Statistics from Altmetric.com
Partial restoration of vision was achieved following vitrectomy in 15 of 17 eyes, ostensibly through removal of diabetic fibrovascular proliferations from the nasal part of the optic disc and relief of vitreopapillary traction which, for between 6 months and 6 years, had caused a reversible functional impairment of the papillomacular bundle via stretching and kinking of ganglion cell axons and additional or consecutive effects on their prelaminar blood supply. Eyes with such features (that is, with traction primarily localised nasally on the disc and unaccountably affecting acuity without any associated disturbance of the central visual field) should be subjected to early vitrectomy in order to prevent irreversible long term damage to central vision. This is the recommendation of Kroll and colleagues in a report which is published in this issue of theBJO (p 261) and which merits the careful attention of all ophthalmologists involved in the management of diabetic eye disease. Indeed, some will already be asking—if this entity is so common, how have I failed to recognise it for all these years?
Diabetic papillopathy (that is, disc swelling without any tractional component) has proved difficult to characterise clinically1; similarly, defining the precise physiological basis of visual loss in eyes with ischaemic/proliferative diabetic retinopathy is frequently problematic. Biomicroscopic signs of subtle but visually significant vitreomacular traction, for example, may be hard to elicit given the difficulties in detecting the consequent minor distortions of the transparent outer retina or the intraretinal disruption representing tractional schisis; reports of visual benefit and reversal of macular oedema after vitreous detachment2or after vitrectomy and peeling of the posterior hyaloid membrane3-5 may bear witness to a “trampoline effect” across the macula in some eyes and may obviate the need to invoke alternative mechanisms of recovery such as physiological “vitreoperfusion”.6 7 Fluorescein angiography may also be less than definitive in diabetic eyes since good correlation between the extent of enlargement of the foveal avascular zone and the deficit in vision is lacking.8 Attributing visual loss unreservedly to diabetic tractional papillopathy, then, is no mean task and any criticism of the achievements of Kroll and colleagues must take this into account. Given that the changes in the appearance of the optic disc after surgery may not be especially informative,9 their case rests largely on the extent to which other mechanisms of visual benefit from surgery were excluded and on electrophysiological data. Regarding the former, scrutiny of their exemplar fluorescein study fails to inspire confidence since there are clear signs of macular retinal traction and perifoveal dye leakage in the preoperative illustration notwithstanding their stated exclusion criteria. Furthermore, removal of mild to moderate vitreous haemorrhage (that is, haemorrhage of less severity than the level warranting exclusion from their study) may well have had a significant influence on the visual benefit deriving from surgery in some eyes; it would have been reassuring if parallel “control” electrophysiological studies had been undertaken in eyes undergoing vitrectomy for mild to moderate vitreous haemorrhage wherein no vitreopapillary traction was evident. More information on the state of the vitreous in these eyes would also have been valuable—was the vitreous attached or detached from the retina?
Electrophysiological data were lacking in six eyes, restricted to visually evoked potentials (VEPs) elicited by flash in eight eyes, and derived from pattern stimulation in only three of the 17 eyes studied by Kroll and colleagues. Contrary to their statement, only a small minority (that is, not “most”) of the 17 eyes met their own evaluation criteria that significant increases in acuity and VEP amplitude, together with a decrease in VEP latency, resulted from surgery—though this standard may be regarded as too stringent. There was no apparent correlation between visual benefit from vitrectomy and improved VEP variables; indeed, the two eyes with the greatest rise in VEP amplitude (and the highest absolute VEP amplitude values postoperatively) enjoyed minimal visual improvement (and had continuing poor vision) after surgery. However, it may well be that changes in VEP latency are more reliable than VEP amplitude changes,10 11 while more use of comparative studies between the operated eye and the fellow eye (the latter always showing a normal disc and flat retina in their series) might have been helpful in defining normative VEP values in these patients, in controlling for the influence of blood glucose changes,10 and in demonstrating repeatability independent of any surgical effect. Extending the range of investigational modalities12 might also have provided more insight into the nature of the residual visual defect seen in the operated eyes.
When contemplating vitreous surgery for proliferative diabetic retinopathy, most attention has so far been focused on vitreomacular traction and there is reason to hope that techniques for better defining retinal structural pathology13 will aid clinical evaluation in future. There is little doubt, however, that ganglion cell axons in severe diabetic eye disease are potentially subject to a variety of metabolic and ischaemic insults; an additional putative mechanical factor is one we surely should not ignore, especially one operating at a point of known vulnerability of axoplasmic transport.14 Whether Kroll and colleagues have provided incontrovertible evidence of such mechanical effects through their report is open to debate; more precisely specified and expressed studies are needed. In the meantime, further extension of the indications for vitrectomy in proliferative diabetic retinopathy should proceed with caution. For example, many eyes with nasal ectopia of the macular neuroretina and underlying pigment epithelium (as a result of fibrovascular contraction nasal to the optic disc) retain excellent visual acuity—perhaps the full thickness retinal ectopia is protective against nerve fibre stretching and visual loss in such eyes.