Intraocular Pressure Changes in the Contralateral Eye After Glaucoma Surgery

Supplemental Digital Content is available in the text. Précis: Glaucoma surgery in 1 eye can result in significant intraocular pressure (IOP) elevation in the fellow eye in a significant percentage of subjects. This effect is more pronounced following glaucoma drainage device implantation and in subjects developing postoperative hypotony in the first eye. Purpose: The aim was to investigate the course and magnitude of IOP changes in contralateral eyes following glaucoma surgery. Patients and Methods: We studied 131 glaucoma patients undergoing surgery. IOP changes in fellow eyes and the number of glaucoma medications were recorded for 6 months. Both normal and glaucomatous fellow eyes were included. Patients on preoperative acetazolamide were analyzed separately. IOP change ≥4 mm Hg or 20% increase from baseline values was considered significant. Results: Baseline IOP in fellow eyes was 13.1±3.3 mm Hg which increased significantly at all-time points with a peak at 1 week (mean increase: 3.1±3.2 mm Hg) gradually decreasing up to 6 months (mean increase 0.9±2.1 mm Hg). The number of fellow eyes with clinically significant IOP elevation at 1 day, 1 week, and 1, 3. and 6 months was 47, 76, 47, 31, and 37 showing mean IOP rise of 4.5±2.3 (38.5%), 5.4±1.8 (48.5%), 4.8±1.6 (44.2%), 3.7±1.3 (36.5%), and 3.2±1.3 (32.6%) mm Hg, respectively. Medications were added to 18 fellow eyes (including 10 initially “normal” eyes) while 7 other fellow eyes required surgery. Significant IOP elevation in fellow eyes was associated with tube shunt surgery (P<0.008 at all-time points), a diagnosis of glaucoma following cataract surgery, IOP ≤5 mm Hg on the first postoperative day (P=0.002) and use of preoperative acetazolamide. Conclusion: Glaucoma surgery was associated with IOP elevation in fellow eyes in a significant number of subjects and suggests that IOP in the unoperated eye should also be checked at postoperative visits.

(J Glaucoma 2021;30:1074-1081) Intraocular pressure (IOP) remains the main modifiable risk factor for glaucoma. 1 It has long been shown that IOP fluctuation in one eye may result in alterations of IOP in the other eye. 2 The precise mechanisms underlying this so-called consensual ophthalmic reaction (COR), first described by Weekers in 1924, are unknown. 3,4 The COR was well documented in a patient with ocular hypertension who received topical beta blockers. The contralateral treatment-naive eye showed IOP reduction which was highly correlated with the amount of pressure reduction in the treated eye. 5 In patients receiving topical pilocarpine or adrenaline in one eye, the COR led to IOP reduction in the fellow eye, most notably in patients with ocular hypertension as compared with normotensive eyes. 2 The COR has also been shown to occur after selective laser trabeculoplasty (SLT) and Argon laser trabeculoplasty resulting in IOP reduction in untreated fellow eyes. 6,7 The literature is inconclusive regarding IOP changes in fellow eyes following glaucoma surgery. While some studies have shown IOP elevation, 3,8 others have reported only early and transient IOP elevation followed by IOP reduction, [9][10][11] or no IOP change at all in fellow unoperated eyes. 12 This study was conducted to evaluate the course and magnitude of IOP change in patients undergoing a variety of glaucoma procedures including trabeculectomy or combined phacoemulsification and trabeculectomy (phacotrabeculectomy) and Ahmed glaucoma valve (AGV) implantation.

PATIENTS AND METHODS
This prospective study was performed at a tertiary eye care center and included consecutive glaucoma patients with uncontrolled IOP on maximum tolerated medical therapy (MTMT) scheduled for glaucoma surgery from January 2018 to September 2019. The study protocol was approved by the ethics committee (equivalent to Institutional Review Board) and adhered to the tenets of Declaration of Helsinki; informed consent was obtained from all participants or their legal guardians.
We included patients with primary or secondary glaucoma and uncontrolled IOP on MTMT who required surgery in 1 eye. MTMT included the concomitant topical use of a prostaglandin analog, a beta-blocker, a carbonic anhydrase inhibitor, and an α-2-agonist. Patients with a variety of glaucoma diagnoses including primary open angle glaucoma (POAG), primary angle closure glaucoma, secondary open angle glaucoma, secondary angle closure glaucoma, Juvenile open angle glaucoma, primary congenital glaucoma (PCG) and glaucoma following cataract surgery (GFCS) in the index eyes were enrolled in the study regardless of whether the fellow eyes were glaucomatous or not, were receiving medical therapy or had previously undergone glaucoma surgery. Monocular subjects, patients on systemic or topical steroids because of ocular or systemic conditions, subjects with significant corneal opacity or pathologies affecting IOP measurements, and patients with <6 months of follow up were excluded from the study.
All subjects underwent a complete preoperative ophthalmic examination for both eyes including LogMAR best corrected visual acuity, slit-lamp examination, IOP measurement with Goldmann applanation tonometry (at the time of planning surgery and again on the day before surgery), gonioscopy and dilated funduscopy on slit-lamp using a 78.0 D lens. The decision to choose the surgical procedure (by F.S. and S.Y.) was individualized based on patients age, life expectancy, drug compliance or allergy, target IOP, type and stage of glaucoma and prior glaucoma surgeries.
Postoperatively, all topical ocular hypotensive medications were discontinued in the operated index eyes while topical glaucoma medications were continued as before for fellow eyes. In subjects taking oral acetazolamide, the medication was discontinued postoperatively. The operated eyes received ciprofloxacin 0.3% (Ciplex, Sina Darou Laboratories Co.) 4 times a day for a week and betamethasone 0.1% eye drops (Betasonit, Sina Darou Laboratories) every 3 hours which was gradually tapered off over 1.5 to 2 months. Antiglaucoma medications were added if IOP exceeded the target pressure range especially after AGV implantation.
Patients had regular follow-up visits and IOP in both eyes was measured on the day before surgery, on postoperative day 1 (POD1), weeks 1 (POW1) and 2 (POW2), and months 1 (POM1), 2 (POM2), 3 (POM3), and 6 (POM6) using a calibrated Goldmann applanation tonometer in the morning (9 to 10 AM). For each eye, 2 consecutive measurements were obtained and the average of 2 readings was used for analysis. If the difference between the 2 measurements exceeded 2 mm Hg, a third reading was taken and the average of the 2 closest readings was recorded. The amount of IOP change in fellow eyes at each postoperative visit was compared with preoperative values. An IOP change exceeding 4 mm Hg or representing more than 20% increase from baseline values was considered as clinically significant. Humphrey visual field (HVF) perimetry (Carl Zeiss Meditec, Dublin, CA) was also performed when possible.

Statistical Analysis
To present data, we used mean, SD, median and range, frequency, and percent values. To assess IOP changes within the groups we used Wilcoxon signed-rank test. To evaluate differences between the groups, Mann-Whitney U and Kruskal-Wallis tests were used. Generalized estimating equations were used to consider possible correlation of the results, whenever needed. Correlation between variables was assessed using Spearman correlation coefficient and Partial correlation coefficient (whenever adjustment for 2 eyes of 1 patient was needed). In order to compare the baseline status of the groups we used ordinal logistic regression.
All statistical analyses were performed using IBM SPSS Statistics, version 25.0 (SPSS Inc., Chicago, IL). P-values <0.05 were considered as statistically significant.

RESULTS
A total of 131 patients with mean age of 58 ± 21.5 years including 73 (55.7%) male subjects were operated in 2 eye while their fellow eyes were studied for possible IOP changes. Thirtyone eyes (23.7%) were scheduled for trabeculectomy, 35 eyes (26.7%) for phacotrabeculectomy and 65 eyes (49.6%) for AGV implantation. Table 1 shows baseline characteristics of enrolled patients. All cases of POAG, Juvenile open angle glaucoma, primary angle closure glaucoma, PCG, and GFCS were bilateral. A total of 49 fellow eyes (37.4%) were normal all of which belonged to subgroups of patients with secondary open angle glaucoma and secondary angle closure glaucoma ( Table 2). Seventeen operated eyes (13%) and 39 fellow eyes (29.7%) had history of prior glaucoma surgery including AGV implantation, phacotrabeculectomy, trabeculectomy, and deep sclerectomy. History of other procedures, including cataract surgery and laser peripheral iridotomy, was present in 41 (31.3%) and 45 (34.4%) of operated and fellow eyes, respectively.
Mean IOP of operated eyes and the number of topical medications decreased from 27.1 ± 7.3 mm Hg and 3.6 ± 0.5 medications preoperatively, to 11.5 ± 2.7 mm Hg and 0.7 ± 1 medications by the end of POM6, respectively (P < 0.001). Mean baseline IOP of fellow eyes was 13.1 ± 3.3 mm Hg which increased after operation on the other eye and reached a maximum level at POW1 (16.3 ± 3.7 mm Hg, range = 8 to 26 mm Hg, mean IOP elevation: 3.1 ± 3.2, P = 0.001). Thereafter, mean IOP of fellow eyes gradually decreased to 13.7 ± 2.8 mm Hg at POM6 with a mean net IOP increase of 0.8 ± 2.1 (P < 0.001) (Fig. 1). The number of topical glaucoma medications in fellow eyes was 0.9 ± 1.2 at baseline which increased to 1.0 ± 1.3 (P < 0.001) at POM1 and reached 1.0 ± 1.2 (P < 0.001) at POM6. A significant positive correlation was observed between IOP reduction in the operated eyes and IOP increment in fellow eyes (r = 0.23; P < 0.001). Forty-two operated eyes (32%) received topical antiglaucoma medications during the postoperative followup course of which 31 eyes underwent AGV implantation. B-blockers, carbonic anhydrase inhibitors and prostaglandin analogs were added consecutively as needed. Only 2 operated eyes received antiglaucoma medications at the end of POM1 and the rest of eyes were prescribed medications by the end of POM2 or later. The number of fellow eyes with clinically significant IOP elevation at POD1, POW1, POM1, POM3, and POM6 was 47 (35.8%), 76 (58%), 47 (35.8%), 31(23.6%), and 37 (28.2%) and mean IOP elevation in this group was 4.5 ± 2.3 (38.5%), 5.4 ± 1.8 (48.5%), 4.8 ± 1.6 (44.2%), 3.7 ± 1.3 (36.5%), and 3.2 ± 1.3 (32.6%) mm Hg, respectively. No correlation was present between the magnitude of IOP elevation and sex, age, and number or class of preoperative topical glaucoma medications in either fellow eyes or operated eyes.
Analysis of IOP changes in fellow eyes based on glaucoma diagnoses revealed that the greatest IOP increase occurred in patients with GFCS (3.6 ± 4.9 mm Hg) while the lowest increase was observed in POAG (1.4 ± 1.2 mm Hg) (Figure, Supplemental Digital Content 1, which demonstrates IOP changes in different glaucoma subtypes, http://links.lww. com/IJG/A583) ( Table 3). None of the patients with GFCS were on preoperative acetazolamide. IOP changes in fellow eyes based on the type of surgery on index eyes showed AGV implantation was associated with greater IOP elevation in fellow eyes over the entire follow-up period (P < 0.008) (Fig. 2). IOP changes correlated with baseline status of fellow eyes (normal, on glaucoma medications or with previous glaucoma surgery) are shown in Figure 3 and Table 4.
Although not statistically significant, fellow eyes with previous glaucoma surgery had the greatest amount of IOP elevation (Fig. 3). IOP levels of fellow eyes at POW1 and POM1 were compared with that of operated eye on POD1. When IOP of operated eye was ≤ 5 on POD1 , which happened in 23 operated eyes (including 17 eyes with AGV implantation, 3 eyes undergoing trabeculectomy, and 3 eyes in the phacotrabeculectomy subgroup), the fellow eye developed greater IOP elevation at POW1 (18.4 ± 3.9 vs. 15.8 ± 3.5, P = 0.002) and at POM1 (16.4 ± 4.3 vs. 14.5 ± 3.6, P = 0.024).
To evaluate the effect of discontinuation of acetazolamide after surgery, patients who were on oral acetazolamide in addition to topical medications preoperatively (54 patients, 41%), were compared with those who were not. Both groups showed peak IOP elevation at POW1, patients on acetazolamide developed greater IOP elevation in the been categorized as normal at baseline. Seven other fellow eyes (5.3%), which had previously controlled IOP on MTMT, developed elevated IOP necessitating glaucoma surgery. Table 5 shows details of the eyes that required surgery. These eyes were excluded from the analyses after surgery of the fellow eyes. There were 6 patients with GFCS secondary to bilateral congenital cataract surgery who were younger than the rest of the patients (age range: 4 to 11 y). All fellow eyes in  this group were on topical antiglaucoma medications. Three fellow eyes underwent tube shunt implantation because of uncontrolled IOP (Table 5).
Overall, reliable preoperative HVF tests were available for 94 operated eyes and 120 fellow eyes . On the basis of VF mean deviation (MD) all operated eyes had advanced glaucoma as well as fellow eyes with prior glaucoma surgery. Medically controlled fellow glaucomatous eyes had mild to moderate glaucoma ( Table 2). There was a significant positive correlation between MD of fellow eyes and IOP changes at POM1 (P = 0.02, r = 0.210). However, subgroup analysis revealed that only normal fellow eyes demonstrated a significant correlation between MD and IOP changes at POM1 (P = 0.045, r = 0.297).

DISCUSSION
Our study showed that following glaucoma surgery and IOP reduction in 1 eye, IOP of the fellow eye increases. The greatest IOP elevation in fellow eyes was observed in eyes with history of previous glaucoma surgery, a diagnosis of GFCS, AGV implantation in the index eye, early postoperative hypotony in the index eye, and preoperative use of systemic acetazolamide. To the best of our knowledge, this is the first report showing greater fellow eye IOP elevation following tube shunts as compared with trabeculectomy or phacotrabeculectomy and that early hypotony is associated with greater IOP elevation of the fellow eye. The greatest amount of IOP elevation was observed at POW1 and gradually weared off over 6 months after surgery but not back to baseline values.
Alterations in the IOP of 1 eye after IOP change of the contralateral eye has been widely described after ocular compression or contusion, procedures such as paracentesis, tonography, filtering surgery, laser trabeculoplasty, and the use of topical antiglaucoma medications which is referred to as the COR. 2,7-12  While in the case of topical antiglaucoma medications, IOP drop in the fellow eye is at least partially attributable to systemic absorption of the drug, 5,13-15 the effect of IOP lowering procedures of the index eye on the IOP of the fellow eye is still debatable. Studies have shown IOP increase, 3,16,17 decrease, 7,18-20 or no change 10,21 in the fellow eye following glaucoma surgery or SLT on the first eye. Wilmer first noticed IOP reduction of one eye following fistulization surgery on the other eye in an animal study. 22 After treatment with SLT, vasoactive and chemotactic agents, including interleukins (interleukin-1a, interleukin-1b), and tumor necrosis factor-a, are released which could induce repopulation of the angle with normal functioning trabecular meshwork cells, resulting in improved trabecular outflow and IOP reduction. Following their systemic dispersion, similar albeit smaller effects may occur in the untreated fellow eye causing IOP reduction. 7 A number of publications have reported different results from ours. In a study by Vysniauskiene et al, 20 on 24 glaucomatous eyes with POAG, normal tension glaucoma and psuedoexfoliative glaucoma undergoing trabeculectomy, the authors reported IOP reduction in the fellow eye 1 day and 1 month after surgery. Eleven fellow eyes (45%) in their study were on topical glaucoma medications and the remaining 13 eyes were medication free, of which 12 had previously undergone trabeculectomy. In another study on 61 patients with bilateral POAG, Detorakis et al 10 reported no significant change in the IOP of fellow eyes after trabeculectomy in the index eye during the 3 months follow-up.
Some studies in the literature have described findings which are in line with ours. A recent report of 3 infants with PCG showed that after lowering IOP in 1 eye (1 with medication and 2 with surgery), acute corneal hydrops developed in the other eye because of acute IOP rise. 23 Kaushik et al, 3 evaluated fellow eyes of 71 patients who underwent glaucoma surgery including trabeculectomy or glaucoma drainage device implantation in 1 eye. All fellow eyes had IOP elevation during the entire follow-up period; peak IOP occurred on POW 6 (4.7 ± 3.3 mm Hg) and remained elevated until POM 6 (3.3 ± 3.6 mm Hg) which is in agreement with our study, however, maximum IOP elevation in our patients was observed at POW1 (3.1 ± 3.2 mm Hg) and decreased gradually to POM6 (0.9 ± 2.1 mm Hg) (Fig. 1).
Kaushik et al, 3 analyzed patients on systemic acetazolamide separately and found that mean IOP change was not significantly different from patients who did not consume the medication preoperatively. In our study, these 2 subgroups also followed a similar pattern of IOP change, but there was significantly greater IOP elevation in patients taking systemic acetazolamide preoperatively at all-time points (Fig. 4). Upon discontinuation of preoperative acetazolamide, rebound IOP elevation in fellow eyes might happen because of withdrawal of this potent agent and play an additional role.
A retrospective study on 235 fellow eyes naive to any previous glaucoma surgeries, showed different intereye IOP relationship following trabeculectomy, PreserFlo microshunt implantation or filtering canaloplasty in the operated eye. The authors found a significant IOP decline in fellow eyes, that were previously on antiglaucoma medication, following trabeculectomy but this effect was not observed in fellow eyes not on topical therapy and it could be because of enhanced compliance for appropriate drop application. Most fellow eyes in the 2 other groups did not show significant IOP change, however, IOP elevation occurred in 12.7% of all fellow eyes and it happened mostly on POD1, necessitating starting antiglaucoma medication in 2 eyes. 24 There is a complex intereye IOP relationship in different surgical procedures and the same is true upon using antiglaucoma medications, as 1 cannot expect the same IOP lowering effect by using an antiglaucoma agent in both eyes of 1 patient. 25 The mechanism for the COR is not well understood. A compensatory central neural mechanism has been suggested, 2 but it is not known how circuit may work. Sympathetic, parasympathetic and trigeminal nerve pathways have been proposed to constitute either the afferent or efferent arms of this reflex. 16,22,26 In a study on 14 patients with POAG, psuedoexfoliative glaucoma or pigment dispersion glaucoma who underwent trabeculectomy in 1 eye, aqueous production measured by flurophotometry in the fellow eye increased on POD5. This increased aqueous production was attributed to an ocular-central nervous system reflex which was set up to counterbalance the induced hypotony in the operated eye and since the reflex affects both eyes, the fellow eye demonstrates an IOP increase. 27 In our series, patients with GFCS had the greatest amount of IOP increment in the fellow eyes which might be attributed to the younger age of these patients. It is known that aqueous production decreases with aging, therefore younger patients with greater reserve for aqueous production could probably demonstrate more significant IOP rise caused by the COR. 28 In our study, patients who experienced hypotony in the operated eye on POD1 (IOP ≤ 5), had significantly higher fellow eye IOP at POW1 and POM1, indicating a stronger COR and the need to closely monitor IOP in fellow eyes of such patients. Most of eyes with early postoperation hypothony had AGV implantation (17 of 23 eyes), which could explain the greater IOP elevation we observed in the fellow eyes of this subgroup. It may therefore be prudent to avoid hypotony in subjects with advanced glaucoma to prevent or blunt IOP elevation in the fellow eye.
Yarangumeli and colleagues, proposed that the COR might take occur because of increased outflow resistance. Overall, 107 fellow eyes included in their study were normal or had medically controlled glaucoma. They postulated that performing trabeculectomy may have a detrimental effect on the trabecular meshwork because of aqueous hypoperfusion and causes trabecular meshwork densification in the operated eye , a reflex mechanism might also happen in the fellow eye resulting in IOP elevation. 16 However, this theory cannot be supported by our study because greater IOP elevation was observed in fellow eyes with previous glaucoma surgery as compared with normal or medically controlled eyes.
The COR is possibly independent of the aqueous outflow resistance and is thought to act mainly through increased aqueous production rather than altered outflow facility. 29 This central mechanism may be controlled by the supraoptic nucleus and mediated through the optic nerve as suggested by an animal study on rabbits with unilateral optic nerve section who showed a blunted IOP response after water drinking. 29 In addition, Gloster and colleagues found that the diencephalon has a significant contributory role in IOP control mechanisms. 30,31 One interesting observation in our study was that glaucoma medications were added to 18 fellow eyes, of which 10 eyes had been considered as normal. This is in contrast to the assumption that a normal trabeculum can compensate increased aqueous secretion, implying a more complex IOP control system. While discontinuation of acetazolamide and topical medications in the first eyes can explain part of the IOP elevation observed in the fellow eyes, considering the wash out period of glaucoma medications, IOP elevation on POD1 could not be fully explained by the withdrawal of their effects. Amongst glaucoma medications, timolol and brimonidine may slightly decrease IOP in the fellow eye after treatment of the first eye. However, in our study no surgery eye received any glaucoma medication during the first POM, the period during which the fellow eyes showed the greatest IOP elevation.
Limitations of our study limited sample size for some subtypes of glaucoma. However, it has the "real world" advantage of studying the effect of different types of glaucoma surgeries, investigating fellow eyes both with and without prior glaucoma surgery and including different subtypes of glaucoma.
While the precise mechanism of COR is still unknown and complex, the clinical implication of this study is to check the IOP in the contralateral eye at all postoperative visits. Fellow eyes with GFCS and patients on systemic antiglaucoma medications are at greater risk for significant IOP elevation. AGV implantation and postoperative hypotony of the first eye are also associated with greater IOP elevation in the fellow eye.