LETTER TO EDITOR : Effect of Mannitol on Intraocular Pressure in Vitrectomized and Nonvitrectomized Eyes: A Prospective Comparative Study.

IOP reduction. We would also like to note that the goal of our paper was to demonstrate that insertion of a GDI was beneficial to the patient long-term and did not aim to compare AV versus BGI. Tables 2 and 4 list in detail the complications at various timepoints and also lists the type of GDI that failed. With regards to the possibility of tube location as a confounder, our data did not include tubes placed in the ciliary sulcus or posterior segment.2 Tube location was standardized as all eyes underwent tube insertion through the anterior chamber. Also, for those patients that underwent nonvalved GDI surgery, tube fenestrations were employed. However, surgical timing before GDI surgery was not standardized. As this was a retrospective study and various uveitis providers with varying preferred practice patterns were involved in the care of these patients, there was no standard inflammation-free period before surgery. In addition, as the etiology of their uveitis varied from patient to patient, there were likely some eyes that necessitated more urgent surgery to manage their inflammation and IOP. Twenty-two (39.9%) eyes underwent fluocinolone acetonide (FA) implantation concurrently with GDI surgery, and these eyes were more likely to have active inflammation at the time of surgery than those that underwent GDI surgery without FA implants.1 Ultimately, the decision to pursue surgery, whether it is GDI alone versus GDI plus FA implant would likely rest on multiple factors, including the urgency of surgical intervention, patient’s history of inflammation, likelihood of recurrence of inflammation, etiology of uveitis, status of the other eye (if applicable), surgeon preference, as well as logistical issues that may have been out of the surgeons’ control. We agree with the readers’ observation that overfiltration from BGI may be the likely cause of hypotony in the 3 eyes that ultimately failed due to low IOP. In addition, hypotony is not an uncommon outcome of prolonged uveitis, possibly due to ciliary atrophy and aqueous hyposecrection, and this may have played a role in the hypotony seen in nonvalved GDIs.3 In our original paper, we referenced Christakis et al4 and the Ahmed Versus Baerveldt Study, which our readers also cited. While there are several well-established reports of a higher risk of primaryopen angle glaucoma in African Americans compared with Caucasian Americans, these conclusions cannot be generalized to our study as our cohort specifically studies patients with glaucoma secondary to either uveitis and/or steroid-response.5,6 We would like to thank the readers for noting that our table cites 11 failures while our paper cites 10 total failures. This is a typo and review of the data shows we indeed had 11 failures. Of these 11 failed eyes from 7 patients, only 1 patient (14.3%; 2 eyes) was African American. However, we would like to note that the majority of our patients (81.1%) were Caucasian and this may contribute to the low number of failed eyes belonging to African Americans. Regarding the Kaplan-Meier curve, while we did not include the number of patients lost to follow-up, table 3 documents the “number of patients at risk” for each time point. The number of patients at risk represents the number of patients at risk for failure at each time point, which includes those who did not meet criteria for failure and were not yet lost to follow-up. Thus, the survival probability takes into account the number of patients who had follow-up at that time point. There was no capping done for time with respect to enrollment of the last patient. We felt capping would limit our sample size and thus reduce the power of the study. As our study was a retrospective study, future prospective studies are required to assess these difficult cases.

information to draw conclusions regarding IOP reduction. We would also like to note that the goal of our paper was to demonstrate that insertion of a GDI was beneficial to the patient long-term and did not aim to compare AV versus BGI. Tables 2 and 4 list in detail the complications at various timepoints and also lists the type of GDI that failed.
With regards to the possibility of tube location as a confounder, our data did not include tubes placed in the ciliary sulcus or posterior segment. 2 Tube location was standardized as all eyes underwent tube insertion through the anterior chamber. Also, for those patients that underwent nonvalved GDI surgery, tube fenestrations were employed. However, surgical timing before GDI surgery was not standardized. As this was a retrospective study and various uveitis providers with varying preferred practice patterns were involved in the care of these patients, there was no standard inflammation-free period before surgery. In addition, as the etiology of their uveitis varied from patient to patient, there were likely some eyes that necessitated more urgent surgery to manage their inflammation and IOP. Twenty-two (39.9%) eyes underwent fluocinolone acetonide (FA) implantation concurrently with GDI surgery, and these eyes were more likely to have active inflammation at the time of surgery than those that underwent GDI surgery without FA implants. 1 Ultimately, the decision to pursue surgery, whether it is GDI alone versus GDI plus FA implant would likely rest on multiple factors, including the urgency of surgical intervention, patient's history of inflammation, likelihood of recurrence of inflammation, etiology of uveitis, status of the other eye (if applicable), surgeon preference, as well as logistical issues that may have been out of the surgeons' control.
We agree with the readers' observation that overfiltration from BGI may be the likely cause of hypotony in the 3 eyes that ultimately failed due to low IOP. In addition, hypotony is not an uncommon outcome of prolonged uveitis, possibly due to ciliary atrophy and aqueous hyposecrection, and this may have played a role in the hypotony seen in nonvalved GDIs. 3 In our original paper, we referenced Christakis et al 4 and the Ahmed Versus Baerveldt Study, which our readers also cited.
While there are several well-established reports of a higher risk of primaryopen angle glaucoma in African Americans compared with Caucasian Americans, these conclusions cannot be generalized to our study as our cohort specifically studies patients with glaucoma secondary to either uveitis and/or steroid-response. 5,6 We would like to thank the readers for noting that our table cites 11 failures while our paper cites 10 total failures. This is a typo and review of the data shows we indeed had 11 failures. Of these 11 failed eyes from 7 patients, only 1 patient (14.3%; 2 eyes) was African American. However, we would like to note that the majority of our patients (81.1%) were Caucasian and this may contribute to the low number of failed eyes belonging to African Americans.
Regarding the Kaplan-Meier curve, while we did not include the number of patients lost to follow-up, table 3 documents the "number of patients at risk" for each time point. The number of patients at risk represents the number of patients at risk for failure at each time point, which includes those who did not meet criteria for failure and were not yet lost to follow-up. Thus, the survival probability takes into account the number of patients who had follow-up at that time point. There was no capping done for time with respect to enrollment of the last patient. We felt capping would limit our sample size and thus reduce the power of the study. As our study was a retrospective study, future prospective studies are required to assess these difficult cases. The authors cited 2 articles which were pointing central nervous system mechanisms for osmotic agents. However, I think both articles had some methodological points to discuss.
In the first cited article, the authors concluded that transected optic nerves prevented central nervous system signals to reach the receptors in the eye. 2 The transection method of the optic nerve was unclear. We can assume that vascular structures accompanying the eye could be damaged like ciliary arteries and at least the central retinal artery. This would simply prevent the osmotic agents to reach the eye by the vascular path. The authors pointed out that topical anti-glaucoma drugs were still effective in the eyes with optic nerve transection. 2 But optic nerve transection would not prevent ocular absorption of the topical drugs but may simply prevent osmotic agents to reach the eye by the blood supply.
In the second cited article the authors damaged the optic nerve and supraoptic nucleus in the thalamus and concluded that the absence of intraocular pressure rise in those animals was because of the absence of necessary central signals. 3 However, transection of the optic nerve again might have caused vascular damage and prevented hypo-osmotic blood to reach the eye. In addition, the supraoptic nucleus secretes antidiuretic hormone and its damage causes diabetes insipidus which is characterized as significant water loss from kidneys. 4 Supraoptic nucleus damage should have caused hyperosmotic blood and prevented the IOP rise after water drinking.
With the abovementioned reasons, it is still difficult to conclude that osmotic agents have central nervous effects causing IOP to decrease.
Direct osmotic effects could still be a mechanism in silicone-filled eyes. A 100% silicone filling may not be possible in vitreoretinal surgery. The vitreous base may have remaining vitreous and retinal surgeons do not overfill the vitreous with the fear of glaucoma. This may lead to potential space for intravitreal fluid in the eye postoperatively and might be removed with osmotic agents. Or osmotic effects may cause shrinkage of the choroid or ciliary body and cause intraocular pressure to decrease.

Osman A. Polat, MD
Department of Ophthalmology, School of Medicine, Erciyes University, Kayseri Turkey