Immediate Postoperative Use of a Topical Agent to Prevent IntraocularPressure Elevation After Pars Plana Vitrectomy With Gas Tamponade

Objective  To determine whether a single topical aqueous suppressant applied immediatelyafter pars plana vitrectomy with long-acting gas tamponade prevents intraocularpressure (IOP) elevation.

Methods  Fifty patients who met the inclusion criteria and underwent pars planavitrectomy with long-acting gas tamponade were randomized to receive a combinationof timolol maleate and dorzolamide hydrochloride, long-acting timolol alone,dorzolamide alone, or placebo at the conclusion of surgery. The IOP was checkedby a portable, handheld tonometer (Tono-Pen) at the conclusion of surgeryand at 5 hours, 1 day, and 1 week after surgery.

Results  There were no significant differences in IOP among the groups at theconclusion of surgery. The IOP at 5 hours after surgery (27.0 vs 17.4 mm Hg; P<.001) and 1 day after surgery (26.1 vs 19.9 mm Hg; P = .01) showed a statistically significant differencebetween the placebo and timolol-dorzolamide groups. The timolol-dorzolamidegroup showed greater IOP control than either the timolol alone or the dorzolamidealone groups at 5 hours (P = .04 for both).

Conclusion  The use of a single topical aqueous suppressant (timolol-dorzolamide)given after pars plana vitrectomy with long-acting gas tamponade effectivelyprevents significant postoperative IOP elevation at 5 hours and 1 day aftersurgery.

Pars plana vitrectomy is associated with a significant risk of postoperativeelevation in intraocular pressure (IOP). Reasons for IOP elevation includesignificant intraocular inflammation, postoperative hyphema, and forward rotationof the ciliary body with anterior displacement of the lens-iris diaphragmsecondary to intraocular tamponade with either silicone oil or intraoculargases.¹ Fluorinated intraocular gases are usedin many types of vitreoretinal surgery, most typically for postoperative tamponadeof retinal breaks. Excessive expansion of an intraocular gas bubble can leadto postoperative IOP elevation, glaucomatous damage to the optic nerve, andeven central retinal artery occlusion.^2-4

Many vitreoretinal surgeons check the IOP postoperatively, and someroutinely admit patients to the hospital for IOP monitoring. Other surgeonsgive aqueous suppressants (topical medications or oral or intravenous acetazolamidesodium) at the conclusion of surgery to those patients they believe will bemost at risk of developing a postoperative elevation in IOP.

A previous study⁵ showed that the administrationof multiple topical aqueous suppressants at the conclusion of pars plana vitrectomywith long-acting gas tamponade was effective in preventing significant IOPelevation in most patients. A separate study⁶ showedthat intravenous acetazolamide was ineffective in reducing the risk of significantpostoperative IOP elevation.

We designed and completed a prospective, randomized, controlled clinicaltrial to determine whether a single topical aqueous suppressant given immediatelyafter surgery would prevent significant IOP elevation and to determine whetherprophylactic use of such an agent would control postoperative IOP. We electedto evaluate the combination of 0.5% timolol maleate and 2% dorzolamide hydrochlorideas the topical aqueous suppressant because of its proved efficacy, its rapidonset of action, and the additive effect of the combination medication comparedwith the concomitant administration of timolol and dorzolamide.^7,8 Inaddition, we evaluated long-acting 0.5% timolol and 2% dorzolamide individuallyto determine whether either of these agents alone would produce a significanteffect on postoperative IOP.

The Institutional Review Board of the University of Miami School ofMedicine approved this project. A prospective, randomized, controlled clinicaltrial was performed at the Bascom Palmer Eye Institute on the patients of3 vitreoretinal surgeons (M.S.B., T.G.M., and C.W.G.E.); these patients underwentpars plana vitrectomy with long-acting gas tamponade, met inclusion and exclusioncriteria, and consented to enroll in the study. The patients were 18 yearsor older and had an IOP on preoperative examination higher than 5 mm Hg andlower than 22 mm Hg. Exclusion criteria included a current or prior diagnosisof glaucoma, current use of a glaucoma medication, chronic or recurrent uveitis,a history of corticosteroid response, the presence or planned placement ofan anterior chamber intraocular lens, a history of hypersensitivity to anyof the study medications, obstructive airway disease, second- or third-degreeheart block, or current use of tricyclic antidepressants. Patients were randomizedpreoperatively with a randomly permuted block scheme in which the block sizevaried between 1 and 4, with an equal number of patients randomized to eachof the 4 treatment groups at the end of each block. After patient eligibilitywas determined, the examining physician called the Biostatistics Center, whichassigned the patient to 1 of 4 treatment groups: (1) 0.5% timolol–2%dorzolamide (Cosopt; Merck & Co, Inc, Whitehouse Station, NJ), (2) 0.5%timolol gel-forming solution (Timoptic XE; Merck & Co, Inc), (3) 2% dorzolamide(Trusopt; Merck & Co, Inc), or (4) a placebo drop (artificial tears) (acombination of dextran 70 and hydroxypropyl methylcellulose [Tears NaturaleII]; Alcon Laboratories, Inc, Fort Worth, Tex). The baseline characteristicsof the treatment groups are given in Table1.

The patients underwent surgery for 1 of 3 diagnoses: rhegmatogenousretinal detachment with or without proliferative vitreoretinopathy, macularhole, or tractional retinal detachment. A standard 3-port pars plana vitrectomywas performed in all patients. On completion of an air-fluid exchange andclosure of the superior sclerotomies, a mixture of long-acting gas was injectedthrough the posterior infusion cannula while venting through a 27-gauge needleplaced through the pars plana. A gas mixture, 50 mL, was instilled in thismanner, ensuring a complete exchange of gas for air. Either perfluoropropane(C₃F₈) or sulfahexafluoride (SF₆) was used,in minimally expansile or nonexpansile concentrations (C₃F₈, 14%-18%; or SF₆, 16%-20%). The pure gases were dilutedwith filtered air to reach the anticipated concentrations. If the patientsunderwent general anesthesia, no nitrous oxide was administered.

After closure of the final sclerotomy, the IOP was checked with a portable,handheld tonometer (Tono-Pen; Mentor Ophthalmics, Norwell, Mass) with a steriletip. If the IOP was lower than 10 or higher than 21 mm Hg, a gas mixture waseither added or removed from the eye to bring the IOP between 10 and 21 mmHg. After the IOP was confirmed by Tono-Pen to be between 10 and 21 mm Hg,the conjunctiva was closed. The eyes then received a subconjunctival injectionwith a combination antibiotic-corticosteroid, 0.25% scopolamine hydrobromidetopically, and a combination of tobramycin sulfate and dexamethasone topically(Tobradex; Alcon Laboratories, Inc). According to randomization into 1 ofthe 4 groups, the patient then received 0.5% timolol–2% dorzolamide,0.5% timolol gel-forming solution, 2% dorzolamide, or artificial tears beforeplacement of a sterile patch and shield over the eye.

The IOP was checked by Tono-Pen in a masked fashion at 5 hours, 1 day,and 1 week after surgery by trained personnel. If the IOP was elevated atany measurement point, a standard treatment protocol was followed. We consideredan IOP by Tono-Pen higher than 25 mm Hg to be elevated. For an IOP of 26 to30 mm Hg, the patients received 0.2% brimonidine tartrate (Alphagan; Allergan,Inc, Irvine, Calif). For an IOP of 31 to 40 mm Hg, the patients received 2of the following 3 medications: 0.5% timolol, 2% dorzolamide, or 0.2% brimonidine.For an IOP higher than 40 mm Hg, the patient underwent anterior chamber paracentesis.

Patient age and sex, diagnosis, lens status, and prior surgical proceduresperformed were recorded. The type of surgery performed and the occurrenceof significant intraoperative complications were also documented. Outcomedata collected included postoperative IOP measurements immediately after thesurgery and 5 hours, 1 day, and 1 week after surgery.

Continuous baseline characteristics, such as age, were compared using t tests and an analysis of variance to test for differencesamong the treatment groups. Categorical variables, such as sex, were comparedusing χ² tests. Postoperative IOP measurements and change inIOP at 5 hours, 1 day, and 1 week after surgery were compared among treatmentgroups using an analysis of variance and t tests.Paired t tests were used to compare 5-hour, 1-day,and 1-week IOPs with immediate postoperative IOPs within treatment groups.

There were 52 eyes that met the inclusion and exclusion criteria andconsented to participate in the study between September 1, 2001, and December31, 2002. Two patients did not complete the study because of intraoperativereasons (use of air or silicone oil instead of long-acting gas). The 50 remainingpatients were divided into 4 groups: (1) group 1 (n = 12), timolol-dorzolamide;(2) group 2 (n = 13), timolol alone; (3) group 3 (n = 13), dorzolamide alone;and (4) group 4 (n = 12), placebo. The mean age of the 50 participating patientswas 57.6 years (range, 25-83 years). The diagnoses for which surgery was performedon the 50 patients were as follows: rhegmatogenous retinal detachment (withor without proliferative vitreoretinopathy), 40 (80%); macular hole, 5 (10%);and tractional retinal detachment, 5 (10%). There were no significant intraoperativecomplications, such as excessive hemorrhaging, in any of the study patients.According to the study protocol, patients with an IOP higher than 25 mm Hgat the 5-hour point received topical brimonidine: in group 1, 1 of 11 patientsreceived brimonidine at the 5-hour point; in group 2, 2 of 13 patients receivedbrimonidine; in group 3, 2 of 13 patients received brimonidine; and in group4, 9 of 12 patients received brimonidine.

Table 1 provides baselinecharacteristics, such as age, sex, study eye, lens status, and history ofprior surgery, for each of the 4 treatment groups. It also includes diagnosisand preoperative IOPs taken in the clinic at enrollment in the trial. Thetreatment groups were similar, with no statistically significant differences.

Table 2 provides the postoperativeIOP course, with P values comparing IOP outcomesamong the 4 treatment groups.

With timolol alone, dorzolamide alone, and placebo, there was a statisticallysignificant IOP increase at the 5-hour postoperative point (P = .01, P = .001, and P<.001 by paired t test, respectively).This IOP increase persisted at 1 day for these treatment groups (P = .01, P<.001, and P<.001, respectively). There was no statistically significant (P = .91) IOP increase with timolol-dorzolamide. It wasalso true that patients treated with timolol or dorzolamide alone had lessof an IOP increase at the 5-hour point than patients treated with placebo(P = .02 and P = .009 by t test, respectively).

Figure 1 graphically showsthe mean IOP in the 4 treatment groups at the immediate postoperative, 5-hour,1-day, and 1-week points.

Table 3 provides the numberand percentage of patients in each treatment group with an IOP higher than25 mm Hg at 5 hours and 1 day postoperatively. When comparing timolol-dorzolamidewith placebo, there was a statistically significant difference in the likelihoodof an IOP higher than 25 mm Hg at the 5-hour point and a trend toward significanceat the 1-day point.

Numerous researchers^2,3,9-11 havereported previously on the incidence of postoperative IOP elevation afterpars plana vitrectomy with long-acting gas tamponade. When used in vitreoretinalsurgery, these long-acting gases are typically diluted to a minimally or nonexpansileconcentration, usually 18% to 20% for SF₆ and 14% to 18% for C₃F₈. However, there is clear evidence that the IOP can besignificantly elevated after surgery. The risk of IOP elevation after injectionof SF₆ has been reported to range from 6.1% to 67%,^3,9 whilethe risk of IOP elevation after injection of C₃F₈ hasbeen reported to range from 18% to 59%.^10,11 Whilethe maximum size of the long-acting gas bubble is not reached until 24 to48 hours for SF₆¹² and 72 to 96hours for C₃F₈,^13,14 themost rapid rate of volume increase occurs within 5 to 8 hours.¹⁵ Forthis reason, many vitreoretinal surgeons monitor IOP at approximately 5 hoursafter surgery, as we do on our service. Treatment of elevated postvitrectomyIOP is usually accomplished medically via aqueous suppressants, although invasiveprocedures such as anterior chamber paracentesis may occasionally be necessary.

Prophylactic treatment of elevated postvitrectomy IOP has been proposedand studied before. The use of intravenous acetazolamide during surgery hasbeen shown to be ineffective,⁶ while the combineduse of multiple topical aqueous suppressants has been shown to be effectivein the prevention of postvitrectomy IOP elevation.⁵ Inthe present study, we investigated whether a single topical aqueous suppressantwould be sufficient to prevent significant postvitrectomy IOP elevation. Webelieved that the use of a single agent would be more practical in the prophylactictreatment of IOP elevation in this setting.

We chose to evaluate the combination of timolol-dorzolamide becauseit was the single topical aqueous suppressant with the greatest IOP-loweringeffect during the critical 5- to 8-hour postoperative period.⁷ Inaddition, we investigated the effectiveness of timolol and dorzolamide usedseparately. Combination timolol-dorzolamide has been shown previously to bemore effective than monotherapy with either timolol or dorzolamide.¹⁶ Although timolol and dorzolamide alone showed a significanttreatment effect at the 5-hour point when compared with placebo, this treatmenteffect was significantly less than that of the combination of timolol-dorzolamideat the 5-hour point. In patients with no medical contraindications, a combinationtopical medication, such as timolol-dorzolamide, is preferred for decreasingthe risk of an elevated IOP after pars plana vitrectomy with use of a long-actinggas tamponade.

We used the Tono-Pen to measure IOP because it has several advantagesin the current setting. The Tono-Pen is portable and requires no other instrumentation,facilitating its use in the operating room, the recovery room, and the clinic.The latex tip of the Tono-Pen can be sterilized for use during surgery, andthe Tono-Pen is accurate even with an irregular corneal epithelium. Althoughthe Tono-Pen has been shown to be similarly accurate to Goldmann applanationtonometry for eyes with an IOP within the physiologic range, some reports^17-19 have indicated thatit overestimates low pressures and underestimates high pressures.

In gas-filled eyes, the compressible nature of the gas leads to underestimationof the IOP by the typical indentation methods of tonometry. In 1990, Lim etal¹⁸ found that the Tono-Pen significantlyunderestimates IOP at pressures of 30 mm Hg or higher, averaging 12.1 mm Hgbelow the true IOP as measured by manometry. In 1993, Badrinath et al¹⁹ expanded on these findings, compared IOP measurementswith the Tono-Pen with manometric pressure, and reported a calibration curvefor use of the Tono-Pen to measure IOP in gas-filled eyes that had undergonevitrectomy. Patients with a Tono-Pen pressure of 25 mm Hg had an IOP by manometryof 30.77 mm Hg, and those with a Tono-Pen pressure of 30 mm Hg had an IOPby tonometry of 36.73 mm Hg. We believe the Tono-Pen is an easy and relativelyaccurate means of measuring IOP in patients after vitrectomy and that it wasthe best way to measure IOP in this clinical trial. We considered a Tono-PenIOP reading of 25 mm Hg or higher significant, because it corresponds to atrue IOP of greater than 30 mm Hg.^18,19 Inaddition, Mittra et al⁵ previously reportedsimilar results using multiple topical aqueous suppressants, and they similarlyconsidered a Tono-Pen measurement of an IOP higher than 25 mm Hg as a significantelevation.

In their study, Mittra et al⁵ showedthat the use of multiple topical aqueous suppressants after pars plana vitrectomywith long-acting gas tamponade was effective in reducing the incidence ofsignificant IOP elevation after surgery, with a significant increase in IOPdefined as 25 mm Hg or higher. They also showed that the treated eyes didnot on average have a significant increase in IOP during the critical 6-hourpostoperative period while control eyes had a significant increase in IOP(an average 8–mm Hg increase) during the early postoperative period.In the present study, we demonstrated similar results in a prospective, randomized,controlled clinical trial with the use of a single topical aqueous suppressant(timolol-dorzolamide). We showed that eyes treated with timolol-dorzolamidehad no difference between the IOP measured immediately postoperatively andat the 5-hour postoperative point. However, control eyes that received placeboat the end of surgery had a mean 12–mm Hg increase in IOP at the 5-hourpostoperative point and the IOP was still elevated at 1 day (P<.001 for both).

This prospective, randomized, controlled clinical trial shows that 0.5%timolol–2% dorzolamide given immediately after pars plana vitrectomywith long-acting gas tamponade can reduce the likelihood of IOP elevationhigher than 25 mm Hg (9% in group 1 vs 75% in group 4; P = .006) at the 5-hour postoperative point. This decreases the riskof visual loss secondary to an elevated IOP, decreases the likelihood of patientdiscomfort, and reduces the need for further therapy, including invasive proceduressuch as anterior chamber paracentesis. The use of a single topical agent ratherthan multiple agents simplifies the treatment process, eliminating the needfor multiple drops and the potential for washout of drops placed one afterthe other.

With the trend toward outpatient vitreoretinal surgery, there is a desireto avoid altogether postoperative monitoring of IOP during the first nightafter surgery. Our results show that 0.5% timolol–2% dorzolamide appliedimmediately after pars plana vitrectomy with long-acting gas tamponade helpsto avoid IOP elevation in many patients. Vitreoretinal surgeons can be confidentthat most of their postoperative patients can avoid the dangers associatedwith an elevated IOP during the early postoperative period. Surgeons may stillwant to check the IOP at the 5-hour point, especially in patients with a historyof an elevated IOP after surgery or a history of glaucoma or when using aconcentration of gas with more expansile qualities than used in the presentstudy.

Corresponding author and reprints: Matthew S. Benz, MD, Cullen EyeInstitute, Baylor College of Medicine, 6565 Fannin, Mail Stop NC-205, Houston,TX 77030 (e-mail: mbenz@bcm.tmc.edu).

Submitted for publication June 4, 2003; final revision received October8, 2003; accepted October 29, 2003.

This study was supported, in part, by an unrestricted research grantfrom Research to Prevent Blindness, New York, NY.

We thank Merck & Co, Inc, for providing the medications used inthis study.