Portal de Periódicos da CAPES

Comparison of factors that influence the measurement of corneal hysteresis in vivo and in vitro

2011; Wiley; Volume: 89; Issue: 5 Linguagem: Inglês

10.1111/j.1755-3768.2011.02150.x

1755-3768

Tariq A. Alhamad, Keith M. Meek,

Ophthalmology and Visual Impairment Studies

Purpose: The purpose of this study was to compare measurements of corneal hysteresis (CH) obtained in vivo, with similar measurements from excised human eyes and from excised human corneas mounted in an artificial anterior chamber. Methods: Corneal hysteresis was measured using an ocular response analyser (Reichert Ophthalmic Instruments) from three groups: 53 healthy normal corneas of fifty-three patients, six excised eyes and 17 excised corneas. Results: In vivo, it was found that CH was independent of gender, age and mean spherical equivalent, but has a significant inverse relationship with intraocular pressure (IOPcc) (r = 0.53; p < 0.0001). However, there was no correlation between CH and IOPG (r = 0.10; p = 461). The same inverse relationship with IOPcc was recorded in intact, excised eyes (r = 0.74; p < 0.0001), with no significant differences between the behaviour each individual eye. Excised corneas also showed an inverse relationship between CH and trans-corneal pressure (r = 0.72; p < 0.0001), but the measured values of CH were lower than those recorded in vivo and from intact globes. In both excised eyes and excised corneas, we found a significant correlation between CH and central corneal thickness [r = 0.86; p < 0.0001 and r = 0.611; p < 0.0005 (respectively)]. Conclusion: The in vitro results indicate that every normal human eye at physiological hydration shows an identical dependence of CH on IOPcc, the same dependence as is observed in vivo. This therefore would appear to be an intrinsic response of the tissue to a change in IOP. However, it is possible that the lower values of CH recorded from excised corneas reflect the influence of the artificial chamber replacing the eye globe, so in vivo values of CH may be influenced to some extent by the presence of the other components of the eye. The ocular response analyser (ORA) is a noncontact applanation tonometer that can provide a measure of intraocular pressure (IOPcc) independent of corneal factors that can influence that derived from Goldman tonometry (IOPG) (Luce 2005). In addition, two biomechanical parameters are provided, corneal hysteresis (CH) and corneal resistance factor. Corneal hysteresis is defined as the difference in IOP recorded during inward (P1) and outward (P2) applanation and is therefore regarded as an indicator for the viscoelastic properties of the cornea, the higher the value of CH, the more energy absorbed by the system during the deformation process therefore the more viscoelastic it is. In recent years, there have been numerous reports describing the use of the ORA for measuring corneal biomechanics as a function of corneal or other variables. Some reports indicate that CH is correlated with central corneal thickness (Luce 2005; Shah et al. 2006; Montard et al. 2007; Kamiya et al. 2008), although the others found no strong correlation (Touboul et al. 2008). It has also been correlated with IOP (Kamiya et al. 2008; Liu et al. 2008), although, again, this has been questioned (Touboul et al. 2008). Corneal hysteresis has been found to be independent of gender (Montard et al. 2007; Kamiya et al. 2008; Liu et al. 2008). It is also independent of spherical equivalent or whether left or right eyes are examined (Montard et al. 2007; Ehongo et al. 2008). Most studies report that CH shows little or no dependence on age (Kirwan et al. 2006; Montard et al. 2007; Fontes et al. 2008; Kamiya et al. 2008; Kida et al. 2008; Liu et al. 2008), but others have reported a negative correlation with age (Moreno-Montanes et al. 2008). To date, all studies using the ORA have been carried out on corneas in vivo, but natural variability can limit the control available in such studies. In this paper therefore we describe experiments using the ORA both in vivo and in vitro. This has allowed us to more clearly demonstrate the relationship between CH and IOP, and has given some insight into other factors that can influence CH measurements. This study included fifty-six eyes of 56 healthy volunteers combined male and female who had no current or history of significant ocular or systemic pathology or any general health problems other than refractive errors. The volunteers were within the age range of 18–70 years and had a range of refractive errors (spherical equivalent) from −3.75 to 2.25 D. This study had local ethical committee approval, informed consent was obtained from each participating subject, and the study protocol was consistent with the tenets of the Declaration of Helsinki. The measurements of IOP and CH were performed using the ORA (Reichert Ophthalmic Instruments, Depew, NY, USA). We used the IOPcc, a cornea-corrected value of IOP produced by the ORA that the manufacturers claim to be less affected by corneal properties than other methods of tonometry, such as Goldmann. Measurements were made four times, and the average was used in the final result. The validity of this approach was confirmed on some data sets by repeating the analysis using the waveform score of the later version (version 2.04) of the ora software (Reichert Ophthalmic Instruments, Depew, NY, USA) (results not shown). Age, gender and refractive error were also recorded. Six intact human eyes from different donors (aged between 65 and 69) were obtained from the National Disease Research Interchange, Philadelphia, USA. The eyes were frozen within 6–8 hrs of death and shipped to the United Kingdom. The eyes were allowed to thaw and showed no signs of corneal swelling. The central corneal thickness was measured and was found to be within the physiological range in all cases. The IOP was adjusted by injecting the eye from the limbus with 0.9% sodium chloride. To keep the front surface of the cornea wet and thus produce a good optically reflecting surface, artificial tears were applied to the front of the cornea. The eye was gently clamped using a retort stand and carefully positioned in front of the ORA machine. When the cornea was in the correct position (only at this point was an ORA signal produced), we carried out CH measurements as described earlier. The same procedure was performed using twenty-one human corneas with at least 2-mm scleral rims, received from the Manchester Eye Bank, Manchester, UK. These corneas had low endothelial cell counts so were deemed unsuitable for transplantation. On arrival in culture medium, the corneas were found to have swollen by different amounts, with central corneal thicknesses (CCT) in the range 780–961 μm [measured using a Pachette 2, Model – DGH 550 ultrasonic pachymeter (DGH Technology Inc., Exton, PA, USA)]. The corneas were de-swelled using a standard equilibration method (Meek et al. 1991). Corneas were clamped within 12 000 MW cut-off dialysis tubing that was carefully smoothed to remove any air bubbles and to ensure close contact between cornea and tubing. They were then dialysed overnight at 4°C against a solution containing 0.154 m NaCl in 5 mm Hepes, pH 7.4, and 2.75% polyethylene glycol (MW 20 000). This procedure produces an osmotic gradient (Hodson et al. 1991) that equilibrates the cornea to normal hydration. This reduced CCT to 495 μm ± 13.0, close to the physiological level (∼520 μm). Each cornea was then removed from the dialysis tubing, carefully centred within an artificial anterior chamber and then clamped gently by its scleral rim. A balanced salt solution was placed within the chamber behind the cornea, and a pressure equivalent to an IOP was maintained by a supporting column of solution. This pressure could be adjusted by varying the height of the supporting column. For each cornea, CH was measured within the normal range of IOP and at an elevated pressure. Each cornea was mounted in front of the ORA in the same way as the whole eyes, bathed with artificial tears, and ORA measurements recorded as above. In all cases, ORA data were scrutinized carefully following procedures suggested in the ORA User’s Guide (Reichert Ophthalmic Instruments, 2005, ISO-9001/13485), and unsuitable readings were discarded. In the experiments with excised corneas, this meant that some data needed to be excluded, so that we were only able to use data from 17 of the original 21 corneas. Linear regression analysis was used to determine trends in the data. Statistical comparisons of the significance were carried out using spss (IBM SPSS statistics 16.0, Chicago, IL, USA). Nonoverlap of the 95% confidence intervals of regression lines was taken as evidence of statistical significance of differences between them. The residuals from the regression analysis were assessed for normal distributions using the Kolmogorov–Smirnov test and indicated that the data met the necessary assumptions for regression analysis. Pearson tests were carried out to confirm the correlation between the data sets. The Pearson correlation coefficient and p-value are indicated on each graph. Figure 1A shows that there is a negative correlation between CH and IOPcc (the Pearson correlation between CH and IOPcc was r = 0.53, p < 0.0001), whereas Fig. 1B indicates no correlation between CH and IOPG.Figure 2 shows the effects of left versus right eyes and male versus female eyes as a function of IOPcc. Comparison of confidence intervals in Fig. 2 showed that there is no statistical difference of this dependency between right and left eyes, or between males and females. Therefore, in all remaining results, data were used from left eyes, and were combined between the males and the females. Accordingly, Fig. 3A,B clearly show that there is no dependence of CH on either the age or the mean spherical equivalent of the subjects. The relationship between corneal hysteresis and IOPcc (A) and IOPG (B) in vivo. IOP, intraocular pressure. The relationship between corneal hysteresis and IOPcc for left/right (E/F), male/female (A & B/C & D) eyes in vivo. IOP, intraocular pressure. The relationship between corneal hysteresis and age (A) and mean spherical equivalent (B) in vivo. As was the case in vivo, the excised human eyes showed a significant negative correlation between CH and IOPcc (Fig. 4A). As expected, the scatter in the data was much less (r = 0.74 compared to r = 0.53 in vivo). The CH values from excised eyes were similar to those observed in vivo, and comparison between the CH-IOPcc dependencies observed in vivo and in vitro (1, 4) showed no statistical difference (p < 0.0001). Figure 4B shows that there is a significant correlation between CH and CCT in the excised eyes (r = 0.99; p < 0.0001). The relationship between corneal hysteresis and IOPcc (A) and central corneal thicknesses (B) in excised eyes (N = 6). IOP, intraocular pressure. It is also useful to deconstruct Fig. 4 to examine the response of the individual eyes, as it is not possible to evaluate the response of an individual eye to a change in IOP or CCT in vivo. The results are presented in 5, 6. The similarity in the gradients of each of the graphs in Fig. 5 that is apparent on visual inspection is confirmed statistically; in each case, every graph falls within the 95% confidence limits of every other graph in the figure. The results for CCT (Fig. 6) do not appear to be identical for every cornea. The relationship between corneal hysteresis and IOPcc in the individual excised eyes used in Fig. 4. IOP, intraocular pressure. The relationship between corneal hysteresis and central corneal thicknesses in the individual excised eyes used in Fig. 4. Data from the excised corneas also showed a negative correlation between CH and artificial chamber pressure, presented as the IOPcc reading from the ORA (r = 0.72, p < 0.0001) (Fig. 7A). However, the CH values recorded were lower than those observed over the same IOPcc range in vivo. This was not because of the slightly lower than normal thicknesses of the corneas. Although in the in vitro experiment there was a positive correlation between CH and CCT (r = 0.61, p < 0.0005) (Fig. 7B), the hysteresis values were consistently lower than values at the same CCT recorded in vivo (Kamiya et al. 2008), and the data show signs of levelling off above about 490 μm. The relationship between corneal hysteresis and IOPcc (A) and central corneal thicknesses (B) in excised corneas (N = 17). Our in vivo studies found no correlation between CH and gender, or whether left or right eyes were examined, as a function of IOP. This is in agreement with most of the previous reports cited earlier, although Fontes et al. (2008) did report gender differences. Several publications have indicated no dependence of CH on age (Kirwan et al. 2006; Montard et al. 2007; Kamiya et al. 2008; Liu et al. 2008) although others have shown a negative correlation (Kida et al. 2008; Moreno-Montanes et al. 2008). The insensitivity of CH to age found in many reports, including ours, is of interest as it is known that older corneas are considerably stiffer than younger ones (ElSheikh et al. 2007). This led Kotecha et al. (2006) to suggest that CH values should be normalized to produce what they term the cornea constant factor (CCF), which they found to increase with corneal thickness and decrease with age. The CCF may thus be related directly to corneal stiffness and viscoelasticity, and recently, these correction methods on ORA data have been improved to estimate IOP values that are independent of corneal stiffness (Kotecha et al. 2006; ElSheikh et al. 2009). Our studies also confirm that in vivo, there is an inverse correlation between CH and IOPcc, although not between CH and IOPG. This is also in accord with previous studies (Kotecha et al. 2006; Hager et al. 2008; Kamiya et al. 2008; Liu et al. 2008; Kopito et al. 2010). However, Kotecha et al. (2006) pointed out that although the IOP dependence may be real, it could also be influenced by the instrument measurement. The ORA uses a different maximum air pressure (Pmax) to achieve applanation in different subjects. The pressure at applanation (P1) depends both on the true IOP and on the resistive properties of the individual eye, which could themselves depend on IOP. Furthermore, the rate of application of the pressure force increases with P1, and the authors indicate that this could also affect hysteresis. Only one study has been reported in which CH values in individual eyes were measured before and after pharmacological reduction in IOP (Kotecha et al. 2006; Kotecha 2007). The study found a weak negative correlation between changes in CH and changes in IOP. The present paper is the first report of the use of the ORA to measure the relationship between CH and IOP in individual eyes ex vivo. This has several advantages, for example it allows much more precise control of the experimental conditions and it allows us to see if postmortem changes affect the measurements of CH. By changing the IOP within a single eye, intersample variability is obviated and, as expected, the data showed less scatter than was obtained from in vivo studies. Furthermore, ex-vivo studies allowed us to examine CH over a much wider range of IOP. The results for all six eyes (Fig. 4A) were very similar to those from a cohort of patients measured in vivo (Fig. 1A). This suggests that freezing/thawing and/or any postmortem changes in the eyes we examined did not affect CH or its dependence on IOPcc. Furthermore, it was interesting that all individual eyes showed the same dependence of CH on IOPcc (Fig. 5). It would appear likely therefore that a given cornea within the normal CCT range, and at a given IOP, has an intrinsic CH. In vivo, several authors have documented that there is an direct relationship between CH and CCT (Fontes et al. 2008; Kamiya et al. 2008; Liu et al. 2008; Schroeder et al. 2008; Mangouritsas et al. 2009). Excised eyes show a similar dependence (Fig. 4B), whereas excised corneas show a weaker correlation (Fig. 7B). However, comparisons with in vivo results may be misleading. In a clinical CCT measurement, it is difficult to distinguish between the two factors that could affect CCT: thickness related to tissue mass (CCTmass) and thickness related to tissue hydration (CCTswelling). Following refractive procedures such as laser-assisted in situ keratomileusis, corneal thickness changes are due mostly to loss of tissue (i.e. a change in CCTmass). This procedure has been found to cause a reduction in CH, although it is interesting that the reduction does not correlate with the amount or percentage of tissue removed (Kirwan & O’Keefe 2008). In our in vitro experiments, variation in CCT is likely dominated by tissue hydration (CCTswelling), particularly for the excised corneas, which needed to be deswelled before use. For this reason, caution is needed when comparing the results to those obtained from in vivo studies of the normal population where changes in both CCTmass and CCTswelling are likely to occur. Nevertheless, our results have shown for the first time that it is possible to obtain meaningful results using the ORA from whole excised eyes, and this opens up the possibility of using the ORA in controlled ex vivo studies that will allow us to better understand the relationship between CH measurements and conventional biomechanical properties such as stiffness and elasticity (Glass et al. 2008). When excised corneas with scleral rims were examined, great care was needed in the tissue preparation, and poor signals needed to be discarded in accordance with the manufacturer’s instructions. Nevertheless, with these provisos, we found that the CH values were reduced (between 8 and 22 mmHg, the mean values were: CHin vivo = 10.1 ± 1.6; CHwhole eye = 11.1 ± 2.16; CHexcised cornea= 7.2 ± 1.5). The inverse relationship with IOPcc was still present, but was also much reduced. These results are in agreement with in vivo studies, which have also indicated that the lower the CH, the lower its correlation with IOPcc (Touboul et al. 2008). However, there are a number of experimental factors whose influence on these results needs to be considered. First, there is the possibility of postmortem changes affecting the in vitro results. The fact that the same results were obtained from whole eyes postmortem and eyes measured in vivo, however, suggests that freezing/thawing and/or any postmortem changes in the eyes we examined did not affect CH or its dependence on IOPcc. This is not surprising as the biomechanical response of the cornea depends almost entirely on the collagenous component, which is highly resistant to proteolysis. Thus, in the absence of any in vivo technique other than the ORA, the body of literature on corneal biomechanics has all been carried out using postmortem eyes or excised corneas. Second is the fact that most excised corneas were originally swollen. There is a significant relaxation of posterior fibres with increasing hydration and this increases the extensibility of swollen corneas (Hjortdal 1995a,b). For this reason, tissues were equilibrated to near-normal hydration prior to CH measurement. Third, it has previously been reported that when a cornea is stressed at physiological pressure as part of an intact globe, it is less elastic than a cornea tested by strip extensiometry (Smolek 1994). Although this seems to support our results, it is not possible to make a direct comparison as the behaviour of a stretched corneal strip is likely to be different from the behaviour of a cornea inflated in an artificial chamber, where collagen fibrils have not been severed. Finally, the elastic properties of the cornea have been shown to be influenced by the presence of the limbus when IOP is varied (Asejczyk-Widlicka et al. 2007; Boyce et al. 2008). In our system, a scleral rim was used to clamp the corneas so that the limbus was retained during the measurements, which mimics the situation in vivo. So why are values of CH from isolated corneas different to those from whole eyes? To answer this one needs to consider what tissue changes might occur during applanation and recovery. Certainly, when the cornea is distorted by inflation, significant changes in the lamellae are observed (Wu & Yeh 2008) and the same may be the case during applanation (Spörl et al. 2009). Therefore, most energy is likely to be stored within the cornea and limbus, and this would be the same in vivo as in vitro. The absence of epithelial, endothelial and keratocyte function is not likely to have influenced our measurements; if it did, one would not have expected identical results in vivo and from excised eyes where cell function was lost. A more likely inference from our results is that the CH value depends to at least some extent, on the presence of the rest of the eye globe, as was suggested by Kacumen et al. (2008). This would explain how phacoemulsification (Kacumen et al. 2008), axial length (Song et al. 2008), primary open angle glaucoma (Schroeder et al.2008; Shah et al. 2008), high myopia (Shen et al. 2008) and deep sclerectomy (Iordanidou et al. 2010) are all associated with changes in CH even though the cornea is not directly involved. The authors thank Dr J. Guggenheim for help with the statistical analyses. This work is funded by the Medical Research Council [to KMM, grant (G0600755)] and by the award of a studentship to Alhamad T from the Ministry of Higher Education, Saudi Arabia. Keith Meek is a Royal Society/Wolfson Foundation Research Merit Award holder.