Conjunctival findings in hyperbaric and low-tension glaucoma: an in vivo confocal microscopy study


Luca Agnifili, MD, PhD
Via dei Frentani, 114
66100 Chieti, Italy
Tel: + 39 0871 358410
Fax: + 39 0871 358794


Purpose:  To analyse the epithelial features of the bulbar conjunctiva in hyperbaric and low-tension glaucoma (LTG) using in vivo confocal microscopy (IVCM).

Methods:  Thirty-six eyes of 36 patients [18 affected by primary open-angle glaucoma (POAG) and 18 with LTG] were studied; control group was constituted by 28 eyes of 28 healthy subjects. All eyes were examined using digital confocal laser-scanning microscopy (HRT II Rostock Cornea Module). The main IVCM outcome measurements were mean density (MMD: cysts/mm2) and mean total area (MMA: μm2) of the epithelial microcysts.

Results:  The mean intraocular pressure level (mmHg ± SD) was 15.1 ± 1.7, 16.3 ± 3.1 and 12.6 ± 1.8 in healthy, POAG and LTG eyes, respectively. Conjunctival microcysts were found in all patients and subjects: for healthy subjects, MMD = 10.9 ± 11.1 cysts/mm2 and MMA = 1501.9 ± 1191.1 μm2; for patients infected with POAG, MMD = 36.8 ± 28.6 cysts/mm2 and MMA = 7904.8 ± 7050.5 μm2; and for patients infected with LTG MMD = 45.6 ± 29.0 cysts/mm2 and MMA =7946.9± 5227.5 μm2. MMD and MMA were not significantly different between patients infected with POAG and those with LTG, whereas they were significantly greater in patients (fourfold and fivefold, respectively) than healthy subjects.

Conclusions:  The present study demonstrated that conjunctival microcysts represent an in vivo feature in all eyes with medically controlled POAG and LTG. Therefore, conjunctiva deserves careful analysis, because its accurate microscopic definition could help clarify the pathophysiology of aqueous outflow in glaucoma.


Anatomo-pathological changes of ocular tissues were widely documented in patients affected by hyperbaric glaucoma, mainly in the neuro-retinal system (optic nerve head neural rim, retinal nerve fibre layer and lamina cribrosa) and in structures of aqueous outflow pathways (trabecular meshwork, Schlemm’s canal, and suprachoroid and sclera) (Quigley 1998; Civan & Macknight 2004; Johnson 2006).

Recently, modifications of outflow pathways were found using laser-scanning in vivo confocal microscopy (IVCM) within the conjunctival epithelium of untreated ocular hypertension (OH) and topically treated primary open-angle glaucoma (POAG) (Ciancaglini et al. 2008b, 2009). These modifications corresponded to intraepithelial microcysts. They were circular hyporeflective structures similar to those initially reported within the bleb wall of functioning trabeculectomy (Labbé et al. 2005; Guthoff et al. 2006a; Messmer et al. 2006b; Ciancaglini et al. 2008a). These epithelial microcysts cannot be considered exclusive feature of glaucoma because they were described in healthy subjects and in diseases other than glaucoma (Messmer et al. 2006a; Efron et al. 2009, 2010; Wakamatsu et al. 2010). They could be considered a sign of trans-scleral aqueous humour outflow that has been activated and enhanced by hyperbaric ocular conditions (Ciancaglini et al. 2008b).

Low-tension glaucoma (LTG) represents a subgroup of open-angle glaucoma, which differs from high-tension forms in terms of pathogenesis, risk profile, clinical aspects, rate of progression and intraocular pressure levels (IOP) (Shields 2008; Park et al. 2009). The IOP may be within the normal range of values even without treatment.

Although several studies investigated the pathogenesis and clinical features of LTG (Tomita 2000; Sowka 2005) with a considerable number highlighting the role of IOP (Cartwright & Anderson 1988; CNTGS 1999), the presence of aqueous outflow pathway modifications was not studied.

The aim of the present study was to evaluate, through IVCM analyses, the epithelial features of bulbar conjunctiva in patients affected by high- and LTG, to verify trans-scleral outflow route modifications.

Material and Methods

The study adhered to the tenets of the Declaration of Helsinki, and informed consent was obtained from all patients prior to their enrolment. Our local ethics board was notified prior to initiating the study and they determined that their approval was not necessary.

Thirty-six glaucomatous patients (one eye per patient) were consecutively enrolled, 18 with POAG and 18 with LTG. Twenty-eight consecutive healthy subjects (28 eyes) were recruited from the out-patients of the Ophthalmic Clinic of University of Chieti-Pescara, Italy and enrolled as controls.

Healthy eyes showed a best corrected visual acuity ≥8/10, refractive error ≤4 dioptres (spherical equivalent), IOP lower than 18 mmHg, central corneal thickness (CCT) ranging from 530 to 560 μm, absence of signs of glaucomatous optic neuropathy and normal visual field (Humphrey 30-2 full-threshold). Exclusion criteria for controls were a reported history of topical or systemic medical therapy and ocular or systemic disease during the previous 6 months.

Primary open-angle glaucoma and LTG eyes showed a best corrected visual acuity ≥20/40, refractive error ≤4 dioptres (spherical equivalent) and CCT ranging from 520 to 560 μm.

The diagnosis of POAG required an untreated IOP ranging from 24 to 37 mmHg (mean of three measurements at 9 AM, 12 noon and 4 PM) at diagnosis, absence of signs indicative for secondary glaucoma (pseudo-exfoliation or pigment dispersion) and absence of classic ophthalmoscopic signs for glaucomatous optic neuropathy (cupping, neural rim notching and saucerization).

A full-threshold visual-field test (Humphrey 30-2; Carl Zeiss Meditec Inc., Dublin, CA, USA) had to show at least three contiguous points on the total deviation probability plot at the <2% level with Glaucoma Hemifield Test results outside normal limits, in at least one eye.

At least two visual-field examinations within the last 6 months with acceptable reliability standards (fixation loss, false-positive rate and false-negative rate <33%) were required.

If the last two visual fields showed evidence of damage progression (extension or deepening of pre-existing scotoma or appearance of new scotoma), the patient was excluded. The aspect of the optic disc had to be consistent with the visual-field alterations.

Patients infected with POAG were in monotherapy with prostaglandin analogues or prostamides (eight with latanoprost 0.005%, four with travoprost 0.004% and six with bimatoprost 0.03%). Mean diurnal IOP (mean of three measurements at 9 AM, 12 noon and 4 PM) was controlled in all cases.

Exclusion criteria were corrected visual acuity <20/40 in either eye, end-stage glaucoma (mean deviation >15 dB), visual-field defects attributable to non-glaucomatous conditions, history of angle closure or an occludable angle by gonioscopy, argon laser trabeculoplasty or laser iridotomy performed <6 months before enrolment, surgical procedure for glaucoma, aphakia, history or signs of inflammatory eye diseases, uveitic glaucoma, ocular trauma, contact lens wearing, presence of other ocular diseases during the previous twelve months except glaucoma and pregnancy.

The diagnosis of LTG required untreated mean diurnal IOP ≤21 mmHg at diagnosis, IOP asymmetry <4 mmHg and open irido-corneal angle (gonioscopy). As for POAG, the absence of signs of secondary glaucoma (pseudo-exfoliation and pigment dispersion) verified after pupil dilation, classic ophthalmoscopic signs of glaucomatous optic neuropathy and glaucomatous visual-field defects in at least one eye on HFA 30-2 full-threshold standard automatic perimetry (Humphrey-Zeiss, San Francisco, CA, USA) were required. Patients also had at least two visual-field examinations within the last 6 months with acceptable reliability standards (fixation loss, false-positive rate and false-negative rate <33%). Criteria of visual-field abnormality were the presence of a cluster of three adjacent points depressed by more than 5 dB, with one of these points depressed by at least 10 dB. At least three points of this cluster, including the 10-dB depressed point, had to be located on one side of the horizontal meridian. There had to be other points elsewhere in the visual field that were at least 10 dB higher than the most depressed point in the scotoma. Also in LTG, if the last two visual fields showed evidence of damage progression (extension or deepening of pre-existing scotoma or appearance of new scotoma), the patient was excluded. The aspect of the optic disc had to be consistent with the visual-field alterations.

The same exclusion criteria were used for POAG, except for the untreated IOP, which fell within normal range. Also patients infected with LTG were in monotherapy with prostaglandins analogues or prostamides (10 eyes with latanoprost 0.005%, four with travoprost 0.004% and four with bimatoprost 0.03%), and mean diurnal IOP was controlled in all cases.

All glaucomatous patients started their unmodified topical hypotensive therapy at least 6 months before IVCM examination. Moreover, none of the patients had a history of topical or systemic diseases or concomitant therapies within the last 6 months, thus avoiding any significant bias because of ocular surface and/or ocular hydrodynamic iatrogenic modifications.

Each enrolled subject underwent IVCM assessment of the superior bulbar conjunctival epithelium using digital confocal laser-scanning microscope (HRT II Cornea Module; Heidelberg Engineering GmbH, Heidelberg, Germany). Because of the technical difficulty in performing the examination, the inferior sectors of the bulbar conjunctiva were not considered in the analysis.

In vivo confocal microscopy

The technical characteristics of this instrument and the details of conjunctival examination were previously described (Mastropasqua et al. 2006). Sequential images derived from automatic scans and manual frames were acquired throughout the superior bulbar conjunctiva (superior nasal and superior temporal quadrants: from 3 to 9 o’clock) at the intermediate layer (10–20 μm) with examined eye in down gaze.

Conjunctiva was evaluated to identify epithelial microcysts defined as intraepithelial and extracellular optically empty spaces (Ciancaglini et al. 2008b, 2009). The main outcome measurements were the mean microcyst density (MMD: cysts/mm2) and the mean total microcyst area (MMA: μm2).

The confocal examinations were performed by a single operator (VF) who randomly selected eight images (300 × 300 μm in size, from 40 images taken throughout the entire upper bulbar conjunctiva). These images were evaluated by a second IVCM operator (LA). Both operators were masked for patient history and status. The surface area of epithelial microcysts was calculated using ImageJ, open source software (; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD, USA), as described elsewhere (Messmer et al. 2006b).

In addition, as pioneered by Patel & McGhee (2005) for the sub-basal nerves in the human cornea, a wide planar reconstruction of the entire superior conjunctiva in two healthy, two POAG and two LTG eyes were reconstructed to validate the findings observed in the sampled fields.

The mean time (months ± SD) of the IVCM examination from the onset of the disease was 29.3 ± 5.8 and 32.8 ± 7.4 in POAG and LTG, respectively.

Statistical analysis

Statistical analysis was performed using spss® Advanced Statistical™ 13.0 software (2005; Chicago, IL, USA). Student’s t-test was used to evaluate mean time on therapy and age differences among healthy, POAG and LTG eyes. Chi-squared test was used to evaluate gender differences. One-way anova with post-hoc Tukey’s test for multicomparison was used to assess differences between the three groups for age, IOP, mean defect (MD), microcysts density and surface.


The demographical and clinical data of each group are reported in Table 1. No significant differences were found in age and gender between the three groups and for mean time on therapy between POAG and LTG groups.

Table 1.   Characteristics of the patients and groups.
  1. SD = standard deviation, M = male, F = female, LTG = low-tension glaucoma, POAG = primary open-angle glaucoma, MD = mean defect, dB = decibel, mo = months, NA = not applicable, IOP = intraocular pressure level.

  2. * p < 0.001 versus healthy.

  3.  p < 0.001 versus LTG.

  4.  p < 0.001 versus healthy.

  5. § p < 0.001 versus healthy.

Mean age (years ± SD)64.07 ± 5.864.2 ± 8.0562.5 ± 11.9
Gender (M/F)12/168/108/10
Mean IOP (mmHg ± SD)15.1 ± 1.6716.3 ± 3.0512.6 ± 1.75*
MD (dB ± SD)+0.8 ± 0.6−5.7 ± 1.2§−6.1 ± 1.3
Mean time on therapy (mo ± SD)NA29.3 ± 5.832.8 ± 7.4

Intraocular pressure level values in LTG eyes were significantly lower than those recorded in healthy subjects and patients infected with POAG.

Conjunctival epithelial microcysts were found in all participants: healthy, POAG and LTG eyes (Figs 1–3, respectively).

Figure 1.

 Planar reconstruction of the superior bulbar conjunctiva in a representative healthy eye, showing the distribution of epithelial microcysts in the entire section. The final map was obtained by juxtaposing (side-by-side) each acquired image obtained by moving the Cornea Module from the nasal to the temporal region (right and left side of the image, respectively) and from the limbus to the upper lid (bottom and top of the image, respectively).

Figure 2.

 Planar reconstruction of the superior bulbar conjunctiva in a representative patient with primary open-angle glaucoma, showing the distribution of epithelial microcysts. The arrow indicates a microcyst filled with amorphous material, whereas the arrowhead indicates inflammatory cells infiltrating a microcyst.

Figure 3.

 Planar reconstruction of the superior bulbar conjunctiva in a representative patient with low-tension glaucoma, showing the distribution of epithelial microcysts in the whole section. The arrow indicates inflammatory cells infiltrating a microcyst, whereas the arrowhead shows a microcyst filled with a hyper-reflective material.

Mean microcyst density (cysts/mm2) and MMA (μm2) were not significantly different between POAG and LTG eyes. Statistically significant differences were found between normal eyes and POAG or LTG, respectively (Table 2). Particularly, patients infected with POAG and LTG presented a MMD of 36.8 ± 28.6 and 45.6 ± 29.0, respectively, which were almost four times greater than that observed in healthy subjects (10.9 ± 11.1) (p < 0.001 for both groups). Regarding MMA, values of 7904.8 ± 7050.5 and 7946.9 ± 5227.5 in patients infected with POAG and LTG, respectively, were found, which were almost five times greater than that observed in healthy subjects (1501.9 ± 1191.1) (p < 0.001 for both groups).

Table 2.   IVCM parameters.
IVCM parametersHealthyPOAGLTG
  1. IVCM = in vivo confocal microscopy, SD = standard deviation, LTG = low-tension glaucoma, POAG = primary open-angle glaucoma.

  2. * p < 0.001 versus healthy.

  3.  p < 0.001 versus healthy.

  4.  p < 0.001 versus healthy.

  5. § p < 0.001 versus healthy.

Mean microcyst density (cysts/mm2 ± SD)  10.9 ± 11.1  36.8 ± 28.6  45.6 ± 29.0*
Mean total microcyst area (μm2 ± SD)1501.9 ± 1191.17904.8 ± 7050.5§7946.9 ± 5227.5


Epithelial microcysts were first described in the conjunctival bleb walls of successful trabeculectomies as features of trans-conjunctival aqueous humour percolation (Labbé et al. 2005). Amar et al. (2008) subsequently hypothesized that bleb-wall microcysts corresponded to goblet cells carrying aqueous humour through the conjunctiva. However, conjunctival microcysts are not exclusive features of filtering bleb, because they were also reported in untreated OH and in medically controlled POAG (Ciancaglini et al. 2008b). Ciancaglini et al. (2008b) hypothesized that they indicated an increased aqueous flow through the sclera when increased trabecular meshwork resistance is present.

The results of the present study were in agreement with those previously reported by Ciancaglini et al. (2008b), because microcysts were found in all POAG eyes on medical therapy. Additionally, microcysts were also observed in LTG without significant differences with respect to POAG. The presence of epithelial microcysts in all glaucomatous eyes supports the hypothesis of a change in aqueous hydrodynamics and the activation of alternative outflow pathways, such as trans-scleral routes, either in hyperbaric or LTG.

As previously hypothesized for hyperbaric conditions (Ciancaglini et al. 2008b), microcysts formation could represent an adaptive mechanism to increase trans-scleral aqueous percolation when trabecular outflow is reduced. Several studies (Emi et al. 1989; Toris et al. 1999; Jackson et al. 2006) demonstrated that the scleral hydraulic conductivity was not IOP dependent, because the intrascleral routes were components of the uveo-scleral pathway, which is unaffected by IOP.

Accordingly, microcysts formation seems not to be induced or sustained by IOP-related mechanical forces because significant correlation between IOP and microcysts density or area was demonstrated neither in untreated OH nor in medically controlled POAG (Ciancaglini et al. 2008b).

Moreover, the absence of differences between POAG and LTG on MMD and MMA may imply the existence of a similar mechanism leading to the scleral outflow enhancement, albeit the two types of glaucoma differ for the untreated IOP. Nevertheless, even though the microcysts formation seems not directly linked to IOP, the phenomenon may be interpreted as an adaptive mechanism that attempts to lower IOP.

These results provide additional information about the pathogenesis of LTG. In fact, although the majority of scientific evidence highlights the importance of ocular blood flow impairment, suggesting an IOP-independent component in this disorder (Kitazawa et al. 1986; Gasser 1989; Hayreh 1999; Plange et al. 2003), the exact role of IOP in the onset and progression of the disease is yet not completely clarified.

The initial results of the Low-Pressure Glaucoma Treatment Study (Greenfield et al. 2007) showed that eyes with higher IOP did not present worse values of MD and correct pattern standard deviation (CPSD) at visual-field examination, suggesting little or no influence of IOP on the pathogenesis of the disease.

Conversely, in studies of asymmetric LTG (Cartwright & Anderson 1988; Crichton et al. 1989; Haefliger & Hitchings 1990), the eye with a higher IOP showed greater glaucomatous damage, thus sustaining a pressure-sensitive component. Accordingly, the CNTGS (1999) reported a positive effect of lowering IOP, because a third reduction of IOP from baseline reduced the risk of progression.

Our results sustain the theory of an IOP-related mechanism also in the pathophysiology of LTG, because the presence of conjunctival microcysts indicated the existence of hydrodynamic modifications, particularly of uveo-scleral outflow pathway.

Nevertheless, microcysts were also described in healthy eyes. In two different studies, Messmer et al. (2006a) and Efron (2009) occasionally reported such structures in the tarsal and bulbar regions of healthy young subjects.

Zhu et al. (2010) analysed the bulbar conjunctiva in different age groups (from 14 ± 4.1 to 72.5 ± 6.6 years) and reported epithelial microcysts in all groups with an increasing rate with age. The authors speculated that microcysts, in healthy subjects, may correspond to degenerated goblet cells or normal intermediate products that result from cellular development and maturation. Our results agree with those of Zhu et al. (2010), because epithelial microcysts were also found in all healthy eyes. However, MMD and MMA were four and five times lower as compared with those found in patients infected with POAG and LTG, respectively.

Furthermore, conjunctival microcysts were occasionally reported in some non-glaucomatous ocular conditions such as pterygium, (Guthoff et al. 2006b), Sjögren’s syndrome (Wakamatsu et al. 2010) and contact lens wearing (Efron et al. 2010). Here, microcysts were probably features of degenerative processes occurring within the conjunctival epithelium, when disturbances in cellular maturation created debris field cystic spaces. However, in these studies, the authors did not provide any information about the IOP of the enrolled subjects.

Therefore, we hypothesize that epithelial microcysts may have more than one connotation, because in the presence of ocular hydrodynamic impairment, they may represent a sign of trans-scleral outflow enhancement, whereas in other conditions, they may be considered an aspect of epithelial disruption. Nevertheless, we cannot rule out that microcysts, in both healthy eyes and non-glaucomatous conditions, could be a sign of physiological trans-scleral aqueous percolation, which normally occurs as the final phase of the uveo-scleral route.

Based on these findings, we can assume that conjunctival microcysts are in vivo biomarkers of uveo-scleral aqueous outflow modifications in different types of glaucoma. As a consequence, although microcysts cannot distinguish glaucoma subtypes, they could have a diagnostic utility (mainly in case of initial suspect) because they are a feature of disease pathophysiology.

The present study has one major limitation. Information concerning conjunctival status in patients infected with POAG and LTG before therapy was not collected. Thus, the role of topical medication (preservative and active compounds) on the microcysts formation process was not investigated. Even though Ciancaglini et al. (2008b) did not find a significant difference between untreated OH and treated POAG, the effect of therapy should be better evaluated by analysing the conjunctiva before and after the therapy.

In conclusion, these findings indicated that adaptive modifications of the trans-scleral aqueous pathway occur in high- and LTG and that they were documentable in vivo through microscopic analysis of the conjunctival epithelium.