SEARCH

SEARCH BY CITATION

Keywords:

  • corneal hydrops;
  • keratoconus;
  • keratoplasty;
  • New Zealand

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Background

The aim was to identify potential factors associated with acute corneal hydrops in a New Zealand population with keratoconus referred to a hospital eye service.

Methods

In a single hospital centre, in a retrospective review, demographic and clinical features of subjects with keratoconus and corneal hydrops over a 17-year period were compared with an age- and gender-matched control group of subjects with keratoconus but no history of corneal hydrops.

Results

One hundred and one eyes of 101 subjects (mean age 24.6 ± 8.4 years) were identified with keratoconus-related corneal hydrops. Subjects were more likely to be of Pacific but less likely to be of New Zealand European ethnicity than control subjects (n = 101). In comparison, Maori ethnicity was not found to have a significantly positive or negative association with hydrops. The pre-hydrops visual acuity (VA) of affected eyes was poorer than that of controls (p < 0.001) at first presentation to our tertiary referral corneal and contact lens service. Hydrops typically developed approximately four years after diagnosis of keratoconus. Subjects with hydrops were more likely to have a history of eye-rubbing (p = 0.011) but less likely to have a family history of keratoconus (p = 0.05). In 31 cases, the acute hydrops event was their first optometric/ophthalmologic contact. There were no statistically significant differences in the prevalence of atopic disease, contact lens wear or overall corneal transplantation rate between the two groups.

Conclusions

Pacific ethnicity, history of eye-rubbing, poor VA at first hospital presentation and lack of family history were statistically associated with developing acute corneal hydrops in keratoconus in a New Zealand population. Greater understanding of such predisposing risk factors may help develop early management strategies to delay or prevent progression of this disease.

Acute corneal hydrops is an incompletely understood complication of keratoconus, occurring in approximately three per cent of patients with keratoconus.[1] It may also occur in other corneal ectasias, such as pellucid marginal degeneration and keratoglobus.[2, 3] Corneal hydrops develops following acute rupture of Desçemet's membrane and overlying endothelium that allows aqueous to enter the corneal stroma and epithelium, resulting in severe corneal oedema. In recent years, modern technology, such as in vivo confocal microscopy and anterior segment optical coherence tomography, has been able to elegantly demonstrate these features of acute corneal hydrops in vivo.[4, 5]

Although acute corneal hydrops is usually self-limiting and clinical signs of oedema typically resolve after two to four months,[6] it often leaves a vision-impairing scar, necessitating or expediting the need for corneal transplantation. Severe complications of corneal hydrops are infrequent but include extensive corneal scarring, severe neovascularisation, epithelial defects, microbial keratitis, corneal perforation and glaucoma.[7-9] A history of hydrops may also predispose subjects to greater likelihood of episodes of endothelial graft rejection after penetrating keratoplasty.[10]

Keratoconus is the most common indication for corneal transplantation in Australasia, accounting for approximately half of the corneal transplants in New Zealand[11] and almost a third of those performed in Australia.[12] It is generally considered a bilateral disease[13] that is suspected to be more prevalent in Maori and Pacific peoples, in whom the condition also appears to be more aggressive.[11] While several risk factors for the development of hydrops have been postulated, including male gender,[1] advanced corneal ectasia,[1] co-existing vernal keratoconjunctivitis,[6] and eccentric (rather than central) cone location,[14] few studies have concentrated on the identification of potential predictors of the phenomenon.

To further understand the development of acute hydrops in the keratoconic population and to identify potential risk factors, we analysed the records of all patients diagnosed with acute corneal hydrops associated with keratoconus, who attended a hospital-based practice over a 17-year period. A particular focus was the demographic background and clinical characteristics of the affected subjects.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Subject recruitment and assessment

The records of all patients with keratoconus diagnosed with acute corneal hydrops, who presented to the Cornea and Anterior Segment Service or the Contact Lens Service at the Department of Ophthalmology, Auckland City Hospital, during the period January 1 1990 to December 31 2007 (inclusive), were reviewed. Cases were identified through a computerised search for the clinical diagnosis of ‘keratoconus’ and/or ‘other corneal hydrops’ under the hospital's electronic clinical coding system. The clinical diagnosis of acute corneal hydrops was defined as the sudden onset of bullous corneal oedema in the context of a subject with keratoconus. The diagnostic criteria of keratoconus were consistent with those of the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) study.[15] Where both eyes had been affected by hydrops, only the first affected eye was included in the analysis. The study adhered to the tenets of the Declaration of Helsinki.

For all subjects, demographic information (age, gender and ethnicity), ocular history, in particular whether there was a history of contact lens use in the year leading up to the development of hydrops, history of atopy (defined as at least two of asthma, atopic dermatitis or allergic rhino-conjunctivitis), and family history of keratoconus in a first- or second-degree relative were recorded. It is not uncommon for people to belong to more than one ethnic group; however, upon first presentation to the public hospital services in New Zealand, patients are asked to report their primary ethnicity. This is recorded for all subjects and in this study, was used for ethnicity-related analyses. The Snellen visual acuity was recorded for the first hospital visit for both eyes but converted to logMAR for ease of statistical analysis. The date of onset of acute hydrops, the clinical history of the onset of hydrops including history of eye-rubbing (defined simply as ‘no’, ‘frequently’ or ‘a lot’), complications, management and outcome were also recorded. Where information regarding ocular history or patient demographics was missing, following verbal consent, subjects were interviewed by telephone to retrieve the relevant facts.

Subjects diagnosed with acute corneal hydrops due to other corneal ectasia or following penetrating keratoplasty were excluded. Subjects whose records were incomplete and could not be completed by telephone interview were also excluded.

A group of subjects with keratoconus but no history of acute hydrops, was selected on the basis of similar age (at first presentation to the hospital eye department for tertiary assessment) and gender spread. The control group was identified by a computerised search of the hospital clinical coding database and the Contact Lens Clinic database using the sole search term ‘keratoconus’ (Department of Ophthalmology, Auckland City Hospital). Only one eye per subject (selected at random) was included for analysis. Identical information to that obtained for the study group was obtained for the control group from clinical records and where any information was missing, subjects were interviewed by telephone to retrieve the relevant facts. By study design, this was a selective group with keratoconus, who had been referred to the hospital service for contact lens provision or anterior segment specialist assessment.

Statistical Analysis

All values were entered into a database (Excel; Microsoft, Redmond, WA, USA) and subsequently imported into statistical software for analysis (SPSS, ver. 15 for Windows; SPSS Inc., Chicago, IL, USA) with the assistance of a professional biomedical statistician. Basic descriptive statistics were calculated on all data gathered and are reported as mean and standard deviation or number and percentage, as appropriate. Binary logistic regression multivariate analysis was performed to investigate factors influencing the development of acute corneal hydrops in keratoconus. Explanatory variables included were age, gender, ethnicity (self-defined), age at diagnosis of keratoconus, earliest recorded visual acuity, history of atopy, history of eye-rubbing, past contact lens wear and family history of keratoconus.

All tests were two-tailed and p-values are quoted where appropriate.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In total, 101 eyes of 101 subjects with a history of hydrops, who fulfilled all study criteria, were included for analysis. Fifteen additional patients were excluded from further study due to incomplete data or diagnosis of hydrops secondary to other ectasias and the remainder of the patient records did not contain adequate information to satisfactorily fulfil a diagnosis of acute onset corneal hydrops. A total of approximately 800 patient records were reviewed for inclusion/exclusion purposes.

At the time of study, the mean age of the 101 subjects that met the study criteria was 29.7 ± 9.1 years. The male:female ratio was 55:45. An identical number of patients (n = 101) with a history of keratoconus but without corneal hydrops, constituted the control group. This group was not statistically different in terms of age (p = 0.535) and gender (p = 0.754) compared to the hydrops group.

The mean age at diagnosis of keratoconus in the hydrops group was 20.3 ± 7.8 years (range 10 to 47 years), while the hydrops typically developed four years after diagnosis (mean age 24.6 ± 8.4 years). The mean age at diagnosis of keratoconus did not differ significantly (p = 0.943) between the hydrops and control groups (21.9 ± 7.1 years). Thirty-one of the 101 subjects in the hydrops group made their first presentation to eye-care services with the episode of acute corneal hydrops; 80 per cent of these subjects were of either Maori or Pacific Island ethnicity.

In the corneal hydrops group, 54 per cent of subjects were of Pacific ethnicity, 25 per cent Maori, 13 per cent New Zealand European and eight per cent of other ethnicity. Compared to the control subjects, the association of ethnicity with hydrops approaches significance (p = 0.082). When analysed individually, associations reside particularly in those of Pacific Island or New Zealand European descent, with significantly more Pacific peoples (p = 0.026) and fewer New Zealand Europeans (p = 0.013) in the hydrops group compared to the control group. The number of Maori in each group did not statistically differ (p = 0.134).

Eye-rubbing was significantly (p = 0.011) more common in the hydrops group (75 per cent) compared to controls (53 per cent). Fifty-six per cent had worn contact lenses prior to the onset of hydrops compared to 70 per cent of control subjects (p = 0.139). There was no significant difference in the history of atopy between the groups (p = 0.577). The significance of these three variables is independent of each other; however, when comparing family history, those with a history of hydrops were less likely to have family members with keratoconus (24 per cent) compared to those in the control group (43 per cent) (p = 0.05). Additionally, and perhaps unsurprisingly, those who presented with hydrops without a pre-existing diagnosis of keratoconus were much less likely (p = 0.05) to have worn contact lenses (37 per cent) compared to patients who presented with hydrops with a pre-existing diagnosis of keratoconus (65 per cent) or control subjects (70 per cent). No other differences in clinical features were identified between those who presented with hydrops compared to those who developed hydrops with known keratoconus.

The mean initial logMAR visual acuity (VA) recorded for eyes subsequently affected by hydrops was 1.14 ± 0.63 (Snellen 6/72). This was significantly worse (p < 0.001) than the mean earliest logMAR VA recorded in the control group (0.63 ± 0.46) (Snellen 6/24). Excluding the 31 subjects who presented to the eye-care services for the first time with acute hydrops, those in the hydrops group still had significantly (p < 0.001) worse initial logMAR VA compared to the control group (1.09 ± 0.63) (Snellen 6/75). The mean earliest recorded logMAR VA of the contralateral eye for subjects in the hydrops group was 0.45 ± 0.41 (Snellen 6/18), which did not differ significantly (p = 0.125) from the contralateral eye of subjects in the control group (0.31 ± 0.31) (Snellen 6/12).

At the time of the study, 67 per cent of the eyes with acute hydrops had undergone penetrating keratoplasty, which was not significantly (p = 0.865) different from the control group (61 per cent). Thirty per cent of contralateral eyes in the hydrops group had also undergone penetrating keratoplasty compared to 27 per cent of the contralateral eyes in the control group (p = 0.683). Of those subjects with a history of hydrops requiring penetrating keratoplasty in one or both eyes (n = 79), 43 were of Pacific ethnicity, 18 Maori, 10 New Zealand Europeans and 10 were of other ethnicities. In the control group, 69 subjects required penetrating keratoplasty in one or both eyes, of whom 21 were of Pacific ethnicity, 15 Maori, 25 New Zealand Europeans and eight of other ethnic groups. Therefore, of all subjects requiring penetrating keratoplasty in this study, 65 per cent were either Pacific peoples or Maori and 24 per cent New Zealand Europeans.

These variables and key statistics are summarised in Table 1. Illustrative clinical features of subjects from the acute hydrops group are shown in Figures 1-3 to 4.

figure

Figure 1. Representative cases of severe acute corneal hydrops in keratoconus in: (A) a 32-year-old long-term contact lens wearer and (B) a 16-year-old subject, who had initially presented with keratoconus only six months earlier

Download figure to PowerPoint

figure

Figure 2. Acute corneal hydrops in a 17-year-old subject (A) with severe corneal and stromal oedema highlighted by (B) laser in vivo confocal microscopy. The lower images highlight resolving acute corneal hydrops with extensive central corneal scarring and significant corneal stromal neovascularisation (C) with in vivo confocal microscopy revealing well-formed stromal blood vessels (D) containing red blood cells and highly reflective white blood cells.

Download figure to PowerPoint

figure

Figure 3. Resolved corneal hydrops associated with keratoconus in two subjects: (A) a relatively small, off-axis, corneal scar associated with significant irregular astigmatism and poor contact lens-corrected visual acuity and (B) an extensive central and inferior, anterior stromal, corneal scar reducing visual acuity to 6/36 (early cataract is also observed in this image)

Download figure to PowerPoint

figure

Figure 4. Histology of a corneal button obtained from a subject undergoing penetrating keratoplasty for keratoconus associated with prior acute corneal hydrops. Although oedema has resolved the retracted Desçemet's membrane is clearly visible (large arrow head) and the area of Desçemet's absence in the initial tear is still visible (small arrow) with decreased density of keratocytes in the overlying posterior stroma. The epithelium shows variable thickness with areas of Bowman's layer loss and fibrotic scar extending into the epithelium (small arrowhead). (Congo red stain, original magnification ×40).

Download figure to PowerPoint

Table 1. Demographic and clinical factors for both corneal hydrops study group and the hydrops-free keratoconus control group. VA: visual acuity (logMAR)
Clinical factorsHydropsControlp-value
Gender (%)  0.754
Male5563 
Female4538 
Ethnicity (%)  0.082
New Zealand European1334 
Maori2518 
Pacific Island5438 
Other811 
Age at diagnosis of keratoconus (years)20.3 ± 7.821.9 ± 7.10.943
Eye-rubbing (%)75550.011
Contact lens wear (%)56700.139
Atopy (%)59550.898
Family history of keratoconus (%)24430.050
Mean initial recorded VA of affected eye (logMAR)1.14 ± 0.630.63 ± 0.46<0.001
Mean initial recorded VA of contralateral eye (logMAR)0.45 ± 0.410.31 ± 0.310.125
Penetrating keratoplasty of affected eye (%)67610.865
Penetrating keratoplasty of contralateral eye (%)30270.683
Ethnicities of subjects having had penetrating keratoplasty (%)
New Zealand European1025
Maori1815
Pacific Island4321
Other88

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Keratoconus leading to penetrating keratoplasty appears to be of higher prevalence in New Zealand than in other developed nations[11] and on the basis of clinical experience but not yet published data, the severity of keratoconus is also thought to be greater in New Zealand. This component of the Auckland Keratoconus Study aimed to identify demographic and clinical risk factors in the New Zealand population that may predispose to acute corneal hydrops in keratoconus, by comparing a set of variables in subjects with a history of acute corneal hydrops with an age and gender matched control group of keratoconus subjects without hydrops.

In the study population of 101 subjects with acute corneal hydrops, a trend between ethnicity and acute hydrops was identified, with Pacific peoples more likely and New Zealand Europeans less likely to develop acute hydrops. A history of eye-rubbing, poor VA at first hospital presentation and, paradoxically, a negative family history of keratoconus, were also associated with acute hydrops.

While atopy or atopic syndrome is typically defined as ‘A familial tendency to produce IgE antibodies in response to low doses of allergens and to develop typical symptoms, such as asthma, rhinoconjunctivitis or eczema/dermatitis,’[16] the majority of the subjects encountered in this study had not been serologically assessed for diagnosis of atopic disease. Therefore, to avoid over-inclusion, patients in this study required a history of two of the three common clinical manifestations to be assigned as atopic. An association but not causality has been noted between corneal hydrops and atopic keratoconjunctivitis,[1, 6, 17] yet a history of atopy in the current population did not appear to be associated with hydrops. One possible explanation for this, despite the size of the study groups, is that atopy per se is common in the New Zealand population[18] and particularly in the keratoconic population.[19] Notably, although no statistically significant predisposition could be elucidated in this study, atopic disease was highly prevalent in both groups.

We identified a statistically significant association between eye rubbing and acute hydrops. A history of eye-rubbing has been revealed to be significantly associated with hydrops and to be an independent risk factor from contact lens wear and atopy— both potentially confounding variables that in the past have been linked to eye-rubbing.[20-22] In the current study, we recognise that there could be potential recall bias following an event such as acute hydrops and that the classification of severity of eye rubbing may not have been consistent in clinical records, both potential limitations in terms of subsequent analysis.

The current study demonstrated that hydrops is strongly associated with poor VA at first hospital presentation compared to control subjects, even when the 31 per cent of subjects who first presented to hospital eye services with acute hydrops were removed from analysis. Unfortunately, these data confirm the clinical impression that it is not uncommon in New Zealand for the diagnosis of keratoconus to be made late, when patients present with advanced keratoconus or hydrops despite having had poor VA for several years. Despite a ‘free at point of access’ public health service (similar to the British National Health Service) the exact number of such cases and reasons for such late presentation for treatment have yet to be established.

The genetic predisposition to the development of keratoconus is well-established despite being incompletely understood.[23] Not only is there a familial association in keratoconus, certain ethnic groups have a higher prevalence of the disease than others within the same geographical region.[24-26] Interestingly, the current study suggests that those who develop hydrops are more likely to be of Pacific ethnicity and less likely to be of European descent compared to the control keratoconic population. The proportions of European, Pacific and Maori patients attending the tertiary corneal clinic are comparable to those of the New Zealand population;[29] however, patients diagnosed with keratoconus had significantly higher proportions of Maori and Pacific patients and lower rates of European patients than the total population. This suggests that disease-specific factors play a role in the ethnic distribution of patients with keratoconus and corneal hydrops rather than socioeconomic factors. Despite Pacific peoples making up only around seven per cent of the New Zealand population,[27] 43 per cent of corneal transplants performed in this study were on Pacific subjects (with or without a history of hydrops). The authors believe this to be representative of the overall New Zealand population, as approximately half of all corneal transplantation procedures in New Zealand are performed in the Auckland metropolitan area.[28] Furthermore, the public health service, represented by the Auckland District Health Board Ophthalmology Department in this study, performs the majority (71.8 per cent) of corneal transplants in the Auckland region.[28]

Somewhat counter-intuitively in the current study, corneal hydrops was statistically less common in those with a family history of keratoconus in a first or second-degree relative. This observation is also supported by the New Zealand data published by Jordan and colleagues[29] that highlighted a significant difference in the tomographic characteristics of subjects with a family history of keratoconus and those without. In the context of the current study, it must be noted that first, the assessment of family history was based on self-reporting, and second, although the prevalence of family history was much higher in both groups than in many published studies,[30-32] the majority of subjects did not report a family history, and finally, it is conceivable that a family history of keratoconus may tend to ‘protect’ against more severe keratoconus and hydrops in terms of (evolutionary) survival.

Perhaps somewhat paradoxically, subjects with hydrops were no more likely to have had a history of contact lens wear than those without hydrops. This observation is unusual, as the advanced keratoconus that predisposes to corneal hydrops generally leads to greater reliance on rigid contact lens wear for visual rehabilitation and is also inconsistent with the hypothesis that contact lens wear itself is potentially a cause of mechanical trauma and keratocytic apoptosis that may contribute to disease progression and development of hydrops.[23, 33, 34] Possible explanations for this difference are that in almost one-third of the cases, this was the first presentation to hospital eye services (the main source of subsidised contact lens provision for keratoconus in New Zealand) and the hydrops subjects had more advanced ectatic disease in the year prior to the acute event compared to controls and therefore, may have been more likely to be intolerant of contact lens wear.

As the members of the control study group were referred to hospital services with advanced keratoconus, perhaps unsurprisingly, the hydrops subjects were not more likely to undergo penetrating keratoplasty in the affected eye compared to hydrops-free subjects. In contrast, Tuft, Gregory and Buckley[1] reported the development of acute hydrops to be strongly associated with having a subsequent penetrating keratoplasty (59.2 per cent of hydrops subjects versus 13.1 per cent of hydrops-free subjects, p = 0.00001). The reason for the differences in this study may lie in the fact that the non-hydrops subjects with keratoconus were recruited from the hospital optometry and anterior segment clinics. Therefore, these are typically patients with advanced keratoconus referred from the community for specialist optometric care or consideration of a corneal transplant.

In conclusion, in the current study, acute corneal hydrops in keratoconus was statistically associated with Pacific but not Maori ethnicity and appeared less frequently in those of New Zealand European ethnicity. Hydrops also appeared to be associated with a history of eye-rubbing but not with a family history of keratoconus. Those who progressed to hydrops tended to have significantly worse VA at first hospital presentation; however, late presentation to hospital services with advanced keratoconus was not uncommon. Corneal transplantation did not appear more common in the population with hydrops compared to a matched non-hydrops hospital population with advanced keratoconus. Although we believe this large retrospective study involved a representative sample with assessment of 202 New Zealand subjects with advanced keratoconus, we envisage that a forthcoming, larger, prospective, clinical study of the natural history of subjects with keratoconus in New Zealand may further elucidate key associations.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We wish to thank Joanna Stewart (biostatistician) for her extensive statistical advice on this study and Judy Loh and Associate Professor Trevor Sherwin for providing the histology of the corneal button shown in Figure 4. Clinical photographs are courtesy of Professor Charles McGhee and in vivo confocal microscopic images are courtesy of Associate Professor Dipika V Patel and Dr Jennifer Fan Gaskin. We gratefully acknowledge that Dr Fan Gaskin was supported in part by a Health Research Council (New Zealand) research fellowship.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References