The prevalence of NVG after diabetic vitrectomy was reported in 4.6%-23.6% of cases.(14–16, 19) Concerning changes in intraocular anti-VEGF levels after diabetic vitrectomy, it has been suggested that a high VEGF level could be maintained in the vitreous cavity after vitrectomy for PDR.(20–22) Moreover, vitrectomy with or without a cataract operation can disrupt the structures considered barriers to the diffusion of angiogenic factors to the anterior segment from the posterior segment.(23) Although the effect of vitrectomy (removing the vitreous) or phacovitrectomy itself on the intraocular vasoproliferative substances (VEGF and inflammatory cytokines) and development of NVG has not been established, it is clear that NVG developed after diabetic vitrectomy is clinically important.(24–28)
Few studies have been conducted on the prognosis of surgical treatment (AGVI) for NVG after diabetic vitrectomy.(17, 29) They suggested that AGVI is a safe and effective procedure that enables successful IOP control in patients with NVG associated with diabetic vitrectomy. Other studies related to the treatment of NVG also positively evaluated the therapeutic effect of AGVI.(3, 8–12) However, previous studies included cases with heterogeneous underlying ocular conditions and only focused on the normalization of IOP in evaluating the surgical success of AGVI. Detailed visual outcome has not been sufficiently considered due to generally poor visual outcomes and only the proportion of vision less after AGVI was calculated. Because NVG patients after diabetic vitrectomy tend to have better baseline visual acuity before the diagnosis of NVG compared to other conditions, surgical success rate based on visual acuity could be evaluated. Especially, this study focused on patients with unfavorable visual outcomes, even though IOP was well controlled after AGVI.
In addition to normalization of IOP, this study compared the baseline BCVA "before" NVG diagnosis and after AGVI, and the factors affecting visual prognosis were also considered. The overall surgical success rate based on IOP was comparable to a previous study that reported percentages of 83.8% at one year in the NVG after diabetic vitrectomy group.(17) 43.8% and 26.3% had unfavorable visual outcomes in groups 1 and 2, respectively. It is estimated that consistently high IOP may be associated with poor visual prognosis. Especially in group 1, early surgical timing within a week was significantly related to vision preservation (favorable visual outcome). This suggests that close observation should be performed during the first week after NVG diagnosis, and AGVI should be considered without delay in cases with insufficient response to initial non-surgical treatment. Based on our study's results, we suggest that the IOP change during the first week after NVG diagnosis could be used as a criterion for distinguishing group 1 from group 2. In particular, the change in IOP over the first week can be used as a useful indicator when the standard was set to 10 mmHg (Table 1).
Park et al. suggested that visual deterioration after NVG in patients with PDR was attributed to the progression of diabetic retinopathy or GON.(17) In addition to these factors, considering the relationship between the period of IOP increase and deterioration of central vision and the representative case (Fig. 3), retinal ischemia due to blood flow restriction during severe IOP rise seems to be a reasonable additional cause of visual deterioration. The probable mechanisms for retinal ischemia due to increased IOP are as follows: 1) decreased blood flow in the optic nerve head (ONH), 2) failure of blood flow autoregulation in the ONH and 3) decreased blood flow in the arterioles and capillaries of the inner retina (Fig. 4). Perfusion pressure decrease and retinal blood flow autoregulation failure due to severe IOP rise inhibit blood flow in ONH.(30, 31) In addition, there have been many reports of retinal artery occlusion due to IOP rise (18, 32). When perfusion from the optic nerve head is non-physiologically reduced, inner retinal blood circulation may be more directly affected by increased IOP. Moreover, diabetic peripheral neuropathy, the aging process, and accompanying hypertension or arteriosclerosis/atherosclerosis are additional factors that can accelerate retinal ischemia due to increased IOP.(33, 34)
Figure 3 shows serial optical coherence tomography (OCT) photographs, which reveal inner retinal edema with hyper-reflectivity and subsequent progressive retinal thinning. These findings are similar to the pattern of incomplete CRAO rather than complete CRAO and support our conclusion that early AGVI can prevent vision loss. It can be explained because the mechanism of retinal ischemia is not due to embolism, but because the blood flow is limited by the decrease in perfusion pressure and blood flow autoregulation failure. Because the mechanism of retinal ischemia is different, unlike in embolic CRAO, where the damage starts at the middle retinal layer, ischemic change was observed mainly in the inner layer adjacent to the eyeball cavity.
Cascades from the increased IOP mentioned above are presumed to be the contributors to vision loss, especially in group 1 (Fig. 4). The reason for delayed AGVI was not only due to patient compliance, including rejection of surgical treatment, personal circumstances, or difficulties undergoing emergency operations but also delayed consulting process from retina specialists who diagnosed NVG to glaucoma specialists who did not have much experience in managing NVG after diabetic vitrectomy. It has been difficult to specify IOP, which causes impaired blood flow in the optic nerve head and retinal circulation. In hyphema patients (otherwise, healthy eye), it is accepted that surgical management should be performed on the second day for 60 mmHg, the fifth day for 50 mmHg, and the seventh day for 35 mmHg.(35) Based on our study results, in NVG patients with advanced PDR, the maximum tolerable period of the retina against ischemia due to medically uncontrolled high IOP of more than 30 mmHg is estimated to be less than one week. The consulting process with glaucoma specialists, medical treatment attempts, and the evaluation of its effectiveness should be completed within one week without delay.
Limitations of this study include its retrospective nature, small sample size, and limited follow-up period. However, it is difficult to conduct a well-designed case-controlled study or prospective randomized trials because NVG after vitrectomy is rare, conducted under different systemic and ocular conditions, and the threat of vision. Our study applied specific exclusion criteria despite having an insufficient number of cases to evaluate visual outcomes and its associated factors. Although long-term outcomes over one year of AGVI were not included in this study's results, we aimed to investigate the effect of temporary IOP rise on the retina and visual function of PDR patients, so that the appropriate timing of AGVI could be determined. Furthermore, the time point of NVG diagnosis may not be exactly the time point of the IOP increase. However, this study analyzed the effect of the degree and duration of IOP elevation on visual acuity clinically rather than a detailed review of the pathophysiology, ideal treatment options for each stage, and course of NVG. Theoretically, in many cases of group 2, the open-angle served as a potential reservoir for medical treatment. Future research will be required to provide strong evidence for predicting responsiveness to initial non-surgical treatment through the iridocorneal angle evaluation or alternative methods in NVG.
In conclusion, AGVI was a safe and effective procedure that enabled successful IOP control in patients with NVG associated with diabetic vitrectomy. However, to improve the visual outcomes of AGVI, physicians should be cautious not to delay surgical intervention, especially in patients who do not respond to initial non-surgical treatment. If IOP is not adequately controlled (< 30 mmHg) within a week from the time of NVG diagnosis, early AGVI within six days might be necessary to preserve useful vision.