Ten-Year Clinical Outcomes of Acute Primary Angle Closure Randomized to Receive Early Phacoemulsification Versus Laser Peripheral Iridotomy

Supplemental Digital Content is available in the text. Purpose: To compare the 10-year clinical outcomes of eyes with acute primary angle closure (APAC) randomized to receive either early phacoemulsification or laser peripheral iridotomy (LPI). Methods: Sixty-two APAC patients, who underwent either early phacoemulsification (phaco group) or laser peripheral iridotomy (LPI group) in a previous randomized controlled trial, were invited for assessment 10 years after the interventions. The results of the 2 groups were compared. Results: Forty of 62 patients (64.5%; 19 in phaco group and 21 from LPI group) were examined. None of them underwent additional glaucoma procedure but 15 (71.4%) patients in the LPI group received lens extraction before this assessment. The mean follow-up duration was 10.7±0.7 years. The phaco group used less medication (0.16±0.37 vs. 0.76±1.09 bottle per eye, P=0.028), had less extensive anterior synechiae (120.0±116.12 vs. 244.3±139.8 degree, P=0.010), and greater mean Shaffer gonioscopy grading (1.79±0.84 vs. 1.40±0.87; P=0.021) than the LPI group. Five eyes had persistent intraocular pressure elevation of >21 mm Hg in 2 consecutive visits and 4 eyes had blindness (best-corrected visual acuity worse than 6/60 and/or central visual field of <20 degree) in the LPI group, compared with none in the phaco group (P=0.022 and 0.045, respectively). There was no significant difference in the mean intraocular pressure, best-corrected visual acuity, and the number of eyes with visual field progression. Conclusion: At 10 years, APAC eyes that underwent early phacoemulsification required less medication, less peripheral anterior synechiae, lower incidence of intraocular pressure elevation and a lower incidence of blindness compared with APAC eyes that underwent initial LPI.

A cute primary angle closure (APAC) is characterized by an acute, symptomatic elevation of intraocular pressure (IOP) caused by an abrupt closure of angle resulting in cessation of flow through the trabecular meshwork in the anterior chamber angle, mainly as a result of an exaggerated pupillary block that disturbs the normal posterior-to-anterior chamber aqueous flow. 1 This causes headache, blurring, seeing halos, eye pain, redness, mid-dilated pupil, and corneal edema. 2 The crude incidence rates of APAC were 12.2 and 10.4 per 100,000 people per year in the above 30-year-old Singapore and Hong Kong population, respectively. 3,4 That is higher than the average incidence rate of 3.9 to 4.1 cases per 100,000 people per year in the European regions. [5][6][7] Previous randomized controlled trials (RCTs) demonstrated that, soon after APAC is aborted by medications, primary phacoemulsification with intraocular lens (IOL) implantation was a more effective treatment option than laser peripheral iridotomy (LPI) in terms of preventing IOP elevation, controlling IOP at < 21 mm Hg with fewer glaucoma medications, and achieving a larger degree of open angle within the first 2 years. 2,8 However, the outcomes of the patients who were involved in our previous RCT 2 beyond the 18 months have not been reported. The current study aimed to compare the 10-year clinical outcomes of this group of patients who had been randomized to receive early phacoemulsification or LPI shortly after medically aborted APAC. 2 Furthermore, the outcomes of the contralateral eyes, that received prophylactic LPI soon after the treatment of the APAC eyes, were also documented.

METHODS
The authors obtained prior approval of the study protocol by the ethics committee of the Chinese University of Hong Kong. The research protocol adheres to the tenets of the Declaration of Helsinki. Informed consents were obtained from all patients before recruitment.
The design and methods of the RCT that was conducted at Hong Kong Eye Hospital were described previously. 2 The recruitment criteria and outcome measures are listed in Table 1. In brief, 62 eyes of 62 Chinese subjects who had medically aborted APAC in 1 eye were randomized to receive either early phacoemulsification (phaco group) or LPI (LPI group) between February 2004 and August 2005. There were 31 patients in each group. Phacoemulsification and LPI were performed under local or topical anesthesia, followed by different dosages of prednisolone acetate 1% and ofloxacin 0.3% depending on clinical needs. After the treatments, patients were followed up at day 1; week 1; months 1, 3, 6, 9, 12 and then every 6 months for 18 months. The target IOP was ≤ 21 mm Hg and IOP lowering eye drops were added in the following sequence if IOP was DOI: 10.1097/IJG.0000000000001799 > 21 mm Hg: (1) beta-blockers, (2) prostaglandin analogues, (3) carbonic anhydrase inhibitors, (4) pilocarpine and, (5) adrenergic agonists (combined formulae were not in the scheme). Bestcorrected visual acuity (BCVA) examination, IOP measurement, dark-room gonioscopy, optic disc assessment, visual field (VF) examination, and specular microscopy were performed shortly after successful abortion of attack and in the subsequent followup visits following either phacoemulsification or LPI. The procedure-related complications in the first 18th month were mentioned in the previous publication. 2 Intraoperative complications for the phaco group included 12 eyes with corneal edema, 1 eye with posterior capsular rupture, and 1 eye with iris bleeding. Postoperatively, 7 eyes had fibrinous anterior chamber reaction, and 5 eyes had posterior capsular opacification. For the LPI groups, 4 eyes required supplementary laser because of closed iridotomy (1 eye) and small iridotomies (3 eyes). All complications were manageable and there were no severe complications. 2 The study ended at 18-month of follow-up and patients were then referred to either general ophthalmology or glaucoma clinic for standard ophthalmological management by the on-duty ophthalmologist. Additional interventions were determined by glaucoma specialists according to the hospital policy-addition of IOP-lowering medications, laser or surgical procedures (including phacoemulsification) were offered if IOP was persistently > 21 mm Hg or in case of detectable glaucomatous progression. Phacoemulsification was also performed if BCVA was ≤ 0.3 or visual disturbance affecting the patient's daily activity unless the patient refused to undergo the operation.

Ophthalmic Examination
In this retrospective study, 62 patients were invited for a comprehensive ophthalmic assessment. The full ophthalmic assessment of both the affected APAC eyes and fellow unaffected eyes were carried out by a single glaucoma specialist (P.P.C.), including BCVA, IOP measurement (Goldmann applanation tonometry), dark room gonioscopy, anterior segment slit-lamp biomicroscopy, and examination of the fundus and optic disc with dilation of the pupil. Two IOP measurements were recorded for each eye. If the 2 readings differed by ≤ 2 mm Hg, the mean was recorded as the IOP measurement, otherwise, a third reading was performed, and the mean value was recorded. 9 Dark-room gonioscopy was performed according to the standard protocol 10 and the 4 quadrants were graded according to the modified Shaffer grading (grade 0 to 4). The mean gonioscopy angle width was equivalent to the average Shaffer grading of all 4 quadrants. Indentation gonioscopy was performed using a Posner lens at the slit-lamp under topical anesthesia to look for peripheral anterior synechiae (PAS) and the extent of PAS is documented in the number of clock hours. The optic disc was examined through a dilated pupil, using a 90-D lens under the slit-lamp. The vertical cup-disc ratio (VCDR) was taken to be the longest vertical cup diameter divided by the longest vertical disc diameter measurement, using a magnification factor of 1.3. 11 Features consistent with glaucomatous optic neuropathy were noted.
Central corneal thickness (CCT), specular microscopy, optical coherence tomography (OCT) and VF examination were performed by an experienced examiner (F.Y.T.). CCT was determined using an A-scan ultrasonic pachymeter (Pachette 500, DGH Technology Inc, Frazer, PA) for all subjects with a standard protocol, the median of 5 readings was taken as the CCT. 12 Retinal nerve fiber layer (RNFL) thickness and other optic nerve head parameters were measured by OCT (SPECTRALIS, Heidelberg Engineering, Germany) through a maximally dilated pupil. Only OCT images with signal strength ≥ 6 were accepted. VF examinations were performed using the standard automated white-on-white threshold perimetry (SITA Standard 24-2, Humphrey Field Analyzer, Carl Zeiss Meditec, Dublin, CA); only VFs with fixation losses, false-positive and falsenegative errors ≤ 20% were accepted. The examination was repeated twice within 1 month and the VF indices [mean deviation (MD), pattern standard deviation (PSD), and visual field index] were recorded. Glaucoma VF defect is defined as per published criteria for glaucoma: glaucoma hemifield test outside normal limits, abnormal PSD with P < 5% occurring in the normal population and/or a cluster of ≥ 3 points in the PSD plot in a single hemifield with P < 5%, one of which must have a P < 1%; any one of these  13 VF progression was defined as 2 or more points within or adjacent to an existing scotoma worsened by at least 10 dB or 3 times the average of the short-term fluctuation, whichever was larger, and confirmed on 2 subsequent fields. 13

Assessment of Visual Impairment
Blindness of a single eye was defined as BCVA that was worse than 0.1 and/or central VF of < 20 degree. 14 The major cause of blindness was determined by clinical examination and reviewing of the charts. For the assessment of visual impairment, the VF loss in both eyes for each patient were considered, based on the suggestion that a damaged VF in 1 eye will not necessarily impact on the overall visual function of a patient if the other eye remains healthy. 15 "Statutory blindness" is defined as an MD equal or worse than −22 dB in the better eye, as recommended by The United States Social Security Administration (SSA). 16 "Visual impartment" is defined as a patient with an MD that is equal or worse than −14 dB in the eye with the least-damaged VF, based on the suggestion that such patient is highly unlikely to satisfy the vision component of legal fitness to drive in the United Kingdom. 17 The scheme is adapted from a previous study that aimed to estimate the proportion of 3790 glaucoma patients in the United Kingdom that could progress to serious visual impartment in their expected lifetime. 18

Statistical Analysis
Analyses were performed using SPSS for Windows version 24.0 (SPSS Inc., Chicago IL). Continuous variables were expressed as ( ± SD) and categorical variables as individual counts and proportions. Data of the 2 treatment groups at 10 years were compared. Both eyes of each patient were assessed but only the affected eyes were analyzed. Independent t test was used to compare the statistically significant difference in the continuous outcome parameters between the 2 groups, whereas χ 2 test was used to compare categorical variables. A P-value of < 0.05 was taken as statistically significant.

RESULTS
The baseline demographics of the original study (62 eyes of 62 patients) 2 are summarized in Table 2. Twenty-two of the 62 patients did not join this final assessment: 5 patients (8.1%; 1 in the phaco group and 4 in the LPI group) could not be contacted or refused to join the study; 17 patients (27.4%) died within 10 years after the APAC (11 patients in the phaco group and 6 patients in the LPI groups, P = 0.155, χ 2 test). Details of the deceased patients are listed in supplementary Table 1s (Supplemental Digital Content 1, http://links.lww. com/IJG/A524). The mean time interval between the APAC episode and death was 62.6 ± 37.4 months (range, 3 to 108 mo) and the mean age of death was 79.1 ± 5.8 of age (range, 70 to 89 of age). We cannot relate the cause of death to APAC or ophthalmic interventions.
A total of 40 patients (64.5%; 19 and 21 patients in the phaco and LPI group, respectively) were successfully invited for the current assessment and their baseline demographics are also listed in Table 2. The average follow-up period was 10.6 ± 0.7 years (range, 9.3 to 11.7 y; 10.6 ± 0.7 y in the phaco group and 10.8 ± 0.6 y in LPI group, P = 0.383, χ 2 test). There were no significant differences in the mean age (P = 0.371, t test), female-to-male ratio (P = 0.583, χ 2 test), and OD:OS ratio (P = 0.752, χ 2 test), between the 2 groups. For the presenting clinical features, there was no significant difference in the maximum IOP at presentation, IOP after aborted APAC and immediately before intervention, mean Shaffer grading on gonioscopy, VCDR, number of eye drops use immediately before intervention, preoperative spherical equivalent, axial length, anterior chamber depth, and lens thickness between the 2 groups ( Table 2). None of these patients underwent trabeculectomy or any other additional glaucoma procedure although 15 of 21 patients (71.4%) in the LPI group had undergone uneventful phacoemulsification before the current assessment. The mean period between APAC and cataract operation was 49.0 ± 41.7 months (range, 19 to 115 mo). These patients were operated because of reduced visual acuity (10 patients), elevated IOP (3 patients), or both (2 patients). Table 3 summarizes the primary and secondary outcomes of eyes that had APAC. The phaco group required fewer IOP-lowering eye drops (0.16 ± 0.37 vs 0.76 ± 1.09 bottle per eye, P = 0.028), had less extensive PAS (120.0 ± 116.2 vs. 244.3 ± 139.8, P = 0.010) and a greater mean Shaffer grading (1.79 ± 0.84 vs. 1.40 ± 0.87; P = 0.021) than the LPI group. There was no significant difference in the mean IOP (14.6 ± 2.7 vs. 14.2 ± 3.9 mm Hg; P = 0.711), MD (−11.31 ± 8.56 vs. −10.02 ± 8.15 dB, P = 0.296), PSD (5.08 ± 3.10 vs. 4.28 ± 2.69 dB, P = 0.303), and visual field index (73.7 ± 29.7% vs. 78.1 ± 30.5%; P = 0.239) between the phaco group and the LPI group. There were also no significant differences in the VCDR, BCVA, spherical equivalent, CCT, central endothelial cell density, average RNFL thickness on OCT, and the number of VF progression between the 2 groups. Three patients were not included in the evaluation of VF and/or OCT with reasons stated in Table 3. These significant differences were similar when we compared patients in the phaco group with patients in the LPI group that underwent subsequent phacoemulsification after the first 18 months of follow-up (n = 15) ( Table 3).
The LPI group have a significantly higher incident of persistent IOP elevation ( > 21 mm Hg in 2 consecutive visits) and blindness (BCVA of worse than 0.1 and/or central VF of < 20 degree) 14 compare to the phaco group (5 vs 0 eyes; P = 0.022 and 4 vs 0 eyes; P = 0.045, respectively, χ 2 test). One of the blind eyes did not have evidence of glaucomatous optic neuropathy (the patient had bilateral macular edema due to bilateral diabetic maculopathy and central retinal vein occlusion on the APAC eye). Another 3 blind eyes had evidence of glaucomatous optic neuropathy. One of the eyes had persistent IOP elevation underwent posterior vitrectomy and membrane peeling surgery in the 21st month after the APAC episode. Another 2 eyes had severe visual loss (BCVA of 0.1 and MD of −28.81 dB) and loss of light perception (with corneal decompensation after the lens extraction); both patients refused to undergo phacoemulsification until the eighth year despite persistently poorly controlled IOP. There were no statistically significant differences in the number of eyes with glaucomatous VF defect (11 vs. 14 eyes, P = 0.457, χ 2 test) and VF progression (4 vs. 8 eyes, P = 0.240, χ 2 test) between the 2 groups. Overall, they all eventually underwent lens extraction because of poor controlled IOP, glaucoma progression, or worsening of visual acuity. For the contralateral fellow non-APAC eyes, 11 (57.9%) and 9 (42.9%) of the patients in the phaco group and the LPI group underwent phacoemulsification during the 10 years follow-up, respectively. Three (15.8%) eyes in the phaco group and 4 (19.0%) eyes of the LPI group developed glaucomatous VF defect. The MD of both eyes of each patient is illustrated in the scatterplots of Figures 1A  and B. In the phaco group, 2 patients had "visual impairment." In the LPI group, 1 patient had "visual impairment" and 1 patient had "statutory blindness"-the latter was the patient who had bilateral diabetic maculopathy and central retinal vein occlusion on the APAC eye, without glaucomatous optic neuropathy (as mentioned above). Otherwise, most patients had an acceptable overall visual function when both eyes were taken into consideration.

DISCUSSION
Several studies that described long-term results of APAC eyes mostly involved patients who underwent LPI as the initial treatment modalities 14,19-23 without a definite timing for phacoemulsification. In a retrospective study performed by Baek et al, 20 62 APAC eyes underwent LPI followed by phacoemulsification within 1 month of the attack and demonstrated a cumulative rate of IOP rise of 22.56% at 12 months. Our study is different in that it provided the 10-year results of a group of APAC patients who were previously recruited in an RCT and were randomized to undergo either primary phacoemulsification and LPI shortly after abortion of APAC 2 ; phacoemulsification was not performed for the LPI group during the 18 months of follow-up.
Despite the relatively small sample size, our cohort showed better IOP control, lower incidence of blindness, and requiring fewer medications in the phaco group compared with the LPI group. In the previous RCT (the first 18 months), only 1 eye (3.2%) in the phaco group had IOP rise and was controlled by IOP-lowering medications, compared with 14 eyes (46.7%) in the LPI group (P < 0.001). 2 After the 18-month period, our current follow-up study reviewed a zero-incidence of IOP elevation and blindness in the phaco group, compared with the respective rate of 23.8% (5/21) and 19.0% (4/21) in the LPI group. The definition of blindness was adopted from a previous cross-sectional study by Aung et al 14 (VA < 6/60 and/ or VF < 20 degree), that reported a similar incidence rate of blindness (17.8%) after an average follow-up period of 6.3 years of 90 APAC eyes that underwent LPI. The requirement of fewer medications in the phaco group was also important because the cost of medication is the key contributor to the total cost of treating primary angle-closure glaucoma (PACG) according to cost-effectiveness analysis. 24 Medication side effects could also lead to substantial utility loss-as worse as 0.11 of the utility score. 25 Improvement of visual acuity and IOP control by early phacoemulsification should be beneficial, given the demonstrated improvement of visual and overall health status after clear lens extraction in the EAGLE study, 26,27 who had better preoperative visual acuity compared with our cohort.
The phaco group also had wider drainage angles and lesser degree of PAS. Theoretically, it is preferable to open up the angle as soon as possible. Chronic synechial closure causes irreversible damage to the trabecular meshwork, with proliferation of iris or fibrous tissue into the intertrabecular space. 28 However, early lens extraction shortly after medically aborted APAC could be technically demanding and some older age patients might not be able to cope with the stress of the procedure. LPI is still a very helpful technique to relieve pupil block and should be considered as the treatment of choice if primary phacoemulsification is difficult. A successful initial LPI probably would not jeopardize the effect of the subsequent lens extraction procedure, 1 given that patients in the LPI group (the vast majority also underwent phacoemulsification after completing the 18-month follow-up of the last RCT) did not show any significant difference in other outcomes, including IOP, the VF indices, BCVA and RNFL thickness measured by OCT. We included the patient who have blindness with no light perception for the affect eye and label the MD as −30 dB, given that she could perform a good quality visual field with the fellow eye. A, Scatterplots of MD of the phaco group. B, Scatterplots of MD of the LPI group. The patient that was classified as "statutory blindness" had bilateral diabetic maculopathy and central retinal vein occlusion in the APAC eye, without any structural evidence of glaucomatous optic neuropathy. APAC indicates acute primary angle closure; IOL, intraocular lens; LPI, laser peripheral iridotomy; MD, mean deviation.