Aims To investigate the incidence of macula displacement and symptoms of distortion following rhegmatogenous retinal detachment (RRD) repair, quantify the displacement where seen and further characterise the nature of the displacement.
Methods Consecutive patients undergoing primary RRD repair were assessed postoperatively with fundus autofluorescence and optical coherence tomography imaging, and the extent of macula displacement quantified using a novel means. Findings were examined for correlations with symptoms and pre-operative features.
Results Macula displacement was evident postoperatively in 72% of 32 consecutive fovea-involving detachments treated with vitrectomy and gas. It was also evident in 5/17 foveal-sparing cases treated with vitrectomy and gas and in two of two patients with fovea-involving detachments treated with vitrectomy and oil. There was a significant correlation between the presence of macula displacement and symptoms of distortion in the early postoperative period (p=0.013). Symptomatic patients described bending of lines with or without objects appearing smaller or narrower in the operated eye. Quantifying the displacement demonstrated that the extent of displacement was associated with distance from the optic disc (p=0.005) and the extent of retinal detachment.
Conclusions Displacement of the macula is common following RRD repair and heterogeneous in nature. Most affected patients are symptomatic in the early postoperative period.
- Treatment Surgery
Statistics from Altmetric.com
Symptoms of distortion are prevalent among patients who have undergone anatomically successful retinal detachment repair, and may occur despite good recovery of visual acuity.1–5 The symptoms can be considered as distortion of image size (dysmetropsia) and/or form (metamorphopsia). In patients with good vision in the fellow eye, dysmetropsia and metamorphopsia in the operated eye can impair fusion and compromise binocular single vision.6 Unless one of the images is suppressed, the patient is at risk of binocular diplopia, asthenopia or retinal rivalry, which may be poorly tolerated. In extreme cases, affected patients may choose to partially occlude the operated eye despite good postoperative acuity.
It is likely that distortion following rhegmatogenous retinal detachment (RRD) repair results from ectopic replacement of macula photoreceptors.7 Shiragami et al8 described how displacement of the retina can be detected post-operatively by fundus autofluorescence (FAF) imaging. Hyper-autofluorescent lines were detected running parallel to selected retinal blood vessels; it is hypothesised that the hyper-autofluorescence is due to increased metabolic activity of retinal pigment epithelium (RPE) that had previously been located under major retinal blood vessels and is postoperatively exposed to light. Comparison of the retinal vasculature and the parallel hyper-autofluorescent ‘RPE vessel ghost’ lines thus gives a indication of retinal displacement.
The aims of this study were to determine the incidence of macula displacement and association with symptoms of distortion following retinal detachment repair, to develop a means of quantifying the displacement when seen, and to further characterise the nature of the displacement.
Over an 11-month period patients undergoing RRD repair were routinely evaluated with FAF and optical coherence tomography (OCT) imaging and asked about symptoms of distortion of image size or form on postoperative review. The findings of those undergoing primary surgery, with no history of previous vitrectomy, RRD repair or macula disease, were examined in this study. The study was approved by the local research and development review board.
FAF imaging was performed with the Topcon TRC-50IX fundus camera and Spaide autofluorescence filters (Topcon, Tokyo, Japan) following pupillary dilation.9 The excitation and emission bandwidths of these filters are 500–610 and 675–715 nm, respectively. In brief, 50° field images were taken from both eyes centred on the fovea. Flash intensity, aperture and gain were adjusted to accommodate differences in media clarity and pupil size. Cross-sectional macular images with a 6 mm width were also acquired with spectral domain OCT imaging (3D OCT-1000; Topcon) from both eyes and centred on the fovea at the same visit as the FAF imaging was performed. At the same visit patients were asked if they were aware of distortion of straight lines when looking with just the operated eye, or a difference in image size between the two eyes. Visual acuities were tested using COMPlog, a validated, computerised and semi-automated visual acuity measurement device, and measured in number of ETDRS letters.10
FAF images were further analysed using Photoshop CS4 (Adobe Systems, San José, California, USA). ‘RPE vessel ghosts’ were defined as hyper-autofluorescent lines running approximately parallel to retinal blood vessels and with a similar contour but separate from the blood vessel. The presence of RPE vessel ghosts was taken as evidence of retinal displacement only in the absence of confluent subretinal fluid (SRF) on the corresponding OCT scan (to exclude potential artefact from a parallax effect). Affected images were overlaid with concentric measuring circles, centred on the midpoint of the optic disc, with diameters of 1.0–10.0 disc diameters (figure 1). RPE vessel ghosts were highlighted, and where these crossed one of the concentric measuring circles, the distance was measured from that intersection to the intersection of the same measuring circle and the corresponding blood vessel. Measurements were expressed as a percentage of disc diameter in the same eye. For each sample point the distance from the current vessel position to the centre of the optic disc (r) was measured, along with the angle of the interconnecting line. The x and y co-ordinates of each sample point could then be calculated using trigonometry (figure 1). Positive x and y co-ordinates corresponded to retina temporal and superior to the optic disc, respectively. The number of sample points identified in each image using this standardised approach was used as an indicator of how widespread the displacement of blood vessels was in that subject's macula, and used to test for an association with the presence or absence of postoperative symptoms of distortion.
Patient characteristics and outcomes were obtained by a review of the case notes and a prospectively completed electronic database used to record operation findings and procedures at the time of surgery (VITREOR, Microsoft Access, available with the book on vitreoretinal surgery published by Springer).11 The operation note was used to determine whether the detachment was fovea-involving or fovea-sparing at the time of surgery. Operation drawings were analysed in Photoshop to measure the maximal extent of the detachment, in the temporal half of the retina, above and below the fovea (0–90°).
Fisher's exact test was used to examine for associations between the presence or absence of displacement on FAF imaging and lens status at the time of imaging (phakia or pseudophakia) and recorded symptoms of distortion. A t test was used to test for differences in visual acuity between patients with and without displacement. The relationship between the amplitude of post-operative retinal vessel displacement and position within the macula was studied using a multi-level mixed effects linear regression model which accommodates different measurements taken from individual patients and adjusts for potential confounding variables. The Stata V.12.1 (College Station, Texas, USA) xtmixed module was used for this. For outcomes that were not normally distributed we used a normal logarithmic transformation prior to analysis. The level of probability taken as significant was p=0.05 or less.
Demographics and clinical characteristics of the patient sample
There were 87 eligible patients during the study period of whom 27 were excluded from further analysis due to incomplete follow-up data. Specifically, 10 patients were transferred back early to referring units, did not attend follow-up appointments or had investigations inadvertently missed, seven patients had media opacities preventing adequate postoperative imaging, and 10 patients had intra-operative or early postoperative complications such as retinal re-detachment. Of the 60 patients whose follow-up data were further analysed, 49 were treated with vitrectomy, retinopexy and gas tamponade (of whom 32/49 had fovea-involving detachments at the time of surgery), two were treated with vitrectomy, retinopexy and oil tamponade (both fovea-involving detachments), and nine were treated with cryo-buckle non-drain procedures (of whom three had fovea-involving detachments). FAF and OCT imaging, and symptom evaluation, were typically performed at the first postoperative visit following absorption of intraocular gas when used; the median interval between surgery and imaging in these 60 patients was 29 days. Subsequent follow-up images and functional data were not included in this study. Postoperative findings are further evaluated in relation to the type of operative procedure performed and foveal status recorded at the time of surgery.
Following vitrectomy and gas, RPE vessel ghosts indicating displacement were seen in 23 of 32 eyes with fovea-involving detachments (72%). The detachment and surgical details for these patients are shown in the online supplementary data table. Analysis of the fovea-involving detachments treated with vitrectomy and gas (n=32) found no association between the presence or absence of distortion and lens status (phakia or pseudophakia) or post-operative visual acuity (p=0.648 and p = 0.678, respectively). However, the presence of postoperative symptoms of distortion did correlate with number of displaced blood vessel sampling points identified in each autofluorescence image (an indicator of how widespread the displacement of vessels was in that subject's macula) (r=0.565, p<0.001).
Of the 17 eyes with fovea-sparing retina detachments treated with vitrectomy and gas, five had detached retina confined to the superonasal quadrant only. None of these showed evidence of macula displacement following surgery. The remaining 12 fovea-sparing detachments treated with vitrectomy and gas had detached retina directly superior or superotemporal to the macula. Five of these had postoperative shift evident within the macular arcades and in three this was confined to the area of detachment seen clinically, but in two cases there was a shift of blood vessels directly above and below the fovea suggesting that the previously uninvolved fovea had become involved. Both these patients experienced symptoms of distortion postoperatively. In all cases with vitrectomy and gas where macular displacement was present, the vertical component of movement was in a superior to inferior direction with RPE vessel ghosts lying superior to the postoperative position of the corresponding retinal blood vessels (figure 1).
Two fovea-involving detachments were treated with vitrectomy and 1300 cs silicone oil as a primary procedure. Both had near total (>270°) retinal detachments. In one patient this was secondary to a giant retinal tear and the other patient had multiple tears in the inferior, temporal and superior quadrants. Neither patient was advised to adopt a specific posture postoperatively and both had evidence of macula displacement. In the patient with a temporal giant retinal tear, the direction of shift was upwards with RPE vessel ghosts lying inferior to the corresponding retinal blood vessel (figure 2); in the other the observed shift was downwards. Of the nine patients treated with cryo-buckle procedures, none had RPE vessel ghosts evident on imaging.
OCT images were reviewed for evidence of disrupted retinal structure postoperatively. Of the 32 eyes treated with vitrectomy and gas for fovea-off detachments, 47% (15/32) had an OCT abnormality at this early time point of imaging. The abnormalities detected were outer-retinal folds (n=6), multiple blebs of SRF (n=3), a single bleb of SRF (n=1, with the bleb being sub-foveal), outer-retinal photoreceptor layer defects (n=3), an epiretinal membrane (n=1), a full thickness macular fold (n=1) and a full thickness macular hole (n=1, this same eye also had outer-retinal folds).
Of the 17 eyes with fovea-on detachments treated with vitrectomy and gas, four had OCT abnormalities within the macula, namely outer-retinal folds (n=2), cystoid macula oedema (n=1) and epiretinal membrane (n=1). The two eyes with outer-retinal folds both had superotemporal detachments that were recorded as being fovea-sparing. Postoperatively, one had evidence of a single outer-retinal fold passing through the fovea; it is likely the fluid had either extended further pre- or peri-operatively than had been detected clinically, or that the extent of fluid had progressed further at the end of surgery or postoperatively. The patient had 6/6 vision but was aware of a kink in horizontal lines when imaged 48 days following surgery. The FAF image had evidence of displaced retina superior and temporal to the fovea. The other eye with outer-retinal folds had presented with an acute-on-chronic round-hole superotemporal RRD. The postoperative OCT showed multiple diffusely distributed outer-retinal folds affecting the macula when imaged 17 days following surgery. At this time point, vision was limited to 6/36.
Postoperative distortion was reported in 69% (22/32) of those patients with foveal-off detachments treated with vitrectomy and gas. Of those with evidence of macular shift, symptoms were present in 83% (19/23), and conversely, in those with symptoms there was evidence of shift in 86% (19/22), indicating a high level of concordance between FAF evidence of shift and postoperative symptoms of distortion (Fisher's exact p=0.0126; figure 3). Some 45% of patients who were aware of bending of straight lines also responded positively when asked if there was a difference in image size between the eyes; this was variably described as objects appearing smaller or narrower in the operated eye as compared to the unoperated eye. No patients reported images appearing larger in the operated eye.
Of the patients treated with vitrectomy and gas for fovea-sparing detachments, two reported postoperative distortion. These patients accounted for two of the five with evidence of macula shift postoperatively. One had an acuity of 78 ETDRS letters (UK Snellen equivalent 6/8) and a normal OCT but FAF evidence of shift above and below the fovea suggestive of foveal involvement. The other was described previously with OCT evidence of a single outer-retinal fold passing through the fovea, symptoms of distortion and evidence of displaced blood vessels superior and temporal to the fovea.
Further analysis of the FAF images was performed for the largest single group of patients, fovea-involving detachments treated with vitrectomy and gas. The single patient with a full-thickness macular fold was excluded from this analysis as the gross distortion of normal anatomy prevented accurate quantification of displacement. In the remaining 22 patients with fovea-involving detachments treated with vitrectomy and gas and with postoperative retinal displacement evident, mixed model statistical analysis was used to test for associations between the amplitude of displacement and each of x, y and r (distance of the sample point from the optic disc). A significant association was found with r (p=0.005) but not x or y (p=0.156 and p=0.711, respectively). However, it was evident that the displacement is typically heterogeneous within the macula, or along the course of individual blood vessels, as demonstrated in the inferior vessel displayed in figure 1. These findings indicate that rather than there having been a uniform shift of the retina, there has instead been variable distortion or stretch which is only partly determined by position within the macula.
In the 22 patients with quantitative analysis, the extent of retinal detachment in the temporal retina was measured in degrees of involvement above and below the fovea. In our regression model, we found significant associations between the magnitude of displacement and the above (p=0.0003) and below (p=0.02) angles of detachment (table 1). The regression model we applied takes into account the non-independence of multiple observations from the retina of individual subjects, and the results therefore suggest that when present, the displacement was more marked in eyes with larger extents of detached retina at the time of surgery.
This study demonstrates that displacement of the retina is both common following vitrectomy surgery for retinal detachment and has functional significance. Furthermore, we have developed a means of quantifying the displacement and further characterised the changes seen. The ability to quantify displacement permits comparisons between patients and between different techniques and allows measurement of change during longitudinal follow-up; it is therefore likely to facilitate further understanding of the aetiology and steps that can be taken to minimise its occurrence.
The finding that the extent of retinal displacement within the macula of individuals is heterogeneous is of functional and prognostic significance. These results indicate that the retina has not been purely rotated around the disc, or shifted downwards, but instead there is a more complex movement with some regions within the macula being displaced more than others (heterogeneous shift). We propose this is because the retina has been variably stretched rather than there being a uniform displacement. This explanation is consistent with the symptoms of distortion and change in image size, rather than double vision alone, in patients following retinal detachment surgery.1 Furthermore, all patients who had dysmetropia in this series and that of Ugarte and Williamson,1 had micropsia and not macropsia; this would be explained by the fixation target stimulating a smaller number of foveal photoreceptors than in the fellow eye and so further supports the theory that at least some of the displacement seen is due to retinal stretch and hence greater separation between foveal photoreceptors. Where patients do have double vision, the co-incident distortion means that prismatic correction is unlikely to be able to induce fusion of the disparate images. Paradoxically, prismatic correction is therefore sometimes used to increase the disparity such that the patient can ignore the second image, as an alternative to the use of fogging lenses.12
Macular folds are thought to arise from a combination of SRF, intraocular gas and the effect of gravity on the SRF.13 Hypotony, scleral redundancy and undulations in the detached retina may also play a contributory role.14 They may be formed deliberately in limited macular translocation surgery15 or inadvertently following retinal detachment surgery.16–18 It is likely that the retinal displacement detected here, which sometimes co-exists with partial thickness retinal folds, is a result of the same processes but to a lesser extent.5 ,8 ,14 ,19 It will therefore be interesting to see if techniques to maximise drainage (eg, posterior retinotomies and the use of perfluorocarbon liquids) or techniques to reduce the propensity for folds to form in the presence of SRF (eg, small volume 100% gas bubbles rather than complete gas fills, specific post-operative posturing regimes, or pneumatic retinopexy) are effective at reducing displacement. At present there is no consensus or proof of the effectiveness of these steps in reducing postoperative displacement and distortion.
In the study by Shiragami et al8 of 43 consecutive patients treated with vitrectomy and gas, perfluorcarbon liquids were used in the majority of cases (30/43) and in all cases drainage was through the break rather than via a posterior retinotomy. Immediately following surgery, patients were sat up for several minutes before face-down or other positioning was commenced. Displacement was detected in 27 of the 43 eyes (63%). In a smaller series by Codenotti et al,19 five out of five consecutive macular-off RRDs treated with vitrectomy and gas had evidence of displacement; all of these patients were positioned face-down immediately following surgery for an unspecified duration and then advised to avoid the sitting position. In our series there were differences in drainage techniques, but almost all patients were positioned macula down for over 60 min immediately following surgery. Despite this, there was still inferior displacement in 72% of the fovea-involving detachments treated with vitrectomy and gas. One explanation is that longer periods of postoperative posturing could be required for full fluid reabsorption. However, the temporal displacement seen in study patient no. 30 (figure 2 and see online supplementary data table) may indicate a potential risk from prolonged macula down posturing. An alternative approach, in the event of significant residual SRF, may be to position patients supine with residual vitreous cavity fluid so as to avoid a fluid–gas gradient across the macula.11 There may be multiple reasons for displacement of blood vessels following surgery and it is likely that a larger study would be required to identify specific risk factors.
To date, and to the best of our knowledge, all reported instances of retinal displacement after vitrectomy and gas have been in a downwards direction, presumably due to the effect of gravity on residual SRF in gas-filled eyes.5 ,8 ,14 ,19 The patient reported here with silicone oil tamponade and upwards displacement confirms a previous findings by Codenotti et al19 who had two patients with silicone oil tamponade in which upwards displacement was seen. The numbers are not sufficient to justify a definite conclusion, but the most likely explanation for the direction of displacement in the patient we report relates to their presentation with a giant retinal tear; the tear extended from 1 to 5 o'clock temporally, resulting in a very mobile retina. If drainage was performed from the superior extent of this tear during silicone oil exchange, it is conceivable that the retina would be displaced in this direction, which we believe is a more likely explanation than theories related to the surface tension of oil.19
One of our aims was to develop a means of quantifying displacement such that comparisons could be made between patients. Image magnification can vary with refractive error, and potentially between visits due to small changes in focussing. All images were therefore measured with reference to the optic disc diameter as a means of standardising the measurements. Another challenge was how to identify points to measure between as unless there is a vessel bifurcation, it is not possible to determine the precise point on a vessel ghost which corresponds to a specific point on the corresponding retinal blood vessel. It is to overcome this problem that the concentric circles were used; measurements were made between ghost vessel and retinal blood vessel intersections on the same sampling circle. This method therefore involves an assumption that the shift of retinal blood vessels is primarily a rotation around the optic disc, in keeping with the original description by Shiragami et al.8 However, the results from our study indicate that there is more to this movement than a simple rotation. Furthermore, although shift of macula vessels was seen to be in the superior–inferior axis, it is to be noted that the vessels are predominantly horizontal and so shift in the horizontal axis would be harder to detect. In selected cases, shift in other directions could be seen outside the macula (figure 2). The absolute values of displacement measured should therefore be used with caution but do allow comparisons over time, and have been sufficient to demonstrate here that the extent of movement is heterogeneous within the macula.
Macula displacement is common following vitrectomy retinal detachment repair with gas tamponade, and is associated with symptoms of distortion. It is heterogeneous within individuals, indicating variable stretch of the affected retina rather than a simple rotation. Measuring the extent of displacement will permit comparisons over time and between individuals, and so opens the door for further research to study the aetiology and prognosis.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors All authors contributed to study design and writing the paper. All authors except for PH contributed to data collection. PH led the statistical analysis, and described aspects of the study design.
Competing interests None.
Ethics approval Guy's and St. Thomas’ NHS Foundation Trust R&D Department approved this study (reference RJ112/N127).
Provenance and peer review Not commissioned; externally peer reviewed.
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.