the study and after the goals and methods of the study and the potential side effects that WDT may entail were adequately explained. 

After recording of the baseline IOP, the patients were submitted to the WDT according to Nørskov ²¹ They were instructed not to eat or drink after midnight. IOP was measured 10, 20, 30, 45, and 60 minutes after ingestion of 1 L of water within 5 minutes. 

Pachymetry was performed after noncontact tonometry to prevent IOP reduction due to its corneal indentation. 

Surgery was performed in all patients by one surgeon (CT) with the same technique. Both patients’ eyes were treated at the same session. An uneventful PRK was performed with an excimer laser (Technolas Keracor 217-C; Chiron, Irvine, CA), aiming at emmetropia in all cases. Before photoablation the corneal epithelium was manually removed under topical anesthesia. Postoperative topical antibiotics and artificial tears were routinely used. Topical steroids (fluorometholone acetonide 0.1%, Flarex; Alcon, Fort Worth, TX) were limited to a 7-day therapy at two doses per day. 

All patients were regularly followed up in the postoperative period. At the 6-month visit, pachymetric data as well as IOPs at baseline and during WDT were obtained from the subjects, who were required to follow preoperative dietetic instructions. 

To minimize the influence of different physical pre- and postoperative conditions of the patients’ corneas, all measured IOPs were recalculated according to Orssengo and Pye ²² formula, briefly described elsewhere. ¹⁹ This formula calculates, in mathematical terms, the corrected IOP (IOPC), dividing IOP readings by a complex correction factor that is dependent on CCT, R, applanated area, and Poisson’s ratio of the cornea. 

Absolute IOP and IOPC increases (i.e., the differences between tonometric measurements at each time interval [10, 20, 30, 45, and 60 minutes] during WDT and at baseline), as well as the corresponding percentages were also calculated before and after PRK. 

Statistical Analysis

A multivariate analysis of variance (MANOVA) for repeated measures with post hoc adjusted t-tests was performed to compare IOP and IOPC at baseline and during the WDT, pre- and post-PRK. The same analysis was conducted on the percentage increases. Surgery (i.e., before and after PRK) and time (i.e., at baseline and during WDT) were the within-subject factors. A significant (P < 0.05) interaction effect between these two variables was used as an indicator of differences in the shapes of the curves obtained before and after laser treatment. Percentage differences in mean tonometric readings were evaluated by adjusted t-tests, and P ≤ 0.025 was considered statistically significant to compensate for the multiple comparisons. Statistical analysis was performed on computer (SPSS, ver. 12.0.0; SPSS Science, Inc., Chicago, IL). 

Results

Demographic and clinical data are shown in Table 1 . Mean IOP and IOPC recorded at baseline and at various time intervals during WDT, before and after surgery, are graphically displayed in Figure 1and presented in detail in Table 2 . After PRK, both IOP and IOPC were lower than before laser treatment. MANOVA showed a significant effect of surgery (F_(1,29) = 126.227, P < 0.001, and F_(1,29) = 4.954, P = 0.034 for IOP and IOPC, respectively) and time (F_(5,25) = 50.437, P < 0.001, and F_(5,25) = 20.080, P < 0.001 for IOP and IOPC, respectively) and a significant interaction of surgery and time (F_(5,25) = 12.058, P < 0.001, and F_(5,25) = 2.924, P = 0.033 for IOP and IOPC, respectively). Post hoc adjusted t-tests detected significant (P ≤ 0.025) differences in mean IOP between pre- and postoperative conditions at all time points, and in the mean IOPC at 20, 30, and 45 minutes during WDT (see Table 2for probabilities). 

The IOP and IOPC increases from baseline during WDT expressed as an absolute value and percentage are shown in Table 3 . Postoperative absolute as well as percentages of increase in IOP and IOPC during WDT were lower than the preoperative ones, except for the IOP increases at the 60-minute time interval. MANOVA on the percentages of tonometric increase showed a significant effect of surgery (F_(1,29) = 4.315, P = 0.047, and F_(1,29) = 5.587, P = 0.025 for IOP and IOPC, respectively) and time (F_(4,26) = 34.771, P < 0.001, and F_(4,26) = 22.147, P < 0.001 for IOP and IOPC, respectively), and a significant interaction of surgery and time on IOP (F_(4,26) = 4.365, P = 0.008) but not IOPC readings (F_(4,26) = 1.987, P = 0.126). Post hoc adjusted t-tests did not detect significant differences in the mean percentages of IOP and IOPC increase at any time interval between pre- and postoperative conditions, except for the 30-minute time interval (see Table 3for probabilities). 

Discussion

The relationship between CCT and IOP has been investigated in several studies. ^{3 8 23 24 25 26 27 28 29 30} Under- and overestimate of IOP measurements may respectively occur for thinner or thicker corneas than the “normal” CCT of 520 μm proposed by Ehlers et al., ³ with an average error of 0.7 mm Hg per 10-μm deviation. Such an estimate was derived from a linear relationship between the error of Goldmann applanation tonometry, obtained experimentally by comparison with manometric readings in cannulated in vivo human eyes, and CCT as measured by optical pachymetry. However, different correction factors ranging from 0.11 to 0.71 mm Hg per 10 μm of CCT and different normal CCTs have also been proposed. ^{1 3 4 8 25 26 27 28 29} For instance, Whitacre et al. ⁴ based their formula on CCT data from a regression analysis of 15 eyes, whereas Doughty and Zaman ¹ examined the meta-analysis of 300 data sets to suggest, for healthy eyes, a correction of 1.1 mm Hg for a 50-μm difference in CCT from the “normal” value of 535 μm. Recently, Shimmyo et al. ³¹ in an observational retrospective cross-sectional study argued that IOP adjustment could not be made with a single ratio or a linear formula, because of a nonlinear relation between Goldmann applanation tonometry readings and its errors; thus, extrapolating from their observations and the data of Ehlers et al., ³ they presented statistical approximation formulas with exponential function to calculate IOP, taking into account CCT by ultrasonic pachymetry, with 550 μm as a statistical norm (average), and corneal curvature. From a different standpoint, Orssengo and Pye ²² developed a mathematical model in which the cornea was modeled as a shell, and the theoretical equations for the deformations of a shell due to internal and applanating pressures were combined to model the behavior of the cornea during applanation tonometry. In the past year, two groups have used such a purely theoretical nomogram that takes into account some corneal features, including CCT, to correct tonometric readings. ^{2 19}  

Increasing clinical interest on the tonometric readings of changes in actual IOP at different CCTs has found in refractive surgery an excellent in vivo model to decrease the CCT and to evaluate the dynamic relationship between CCT and IOP measurements. Specifically, a recent paper on myopic patients who were undergoing bilateral PRK treatment showed that the effect of hypotensive drugs on IOP readings may be underestimated because of measurement errors due to the laser-induced CCT reduction. ¹⁹ However, to date, no study has been undertaken to investigate the influence of different CCTs on tonometric readings in the opposite condition, (i.e., at increasing actual IOPs). To this end, in a sample of myopic patients who were undergoing excimer laser treatment, we decided to use the WDT as a mean to increase IOP, regardless of its diagnostic or prognostic value for glaucoma detection, because it is a practical test and not related to initial IOP. ^{32 33 34 35} However, several studies have revealed that the WDT results are not reproducible because some factors such as food and fluid intake, volume of fluid to be ingested at the time, scleral rigidity, and the patient’s age, may influence the WDT. ^{36 37} Nevertheless, in the present study we can consider our results reliable, because IOP readings were determined with the same test modalities at each time interval and compared in the same patient before and after PRK. 

Although Goldmann applanation tonometry is still considered the standard method of IOP measurement, we used noncontact tonometry because it is practical, it avoids epithelial corneal alterations related to the applanation, and it needs neither topical anesthesia nor skilled physicians. In addition, this method has been demonstrated to be reproducible when the mean of three measurements is used ³⁸ and accurate when compared with the Goldmann tonometer in normal eyes. ³⁹ In this regard, a recent report has shown no significant difference between measurements even in patients undergoing PRK and LASIK, and for each degree of treated myopia, even if noncontact tonometry readings were slightly higher. ⁴⁰  

In agreement with previous studies on decreased IOP measurements after PRK and LASIK, ^{11 12 13 14 15 16 17 18} this study showed a statistically significant IOP decrease after excimer laser treatment at baseline as well as during WDT. However, in our patient sample, this reduction was not constant at the different time intervals during WDT, with increasing reductions toward the peak at the 20-minute time interval, and successive decreasing reductions up to the 60-minute time interval (see Fig. 1 ), which is when the effect of WDT should end, as suggested by previous studies. ^{34 36} Similarly, when evaluating the absolute IOP increases during WDT, the mean postoperative increases were lower than before PRK, except for the values obtained when the IOP returned to baseline after 60 minutes. After individual normalization of IOP increases by comparison with baseline, the percentages of increase were still significantly different between the pre- and postoperative conditions (see MANOVA results), in particular, at the 30-minute time interval (see post hoc analysis in Table 3 ). Finally, our findings show that the corneal changes induced by PRK for myopia may underestimate IOP increase, mainly at high values. 

Despite data correction for both CCT and R, the mean postoperative IOPCs remained significantly lower than preoperative ones across the time points (see MANOVA results), differing at 20, 30, and 45 minutes during WDT (P ≤ 0.025, see post hoc analysis in Table 3 ). More interesting and different from the IOP increases, the mean percentage of IOPC increase revealed a similar behavior across the time intervals in the pre- and postoperative conditions (see MANOVA results and Table 3 ). Nonetheless, the significant effect of the isolate surgery factor on these data seems to indicate the inability of the Orssengo and Pye ²² nomogram to fully correct the tonometric readings when actual IOP increases (see Fig. 1 ). In this respect, the lack of statistical significance (by post hoc analysis) at the peak values, apart from the 30-minute time interval, might be accounted for by too conservative statistics (see Table 3 ). 

In the present study, the purely theoretical formula of Orssengo and Pye ²² was chosen because formulas derived from experimental or retrospective analysis of normal corneas may not be fully applicable to our study population, whose CCT and R were significantly modified by PRK. The inadequacy of this formula when the cornea is placed under dynamic actual IOP variation may depend, at least in part, on decreased corneal thickness and related changes in biomechanical properties ¹⁴ as well as the high IOP level induced by WDT. 

Physical corneal properties are largely governed by the structure of the stromal extracellular matrix, the bulk of which in human cornea comprises collagen fibrils arranged in approximately 300 to 500 parallel lamellae. ⁴¹ Fibrils within a lamella are parallel to each other and to the corneal surface, but run uninterrupted from limbus to limbus at angles in relation to fibrils in adjacent lamellae. ⁴² This collagen network, specifically the diameter of collagen fibrils, their orientation in relation to the applied force, and the collagen content of the tissue, all determine the cornea’s tensile strength. ^{43 44} In a mathematical formula, ⁴³ the tensile strength of the tissue (σ_t)—that is, the stress at which the tissue breaks—is determined by  

\[{\sigma}_{\mathrm{t}}\ {=}\ {\beta}{\sigma}_{\mathrm{f}}\ {+}\ (1\ {-}\ {\beta})\ {\sigma}_{\mathrm{g}}\]

where β is the volume fraction of the tissue that is occupied by collagen, (1 –β) is the volume fraction occupied by the ground substance (i.e., stromal matrix elements other than fibrillar collagen, ignoring the volume of the cells), and σ_f and σ_g are the tensile strengths of the fibrils and of the ground substance, respectively, at the fibril critical length. Such a length represents the minimum fibril length required for effective tissue reinforcement, and it is related to the σ_f, the fibril radius and the shear stress exerted on a fibril by the ground substance. ⁴³  

The higher packing density of collagen fibrils physiologically observed in the prepupillary cornea, and thus the increased collagen volume fraction with reduced fibril spacing, is necessary to maintain tissue strength, ⁴⁵ bearing in mind that the cornea is thinner centrally. ^{46 47} In this way, for σ_f > σ_g in the equation, increasing the volume fraction of collagen produces a proportional increase in the mechanical strength of the tissue. Such a mechanism could help to preserve dioptric stability in the cornea by helping to maintain surface curvature in the presence of variations in tissue thickness—of course, assuming that corneal collagen fibrils are at least as long as their critical length. ⁴⁵  

During myopic PRK, the packed collagen fibrils are photoablated in the central anterior third of the corneal stroma, and morphologic changes in subepithelial keratocytes and in the extracellular matrix occur, even in biomicroscopically clear corneas. ⁴⁸ This postoperative reduction of CCT, mainly related to a decreasing volume fraction of collagen, by determining σ_f < σ_g in the equation, produces a proportional decrease in the mechanical strength of the tissue. When an applanating force is applied to the cornea, the tonometric reading occurs at the balance between the external force and the sum of the actual IOP and the restoring force due to the stretched fibrils and the ground substance. Thus, we can hypothesize that the decrease in tensile strength after PRK reduces the restoring force that counteracts the applanation during tonometry. Nonetheless, this corneal behavior seems not to depend linearly on CCTs, and other currently unpredictable factors may affect this relation. Hardness or softness of the tissue may be involved and, although no significant correlation has been found between CCT and overall ocular rigidity, alterations in topical corneal rigidity–elasticity properties occurring by photoablative CCT reduction over the applanation area may influence IOP assessment. ⁴⁹ In addition, as ocular rigidity seems not to be significantly altered by refractive status, ⁴⁹ the myopic refraction of our study population should not have implications for our tonometric findings. Moreover, based on our data, this nonlinear effect may be amplified by a high IOP that increases the stress and stretch of an already altered corneal mechanical strength and further decreases the restoring force—thus explaining, at least in part, the increasing reductions of the postoperative increase in IOP during WDT in our study population (see Fig. 1 ). Similarly, an underestimate of IOP in the high-pressure range was recently detected in experimental conditions on human cadaveric eyes. ⁵⁰ In the future, a better comprehension of biomechanical properties of the cornea at different CCTs under different IOP levels may aid in accounting for our results and provide further factors to ameliorate the theoretical Orssengo and Pye’s nomogram. 

Finally, our data show a nonlinearity of the tonometric readings at higher IOPs after corneal thinning by PRK for myopia. Nevertheless, this finding cannot be transferred to subjects with physiologically “thin” corneas. Moreover, if further studies on untreated thin corneas confirm our data, significant clinical implications in tonometric assessment of normal subjects and patients with glaucoma may be drawn—that is, eventual IOP peaks may be underestimated or even unrecognized in eyes with low CCT and high R. 

 Submitted for publication October 20, 2004; revised April 8, 2005; accepted April 13, 2005.

 Disclosure: C. Tamburrelli, None; A. Giudiceandrea, None; A.S. Vaiano, None; C.G. Caputo, None; F. Gullà, None; T. Salgarello, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Demographic and Clinical Data

Sex (F/M)	12/18
Age (y)	33.87 ± 7.63 (22–51)
Refractive error (D)
Preoperative	−6.57 ± 2.39 (−2.25 to −11.00)
Postoperative	−0.12 ± 0.34 (−1.00 to 0.25)
Anterior corneal radius (mm)
Preoperative	7.84 ± 0.20 (7.54–8.27)
Postoperative	8.76 ± 0.34 (8.23–9.50)
Pachymetry (μm)
Preoperative	544.83 ± 19.69 (519–590)
Postoperative	453.97 ± 29.95 (387–518)

 Data are expressed as the mean ± SD, with the range in parentheses; D, diopter spherical equivalent.

Figure 1.

 

Tonometric readings serially recorded before (time 0) and at different time intervals during the WDT. Error bars, SEM.

Table 2.

 

Measured and Corrected IOP

	Baseline	At 10 Minutes	At 20 Minutes	At 30 Minutes	At 45 Minutes	At 60 Minutes
IOP
Before PRK	15.05 ± 2.78 (9.90–21.00)	18.32 ± 3.42 (12.80–25.30)	18.59 ± 2.99 (14.60–24.30)	17.80 ± 2.85 (13.30–23.20)	16.35 ± 3.02 (11.80–22.70)	14.90 ± 2.52 (10.40–21.00)
After PRK	9.83 ± 2.56 (6.00–18.00)	11.42 ± 3.10 (6.00–20.70)	11.54 ± 2.54 (7.70–19.00)	10.87 ± 2.22 (7.90–16.30)	10.26 ± 2.21 (7.00–16.80)	9.81 ± 2.32 (6.90–17.00)
	P < 0.0001*	P < 0.0001*	P < 0.0001*	P < 0.0001*	P < 0.0001*	P < 0.0001*
IOPC
Before PRK	13.64 ± 2.33 (9.61–19.27)	16.61 ± 2.77 (12.80–23.21)	16.96 ± 2.69 (13.04–22.80)	16.10 ± 2.50 (12.70–21.60)	14.92 ± 2.72 (11.07–20.83)	13.82 ± 2.27 (10.10–19.60)
After PRK	13.05 ± 2.98 (8.33–19.57)	15.08 ± 3.59 (8.33–22.50)	15.33 ± 2.96 (10.26–20.65)	14.42 ± 2.60 (10.61–18.48)	13.62 ± 2.65 (8.54–18.26)	13.05 ± 2.55 (8.78–18.48)
	P = 0.374	P = 0.041†	P = 0.015*	P = 0.004*	P = 0.021*	P = 0.151

 Data (in mm Hg) are provided for baseline and time points during WDT, before and after PRK (mean ± SD, with the range in parentheses).

  P indicate the t-test’s statistical significance between pre- and postoperative results at the adjusted P of 0.025 (*) and at the common P of 0.05 (†).

Table 3.

 

Measured and Corrected IOP Increase from Baseline at Different Time Intervals during WDT

	At 10 Minutes	At 20 Minutes	At 30 Minutes	At 45 Minutes	At 60 Minutes
IOP
Before PRK	3.27 ± 1.77 (−0.90 to 6.90)*	3.54 ± 2.18 (−0.20 to 8.20)	2.75 ± 1.91 (−1.00 to 6.50)	1.31 ± 1.6 (−2.10 to 4.70)	−0.15 ± 0.64 (−2.00 to 0.90)
	22.39 ± 12.98 (−6.29 to 48.23)†	25.18 ± 16.88 (−1.05 to 65.60)	19.62 ± 14.19 (−5.26 to 52.00)	9.30 ± 10.97 (−13.04 to 36.00)	−0.63 ± 4.22 (−10.53 to 6.38)
After PRK	1.59 ± 1.37 (−2.40 to 3.60)	1.71 ± 1.29 (−1.70 to 4.60)	1.05 ± 1.19 (−1.70 to 3.00)	0.43 ± 0.95 (−1.40 to 2.30)	−0.02 ± 0.63 (−1.40 to 1.20)
	16.71 ± 15.22 (−22.43 to 50.79)	19.31 ± 16.01 (−15.89 to 55.42)	12.53 ± 12.87 (−13.08 to 36.14)	5.78 ± 10.62 (−13.08 to 26.67)	0.58 ± 6.98 (−14.44 to 19.05)
	P = 0.059	P = 0.075	P = 0.023‡	P = 0.129	P = 0.440
IOPC
Before PRK	2.97 ± 1.63 (−0.90 to 6.47)	3.33 ± 2.04 (−0.20 to 7.74)	2.47 ± 1.77 (−0.90 to 6.13)	1.28 ± 1.51 (−1.85 to 4.35)	0.18 ± 1.09 (−1.70 to 4.55)
	22.66 ± 13.57 (−6.29 to 52.04)	25.83 ± 17.03 (−1.27 to 65.65)	19.31 ± 14.45 (−5.73 to 51.99)	9.81 ± 11.02 (−12.59 to 36.05)	1.78 ± 7.71 (−10.83 to 30.23)
After PRK	2.03 ± 2.86 (−3.33 to 5.38)	2.28 ± 1.74 (−2.36 to 6.13)	1.37 ± 1.57 (−1.94 to 4.00)	0.57 ± 1.29 (−1.94 to 3.07)	0.00 ± 0.88 (−2.29 to 1.65)
	16.19 ± 15.43 (−22.41 to 50.74)	19.26 ± 16.02 (−15.88 to 55.37)	12.26 ± 13.06 (−13.06 to 36.13)	5.50 ± 10.76 (−13.06 to 26.69)	0.85 ± 6.78 (−13.06 to 18.86)
	P = 0.042§	P = 0.050§	P = 0.025‡	P = 0.076	P = 0.637

 Data are the difference between results at the different time intervals during WDT and at baseline, before and after PRK.

 Absolute (in mmHg; *) and percentage (†) values are expressed as the mean ± SD, with the range in parentheses. P indicate the t-test’s statistical significance between pre- and postoperative percentage results at the adjusted P of 0.025 (‡) and at the common P of 0.05 (§).