You are here

  • Home
  • Archive
  • Volume 99, Issue 5
  • The role of retrobulbar and retinal circulation on optic nerve head and retinal nerve fibre layer structure in patients with open-angle glaucoma over an 18-month period

Article Text

Article menu

Download PDFPDF

Clinical science
The role of retrobulbar and retinal circulation on optic nerve head and retinal nerve fibre layer structure in patients with open-angle glaucoma over an 18-month period
  1. Leslie Abrams Tobe1,
  2. Alon Harris1,
  3. Rehan M Hussain1,2,
  4. George Eckert1,
  5. Andrew Huck1,
  6. Joshua Park1,
  7. Patrick Egan1,
  8. Nathaniel J Kim1,
  9. Brent Siesky1
  1. 1Department of Ophthalmology, Indiana University School of Medicine, Indianapolis, Indiana, USA
  2. 2Department of Internal Medicine, Henry Ford Hospital, Detroit, Michigan, USA
  1. Correspondence to Professor Alon Harris, Department of Ophthalmology, Indiana University School of Medicine, 1160 West Michigan Street, Indianapolis, IN 46202, USA; alharris{at}indiana.edu

Abstract

Background/aims Evidence suggests that vascular abnormalities play a role in the pathogenesis of open-angle glaucoma (OAG) in some patients. This study aims to assess changes in retrobulbar and retinal blood flow over time in patients with glaucoma and examine their relationship to glaucomatous progression, as determined by retinal and optic nerve structure.

Methods In this observational study, 103 patients with OAG were examined at baseline and 18 months follow-up. Retrobulbar blood flow was measured by colour Doppler imaging in the ophthalmic, central retinal and temporal posterior ciliary artery (TPCA) and nasal short posterior ciliary artery. Retinal capillary blood flow was measured by confocal scanning laser Doppler. Peripapillary retinal nerve fibre layer thickness was assessed by optical coherence tomography. Non-parametric Wilcoxon signed ranks tests were used to assess for any statistically significant changes between the baseline and 18-month visits for the retrobulbar and retinal flow, as well as the structural parameters.

Results In general, retinal and retrobulbar blood flow parameters decreased over 18 months. Thinning of the optic disc rim and increase in cup area were associated with a higher resistance index (p=0.0334) and lower peak systolic velocity of TPCA (p=0.0282), respectively. A higher amount of retinal zero pixel blood flow correlated with a greater increase in cup/disc ratio (p=0.0170).

Conclusions Reductions in retrobulbar and retinal blood flow over time were associated with structural glaucomatous progression, as indicated by retinal and optic nerve changes.

  • Glaucoma
  • Optic Nerve
  • Physiology
  • Retina
  • Imaging

https://doi.org/10.1136/bjophthalmol-2014-305780

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Primary open-angle glaucoma (OAG) is a multifactorial optic neuropathy characterised by progressive retinal ganglion cell death and visual field loss. Despite being one of the leading causes of impaired vision worldwide, the pathogenesis of OAG is not well understood. The only currently available management of OAG is based upon reduction of intraocular pressure (IOP). However, many studies have shown that some patients experience disease progression despite significantly lowered IOP.1–4

    Over the past several decades, evidence has suggested that vascular pathologies play an important role in the aetiology and progression of OAG. Many population-based studies have found decreased retinal, choroidal and retrobulbar blood flow to be associated with increased glaucoma prevalence and incidence.5–11

    While the relationship between ocular blood flow and glaucoma is relatively well established, the nature of that relationship is not well understood. The reduced ocular blood flow seen in patients with OAG may be secondary to elevated IOP, retinal ganglion cell death or of primary vascular origin.6 ,7 Some theorise that decreased blood flow causes ischaemic insult to the optic disc and retina, which may play a role in the structural and functional damage that is characteristic of glaucoma.10 Others theorise that elevated IOP in glaucoma leads to tissue injury, thus obviating the need for blood flow to that region.11

    While the mechanism is not well established, there is growing evidence that glaucomatous eyes have impaired blood flow. Previous studies have found decreased ocular blood flow to be correlated with glaucomatous structural and functional progression.12–14 However, these studies are small in number, often measure only one vascular bed, and many do not address the change in both parameters over time. The purpose of this study is twofold. First, this study will assess a large cohort of patients with OAG to see how ocular blood flow in both retrobulbar and retinal vessels, as well as optic nerve head (ONH) structure, changes over time. Second, this study will also assess how changes in ocular blood flow over time are related to glaucomatous structural progression over time.

    Methods

    A total of 103 patients with primary OAG were assessed at baseline and prospectively followed every 6 months thereafter for ocular perfusion and blood flow as well as markers of structural glaucomatous progression. A comparison of results at baseline and 18 months was performed. All patients signed an informed consent after explanation in accordance with the Declaration of Helsinki, and the study protocol was approved by both the clinical and reading centre Institutional Review Boards.

    The diagnosis of OAG was confirmed by a certified ophthalmologist. Exclusion criteria included the following: extensive Humphrey visual field damage consisting of either a mean deviation (MD) <−15 decibels or a clinically determined threat to fixation in both hemifields, evidence of exfoliation or pigment dispersion, history of acute angle closure or an occludable anterior chamber angle by gonioscopy, history of chronic or recurrent inflammatory eye diseases, history of intraocular trauma, severe or potentially progressive retinal disease, any abnormality preventing reliable applanation tonometry, current use of any ophthalmic or systemic steroid, cataract surgery within the past year, resting pulse <50 bpm and uncontrolled cardiovascular, renal or pulmonary disease.

    All patients were questioned for their clinical history, assessed for IOP by Goldmann applanation tonometry and assessed for visual acuity. Blood flow in the retrobulbar vessels was assessed by a Philips HDI 5000 colour Doppler imaging (CDI) system (Philips Ultrasound, Bothell, Washington, USA) with a 7.5 MHz linear probe, used as described in detail previously.14 In brief, samples of pulsed-Doppler signal from within a 0.2×0.2 mm sample area were analysed to calculate blood velocities in the retrobulbar vasculature. CDI measurements were taken in the OA, central retinal artery (CRA), and nasal posterior ciliary artery (NPCA) and temporal posterior ciliary artery (TPCA). In each vessel, peak systolic velocity (PSV) and end diastolic velocity (EDV) were determined, and Pourcelot's resistive index (RI) was calculated (RI=(PSV – EDV)/PSV). Appropriate angle correction was used for the vessels measured, including OA as previously described.14

    Perfusion within peripapillary retinal capillary beds was assessed by confocal scanning laser Doppler flowmetry (Heidelberg Retinal Flowmeter (HRF), Heidelberg Engineering, Heidelberg, Germany). This technique has been described in detail previously.15 Briefly, HRF uses an infrared laser that scans the retina. The frequency and amplitude of Doppler shifts in the reflected light allow for calculation of total blood flow and create a physical map of flow values contained in the retina. This analysis also differentiates avascular tissue (measured as zero flow pixels) from perfused tissue, thereby describing the degree of vascularity of the fundus.16 ,17

    Peripapillary retinal nerve fibre layer (RNFL) thickness was assessed by optical coherence tomography (OCT) (Stratus software V.4.0, Zeiss Meditec, Dublin, California, USA). The examination was performed with an undilated pupil unless the pupil size precluded good quality analysis (signal strength <7) being achieved. To assess RNFL thickness, measurements were made along a circle concentric with the optic disc (Fast RNFL Thickness acquisition protocol). For each eye, we recorded the mean RNFL thickness and superior, inferior, nasal and temporal quadrant thicknesses, all calculated by OCT using existing software. Patients were followed up with repeat examinations 18 months after the first visit.

    Non-parametric Wilcoxon signed ranks tests were used to assess for any statistically significant changes between the baseline and 18-month visits for the retrobulbar and retinal flow, as well as the structural parameters. The change from baseline to 18 months was calculated for all measurements. Multivariable linear regression models were analysed for the change from baseline for each measurement with the following variables included in all models: sex, race (white vs non-white), age, baseline IOP, the baseline value for the measurement and a baseline perfusion pressure measurement (ocular perfusion pressure (OPP) if none of systolic perfusion pressure (SPP), diastolic perfusion pressure (DPP) or mean perfusion pressure (MPP) were significant when added to the other covariates and more significant than OPP). Stepwise model selection procedures were then used to evaluate HRF and CDI measurements for predicting OCT changes, with only measurements significant at p<0.05 retained in the models. Baseline values of the following measurements were log-transformed: OA, CRA, NPCA and TPCA systolic and diastolic velocities, HRF superior and inferior mean retinal capillary blood flow, OCT vertical integrated rim area, horizontal integrated rim width, disc area, cup area and rim area. All analyses were performed using SAS V.9.2 (SAS Institute, Cary, North Carolina, USA).

    Results

    A total of 103 patients were included in the analysis (mean age 67±10.6 years; 60 women). The ratio of Asian:black:Hispanic:white was 1:25:1:76. Table 1 includes information regarding the characteristics of the study population at baseline and 18-month follow-up.

    View this table:
    Table 1

    Patient characteristics at both baseline and 18-month follow-up

    In the study population as a whole over time, there was a decrease in retinal capillary mean flow, as measured by HRF, in the superior and inferior retinal regions (−19.4±13.7, p=0.0605; −18.8±12.1, p=0.0648, respectively). There was also a statistically significant increase in inferior retinal area of non-perfusion (amount of zero blood flow pixels) (0.015±0.006, p=0.0048). The mean retrobulbar blood flow values for PSV, EDV and RI are included in table 2.

    View this table:
    Table 2

    Retrobulbar blood flow velocity parameters at baseline and the 18-month visit, as measured by a colour Doppler imaging device

    In the study population, the ONH structural parameters and the RNFL thickness deteriorated from baseline to the 18-month visit. There was an increase in the cup/disc horizontal ratio (0.034±0.013, p=0.0168). We also found decreases in the horizontal integrated rim width (−0.09 mm2±0.03, p=0.0035), rim area (−0.11 mm2±0.04, p=0.0454), RNFL thickness in the inferior and nasal regions, and mean RNFL thickness (−5.1 µm±1.3, p=0.0001; −6.3 µm±1.5, p=0.0001; −2.9 µm±0.6, p<0.0001).

    In a multivariable linear regression model analysing the change of the horizontal integrated rim width from baseline, a greater decrease in the rim width was associated with a higher TPCA RI and a greater increase in OA PSV at 18 months, along with a higher rim width at baseline (table 3). A second model assessing change in vertical cup/disc ratio (CDR) as a dependent variable found a greater increase in CDR for a larger number of superior zero pixels at 18 months and for a lower CDR at baseline (table 3). A third model assessing the change in cup area from baseline found that a greater increase in cup area was associated with a greater decrease in TPCA PSV and a lower cup area at baseline (table 3). There was no statistically significant variation by age, gender or race (white vs non-white).

    View this table:
    Table 3

    Multivariable correlational model for retinal/retrobulbar flow and change in horizontal integrated rim width, cup/disc vertical ratio and cup area

    Conclusion

    While many studies in patients with glaucoma have found a correlation between diminished ocular blood flow and glaucomatous optic nerve changes, our study is unique in that it examined the temporal relationship of changes in retrobulbar and retinal blood flow in comparison with changes in optic nerve structural parameters over 18 months.

    As expected with primary open angle glaucoma, our results showed that the ONH morphology deteriorated over 18 months, with increased CDR. There were also statistically significant decreases in mean RNFL thickness, horizontal integrated rim width and rim area. These progressive glaucomatous changes were associated with an increase in areas of retinal non-perfusion over the 18-month period, with statistical significance obtained specifically in inferior retinal non-perfusion. This evidence supports the notion that reductions in peripapillary retinal blood flow at least occur simultaneous to optic nerve structural degeneration, though it is undetermined whether reduced retinal blood flow is contributory to the disease process or serves as a marker of pathogenesis. As would be expected with glaucomatous nerve head damage, the patients in the study also had visual field deterioration, with MD progressing from −3.1 to −3.4 over the 18-month course of the study.

    Several other studies have explored the relationship between altered retinal blood flow and optic nerve structural damage in patients with glaucoma, though the results are varied. Similar to our findings, Resch et al12 found ONH retinal capillary blood flow to be inversely correlated with horizontal CDR. Logan et al13 found that retinal blood flow was decreased in rim segments identified as abnormal (by Moorfield's Regression Analysis software) compared with those that were normal. Contrarily, Berisha et al found that lower retinal blood flow in patients with early-stage OAG was correlated with increased RNFL thickness, though no mechanism for this finding could be elucidated. They used the Cannon laser blood flowmeter, which measures different vascular tissue than the HRF, that may partially explain the difference in retinal blood flow results.18 Hwang et al19 found no correlation between retinal blood flow and structural loss of rim area or RNFL thickness, though those patients were not followed over time and were analysed with Fourier-domain OCT.

    In respect to the retrobulbar circulation, our results suggest that decreases in OA blood flow velocity over time are associated with progressive structural damage of the optic nerve in patients with glaucoma. Similarly, Januleviciene et al20 found that RNFL thinning was associated with significantly decreased blood flow in the OA in patients with OAG, though unlike our results, she also observed a statistically significant decrease in CRA blood flow velocity. Plange et al21 found that decreased CRA flow velocities were correlated with increased optic disc damage, though contrary to our results they did not find that OA and PCA flow velocities were correlated with stereometric parameters of the optic disc. Neither of the aforementioned studies evaluated these changes over a prospective time period, however.

    While OA EDV decreased in our study, NPCA and TPCA both had significantly increased PSV and EDV over the same time span, and CRA had no significant change. The mechanism behind this finding is unclear, but may be related to an early compensatory mechanism to shunt blood flow to the PCAs, which serve as the main blood supply to the optic nerve.22 With further glaucomatous progression and imbalance of vascular regulation, we might expect results similar to Zeitz et al,23 who found progressive glaucoma to be associated with decreased blood flow velocities of the posterior ciliary arteries. Interestingly, we found decreased PSV and increased RI of TPCA to be correlated with greater increase in cup area and a greater decrease in the rim width, respectively.

    Additionally, it is important to note the relationship between PSV, EDV and RI in interpreting our results. One model found a higher TPCA RI to be significantly correlated with a decrease in optic nerve rim area over the course of the study while another model found a greater decrease in TPCA PSV to be significantly correlated with a greater increase in cup area. RI and PSV are related by the following equation: RI=(PSV – EDV)/PSV. Taken with this equation, our results suggest a relationship between decreasing retrobulbar blood flow and ONH damage characteristic of OAG.

    We recognise that our study is not without limitations. Our study lacks a control group of healthy, non-glaucomatous patients; thus, we cannot know that these correlations between ocular blood flow and ONH structural changes would not be seen in healthy controls. Additionally, participation in this study did not affect glaucoma treatment such that patients were on different types and amounts of ocular antihypertensives, which are outlined in table 1. Additionally, many patients were also on systemic antihypertensives, which we understand could have an effect on ocular blood flow measurements. We also acknowledge that a selection bias may predispose patients with more vascular problems to participate in our study, while it is important to also note that despite a robust sample of over 100 patients it is possible that other relationships not found in this study exist beyond our ability to detect them. Further, in our study we used Stratus OCT to evaluate ONH structural parameters and RNFL thickness, which has lower resolution and slower scanning speed compared with spectral-domain OCT. However, at the time of the study's initiation in 2008, Stratus OCT was the standard in OCT imaging devices with the US Food and Drug Administration approval of the first spectral-domain OCT happening only 2 years prior in 200624 and only OCTs with a signal strength of seven or greater were included in the study. Also, a single experienced operator with over 10 years of experience (BS) performed all measurements at all time points. In addition, retrobulbar blood flow velocities obtained by CDI were standardised by using a printout at each visit to ensure that velocities were taken from nearly the same location in each vessel at each visit. Stalmans et al published a recent paper identifying how proper training significantly improved reproducibility;25 however, we also acknowledge that a higher variability was found in measurements of the SPCAs than OA and CRA.26

    In this cohort of patients with OAG, retinal and retrobulbar blood flow values generally decreased over time, with related glaucomatous changes in optic nerve structure. The large sample size, assessment of multiple vascular beds and observation of significant structural changes in the optic nerve suggest that a vascular component is involved in the progression of OAG. Future studies should include a properly designed predictability analysis.

    Acknowledgments

    AH would like to disclose that he receives remuneration from MSD and Alcon for serving as a lecturer and from Merck, Pharmalight, Sucampo, Biolight, Nanoretina and ONO Pharmaceuticals for serving as a consultant. AH also holds an ownership interest Adom (all above relationships are pursuant to IU's policy on outside activities). None of the other authors have any financial disclosures.

    References

    1. ↵
      1. Heijl A,
      2. Leske MC,
      3. Bengtsson B, et al
      ; Early Manifest Glaucoma Trial Group. Reduction of intraocular pressure and glaucoma progression: results from the early Manifest Glaucoma Trial. Arch Ophthalmol 2002;120:1268–79.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      Collaborative Normal-Tension Glaucoma Study Group. Comparison of glaucomatous progression between untreated patients with normal-tension glaucoma and patients with therapeutically reduced intraocular pressure. Am J Ophthalmol 1998;126:487–97.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Parc CE,
      2. Johnson DH,
      3. Oliver JE, et al
      . The long-term outcome of glaucoma filtration surgery. Am J Ophthalmol 2001;132:27–35.
      OpenUrlCrossRefPubMed
    4. ↵
      1. Ehrnrooth P,
      2. Puska P,
      3. Lehto I, et al
      . Progression of visual field defects and visual loss in trabeculectomized eyes. Graefes Arch Clin Exp Ophthalmol 2005;243:741–7.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Weinreb RN,
      2. Harris A, eds
      . Ocular blood flow in glaucoma: the 6th consensus report of the world glaucoma association. Amsterdam, The Netherlands: Kugler Publications, 2009.
    6. ↵
      1. Rojanapongpun P,
      2. Drance SM,
      3. Morrison BJ
      . Ophthalmic artery flow velocity in glaucomatous and normal subjects. Br J Ophthalmol 1993;77:25–9.
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Butt Z,
      2. McKillop G,
      3. O'Brien C, et al
      . Measurement of ocular blood flow velocity using colour Doppler imaging in low tension glaucoma. Eye 1995;9:29–33.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Moore D,
      2. Harris A,
      3. WuDunn D, et al
      . Dysfunctional regulation of ocular blood flow: A risk factor for glaucoma? Clin Ophthalmol 2008;2:849–61.
      OpenUrlPubMed
    9. ↵
      1. Flammer J,
      2. Orgül S,
      3. Costa VP, et al
      . The impact of ocular blood flow in glaucoma. Prog Retin Eye Res 2002;21:359–93.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Harris A,
      2. Sergott RC,
      3. Spaeth GL, et al
      . Color Doppler analysis of ocular vessel blood velocity in normal-tension glaucoma. Am J Ophthalmol 1994;118:642–9.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Sigal IA,
      2. Ethier CR
      . Biomechanics of the optic nerve head. Exp Eye Res 2009;88:799–807.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Resch H,
      2. Schmidl D,
      3. Hommer A, et al
      . Correlation of optic disc morphology and ocular perfusion parameters in patients with primary open angle glaucoma. Acta Ophthalmol 2011;89:544–9.
      OpenUrlCrossRefPubMed
    13. ↵
      1. Logan JF,
      2. Rankin SJ,
      3. Jackson AJ
      . Retinal blood flow measurements and neuroretinal rim damage in glaucoma. Br J Ophthalmol 2004;88:1049–54.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Williamson TH,
      2. Harris A
      . Color Doppler ultrasound imaging of the eye and orbit. Surv ophthalmol 1996;40:255–67.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Kagemann L,
      2. Harris A,
      3. Chung HS, et al
      . Heidelberg retinal flowmetry: factors affecting blood flow measurement. Br J Ophthalmol 1998;82:131–6.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Jonescu-Cuypers CP,
      2. Harris A,
      3. Wilson R, et al
      . Reproducibility of the Heidelberg retinal flowmeter in determining low perfusion areas in peripapillary retina. Br J Ophthalmol 2004;88:1266–9.
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Jonescu-Cuypers CP,
      2. Harris A,
      3. Bartz-Schmidt KU, et al
      . Reproducibility of circadian retinal and optic nerve head blood flow measurements by Heidelberg retina flowmetry. Br J Ophthalmol 2004;88:348–53.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Berisha F,
      2. Feke GT,
      3. Hirose T, et al
      . Retinal blood flow and nerve fiber layer measurements in early-stage open-angle glaucoma patients. Am J Ophthalmol 2008;146:466–72.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Hwang JC,
      2. Konduru R,
      3. Zhang X, et al
      . Relationship among visual field, blood flow, and neural structure measurements in glaucoma. Invest Ophthalmol Vis Sci 2012;53:3020–6.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Januleviciene I,
      2. Sliesoraityte I,
      3. Siesky B, et al
      . Diagnostic compatibility of structural and haemodynamic parameters in open-angle glaucoma. Acta Ophthalmol 2008;86:552–7.
      OpenUrlCrossRefPubMed
    21. ↵
      1. Plange N,
      2. Kaup M,
      3. Weber A, et al
      . Retrobulbar haemodynamics and morphometric optic disc analysis in primary open-angle glaucoma. Br J Ophthalmol 2006;90:1501–4.
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Harris A,
      2. Jonescu-Cuypers C,
      3. Kageman L, et al.
      , eds. Atlas of ocular blood flow. 2nd edn. Philadelphia, PA: Butterworth Heinemann Elsevier, 2010:5–8.
    23. ↵
      1. Zeitz O,
      2. Galambos P,
      3. Wagenfeld L, et al
      . Glaucoma progression is associated with decreased blood flow velocities in the short posterior ciliary artery. Br J Ophthalmol 2006;90:1245–8.
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Yi K,
      2. Mujat M,
      3. Park BH, et al
      . Spectral domain optical coherence tomography for quantitative evaluation of drusen and associated structural changes in non-neovascular age-related macular degeneration. Br J Ophthalmol 2009;93:176–81.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Stalmans I,
      2. Vandewalle E,
      3. Anderson D, et al
      . Use of colour Doppler imaging in ocular blood flow research. Acta Ophthalmol 2001;89:e609–30.
      OpenUrl
    26. ↵
      1. Harris A,
      2. Williamson T,
      3. Martin B, et al
      . Test/Retest reproducibility of color Doppler imaging assessment of blood flow velocity in orbital vessels. J Glaucoma 1995;4:281–6.
      OpenUrlPubMed

    Footnotes

    • Contributors BS and AH participated in the conception and design of the study, obtaining funding, supervision, data collection, analysis and interpretation, and critical revision of the manuscript. LAT, RMH, AH, JP, PE and NK all contributed to data collection, analysis of data and writing of this manuscript. GE assisted in analysis and interpretation, statistical expertise and revision of the manuscript. All authors approve of the final version of this manuscript.

    • Funding Supported by an unrestricted grant from Research to Prevent Blindness. The funding party did not have any role in the study design, collection of data, analysis of data, writing of the manuscript or decision to submit the manuscript.

    • Competing interests None.

    • Ethics approval Institutional Review Board, Indiana University School of Medicine.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles

    • At a glance
      Highlights from this issue
      Keith Barton James Chodosh Jost Jonas
      British Journal of Ophthalmology 2015; 99 i-i Published Online First: 17 Apr 2015. doi: 10.1136/bjophthalmol-2015-306989