Article Text

Download PDFPDF

Original article
Subjective and functional deterioration in recurrences of neovascular AMD are often preceded by morphologic changes in optic coherence tomography
  1. Robert Hoerster,
  2. Philipp S Muether,
  3. Manuel M Hermann,
  4. Konrad Koch,
  5. Bernd Kirchhof,
  6. Sascha Fauser
  1. Centre of Ophthalmology, University of Cologne, Cologne, Germany
  1. Correspondence to Philipp S Muether, Department of Vitreoretinal Surgery, Center of Ophthalmology, University of Cologne, Kerpener Strasse 62, 50924 Cologne, Germany; philmuether{at}


Background Different tests were applied to test the sensitivity of patient self-control; Amsler grid and visual acuity (VA) assessment, as well as fundus examinations to reveal recurrent choroidal neovascularisation (CNV) activity in age-related macular degeneration as detected by spectral domain optical coherence tomography (SD-OCT) in monthly controls.

Methods A prospective interventional case series of patients with exudative age-related macular degeneration was examined, which received ranibizumab injections until complete resolution of fluid in SD-OCT. Analysis of changes in subjective perception, Amsler grid, early treatment diabetic retinopathy study (ETDRS) VA, Radner reading VA and fundus examination was conducted in the case of OCT-confirmed CNV recurrences.

Results Out of 40 morphological recurrences determined by SD-OCT, six (15%) were noticed by subjective patient perception. Amsler grid testing revealed deterioration in 12 cases (30%); 11 recurrences (28%) were accompanied by loss of ≥5 letters in ETDRS VA and/or ≥1 line in Radner VA; fundus examination showed signs of novel CNV activity in 10 out of 40 recurrences (25%). The combined sensitivity of all diagnostic methods compared to SD-OCT for recurrence detection was 67.5% (27 out of 40 recurrences).

Conclusion Subjective patient perception, Amsler grid, VA as well as fundus examination lead to pronounced underestimations of CNV recurrences. Morphologic recurrences can be detected prior to functional deterioration. As any delay of treatment can result in irreversible vision loss, attempts should be made to provide monthly OCT controls to detect recurrences as early as possible.

  • Wet macular degeneration
  • optical coherence tomography
  • visual acuity, neovascularisation
  • treatment medical
  • visual perception

Statistics from

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.


Neovascular age-related macular degeneration (AMD) is a leading cause of severe and irreversible vision loss in the developed world among patients who are of 50 years or more.1–5 The clinical ANCHOR and MARINA trials for treatment of exudative AMD have demonstrated the efficacy and safety of ranibizumab for the preservation and improvement of visual acuity (VA) by using a fixed-dosing regimen, which requires monthly injections over 2 years.6 7

The CATT trial showed equivalent results for optical coherence tomography (OCT)-guided pro re nata (PRN; as needed) treatment when compared to a monthly fixed dosing regimen with ranibizumab. In the PRN study arm, repeated treatment was performed whenever persistent or recurrent choroidal neovascularisation (CNV) activity was observed, including fluid accumulation on OCT or leakage on fluorescein angiography (FA).8 Similar results were obtained in the earlier but smaller uncontrolled PrONTO study9 10 as well as in the controlled and prospective SUSTAIN study.11

All PRN regimen-based studies required monthly visits and OCT controls and FA if needed. Monthly controls for the detection of recurrence in specialised centres impose a considerable burden on health systems. Therefore, it may be tempting to conduct standard diagnostics such as VA testing and fundus examinations instead. Patients may even be instructed to return to specialised centres for OCT control only if they notice a decrease in their VA or perception of new metamorphopsia.

The objective of this study was to test the sensitivity of screening methods, in particular, patient self-control, Amsler grid and VA assessment, as well as fundus examinations to reveal recurrent CNV as detected by monthly spectral domain optical coherence tomographies (SD-OCT) or fluorescence angiographies.

Materials and methods

This prospective interventional case series was conducted at the Department of Vitreoretinal Surgery, Center of Ophthalmology, at the University of Cologne. The protocol followed the German product characteristics for ranibizumab treatment and the tenets of the Helsinki protocol. Informed consent was obtained from all patients.

Following primary diagnosis of exudative AMD based on FA (fluorescein and indocyanine green) and SD-OCT, 31 patients with a total of 40 recurrences (31 primary recurrences, 9 secondary recurrences) received three consecutive monthly intravitreal injections of ranibizumab. After these three injections, FA and SD-OCT were performed again to evaluate CNV activity. If needed, therapy was continued until complete resolution of CNV activity. Eyes with persistent fluid retention, or sustained resolution of fluid requiring continuous injections and thus without a recurrence of CNV activity, were excluded from the analysis.

After complete resolution of fluid retention, monthly controls including SD-OCT (Spectralis HRA+OCT, Heidelberg Engineering GmbH, Dossenheim, Germany) were performed according to the following specifications on all visits: scan area 20°×15° (5.8×4.4 mm), 37 B-scans, distance between B-scans 121 μm, 29 images averaged per scan. FA at follow-up was only performed whenever CNV activity based on SD-OCT images was questionable. At all visits, subjective perception was assessed. Metamorphopsia and scotoma were tested with the Amsler grid. Standardised early treatment diabetic retinopathy study (ETDRS) VA testing, Radner visual-reading acuity, fundus examination and fundus photography were performed.

Subjective patient perception was determined utilising the following questions:

  • ‘Did the reading ability of your affected eye increase, decrease or remain stable?’

  • ‘Did metamorphopsia of your affected eye improve, deteriorate or remain stable in daily life?’

  • ‘Regarding the overall impression of your affected eye, did it improve, worsen or remain stable?’

Recurrent activity of CNV was defined as described by Bashshur et al and by Lalvani et al: (1) a recurrence of any subretinal fluid or cystic maculopathy on OCT in a previously dry macula; (2) an increase in CRT by more than 100 μm from the lowest recorded value, especially for eyes with persistent subretinal fluid of cystic maculopathy on OCT; (3) the appearance of new areas of classic CNV; (4) the appearance of new haemorrhages or (5) a decrease of five letters from the highest recorded BCVA, associated with leakage on FA or fluid on OCT.10 12

As signs for possible recurrence, we considered subjective deterioration with regard to any of the questions: (1) increase in metamorphopsia or scotoma in Amsler grid testing; (2) loss of ≥5 letters in ETDRS VA or (3) loss of ≥1 line in Radner VA. Any change of these functional or subjective parameters at the time of OCT-detected recurrence was compared to the previous examination with a dry OCT status.

Statistical analyses were performed using commercially available software. Significance levels were calculated with the Fisher exact test (SPSS V.18.0, IBM).


In this prospective interventional case series, 40 AMD-related CNV recurrences in 31 patients were analysed. Patient characteristics are depicted in table 1. At baseline examination before treatments, 19 of the later recurrences were assessed to be due to occult CNV lesions, another 19 were associated with predominantly classic lesions and 2 cases showed retinal angiomatous proliferation. Two subsequent recurrences in the same patients were observed in nine cases.

Table 1

Characteristics at baseline and recurrence

Of the 40 morphological recurrences determined by SD-OCT, six (15%) were noticed by subjective patient perception. Of these patients, three (7.5%) noticed increased metamorphopsia or scotoma in daily life, four (10.0%) noticed a decrease in reading ability and five (12.5%) noticed a decrease in overall condition of the affected eye; multiple complaints were possible. All of the patients who perceived a recurrence by subjective VA deterioration were shown to have an objective deterioration either in BCVA, Radner reading VA or Amsler grid or in several tests. In 17 of 40 recurrences (42.5%), the fellow eye generally had a better VA than the study eye; VA was worse in 23 of 40 recurrences (57.5%). Perception of VA deterioration did not depend significantly on the VA of the non-study eye (p=1.0). There was no significant difference in the perception of recurrences in patients with one or two recurrences (p=0.6726). Moreover, there was no difference between the perception of first-time and second-time recurrences (p=1.0).

Upon ophthalmologic examination, Amsler grid testing revealed deterioration in 12 cases (30%). Eleven recurrences (28%) were accompanied by a loss of ≥5 letters in ETDRS VA and/or ≥1 line in Radner VA assessment. In five cases (12.5%), a decrease in ETDRS distant VA was observed, and in nine cases (22.5%), a decrease in Radner reading VA was observed. The sensitivity of the Radner reading VA and the ETDRS distant VA did not differ significantly (p=0.378). Of the 11 recurrences accompanied by a VA loss, only two patients perceived this loss (18%). During the absence of fluid, ETDRS VA was 59.7 letters, and the Radner VA was 0.70 logRAD. At the time of CNV recurrence, ETDRS VA was 60.0 letters, and the Radner VA was 0.73 logRAD (table 1).

Fundus examinations showed signs of novel CNV activity in 10 of 40 recurrences (25%). The combined sensitivity of all of the above-mentioned diagnostics, compared to SD-OCT detection of recurrence, was 67.5% (27 of 40 recurrences) (figure 1).

Figure 1

Sensitivity of choroidal neovascularisation (CNV) recurrence detection by subjective and objective diagnostics. White bars represent the sensitivity of single tests, grey bars represent combined sensitivity of tests, black bar represents SD-OCT set to 100% detection rate. ETDRS, early treatment diabetic retinopathy study; meta., metamorphopsia; SD-OCT, spectral domain optical coherence tomography; subj., subjective.


In the treatment of exudative AMD, either monthly ranibizumab injections or PRN treatment with monthly controls ensured the best results,6 7 9–13 whereas fixed quarterly injections14 resulted only in limited benefit. In view of the increasing numbers of patients suffering from exudative CNV, regular controls impose a considerable burden on health systems. We compared the sensitivity of multiple screening diagnostics for detection of recurrences. We could clearly demonstrate that neither patient perception, nor VA assessment, nor fundus examinations are sensitive enough to indicate disease recurrence at a stage at which SD-OCT already clearly visualises retinal fluid (figure 1).

Patients with neovascular AMD have only poor perception of morphological recurrences after initial resolution of intraretinal or subretinal fluid. Only 12.5% of all cases had noticed this decrease when interviewed about their general reading ability, their metamorphopsia in daily life or their overall functional impressions. There was no difference between perceptions of first-time and second-time recurrences. This indicates that patients with two recurrences were unable to perceive their second recurrence more readily because of a learning effect from the first recurrence or less often due to increasing indifference. This is in part in contradiction to reported patient perceptions shortly after ranibizumab therapy, where an increase of VA and subjective measures has been noticed.7 15 16 This may be explained by the more rapid decrease of OCT thickness after injection9 in comparison to its slower increase in cases of recurrence.

When the semi-subjective Amsler grid was presented to the patients in a standardised fashion, this test had a higher sensitivity than any subjective perception (30.0% vs 12.5%). Only 11 early recurrences were accompanied by a decrease of ≥5 letters ETDRS, or ≥1 line of Radner reading VA, and only in two cases was this decrease perceived by the patient. Radner reading VA showed a slightly higher sensitivity (22.5%) compared with ETDRS VA (12.5%), although the difference was not significant. Overall, there was no mean change in VA either in ETDRS and Radner testing, even when OCT already showed obvious signs of novel CNV activity. On average, the detection of fluid in OCT scans was sensitive enough to precede any deterioration in VA.

For early CNV recurrences, fundus examination revealed signs of novel CNV activity in only 25% of cases. Even when all clinical diagnostics except OCT were combined, the sensitivity was only 67.5%. In this study, clinical examination, and in particular VA assessments, was conducted in a highly standardised fashion. It can be assumed that sensitivity would be reduced in less standardised conditions.

Even though impaired visual function may be a hint of recurrent CNV activity in AMD, it does not appear sensitive enough for the detection of early morphologic recurrences. However, standard tests of VA and reading VA do not recognise the full spectrum of macular dysfunction. Microperimetry testing of the central fovea and the macular area has been shown to detect functional correlates of ranibizumab therapy more readily than VA testing.15 16 However, due to the complexity and time requirements of this examination method, microperimetry is not feasible as a clinical screening tool.

Recently established treatment regimens like the treat-and-extend regimen aim to individualise the treatment of neovascular AMD by stepwise extension of injection intervals until recurrence occurs. Subsequently, the recurrence interval minus 2 weeks is chosen as the regular prophylactic re-treatment interval.17 This interesting treatment regimen seems to have comparable outcomes to PRN regimens. However, regular recurrence intervals have to be assumed to occur when applying such treatment protocols.

This study shows that reliance on self-controlled patient perception in neovascular AMD after ranibizumab therapy, as well as on VA testing and fundus examination, leads to pronounced underestimations of CNV recurrences and delays of treatment unless OCT imaging is also used. Morphologic recurrences can be detected prior to functional deterioration. It should be emphasised that recent data underscores the fact that any delay of treatment leads to significant loss of VA and macular function.18

We conclude that if a PRN treatment is regarded as the regimen of choice in exudative AMD, attempts should be made to provide monthly OCT controls to detect early recurrences and avoid irreversible deterioration of VA.



  • RH and PSM contributed equally to this work.

  • Funding Supported by the Koeln Fortune Program/Faculty of Medicine, University of Cologne.

  • Competing interests None declared.

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