Jump to comment:

• Sequelae of diabetic vitrectomy
  Published on: 18 September 2000

• Published on: 18 September 2000

  Sequelae of diabetic vitrectomy

  • , Chief, Retina Division
  • Other Contributors:
    • Allan E. Krieger

  Editor,
  The work of West and Gregor again points out the importance of sclerotomy complications following pars plana vitrectomy. They demonstrate that, even in the hands of a skillful and experienced surgeon, vitreous hemorrhage after vitrectomy for diabetic retinopathy is common and requires vitreous cavity washout (VCWO) in 12% of cases. In their series, over half of the eyes had detectable fibrovascular...

  Show More

  Editor,
  The work of West and Gregor again points out the importance of sclerotomy complications following pars plana vitrectomy. They demonstrate that, even in the hands of a skillful and experienced surgeon, vitreous hemorrhage after vitrectomy for diabetic retinopathy is common and requires vitreous cavity washout (VCWO) in 12% of cases. In their series, over half of the eyes had detectable fibrovascular ingrowth (FVI) as the cause of the hemorrhage. Interestingly, in this case series of 159 eyes, no occurrences of anterior hyaloidal fibrovascular proliferation (AHPV) were noted.

  Definition of the relationship between these two entities has been controversial, to say the least. Part of the controversy is due to a misunderstanding of the nature and pathogenesis of FVI. As McLeod points out in his editorial, FVI is a term that has been used inadvisedly, suggesting the episcleral tissue grows into the eye through the sclerotomy incision [1]. While episcleral tissue, scleral fibroblasts, and ciliary epithelium all contribute, the majority of the fibroproliferative healing of a sclerotomy originates from the uvea of the ciliary body [2]. In normal wound healing, early fibrovascular proliferation in the incision is followed by its involution and contraction, with the result being the small scar seen at the internal aspect of a healed sclerotomy [2].

  Inevitably, because of the proximity of the vitreous base and anterior hyaloid, vitreous strands are adherent to the wound and fibrous tissue extends a short way into the vitreous body. This tissue may contain blood vessels, even with normal healing. From this perspective, all sclerotomy wounds heal with fibrovascular ingrowth. That is, ingrowth of tissue from the eye wall extends into the vitreous cavity. Fortunately, only in unusual circumstances does this process become exaggerated and result in what clinicians have termed FVI with its concomitant intraocular mischief [3].

  McLeod pointed out that ischemia is an important factor in inducing FVI and that it is seen mainly following vitrectomy for ischemic retinopathies. I agree that this is the case if one includes anterior proliferative vitreoretinopathy (APVR) in this group. Patients with APVR who have had previous vitrectomy frequently have an excessive amount of fibrovascular scarring from their sclerotomies that significantly affects the pathologic anatomy of the basal vitreous and its environs. These patients, however, often have had extensive scleral buckling and cryopexy, processes which undoubtedly induce some anterior ischemia in themselves [2]. In the series of West and Gregor, no patient was found to have a retinal detachment utlrasonographically or at the time of VCWO. In the original description of AHFP, most of the patients had retinal detachments that had required scleral buckling [4]. Since retinal detachment and scleral buckling exacerbate anterior ischemia, it is likely that AHFP, which is fibrovascular proliferation into the vitreous base from the retina and ciliary body, is induced by an ischemic drive similar to that casing FVI. The two entities exist on a continuum. When there is a surgical injury such as a sclerotomy, with disruption of tissue and inoculation of blood into the surrounding vitreous, excessive proliferation may occur with less induction than that which causes AHFP. Personally, although I have observed cases of AHFP without having previous vitrectomy, I have never seen a case of post-vitrectomy AHFP without some concurrent FVI.

  Finally, I'd like to make two other points. The first is that West and Gregor used clinical criteria to determine whether or not FVI existed and caused the recurrent vitreous hemorrhage. I have observed vitreous hemorrhage in an autopsy eye from what grossly appeared to be a normally healed sclerotomy wound [5]. Microscopically, that white scar contained numerous capillaries that were the source of the hemorrhage. Therefore, it may be that some of their non-FVI patients might actually have had vitreous hemorrhage from a sub-clinical FVI. Furthermore, FVI can involute with time, becoming less vascular in its appearance. So, the frequency of FVI may be even higher than reported. Lastly, I agree that episcleral sentinel vessels, externally entering the wound site, sometimes, but now always, indicate a possible FVI. These vessels are the result of a high degree of metabolic activity during the healing of sclerotomy wounds and may persist even though wound fibroplasia becomes involutional and clinically unimportant. Similar vessels are seen microscopically in the ciliary body [2]. When present, sentinel[2] vessels should raise our suspicions of FVI; but they do not rule it in, nor does their absence rule it out.

  References
  1. McLeod D. Entry site neovascularisation after diabetic vitrectomy. Br J Ophthalmol 2000;84:810-811.
  2. Kreiger A. The pars plana incision: experimental studies, pathologic observations, and clinical experience. Trans Amer Ophthalmol Soc 1991;89:549-621.
  3. Kreiger A, Straatsma B, Foos R. Incisional complications in pars plana vitrectomy. Mod Prob Ophthalmol 1977;18:210-233.
  4. Lewis H, Abrams G, Williams G. Anterior hyaloidal fibrovascular proliferation after diabetic vitrectomy. Am J Ophthalmol 1987;104:607-613.
  5. Foos R, Kreiger A, Nofsinger K. Pathologic study following vitrectomy for proliferative diabetic retinopathy. Retina 1985;5:101-106.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top