Refractive errors are due to an improper relationship between the refractive power of the cornea and lens, the depth of the anterior chamber, and the length of the eye globe. Myopia represents the most common refractive disorder with a prevalence of approximately 20–25% for Europeans, 50–70% for Chinese, and less than 10% for Eskimos. Despite intense research efforts, the etiology and pathogenetic mechanism(s) responsible for development of myopia, and for the exceedingly wide variation among ethnic groups, remain elusive. Traditionally, refractive errors have been corrected with spectacles or corneal contact lenses. However, the idea of obtaining maximal visual acuity without corrective eyewear has motivated a search for surgical alternatives to manipulate the ocular refraction permanently and predictably. The prime focus has been the cornea that represents the main refractive element of the eye. The anterior corneal surface has a refractive power of approximately 49 diopters (D) as compared to −6 D of the posterior corneal surface. Thus, of the eye's total refractive power of about 60 D, approximately 43 D or 72% is located in the cornea. Alteration of the anterior corneal curvature therefore offers a good opportunity for surgical correction of visual refractive errors. For the treatment of myopia, the overall intention has been to induce a controlled flattening of the central cornea (and a steepening of the periphery), and still maintain a smooth and optically transparent surface. During the last 25 years, many keratorefractive surgical procedures have been designed and tested based on various incisional, lamellar, and thermal principles. However, since the introduction of non-thermal, ultraviolet excimer laser photoablation, the field of refractive corneal surgery has widely expanded. In biophysical terms, excimer laser photoablation is based on ultraviolet radiation from an ‘excited dimer’ of argon and fluorine gas molecules, which exists transiently before dissociating with intense emission of energetic photons with a wavelength of 193 nm. The emitted photons are absorbed within a thin layer of the treated surface and have sufficient energy to break intermolecular bonds leading to ablative decomposition of the tissue into minor fragments that are ejected. Currently, there are two main treatment approaches: (1) photorefractive keratectomy (PRK), where the apical corneal epithelium initially is removed (by manual debridement or by transepithelial photoablation) followed by ablation of the denuded stromal surface (anterior stromectomy); (2) laser in situ keratomileusis (LASIK), where the photoablation is performed in the mid-stroma following temporary displacement of a 130–160 µm hinged (epithelial and stromal) tissue flap, formed with a microkeratome (intracorneal stromectomy). To reduce the anterior corneal convexity and correct 1·00 D of myopia, both procedures require removal of a stromal lenticule shaped as a biological contact lens with a central thickness of approximately 13 µm for a 6-mm diameter optical zone and 9 µm/D for a 5-mm diameter ablation zone (Munnerlyn et al. 1988; Colliac et al. 1994). The predictability of excimer laser surgery for myopia has so far never been comparable to that of corrective eyewear including spectacles and contact lenses. Today as well as in the past, the refractive outcomes of both PRK and LASIK have shown an exceedingly wide variation, with 70–100% of low myopes, 40–85% of moderate myopes, and 20–67% of high myopes being within ±1·00 D of intended refraction by 1 year post-surgery (Dutt et al. 1994; Epstein et al. 1994; Maguen et al. 1994; Talley et al. 1994; Seiler et al. 1994; Sher et al. 1994; Piebenga et al. 1995; McCarty et al. 1996; Schallhorn et al. 1996; Hersh et al. 1998; Tuunanen & Tervo 1998; Shah et al. 1998; Han et al. 2000). There appears to be two main sources of variation to account for this low predictability. Firstly, there is no precise control of the actual photoablation depth during surgery in each individual patient (i.e. the exact amount of stromal tissue removed). Secondly, there is no precise control of the postoperative wound-healing response that tends to distort the induced refractive correction by addition of variable amounts of new repair tissue. The threshold for achieving photoablation of the corneal stroma is about 30 mJ/cm2 (beam energy density) for collagen and 40 mJ/cm2 for keratocytes (Berns et al. 1999). Below these levels, no ablation will occur. Above these thresholds, a linear relationship exists between ablation rate and beam energy density with 0·4–0·5 µm stroma removed per pulse at the energy levels (160–180 mJ/cm2) currently used in most excimer lasers (Dougherty et al. 1994; Huebscher et al. 1996). However, the laser–cornea interaction is dependent on tissue hydration at a given energy level. Dehydrated corneas generally have more dry tissue mass removed per laser pulse (leading to over-correction), whereas edematous corneas show less dry tissue mass removed per laser pulse (leading to under-correction) (Dougherty et al. 1994). Many factors may influence corneal tissue hydration prior to treatment (including time and technique for epithelial denudation, temperature and humidity of environment, use of irrigation solutions, etc.) leading to inter-session variation in the photoablation process in the order of ±10% (Dougherty et al. 1994). Besides, there may exist unique inter-individual variation in the physiological (anterior–posterior) tissue hydration profile (and dry tissue mass components) of the preoperative stroma, leading to differences in the actual photoablation rate. These potential variations are currently not accounted for. It is also important to note that the actual laser beam intensity (in most lasers) is manually adjusted prior to surgery, based on the surgeons' subjective judgement of the ablative performance (e.g. on a piece of metal foil). This calibration procedure may be an important source of inter-session variation. Thus, the nominal photoablation depth displayed on the excimer laser prior to treatment (i.e. the nomogram) is empirical and based on mean population responses. Currently, no valid method exists for the assessment of individual photoablation depths in single patients. There is therefore a fundamental need to develop an accurate corneal biometry method to see if the laser treatment actually has induced the desired thinning immediately post-ablation (I, II). The refractive changes after excimer laser photoablation usually follow a characteristic temporal pattern. Immediately following treatment for myopia, the patient is left with an intentional (hyperopic) over-correction to account for the subsequent regression of treatment effect. During the first year after both PRK and LASIK, the refraction typically undergoes a 0·8–3·0 D myopic regression (gradually becoming less hyperopic and more myopic), especially following higher corrections (Dutt et al. 1994; Epstein et al. 1994; Maguen et al. 1994; Seiler et al. 1994; Sher et al. 1994; Talley et al. 1994; Piebenga et al. 1995; McCarty et al. 1996; Kim et al. 1997; Williams 1997; Chayet et al. 1998; Gartry et al. 1998; Hersh et al. 1998; Tuunanen & Tervo 1998; Han et al. 2000). For this reason, all excimer laser algorithms work with built-in (over-) correction factors to account for the loss of treatment effect (Huebscher et al. 1996). However, these adjustment factors are not based on direct measurements of corneal sublayer wound healing but are empirically modified from the average refractive outcome (mean population response) of clinical trials. Since the magnitude of the initial over-correction is empirical and since the subsequent refractive regression varies considerably among individuals, the procedure may leave the patient emmetropic, hyperopic, or myopic. There is currently no way of predicting those patients who will develop profound refractive shifts after photoablation. Although the refractive changes appear to level off after 1 year, longer term studies of both PRK (2–5-year follow-up) and LASIK (2-year follow-up) have revealed a continued potential for ongoing myopic regression with time (Epstein et al. 1994; Haviv et al. 1997; Kim et al. 1997; Williams 1997; Tuunanen & Tervo 1998; Han et al. 2000), suggesting that corneal wound healing is considerably delayed in humans and may require decades to be completed. Still the permanency and long-term consequences of excimer laser procedures are unknown. Despite the high frequency and clinical importance of postoperative refractive regression following PRK and LASIK, there has been a major lack of clinical studies of the underlying pathogenetic mechanism(s). The limited knowledge available originate predominantly from ex vivo histology of fixed and processed corneal specimens from experimental animals. Although the validity of such observations clearly is questionable, it is generally suspected that both epithelial and stromal wound healing may distort the induced refractive correction by addition (regrowth) of new tissue filling in the photoablated region. Thus, many investigators of PRK-procedures have noted an apparent and most often transient compensatory hyperplasia of the corneal epithelium (up to 60–70 µm in primates and rabbits as compared to 40–50 µm in normals) (Hanna et al. 1989, 1990, 1992; Tuft et al. 1989; Del Pero et al. 1990; Fantes et al. 1990; Beuerman et al. 1994; Amm et al. 1996; Lohmann et al. 1999). Subepithelial deposition of various new and abnormal extracellular matrix components (i.e. stromal fibrosis) has also been noted post-PRK (Tuft et al. 1989; Del Pero et al. 1990; Hanna et al. 1990, 1992; Malley et al. 1990; Fitzsimmons et al. 1992; Latvala et al. 1995; Amm et al. 1996; Weber et al. 1997). However, none of these static ex vivo studies have provided true quantitative or prospective data on the stromal regrowth, why important characteristics such as magnitude, time course, and regulatory stimuli have remained unidentified. As for human in vivo data, a few clinical studies have indicated central corneal re-thickening after both PRK (Ehlers & Hjortdal 1992; Krueger et al. 1995) and LASIK (Chayet et al. 1998; Perez Santonja et al. 1999), symmetric in both time course and magnitude with the induced myopic regression and leading to central corneal re-steepening. However, the magnitude and relative contribution of epithelial and stromal wound healing have largely remained unidentified due to a lack of sensitive corneal biometry methods. It should also be noted that biomechanical distortion of the remaining photoablated tissue (progressive stromal ectasia or ‘bulging’) has been proposed as a third potential mechanism of regression following both PRK (Ramirez-Florez & Maurice 1996) and LASIK (Seiler et al. 1998; Seiler 1999). Due to the linear relationship between attempted correction and keratectomy depth, the biomechanical stability of the residual stromal lamellae may obviously be of concern following high corrections. Currently, the accepted limit for the minimum thickness of the stress-bearing residual stroma is 250 µm (Chayet et al. 1998; Seiler et al. 1998; Seiler 1999), which relates directly to the possible upper limit of myopic laser correction without inducing stromal weakening and refractive destabilization. Clearly, identification of the specific mechanism(s) involved in the development of regression following both PRK and LASIK will be essential to our ability to formulate therapeutic strategies to effectively control and improve the refractive surgical outcomes of these procedures. It is therefore essential to develop an accurate in vivo technique for direct and simultaneous in vivo measurements of epithelial and stromal thickness to specifically assess the role of tissue addition (re-thickening) in explaining refractive regression after excimer laser refractive surgery (I, II, VII). Excimer laser PRK usually causes the very transparent cornea to lose some of its optical clarity and develop diffuse and reticular haze, within the photoablated subepithelial region (Epstein et al. 1994; Maguen et al. 1994; Seiler et al. 1994; Sher et al. 1994; Talley et al. 1994; McCarty et al. 1996; O'Brart et al. 1996; Schallhorn et al. 1996; Hersh et al. 1997; Williams 1997; Kremer et al. 1999). By contrast, LASIK patients show a lower incidence of opacities at the photoablated interface, but the optical quality of LASIK is often variable due to flap-related complications including wrinkles or striae (Hersh et al. 1998; Perez Santonja et al. 1999; Stulting et al. 1999). Clinically on the slit-lamp, the subepithelial haze is apparent several weeks following PRK, peaks at about 3–6 months, and progressively clears up during the following 9–12 months. During this healing period, haze often reduces visual function and limits low contrast visual acuity performance (Lohmann et al. 1991; McCarty et al. 1996; Shah et al. 1998). Most important, however, visually significant haze persists in the optical axis of 3–5% of all PRK-patients, especially following higher corrections. However, there is currently no method available for predicting those patients who might have an exaggerated wound-healing response and develop severe haze. Although multiple factors may be associated with the occurrence of haze, only a few have been identified. It appears that the likelihood and severity of haze is principally related to the maximal photoablation depth, with higher corrections (i.e. deeper ablations) generally being associated with greater corneal opacification (Ehlers & Hjortdal 1992; Carson & Taylor 1995; McCarty et al. 1996; Williams 1997; Shah et al. 1998; Kremer et al. 1999). By contrast, there appears to be no consistent association with the presence of refractive regression (Kim et al. 1997; Gartry et al. 1998). Currently, there is no good hypothesis to explain why haze occurs and specifically why some eyes have persistent haze. Clearly, the biological causes of corneal haze development need to be fully elucidated before rational strategies to eliminate them can be initiated. Although previous studies have attempted to relate the development of postoperative corneal haze to stromal wound repair, the true origin and location of haze remains unknown. Based on histologic ex vivo studies of photoablated human, primate, and rabbit corneas, a number of potential causes of increased light scattering have been suggested. These include: (1) subepithelial deposition of new stromal extracellular matrix components (collagen, proteoglycans, and other macromolecules) with abnormal composition and organization (i.e. stromal fibrosis); (2) appearance of vacuoles or membranous inclusions in the superficial stroma; (3) altered tissue hydration with focal areas of intercellular edema; (4) irregularities at the photoablated stromal surface with disruption of the regular arrangement of stromal lamellae; and (5) increased cellularity of activated wound-healing keratocytes in the anterior stroma (Hanna et al. 1989, 1992; Tuft et al. 1989; Del Pero et al. 1990; Fantes et al. 1990; Malley et al. 1990; Lohmann et al. 1991; Campos et al. 1992; Fitzsimmons et al. 1992; Rawe et al. 1992; Beuerman et al. 1994; Latvala et al. 1995; Amm et al. 1996; Ramirez-Florez & Maurice 1996; Weber et al. 1997). Although these ex vivo studies have provided valuable insight into corneal wound healing, the exact structural and cellular basis for haze after photoablation is disputed and has remained unidentified. Specifically, it is uncertain whether haze predominantly originates from cellular or extracellular structures. This is largely a result of an inability in previous studies to assess the light scattering properties of the observed abnormal histopathologic structures. It is therefore important to identify the exact structural origin of corneal haze after photorefractive surgery by a direct correlation of in vivo histopathology with the amount and z-axis localization of corneal light reflectivity over time (III). In the search for specific factors controlling the magnitude of corneal wound healing following excimer laser photoablation, authors have focussed on the initial keratocyte loss seen after manual epithelial debridement. This interesting observation was first made by Dohlman et al. (1968) and later confirmed by others (Nakayasu 1988; Crosson 1989; Hanna et al. 1989; Fantes et al. 1990; Campos et al. 1994a, b; Szerenyi et al. 1994). Thus, it appears that the anterior keratocytes promptly die after simple epithelial abrasion (scraping) and leave a continuous 50–150 µm acellular stromal zone that subsequently becomes repopulated over the next 4 weeks. Investigators have focussed on possible epithelial–stromal interactions underlying this phenomenon, and it has been proposed that specific mediators (cytokines) may be released from the injured epithelium and may induce apoptosis (i.e. programmed cell death) in the underlying keratocytes (Wilson et al. 1996a, b). These authors also identified slightly lower levels of keratocyte apoptosis following excimer laser transepithelial photoablation (Helena et al. 1998; Kim et al. 1998). Taken together, these ex vivo findings have prompted the idea that a decreased keratocyte loss may lead to a diminished corneal wound-healing response, providing a fundamental rationale for performing transepithelial photoablation, especially for treatment of higher levels of myopia and for retreatment of regression (Wilson 1997; Helena et al. 1998; Kim et al. 1998; Wilson & Kim 1998). However, this hypothesis has never been confirmed by actual comparative measurements of the initial keratocyte loss and subsequent wound-healing parameters. It is therefore important to conduct an in vivo study to establish definitively the relationship between initial depth of keratocyte killing and stromal loss, and subsequent repopulation and development of corneal haze (IV). It is well recognized that many growth factors and cytokines provide important regulatory signals in the complex cascade of wound-healing events after corneal injury. The most potent and direct mediator of new stromal tissue formation appears to be transforming growth factor beta (TGFβ) that promotes the formation of fibrosis in both corneal and dermal tissue. As a multifunctional and auto-inductive cytokine, TGFβ stimulates corneal keratocyte proliferation, migration, and myofibroblast transformation and enhances the synthesis of stromal extracellular matrix components such as fibronectin and collagen, while reducing stromal degradation by inhibiting the synthesis of matrix metalloproteinases (Girard et al. 1991; Grant et al. 1992; Ohji et al. 1993; Jester et al. 1996; Andresen & Ehlers 1998). The critical role of the TGFβ-mediated cytocrine pathway has further been supported by the finding in experimental animal models that topical application of neutralizing antibodies to TGFβ significantly reduces both dermal scar tissue formation and corneal fibrosis following lamellar keratectomy (Shah et al. 1994, 1995; Jester et al. 1997). By analogy, early inhibition of the concentration of TGFβ may prove to be a clinically useful strategy to directly control and modulate corneal wound healing following excimer laser procedures. It is therefore of interest to perform a detailed analysis of the therapeutic potential of anti-TGFtreatment in preventing haze development and regression of photoablative effect following excimer laser PRK (V). According to the well-established lattice theory of Maurice, the optical transparency of the normal cornea depends on the uniform diameter and spacing of collagen fibrils in the stromal extracellular matrix, leading to destructive interference of scattered incident light, but allowing light transmission in the forward direction (Maurice 1957). Attempts to explain loss of transparency, including haze development after excimer laser procedures, have therefore understandably focussed on the disruption of the collagen fiber array (i.e. stromal fibrosis) as the principal source of corneal light scattering. However, it appears from in vivo confocal microscopic studies (II–V, VII) that the major intracorneal structures showing enhanced light scattering after photoablation are the cell body and cell processes of the stromal keratocytes, challenging the current explanations of corneal tissue transparency. There appears to be a close relationship between the loss of corneal transparency (haze) and the presence of highly reflective wound-healing keratocytes which exhibit enhanced cellular-based reflections from their normally invisible cytoplasm. Thus, it seems that the reflectivity of keratocytes can be environmentally altered, suggesting a dynamic regulation of the relative refractive index between the cells and their extracellular matrix. This cell-related loss of corneal clarity is not taken into account by the conventional models of corneal transparency based on the regular ultrastructural array of tightly packed, orthogonally arranged collagen fibers and suggests a novel and previously unrecognized cellular contribution to the maintenance of normal corneal transparency. Interestingly, a similar enhancement of keratocyte reflections can be induced by simple formaldehyde fixation of ex vivo eyes, revealing fine details of the broad cell processes (Jester et al. 1992). This finding suggests that fixation may alter the relative refractive index of either the matrix or the keratocyte contents, which may be explained by the precipitation of intracellular soluble proteins, resulting in a change in the cellular refractive index relative to that of the surrounding matrix. It should be noted that the transparent intraocular lens of the normal eye is known to contain major water-soluble proteins, termed crystallins, responsible for maintaining its optical properties (Benedek 1997). By analogy, the clear lens loses its normal transparency when exposed to temperatures below 15°C due to precipitation of these intracellular lens proteins (i.e. ‘cold cataract’) (Benedek 1997). In an attempt to further explore the ‘stealth-like’ invisibility of normal corneal keratocytes and the cellular-based light scattering of wound-healing fibroblasts, it is therefore relevant to search for unrecognized, water-soluble crystallin-like proteins in the cytoplasm of keratocytes from clear and opaque corneal tissue (VI). Based on the considerations stated above, a parallel series of clinical and experimental animal studies (I–VII) were designed and initiated. The overall intention was to assist in elucidating the biological causes and structural origin of: (1) instability of the intended refractive changes (regression); and (2) loss of corneal optical clarity (haze) following excimer laser PRK. The specific aims of these investigations were: To develop a new biometry method based on in vivo confocal microscopy that allows for quantitative evaluation of corneal sublayer wound healing by providing precise epithelial and stromal thickness measurements and unbiased assessment of corneal light backscattering (haze). To measure the actual photoablation depth in PRK-patients and to assess the role of epithelial and stromal re-thickening in explaining regression of refractive effect. To identify the true location and structural basis for corneal haze development following PRK by correlating in vivo histopathology with the amount and z-axis localization of corneal light scattering at differing temporal intervals in humans and rabbits. To determine how temporal changes in rabbit corneal thickness after PRK relate to stromal acellularity, fibroblast activation and migration, epithelial regeneration, and stromal regrowth. To characterize corneal wound repair following epithelial abrasio in rabbits and to determine whether excimer laser transepithelial photoablation can reduce the initial keratocyte killing. To establish the relationship between initial keratocyte loss, volume of stromal tissue removal, and the magnitude of wound healing and haze development in photoablated rabbit corneas. To evaluate the role of TGFβ in post-PRK stromal wound repair by neutralizing the bioactivity of TGFβ in a rabbit eye model using specific blocking antibodies. To investigate the basis for cellular transparency by analyzing the expression of specific water-soluble crystallin-like proteins in the cytoplasm of keratocytes from clear and hazy rabbit corneas. Due to their non-invasive optical sectioning ability, in vivo confocal microscopes are ideally suited for dynamic evaluation of corneal wound healing in three or four dimensions (x, y, z, time). Using this form of vital microscopy, high-resolution en face images can be obtained in real time from different depths within the intact living cornea without the need for staining or processing. The technique enables direct visualization of temporal changes in corneal wound-healing morphology in the same living eye at the cellular level of magnification. Cellular responses and structural changes can thereby be correlated directly and sequentially over time to clinical observations. During the past few years, several such in vivo confocal microscopes have been developed and applied clinically (Cavanagh et al. 1993; Auran et al. 1995; Wiegand et al. 1995). However, the tandem scanning confocal microscope (TSCM) used in present studies has one major advantage compared to other systems. The TSCM has a specially designed objective in which internal lenses can be accurately moved to vary the distance from the objective tip to the position of the focal plane (optical section) without changing the position of the objective surface (Petroll et al. 1993). The exact z-axis depth from which the image originates can therefore be precisely defined and controlled (in microns) and concurrently displayed during in vivo imaging using a depth encoding system (Petroll et al. 1996). Due to a tissue penetration of 0–1·5 mm (working distance), the TSCM system enables a direct quantitative evaluation of the three-dimensional location (z-axis position) of cellular structures throughout the cornea. Besides the objective lens, the TSCM has the following main characteristics. It is mounted horizontally to operate similar to a standard slit-lamp (Cavanagh et al. 1993). The light source is a 100-W mercury arc lamp (broadband white light) that is filtered using ultraviolet and heat filters to prevent light toxicity to the eye. The optical path contains a 0·25% transmittance, glass Nipkow disc with 35 000 pinholes (20 µm in diameter) arranged densely along multiple Archimedean spirals (Cavanagh et al. 1993; Petroll et al. 1996). This disc pattern has a bilateral symmetry so that light beams passing through (illuminated) pinholes on one side of the disc are focussed into a thin slice of the specimen. Reflected light from the illuminated volume are then brought into focus at the exact conjugated (detector) pinholes on the opposite side of the disc (dual light path). When the disc is rotated at a speed of about 1000 rev/min, the pinhole arrangement covers and rebuilds the whole field of view, although each single pinhole only scans a single line. Most importantly, all light beams that do not originate from the focal plane are intercepted by the opaque areas on the disc and are prevented from reaching the video camera (rejection of out-of-focus signals). Overall, this TSCM design produces serial, two-dimensional images with an approximate field-of-view of 450 × 360 μm, an effective lateral (x, y) resolution of about 1 µm, and an axial (z) resolution of approximately 9 µm (optical slice thickness). Moreover, the technique has a remarkably enhanced sharpness and contrast (high signal-to-noise ratio) that enables visualization of low reflecting structures such as keratocyte nuclei (Petroll et al. 1993, 1996) (I, II). Although TSCM enables direct imaging of living cells and undisturbed physiological processes within the cornea, it is important to validate the in vivo morphological findings by comparison to well-known ex vivo histology. Such comparisons may be compromised by the uncontrollable processing artifacts inherently associated with ex vivo techniques including conventional light and electron microscopy. However, there generally appears to be a good correlation between parallel in vivo and ex vivo observations of all layers in the normal cornea including: various layers of the epithelium, epithelial basal lamina, subepithelial nerve plexus, intrastromal nerves, different subpopulations of keratocytes, and endothelium (Ichijima et al. 1992; Jester et al. 1992; Somodi & Guthoff 1995; Böhnke & Masters 1999). During corneal wound repair, it is similarly important to verify in vivo histopathological observations by coupling to ex vivo techniques. Such documentation has included identification of specific keratocyte repair phenotypes (by immunofluorescently labeled stress fibers) and verification of extracellular changes such as deposition of new matrix components (by specialized stains) (III, IV) (Ichijima et al. 1994; Jester et al. 1995). Still, morphological changes detected by TSCM should be interpreted with caution. To evaluate the biological mechanisms causing refractive instability following PRK, it is essential to obtain accurate information on total corneal and sublayer (epithelial and stromal) thickness. The traditional and most widely used corneal biometry methods include (non-contact) optical pachymetry (Mishima & Hedbys 1968; Olsen et al. 1980a), (non-contact) specular microscopy (Olsen & Ehlers 1984), and (contact) ultrasonic pachymetry (Salz et al. 1983). Over the past few years, new and more sophisticated systems have been introduced including (non-contact) automated video slit-lamp optical pachymetry (McLaren & Bourne 1999), (non-contact) Orbscan optical scanning sl
Frequency doubling technology perimetry (FDTP) is a new unconventional method of visual field testing which analyses selectively the My ganglionic cells, which have a very low redundancy. A small CRT monitor displays stimulus patterns of sinusoid gratings (alternate vertical black and white bars) with low spatial frequency (0.25 c/d) and high temporal frequency counterphase flicker (25 Hz). The threshold is the amount of contrast needed to perceive the stimulus and is measured in 17 visual field locations (four per visual field quadrant plus a central 5° stimulus). The central 20° visual field is tested. Test outcome appears both as a numerical table and as probability maps with five grey tones. A very swift screening programme is also available. A number of methods have been proposed in the last few years to obtain an objective visual field test, unlike automated perimetry, which is a psycho-physical subjective test: pupillo-perimetry, in which either the amplitude or the latency of pupillary responses to a bright stimulus are recorded by linking an automated perimeter to an infra-red electronic pupillometer; multifocal ERG, in which a number of areas (usually 103) within the central retina are stimulated by using white and black hexagonal stimuli presented on a CTR screen. The signals are recorded by a bipolar contact lens and responses are extracted from each stimulus element; multifocal VEP, in which pattern stimuli in a dartboard configuration are used. The recordings are collected by using two electrodes placed on the occipital scalp. Peak-to-trough amplitudes for each wave are determined. Automated flicker perimetry (AFP) is an unconventional testing method which measures the critical fusion frequency (the number of flickering stimuli/s at which the light does not appear to flash). AFP, like FDT, analyses selectively the magnicellular system. This technique should be more sensitive than standard automated perimetry in detecting early glaucomatous defects, but it is a lengthy and tiring test for patients. A special software has recently been created by Matsumoto to perform this technique with the Octopus 1-2-3 perimeter and will be included in the next generation of Octopus instruments (Octopus 301), together with new faster strategies. Medmont M700 automated perimeter has a programme that uses temporal modulation perimetry, where the flicker frequency is fixed and stimuli intensity increases up to perception. The capability to perceive movement is a particular visual ability mediated by the magnicellular system. A number of studies using different techniques have demonstrated that this ability is affected in the early stages of glaucoma. The main techniques which study the perception of movement are: Peripheral displacement threshold (Fitzke et al. 1990) in which two thin bars are moved from side to side above and below the blind spot up to perception; Motion perception perimetry (Silvermann et al. 1990; Bullimore et al. 1993; Trick et al. 1995), in which a number of stimuli have a coherent movement on a screen where random stimuli are moving; Random dot motion perimetry (Wall et al. 1995) in which, against a background of fixed dots, 50% of dots have a coherent movement within a circular area which is enlarged up to perception; Motion coherence perimetry and automated motion perimetry (Joffe et al. 1997; Bosworth et al. 1997; Brusini, 1999) in which an increasing percentage of coherent dots move within areas of various width and shape on a screen where similar dots have a random movement.
Contradicting results concerning IOP control and visual field deterioration are presented. Some of these inconsistencies may be due to the statistical method of analysis. Sixty POAG patients with a perimetric follow-up over 3 years were selected. Mean and maximum IOPs were considered during the same period. The patients were divided into two groups according to the IOP control (well controlled or poorly controlled). Visual field progression was defined as a reduction in sensitivity over the fifth percentile in more than four points. Mean IOPs were not significantly different in the group of patients with a visual field deterioration compared to the stable ones, but the percentage of patients with a visual field deterioration was significantly higher in patients with higher IOPs. This holds especially true if IOP below 16 mmHg (G) is considered the 'target pressure'. IOP reduction seems to play an essential role in visual field progression. In glaucomatous patients, a strict ( 16 mmHg (G)) might be necessary.
To verify and quantify hypertrichosis, if any, of the eyelashes during 6 months of topical treatment with latanoprost. Forty-four patients (mean age 69 years, range 33–91, M/F = 20/24) affected by POAG or ocular hypertension were enrolled and put into two groups. One group was treated with latanoprost 0.005% daily (n = 26) and a control group (n = 18) with timolol 0.5% b.i.d. Eyelid eyelash lengths were measured with a surgical compass before treatment and after 2, 4 and 6 months during treatment. A variance analysis for repeated measurements was used to evaluate the significance of the results. The sample sizes were enough to establish a significance (α = 0.05) of an eyelash length variation above 1 mm with a power (1 − β) of 80%. Two patients in the group treated with latanoprost and none in the control group showed an aesthetically apparent growth (above 1 mm) of the eyelashes in both eyelids. In both groups, eyelash length variation of both upper and lower eyelids at the checks were not statistically significant (all P > 0.2), although the latanoprost-treated patients showed a slight tendency towards an increase (with a mean of 0.2 mm) in eyelash length in both eyelids at the 6-month check. After 6 months of taking latanoprost, the hypertrichotic effect is negligible and not significant statistically, being aesthetically evident in only a modest percentage (7%) of subjects so treated.
Standard threshold perimetry is based on the detection by the patient of localized light spots shown on a homogeneously illuminated background. Many different techniques are used to assess visual field. High pass resolution perimetry (HRP) (Nikon-High Tech Vision, Malmö, Sweden) is based on resolution or acuity targets and has been proposed as an alternative to standard threshold perimetry in order to identify early visual field defects by detecting P cell loss (Frisén 1987; Frisén 1993). HRP showed a better reproducibility and also a strong capability of detection of early visual field dishomogeneity in the perimetric glaucomatous damage. Frequency doubling technology (FDT) (Welch Allyn, Skaneateles, NY; Zeiss Humphrey, San Leandro, CA, USA) is a perimetric technique recently introduced into the clinic. It is based on a different type of stimuli that are detected by retinal ganglion cells which have the larger mean axon diameter sizes associated with magnocellular cells (M cells) (Johnson & Demirel 1997; Quigley 1998). The aim of this study is to evaluate the correlation between two different perimetric techniques, HRP and FDT, which theoretically detect the loss of different patterns of retinal ganglion cells: P-cells and M-cells, respectively. Eighty-two eyes were consecutively included in the study. Patients were not excluded on the basis of media opacity, gender, age or race. Visual fields were assessed by a Humphrey Field Analyser 750 (HFA), 30–2 program, which tested the central 30° of the visual field. Patient refractive errors ranged from −7 to +7 dioptres. Only one eye per patient was randomly selected for data analysis. Patients were classified as having POAG when they had a typical glaucomatous visual field, a typical abnormal ONH/RNFL, or both, open angle on gonioscopy and no clinically apparent secondary cause for their glaucoma (EGS 1998). Patients with ocular hypertension were identified if they had high intraocular pressure (>21 mmHg untreated), normal visual field and normal ONH and RNFL (EGS 1998). All the patients had their visual fields assessed by HRP and FDT. In HRP, thresholds are determined by varying the size of the stimuli, which have fixed luminance characteristics. The simulus is ring-shaped. Each ring is made up of a bright core with darker boundaries with a luminance of 25 cd/m2, respectively. Stimuli below threshold are not visible because they vanish into the 20 cd/m2 background. With this technique, as soon as the patient sees the stimulus, he also perceives its shape. In this technique, thresholds for perception (the identification of a stimulus without recognition) and for recognition (the identification of the form and margins of the stimulus) are calculated simultaneously. Therefore, the perception of a stimulus coincides with the recognition of the stimulus, a function which involves the neuronal system sensitive to the high frequency levels of the parvocellular chain (Frisén 1989). The FDT provides a rapid means for detecting glaucomatous and neurological visual field defects. The technique demonstrates high sensitivity and specificity and can quantify visual field loss accurately (Johnson & Samuels 1997; Iester 2000; Iester et al. 2000). The FDT presents stimuli on a black and white videomonitor with specialized control circuitry interfaced to a microprocessor. An optical system is used to display the stimulus at optical infinity, with an eye-piece adjustment provided to correct for spherical refraction errors up to 7 dioptres. During program C-20, full threshold, 17 points are tested, one round centrally and 16 square ones in the periphery up to 20° eccentricity. The stimulus angular width is approximately 10° × 10° peripherally and 5° × 5° in diameter for the centre one. Each stimulus consists of a 0.25 cycles/degree sinusoidal grating undergoing 25 Hz (50 times/s) counterphase flicker (contrast reversal of light and dark bars). For all stimuli, the total exposure time is 2 s with a 1-s interval between trials. Test time ranges between 4.5 and 5 min. The location for each stimulus presentation is selected randomly and the contrast betwee black and white bars is modified according to the conventional ‘bracketing’ threshold strategy of the automated standard perimetry. The threshold value for each test location is defined by the minimal contrast of the pattern that is perceived. All three functions are very important in glaucoma for early diagnosis and a test exploring them is of strong clinical interest (Johnson & Demirel 1997; Johnson & Samuels 1997). The HRP indices global deviation (GD) and local deviation (LD), and the FDT indices FDT–mean deviation (FDT-MD) and FDT–pattern standard deviation (FDT-SD) were considered for analysis. All the data were analysed by Student's t-test and Pearson's r-coefficient when the distribution of the data was normal, and by Spearman coefficient correlation and Mann–Whitney test when the distribution of the data was not normal. A P < 0.005 was considered statistically significant. Table 1 lists the demographic data of the patients considered. When the entire group was analysed, significant ( P < 0.001) correlation was found between the HFA indices and those of HRP and FDT, as it also was when the glaucomatous group alone was considered. When the correlation between HRP and FDT indices was analysed, strong correlation was found between GD and FDT-MD (−0.57) and between LD and FDT-PSD (0.55) both for the entire group and for the glaucomatous group. Significant correlation was found in the ocular hypertensive group between LD and FDT-PSD. The correlation between standard threshold perimetry indices and the indices of the other perimetric techniques are of great interest in the studies concerning the early detection of glaucomatous visual field loss. It has been shown that the glaucomatous alterations of ONH and RNFL caused by nerve fibre loss precede the perimetric defect (Pederson & Anderson 1980). As it is clinically relevant to detect visual field glaucomatous damage at an early stage, an array of psychophysical tests is being investigated for this purpose. Furthermore, many new perimetric techniques are also available to detect very early visual field damage, including HRP, FDT, short-wave automatic perimetry, motion perimetry and flicker perimetry, each of which is based on a different theoretical concept (Wiesel & Hubel 1966; Breton & Drum 1996). Many studies have shown correlation between standard threshold perimetry and some of these new techniques. FDT was found to have very good sensitivity and specificity (Johnson & Samuels 1997), while a good correlation of sensitivity and specificity was demonstrated between Humphrey and FDT indices (Quigley 1998). A good correlation between FDT and Humphrey 30–2 perimetry results has been shown (Sponsel et al. 1998) and also a significant correlation between FDT and Octopus visual field (Iester et al. 2000). A strong correlation exists between the indices of Octopus perimeter, program G14, and the indices of HRP program Ring (Dannheim et al. 1988/9) and between HFA and HRP indices (Chauhan et al. 1993). Also in this study, a significant correlation was found between FDT-MD and HRP-GD, and between FDT-PSD and HRP-LD in glaucomatous patients. A further significant correlation was found between the indices FDT-PSD and HRP-LD, which showed localized damage of visual field in the ocular hypertensive group. The significant correlation suggests that early glaucomatous damage of the visual field generally starts with localized sensibility depression. In conclusion, these data suggest that both FDT and HRP are very useful for screening populations and detecting early glaucomatous visual field progression in early and moderate stages of the disease. However, not very evident from these data are the different pathways used by the two techniques, but both new techniques seem to measure ganglion cell loss.
To compare the 24-h pattern of intraocular pressure (IOP) in three groups of white individuals: healthy young, healthy aged subjects and untreated patients with primary open-angle glaucoma (POAG) or ocular hypertension (OHT). Twenty young volunteers (aged 23–27 years), 20 older subjects (aged 55–75 years) and 20 untreated patients with diagnosis of POAG or OHT (aged 55–75 years) were hospitalized and their IOPs were measured at 3, 6, 9 a.m. and 12 noon, and at 3, 6, 9 p.m. and 12 midnight with a handheld electronic tonometer with the patient in the supine and sitting positions, and then with a Goldmann tonometer, by two well-trained evaluators. Systemic blood pressure was recorded at the same intervals. Supine position tonometric readings were higher during the night in the young group (P < 0.04), while this pattern could not be observed in the other two groups. A highly significant (P < 0.001) difference between supine position and Goldmann IOPs was found at all points of the curve, although in the young group, more evident during the night. In the older, healthy group, no significant IOP peak or trough could be observed in supine position or Goldmann readings, which were significantly higher than measurements in the young group. POAG/OHT patients showed significantly higher IOPs (P < 0.001) than the other two groups, and a morning-type pattern was observed with a peak of IOP at 9 a.m. A nocturnal elevation of IOP was found only in the young group and was observed only in the supine position. Posture seemed to have a limited effect in older people, both healthy and diseased. Goldmann IOPs were higher in older than in young volunteers. A morning peak of IOP was observed in POAG or OHT untreated patients.
Primary angle closure glaucoma (PACG) is determined by the apposition of iris tissue on the angular region, which, in turn, causes a sudden increase of intraocular pressure (IOP). This clinical situation is characterized by pain and a decrease in vision. The classical treatment for PACG is medical therapy, either topical or systemic, associated with peripheral iridotomy. The crucial part of PACG treatment is IOP reduction quickly, in order to avoid permanent damage to the optic nerve. PACG medical therapy includes pilocarpine, beta-blockers, hyperosmotic agents and intravenous carbonic anhydrase inhibitors. All these systemic hypotensive drugs can have serious side-effects or be contraindicated in patients suffering from renal or hepatic failures, or congestive heart failure. Their use, in consequence, should be limited in hospitalized patients. In spite of the use of systemic and topical drugs, IOP can remain elevated and difficulties encountered in performing peripheral iridotomy. In these cases, it is important to try to open the iridocorneal angle mechanically. Argon laser peripheral iridoplasty (ALPI) has been proposed to try and 'remove' iris tissue from the angular region and thus try to restore normal levels of IOP in cases of acute PACG attack. The aim of this study was to determine if ALPI was a effective treatment for the relief of an acute PACG attack. Twenty-two consecutive patients were included in the study, with inclusion criteria: (a) first attack of PACG; (b) IOP ≥ 40 mmHg; (c) no previous ocular surgery; and (d) no previous administration of ocular hypotensive drugs before ALPI. ALPI was performed under topical anaesthesia with ossibuproparacaine 1% eyedrops. The argon laser was set at energy level 250 mW to be increased, spot size 500 mm and duration time 0.2 s. All the patients received one drop of 0.5% timolol and 4% pilocarpine before ALPI. IOP and corneal oedema were evaluated 1, 2 and 4 h after ALPI. On corneal clarity restoration, Nd:YAG iridotomy was carried out on all the eyes of the study. ALPI significantly reduced IOP within the first postoperative hour in all the eyes examined. Corneal oedema resolved in all the cases within 2 h of ALPI. In all cases, it was possible to perform Nd:Yag laser iridotomy without any difficulty or complications (Fig. 1). Gonioscopic evaluation of the angle after ALPI showed that, in 16 eyes, iridocorneal angle was visible only for small portions, even when IOP was normalized. Behaviour of IOP Argon laser peripheral iridoplasty. Our results suggest that ALPI seems to be a viable and safe procedure in the management of an acute PACG attack. This procedure is able to promptly reduce IOP and allow an optimal visualization of peripheral iris in order to perform peripheral iridotomy without complications in a short period of time.
Experience over many years points to extremely variable therapeutic success in different individual eyes due essentially to the cicatrizing reactions. Several drugs have been evaluated as to the pharmacological modulation of this response. All these compounds have necrotic effects. This in vitro study aimed to evaluate various drugs. Three different agents: 5-FU, interferon α2 (IFN) and cyclosporin A (CsA) were assayed as to their ability to interfere with cell proliferation and survival. In addition, apoptosis can be inhibited in these cells by trolox. This is a nontoxic potent antioxidant water-soluble drug analogue to vitamin E. Trolox is known to penetrate biological membranes with ease and behaves as a powerful peroxyl radical scavenger. The authors also present here the data obtained with heparin in the same model system. This drug is known also for its antiproliferative activity and therefore it was evaluated as to its ability to induce apoptosis and/or necrosis. The antifibrotic activity of CsA, IFN, 5-FU and heparin was investigated on 3T6 cells in culture. Cell viability and proliferation was assessed after drug treatment. Molecular analysis of DNA degradation was evaluated by means of radioactive labelling and gel electrophoresis. Treatment with trolox lasted for 24 h with a final concentration of 1 mm. All three drugs (CsA, IFN and FU) were shown to affect cell proliferation and viability in a differential fashion (Fig. 1). However, only cyclosporin A was able to control cell proliferation by inducing apoptosis. This phenomenon was controlled by contemporary supplementation of trolox, a compound known to inhibit programmed cell death. These results strongly suggest that this model system might be useful as a test of pharmacological functionality. Concerning heparin, results reported here give evidence that, at a concentration between 5 mm and 10 mm, the aggregation of cells, possibly due to membrane modification, occurs. In addition, care should be taken since at higher drug concentrations (20 mm) extensive necrosis may occur. A rapid and efficient model system is described for the assessment of cell viability and proliferation after treatment with agents of potential pharmacological use. Cyclosporin A induces a significant apoptosis. This is important for the negative control of fibrotic degeneration in post-trabeculectomy that is required for successful surgery in glaucoma patients. Therefore, cyclosporin A might become a clinically interesting drug for the antifibrotic treatment of post-trabeculectomy. In addition, heparin shows interesting features since it is able to control proliferation via stimulation of necrosis. However, the authors urge an experimentation in vivo of the action of this drug, possibly on animal models subjected to trabeculectomy. In particular, the study of the mechanisms triggering cell necrosis with respect to those mediating the cell fusion phenomena observed at low concentration will be of relevance. It can, however, be concluded that the value of these drugs consists in their moderate and controllable effect in the production of necrosis.
Increased intraocular pressure in glaucoma causes degeneration of the optic disk and the retinal ganglion cell axons (Garcia-Valenzuela et al. 1995) with the result of a marked decrease in the visual field. Retinal damage may be responsible for neurone loss in the lateral geniculate nucleus as well (Yucel et al. 2000). A concomitant increase in nitric oxide synthase (NOS) activity in retinal cells (especially in amacrine neurones) (Neufeld 1999) and in caspase activity (Lam et al. 1999) have been reported. Both increased NOS activity and caspase activation have been shown to play an important role in the apoptotic mechanism affecting retinal cells in diseases such as glaucoma or ischaemia (Katai & Yoshimura 1999; Lam et al. 1999) and after axotomy (Garcia-Valenzuela et al. 1994); Klocker et al. 1998). Inhibition of caspase activity reduces retinal ganglion cell death (Kermer et al. 1998; Chaudhary et al. 1999). Several other factors are also involved in retinal cell death (see Linden et al. 1999 for a review), such as reactive oxygen species (Kortuem et al. 2000). In our study, we were interested in the temporal relationship between the increase in NO synthesis and neuroneal cell death in the adult mammaliam retina in an experimental model of acute glaucoma in the rat, obtained by increasing intraocular pressure for 1 h (Katai & Yoshimura 1999). NOS acitivity was detected by means of NADPH-diaphorase (NADPH-d) activity, which colocalizes with NOS (Dawson et al. 1991) and dying neurones were identified by the ‘terminal transferase method’ (Gavrieli et al. 1992). All experimental procedures on live animals were performed under the supervision of a licensed veterinarian in accordance with the guidelines for care and use of laboratory animals as published by the Italian Ministry of Health (DDL 116/92). Ten adult Wistar rats were deeply anaesthetized with intraperitoneal injections of ketamine hydrochloride (Inoketam, Virbac, Italy; 100 mg/kg body weight). A cannula was inserted in their left eye, delivering a saline solution at a pressure of 110 mmHg for 1 h. The right eye was sham-operated (a cannula being inserted without delivering saline) and served as control. After surgery, the rats were returned to their cages and allowed to survive different time intervals (1, 2, 4, 7 and 10 days). The rats were sacrificed with an overdose of anaesthetics and perfused through the left ventricle with saline followed by 4% PAF. Experimental and control retinas were reacted whole-mount either for NADPH-d histochemistry (Vincent & Kimura 1992) to detect NOS-positive neurones (4 rats), or reacted for TUNEL (Gavrieli et al. 1992) in 10 mm-thick paraffin sections to detect DNA fragmentation in apoptotic neurones (6 rats). In the first case, retinas were reacted for 1 h in a solution of 1 mg/ml NADPH (Sigma) and 0.2 mg/ml nitroblue tetrazolium (Sigma) made up in PB containing 0.5–1% Triton X-100. In the second case, retinal sections were reacted using kit purchased from Roche and following the manufacturer's instructions. The distribution of NADPH-d-positive amacrine profiles in the retina was analysed by means of an Eclipse 600 light microscope equipped with a motorized stage interfaced to a personal computer, using program Neurolucida (Microbrightfield Inc., VT, USA; Glaser & Glaser 1990). Maps of NADPH-d-positive neurones in the retinas were edited, and the nearest-neighbour distance obtained with the program NeuroExplorer. Data were compared with the one-tailed paired t-test. The expression of NADPH-d activity was increased in the retinas following experimental glaucoma, both as total number and density (nearest-neighbour distance decrease) of positive neuronal profiles (amacrine neurones, Neufeld et al. 2000) and as endothelial labelling in blood vessels. NADPH-d-positive neurones were found in the inner nucleus layer and in the ganglion cell layer, presumably amacrine cells. As glaucomatous retina were more fragile and showed some damage due to histological processing, their areas were usually smaller than those of the controls. Nevertheless, the number of NADPH-d-positive neurones was higher than in the controls at 7 and 10 days after surgery. We therefore decided to use cell density (measured as nearest-neighbour distance) instead of the overall number of neurones as a marker for NADPH-d-positivity in the retinas. In all the animals considered, the nearest-neighbour distance among NADPH-d-positive cell profiles was decreased (P = 0.04), with the highest difference between glaucomatous and sham-operated retines at 7 days after surgery (74.6 mm vs 100.2 mm) (1, 2). Computer-reconstructed maps of NADPH-d-positive neurones in experimental glaucomatous (a) and sham-operated retinas of adult rats, 7 days after surgery. Scale bar = 1 mm. Histogram of nearest-neighbour distance (mm, y -axis) among NADPH-d-positive neurones in experimental glaucomatous retinas (hatched bars) and control retinas (dotted bars) at different time intervals after surgery ( x -axis). On the control side, there were no TUNEL positive neurones (data not shown): TUNEL positive cell profiles, on the contrary, were markedly increased both in the ganglion cell layer and in the inner nuclear layer during the first 2 days after surgery (Fig. 3a), whereas they were rare at 7 days after surgery (Fig. 3b). TUNEL staining in experimental glaucomatous retinas at different time intervals after surgery (2 days in (a) , 7 days in (b) ). Arrows point to TUNEL-positive profiles. We have demonstrated that, in an experimental model of acute glaucoma in rat, the increase in NADPH-d activity and apoptotic cell death peak as shown by TUNEL are not coincident, but high levels of apoptotic cell death occur when NADPH-d activity is only slightly if at all increased. These results suggest that our experimental model of ischaemia-reperfusion in glaucoma causes a significant increase in apoptotic cell death in the early period after surgery, whereas the increase in NADPH-d expression occurs later and is protracted over time. Therefore, NO synthesis increase may not be the only factor responsible for apoptotic cell death in the glaucomatous retina. In fact, other authors have demonstrated that caspase-1 and −3 are up-regulated and TUNEL-positive neurones are frequent 24 h after surgery, and that intravitreal injections of caspase inhibitors decrease the amount of TUNEL-positive neurones (Katai & Yoshimura 1999). The significant and long-lasting increase in NADPH-d activity in blood vessels and in amacrine neurones could have different outcomes. Blood vessels show an enhanced NADPH-d activity as early as 1 day after surgery, thus creating a vasodilation in the reperfusion phase. The increase in the density of NADPH-d-positive amacrine neurones following BDNF treatment has been described previously (Cellerino et al. 1999). BDNF in transient retinal ischaemia also decreases the number of caspase-2 immunoreactive neurones (Kurokawa et al. 1999). Therefore, several factors are involved in retinal damage following acute glaucoma. This work has been supported by the Piedmont Region, MURST (40% 1998), the CRT Foundation and the San Paolo Company grants to AV.
Glaucoma is defined as progressive optic nerve cupping and visual field damage which, ultimately, may determine loss of vision. An appropriate management of glaucoma includes the diagnosis of early changes both at the structural and functional levels. Among the many imaging and functional techniques, microperimetry is peculiar because it allows retinal sensitivity while directly examining the fundus of the eye to be assessed. It consists in presenting the visual target in any retinal topographic region, depending on the area of interest. In general terms, microperimetry is (and must be) a manual procedure depending on the direct interaction between the investigator and the real-time imaging of the fundus. For this reason, the clinical usefulness of this technique is limited to the evaluation of retinal sensitivity of localized retinal nerve fibre layer defects in patients with a normal standard visual field test, as its application in cases of diffuse nerve fibre layer defects or cases with established visual field changes does not add to the standard diagnostic techniques. With microperimetry, early loss of retinal sensitivity can be detected within localized nerve fibre defects, suggesting an early loss of visual function in eyes pronounced normal with standard visual field examination. Other applications in the field of glaucoma are unknown so far. Although custom automatic grid strategies have been developed with the aim of conducting an automatic threshold microperimetric examination, their clinical results have shown no advantages over standard techniques.
To evaluate the safety and the efficacy of a stainless steel miniature glaucoma device (ExPress™, manufactured by Optonol, Israel) in reducing the intraocular pressure in eyes affected by primary open-angle glaucoma and cataract when implantation was combined with phacoemulsification. Prospective multicentre study. The protocol was approved by each Institution's Ethics Committee. Selection criteria were: visually significant cataract with uncontrolled POAG, wide open-angle, Caucasian race. Phacoemulsification was performed first, with a clear cornea approach. After placement of the IOL in the capsular bag, the implant was introduced through the sclera at the limbus via a small conjunctival opening 10–15 mm posterior to it. Routine FU were scheduled at 1 day, 1 week, 1 month and then every 2 months. All patients were implanted using a stainless steel device with 27G external diameter and a 50 µ internal opening. All the patients included were implanted successfully. At the mean follow-up of 6 months (min. 3; max. 16), the mean IOP was reduced from 24 mmHg (±2) to 15.5 mmHg (±2). Mean IOP reduction was 36% from baseline. The probability of maintaining an IOP of 16 mmHg on no medication at 9 months, estimated with a survival curve, is 67%. Complications encountered included 1 case of conjunctival erosion, 1 case of hypotony, which was treated with removal of the implant and 2 cases of transient flat anterior chamber with choroidal detachment. One of these 2 cases was a failure and required reoperation for further IOP reduction. Bleb needling for early bleb failure was required in 6 eyes. The early results obtained with this implant support the continuation of clinical trials to assess its role as a substitute for filtration surgery.
Recent electrophysiological evaluations, performed by pattern electroretinogram (PERG) and visual evoked potential (VEP) recordings, suggest that patients affected by open angle glaucoma (OAG) present a retinal dysfunction and a delay in neural conduction in the postretinal visual pathways (Parisi 1997). Although an impairment of the innermost retinal layers (ganglion cells and their fibres) is well documented (Quigley et al. 1995), only recently has an impairment of the lateral geniculate nucleus (LGN) been observed in animals in which experimental glaucoma was induced (Weber et al. 2000). Therefore, it is supposed that the abnormal visual cortical responses observed in glaucoma may result from both retinal impairment and the involvement at the LGN level. In our previous study (Parisi 1997), the VEP responses were derived by means of a single electrode placed over both occipital cortexes and therefore did not allow a separate evaluation of the neural conduction along the crossed and uncrossed fibre visual pathways. On the basis of recent evidence regarding postretinal involvement in glaucoma, the aim of our work now is to evaluate the neural conduction in crossed and uncrossed visual pathways in OAG patients. Twenty-two OAG patients (mean age 65.6 ± 9.5 years, refractive error between +2 and −2 sph) with mean deviation (MD) between −2 and −27 dB and corrected pattern standard deviation (CPSD) between +2 and +13.5 dB of 24/2 Humphrey computerized static perimetry, and with an IOP < 21 mmHg in at least one eye (average of the two highest values of the daily curve in medical treatment with β-blockers only), were enrolled. They were compared to 16 age-matched controls. In OAP and control subjects, VEP recordings were performed as follows. The subjects under examination were seated in a semidark, soundproof room in front of a display with a uniform field of luminance of 5 cd/m2 surround. Visual stimuli were full field checkerboard patterns (each square subtended an angle of 15′ of visual arc; contrast 80%, mean luminance 110 cd/m2) generated on a TV monitor subtending 18°, and reversed in contrast at the rate of 2 reversal/ s. The refraction of all subjects was corrected for the viewing and no mydriatic or miotic drugs were used. Cup-shaped Ag/AgCl electrodes were fixed with collodion in positions: active electrode in O1 (left occipital cortex) and O2 (right occipital cortex (Jasper 1958), reference electrode in Fpz, with ground in the left arm. The interelectrode resistance was kept <3 kΩ. The bioelectric signal was amplified (gain 20 000), filtered (bandpass 1–100 Hz) and averaged (200 events free from artifacts were averaged for each trial) by BM 6000. Analysis time was 200 ms. Stimulation was monocular after occlusion of the other eye and the bioelectrical cortical responses were recorded simultaneously in the homolateral visual (HC) cortex and in the contralateral visual cortex (CC), with respect to the stimulated eye. The transient VEP response is characterized by a number of waves with three subsequent peaks of negative, positive, negative polarity: N75, P100, N145. We accepted VEP signals with signal-to-noise ratio >2. For all VEPs, the implicit time and the peak-to-peak amplitude of each of the averaged waves were measured directly on the displayed records by means of a pair of cursors The differences betweenOAG patients and controls were evaluated by anova, and linear regression analyses (Pearson's test) were used to establish the correlation between perimetric and VEP parameters. In OAG patients, the VEP P100 implicit times observed in HC and CC were both significantly (P < 0.01) delayed compared to those of the controls, and were significantly related (P < 0.01) to the MD observed in the nasal and the temporal hemi-fields, respectively (Table 1). We found an asymmetry in the bioelectrical responses obtained in HC and CC of all the OAG patients, and the intracortical differences (ID:P100 implicit time in HC–P100 implicit time in CC) converted to absolute values, were significantly higher than those of the controls (4.49 ± 3.72 ms and 1.16 ± 1.04 ms, respectively, P = 0.001). In 11 (50%) OAG patients, we observed an ID with a negative value (between −1 and −13.6 ms), indicating VEP P100 implicit times longer in CC than in HC, while in 11 (50%) OAG patients, we observed an ID with a positive value (between 1 and 8.8 ms), suggesting VEP P100 implicit times longer in HC than in CC. The ID observed in OAG patients (ranging from −13.6 to 8.8 ms) was correlated significantly with the CPSD (r: 0.58, P = 0.004) (Fig. 1), but not with the MD (r: 0.18, P = 0.409). Intracortical differences (ID:VEP P100 implicit times observed in homolateral cortex-VEP implicit times observed in Contralateral Cortex to the stimulated eye) of OAG patients plotted vs CPSD of Humphrey 24/2. Statistical analysis is reported in Table 1 . The observed asymmetry in visual cortical responses suggests that crossed and uncrossed visual pathways could be impaired differently in OAG patients.
OBJECTIVE:To compare the circadian intraocular pressure (IOP) reductions induced by latanoprost, brimonidine tartrate, and a fixed combination of timolol maleate and dorzolamide hydrochloride in patients with primary open-angle glaucoma (POAG) or ocular hypertension (OHT).METHODS:In this crossover study, 10 patients with POAG and 10 with OHT were treated with latanoprost once a day, brimonidine twice a day, and a fixed combination of timolol and dorzolamide twice a day for 1 month. Four 24-hour tonometric curves were obtained for each patient. Intraocular pressure (IOP) was measured at 3, 6, and 9 AM, and at noon and at 3, 6, and 9 PM, and at midnight, using a handheld electronic tonometer with the patient in supine and sitting positions and a Goldmann applanation tonometer with the patient sitting at the slitlamp.MAIN OUTCOME MEASURE:Reduction of circadian IOP.RESULTS:All the drugs significantly reduced IOP compared with the baseline at all times, except for brimonidine at midnight, 3 AM, and 6 AM. Latanoprost was more effective than brimonidine in lowering IOP at 3 and 6 AM and at 3 PM (P=.03), and the combination of timolol and dorzolamide was more effective than brimonidine at 3 and 9 AM (P=.04) and at 3 and 6 PM (P =.05) and more effective than latanoprost at 9 AM (P=.05).CONCLUSION:Latanoprost and the fixed combination of timolol and dorzolamide led to similar circadian reductions in IOP, whereas brimonidine was less effective, particularly during the night.
Brimonidine is a less lipophilic analogue of clonidine (Chien et al. 1990) and, like apraclonidine (Robin 1988), provides clinically significant reduction in intraocular pressure (IOP) (Shin et al. 1999). Preclinical studies on brimonidine have documented that it is a potent α-2 adrenoreceptor agonist, 1000-fold more selective for the α-2 vs α-1 adrenoreceptor. Brimonidine α-2 selectivity is 7- to 12-fold greater than clonidine and 23–32-fold greater than apraclonidine (Burke & Schwartz 1996). The aim of our study was to determine the effects of clonidine 0.125%, apraclonidine 1.0% and brimonidine 0.2% on visual field parameters and ocular perfusion pressure in patients affected by primary open-angle glaucoma (POAG). A randomized, double-masked clinical trial was carried out. Thirty-two POAG subjects (15 men, 17 women, mean age 65.25 ± 5.69 years; range 55–76 years) were enrolled. Informed consent to participate in this clinical trial was given by each patient. Visual field (30–2 full threshold program) was measured by the Humphrey field analyser. The best corrected visual acuity was 20/25 or better, with a range of ametropia between −3 and +2 spheric diopters and ± 2 cylinder diopters, with a pupil diameter of at least 3 mm (Johnson 1996). A baseline determination of brachial artery systolic and diastolic blood pressure (SBP and DBP, respectively), heart rate (HR), bilateral IOP and visual field was performed in each subject. The patients were separated randomly into 4 study groups of 8 POAG subjects (16 eyes) to receive in both eyes either clonidine 0.125% (group A), or apraclonidine 1.0% (group B), or brimonidine 0.2% (group C), or placebo (group D) 2 h before the repetition of the clinical evaluations previously described (Table 1). All tests were performed, between 7:30 a.m. and 10:30 a.m., in two separate sessions, 48 h apart. The comparisons between the baseline and α-agonist-induced values were made for the following parameters of each patient or each eye: mean blood pressure [MBP = DBP + 1/3(SBP − DBP)], HR, IOP, ocular perfusion pressure [PP = 2/3(MBP − IOP)] (Mastropasqua et al. 1998a); moreover, a comparison among the visual field indices [mean defect (MD), pattern standard deviation (PSD), short-term fluctuation (SF) and corrected pattern standard deviation (CPSD)] before and after the treatments was performed. A non-parametric test for paired data was used to compare MBP, HR, IOP, PP and visual field parameters at baseline and 2 h after α-agonist acute administration. A probability of P < 0.05 was considered statistically significant. The topical administration of α-agonist compounds significantly reduced IOP in all groups (P < 0.01). At the IOP lowering effects, topical clonidine administration also exhibited a marked reduction in MBP (P < 0.01) which, in turn, further affected the PP (P < 0.001). Consequently, most visual field parameters (MD, SF and CPSD) were modified negatively by the treatment (Table 1). On the contrary, acute application of apraclonidine not resulted in significant PP modification, whereas brimonidine eyedrops induced a slight positive tendency between the statistical comparison of PP values (P = 0.16). Moreover, while apraclonidine eyedrops caused a significant worsening in MD (P < 0.05), SF (P < 0.05) and CPSD (P < 0.01), the acute brimonidine instillation did not induce any significant modification of the visual field indices (Table 1). Lastly, no significant variation in the ocular and systemic parameters was observed after placebo administration (Table 1). Neither placebo nor α-agonists topical administration induced any HR significant variation. Ocular blood flow, including the haemodynamic condition of the optic nerve head, depends on the relationship between IOP, PP, vascular resistance and the presence or absence of blood flow autoregulation. The comparison between baseline and drug-induced values of PP and visual field parameters, represents an indirect evaluation of the retinal microvascular blood flow modification after administration of different α-agonist antiglaucomatous eyedrops. Our findings demonstrate that the acute administration of clonidine, apraclonidine and brimonidine produces a significant IOP reduction in POAG patients treated chronically with other antiglaucomatous eyedrops. Moreover, our data show a worsening of PP and visual field parameters after acute administration of clonidine, which is related to both central and peripheral actions of this less selective α-agonist compound (Krieglstein et al. 1978; Yuksel et al. 1992). Also the apraclonidine instillation induces a visual field deterioration, despite the IOP lowering effect and the consequent PP increase, indicating a possible α-mediated vasoconstriction of the posterior pole microvasculature (Mastropasqua et al. 1998a). The comparative evaluation of MBP and HR data confirms the results of previous studies (Jampel et al. 1988; Mastropasqua et al. 1998a), demonstrating the absence of significant cardiovascular changes after topical administration of apraclonidine, yet systemic blood pressure cannot always be considered a true reflection of blood pressure in the intraocular vessels (Hayreh 1989). On the other hand, brimonidine application does not result in any significant visual field parameters worsening, as already demonstrated, employing the blue-yellow perimetry (Mastropasqua et al. 1998b). Moreover, the significant IOP reduction secondary to brimonidine instillation was not associated with MBP modifications, contrary to that occurring with the application of clonidine and apraclonidine. The moderate PP increase observed after brimonidine administration (P = 0.16) emphasizes the critical role of its peculiar α-2 selectivity when brimonidine effects are compared to those of the other α-adrenoreceptor agonists. Although brimonidine is a relatively selective α-2 agonist, our results suggest that its acute administration do not critically affect the blood flow or the vasomotor activity at the level of the posterior pole. The knowledge of the systemic and ocular effects secondary to the administration of α-agonist eyedrops is crucial as POAG-treated patients undoubtedly have drug-induced alteration of their basal adrenergic tone. Thus, clonidine or apraclonidine prescription should be avoided so as not to cause further blood perfusion reduction or a critical vasoconstriction at the level of ocular posterior segment, whereas brimonidine selectivity circumvented such side-effects.
Only a very brief survey will be given, as Iceland was a part of Denmark for only 18 years at the beginning of the century. The first ophthalmologist in Iceland was Björn Ólafsson (1862–1909). He studied medicine in the Icelandic medical school. After graduation in 1888 he went to Copenhagen, where he studied ophthalmology. He worked at the clinic of Professor Edmund Hansen-Grut, who was influenced by visits to Albrecht von Graefe in Berlin and Desmarres in Paris. In 1890 Björn Ólafsson returned to Iceland as district medical officer in Akranes. However, his skills in eye diseases became known rapidly and many patients also began to see him for surgical treatments (cataract, glaucoma, squint, eyelid surgery). In 1894 he received economic support from the Althing in order to undertake full-time ophthalmology in Reykjavik and to continue as travelling consultant around the country. The population of Iceland in 1900 was 78 000, of whom approximately 6700 lived in Reykjavik. Further, he was obliged to teach eye diseases at the medical school. He died in 1909. His reputation as a gifted eye surgeon and founder of Icelandic ophthalmology is well known, as described by Björnsson (1986). Andrés Fjeldsted opened an eye clinic in 1910. He studied ophthalmology in London, Copenhagen, Edinburgh, Oslo and Vienna. A limited number of specialists had their training in Copenhagen; for example, the two above-mentioned who also had worked in the Faroe Islands during the Second World War. Besides, no major Danish contribution seems to have taken place in this field. In June 2000 the 100-year anniversary of Nordic Ophthalmology was celebrated in Reykjavík, at the XXXIV Congress (see Prause, this Supplement, p. 51, Fig. 3). The high status and international orientation of Icelandic ophthalmology was evident. Of special relevance to the present paper was a paper given by Örn Sveinsson in a session dealing with outreach eye service in the Nordic countries (Sveinsson 2000). The historical Icelandic tradition for outreach service, going back to Björn Ólafsson, was clearly illustrated. The history of Icelandic ophthalmology is in press (Björnsson 2002). Örn Sveinsson is thanked for details of the above notes on Iceland Ophthalmology and their participation in the Faroe Islands.
To assess intra- and interobserver reproducibility of optic disc evaluation by means of HRT. Fifty-five volunteers underwent three sessions of HRT examination of the optic disc performed on three consecutive days. During each session, five single images were randomly acquired by two independent well-trained residents in ophthalmology. One mean topography image (MTI), based on three single images, was then constructed at each session. For the intra-observer, intra-image evaluation, two independent observers traced their own contour line on the MTI of the first session. This procedure was repeated three times. For the intra-observer interimage and interobserver intra/interimage evaluations, the same two observers traced their own contour line on the MTI of the first session, which was then automatically superimposed on the MTls of the other two sessions. Reproducibility assessment for the 12 stereometric parameters was then calculated for each observer by means of intraclass correlation coefficient (ICC) and the Altman plots of two randomly chosen tests. The ICC ranged between 0.60 and 1 for intra-observer intra-image, and between 0.31 and 1 for intra-observer interimage evaluation. The ICC ranged between 0.39 and 1 for interobserver intra-image, and between 0.23 and 0.99 for the interobserver interimage evaluation. Intra- and interobserver reproducibility were substantial to perfect for planimetric measures (0.61 < ICC < 1), almost perfect to perfect for volumetric and cup measures (O.81 < ICC < l), and fair to almost perfect for RNFL related measures (0.21 < ICC < 0.99). In the case of interobserver reproducibility assessment, the Altman plots of interimage evaluation showed a higher variability than intra-image evaluation (P = 0.012, Wilcoxon test). Optic disc measurements by HRT are highly reproducible for most of the stereometric parameters. However, the use of RNFL related parameters should be taken cautiously. The image acquisition induced variability seems larger than operator induced variability.
