Characteristics of Visual Field Progression in Medically Treated Normal-Tension Glaucoma Patients with Unstable Ocular Perfusion Pressure

Abstract

Purpose.: To investigate the characteristics of visual field (VF) progression in medically treated normal-tension glaucoma (NTG) patients (Koreans) with unstable ocular perfusion pressure (OPP).

Methods.: One hundred one eyes of 101 NTG patients followed up for more than 4 years (mean follow-up, 6.2 years ± 12.1 months) were included. Modified Anderson criteria (MC) and linear regression analysis (LA) of VF mean deviation (MD) within the central 10° and 10° to 24° area were assessed for determining VF progression in groups with lowest (LMF) and highest (HMF) 24-hour mean OPP [MOPP = ⅔(mean arterial pressure − IOP)] fluctuation. Kaplan-Meier analyses were used to compare the elapsed time of confirmed VF progression in the two groups. Hazard ratios (HRs) for the association between clinical risk factors including 24-hour MOPP and central VF progression were obtained by using Cox proportional hazards models.

Results.: Three of 33 eyes in the LMF progressed, whereas 12 of 34 eyes in the HMF progressed within the central 10° according to the MC; the between-group difference was significant (P = 0.010). By LA within the central 10°, two eyes from the LMF and nine from the HMF groups showed progression (P = 0.025). The HMF showed a greater cumulative probability of central VF progression than the LMF, by both LA and MC (Kaplan-Meier analysis, P = 0.003, 0.015, log-rank test). In multivariate analysis, only 24-hour MOPP fluctuation was significantly associated with central VF progression (P = 0.014).

Conclusions.: The 24-hour MOPP fluctuation was the most consistent prognostic factor among various IOP, blood pressure, and clinical factors for central VF glaucomatous progression in our series of NTG eyes.

According to the Collaborative Normal Tension Glaucoma (CNTG) report, 20% of NTG patients revealed visual field (VF) progression, even though an intraocular pressure (IOP) reduction of at least 30% from baseline measurement had been attained. ¹ Thus, non-IOP related elements have been investigated as contributors to NTG progression. 

In two studies, we have reported that 24-hour ocular perfusion pressure (OPP) instability correlates with the severity of glaucomatous VFs at initial presentation of NTG. ^2,3 In a subsequent longitudinal study, 24-hour unstable OPP was reported to be strongly associated with glaucomatous VF progression in NTG patients. ⁴ In that study, 24-hour OPP fluctuation was the most consistent prognostic factor for glaucoma progression in medically treated NTG eyes. ⁴  

Past reports have shown that NTG eyes have VF defects that are dense and more central than those in patients with primary open-angle glaucoma, as determined by computerized threshold perimetry. ^{5 –7} Ahrlich et al. ⁸ showed in their recent publication that NTG eyes progress more often in the central VF. However, no investigators have sought to determine which locations or areas within the VF show the most progression in medically treated NTG eyes with 24-hour unstable OPP. The purpose of the present study was to investigate the characteristics of VF progression in such eyes by examining differences in the location of VF progression between eyes, with and without relatively stable 24-hour OPP over a long-term period. 

Methods

Subjects

We retrospectively reviewed the medical records of 330 consecutive NTG patients seen by a glaucoma specialist (MSK) from March 1996 to June 2003 at the glaucoma service of the Asan Medical Center, Seoul, Korea. At initial NTG work-up, each patient received a comprehensive ophthalmic examination including a review of medical history, measurement of best corrected visual acuity (BCVA), slit lamp biomicroscopy, Goldmann applanation tonometry (GAT), gonioscopy, dilated funduscopic examination using a 90- or 78-D lens, stereoscopic optic disc photography, VF examination using standard automated perimetry (SAP), and 24-hour in-hospital IOP measurement. SAP testing was performed using the full-threshold strategy of program 24-2 (Humphrey Visual Field Analyzer; Carl Zeiss Meditec, Dublin, CA). The central corneal thickness (CCT) of each patient was also obtained during initial presentation using ultrasound pachymetry (DGH-550 instrument; DGH Technology, Inc., Exton, PA). 

For NTG diagnosis, patients had to have an optic disc of glaucomatous appearance and glaucomatous VF loss, both confirmed and agreed on by two glaucoma specialists (KRS, MSK); best corrected visual acuity (BCVA) better than 20/30; maximum IOP less than 22 mm Hg during in-hospital 24-hour IOP monitoring with GAT, as well as at the outpatient clinic; a normal anterior chamber; and an open angle on gonioscopic examination. Patients with evidence of intracranial or otolaryngeal lesions, histories of massive hemorrhage or hemodynamic crisis, previous use of antiglaucoma medication, any other ophthalmic disease that could result in VF defects, or histories of diabetes mellitus or eye surgery/laser treatment were excluded from NTG diagnosis. Eyes with glaucomatous VF defects were defined as those that met two of the following criteria, as confirmed by more than two reliable consecutive tests, in addition to compatibility with optic nerve appearance: (1) a cluster of three points with a probability of less than 5% on a pattern deviation (PD) map in at least one hemifield and including at least one point with a probability of less than 1% or a cluster of two points with a probability of less than 1%; (2) a Glaucoma Hemifield Test (GHT) result outside normal limits; and (3) a pattern SD (PSD) outside 95% of the normal limit. Reliable VF assessment was defined as a VF test with a false-positive error <15%, a false-negative error <15%, and a fixation loss <20%. The first perimetric result was excluded from analysis, to obviate learning effects. 

All patients had to meet the following additional inclusion criteria to enter into the current retrospective study: newly diagnosed NTG without previous treatment; age >40 years; in-hospital 24-hour monitoring of IOP and blood pressure (BP); follow-up at our clinic of at least 4 years with visits at 4- to 6-month intervals; good adherence to antihypertensive glaucoma treatment; availability of at least five reliable VF datasets obtained by VF analysis with the program central 24-2 during follow-up (thus excluding the first perimetry dataset); and mean deviation (MD) better than −20.00 dB without threat to fixation, by field analyzer measurement. If surgical or laser treatment had been considered to treat VF progression, those eyes were defined as VF progressors, and only the data obtained in the period before the procedure were analyzed. If both eyes of a patient had NTG and met inclusion criteria, one eye was randomly included in the analyses. Subjects on systemic antihypertensive or other hemodynamically active medications at baseline and during the follow-up period were not excluded. 

Institutional Review Board approval was obtained from the Asan Medical Center. The design of this study adhered to the principles of the Declaration of Helsinki. 

Twenty-Four-Hour Monitoring of IOP and Blood Pressure

All participants were admitted and went through in-hospital 24-hour monitoring of IOP and BP at initial work-up. The method has been described in detail elsewhere. ⁴ Briefly, BP and IOP were evaluated in the hospital for 24 hours in each patient, with measurements taken every 2 hours between 12 PM and 10 AM the following day, except for the period between 12 AM and 6 AM, during which measurements were taken every 3 hours. Systolic and diastolic BPs (SBP and DBP, respectively) were measured with a brachial Riva-Rocci sphygmomanometer on the upper left arm after the patient had been seated for at least 5 minutes. IOP was next measured after the subject had been seated in the slit lamp chair for at least 5 minutes. During the sleep period (midnight to 6 AM), the patients were awakened and seated for at least 5 minutes before BP measurements were taken (all measurements were completed within a few minutes). The patients were asked to refrain from any physical activity that could affect BP during admission. Meals were provided at 6:30 PM and 7:30 AM and did not include any alcohol or caffeine. 

MOPP Estimation

For calculation of MOPP, mean arterial pressure (MAP) was calculated as follows: MAP = DBP+[⅓(SBP − DBP)]. ^9,10 It was thus possible to calculate MOPP at any specified time from the difference between MAP and IOP (substituting for venous pressure) as follows: MOPP = ⅔(MAP − IOP). ^{11 –13} Twenty-four-hour MOPP fluctuation was defined as the amplitude of the nocturnal OPP fall and calculated as diurnal average MOPP minus nocturnal lowest MOPP. ^2,3,14,15  

Definition of Highest, Middle, and Lowest MOPP Fluctuation Groups

Sample size calculation in each group had been previously described based on the assumption that a 20% difference in VF progression between control eyes (those with stable MOPP fluctuation) and eyes at risk (those with high MOPP fluctuation) was clinically relevant. ⁴ Therefore, approximately 33 patients were required in each group, and thus NTG patients were categorized into three groups according to the level of MOPP fluctuation during initial 24-hour IOP measurements (HMF group; the highest tertile, MMF group; middle tertile, LMF; the lowest tertile of MOPP fluctuation). 

Determination of VF Defect Location and Progression

To determine the location of VF defects and progression thereof, the 24-2 field was divided into two areas, within the central 10° region, and the 10° to 24° area. Two test locations within the blind spot were excluded (Fig. 1). 

Figure 1.

 

The central 10° region of the Humphrey 24-2 visual field was determined as illustrated. Two test locations within the blind spot and 10° to 24° region (shaded area) were excluded.

First, the location of VF defect clusters at baseline and at the final VF test was distinguished on the basis of location within the central 10° or outside the central 10° (thus in the 10–24°) region. To determine VF defect location at baseline and at the final VF test, we regarded eyes with clusters of three significant points within the central 10° (with P < 5% on the PD map) were regarded as having a central VF defect and those outside the central 10° (thus in the 10°–24°) were regarded as having a peripheral VF defect. 

Next, VF progression was determined using two methods: the modified Anderson criteria (MC) and linear regression analysis (LA) of VF MD. The MC for VF progression has been described elsewhere. ¹⁶ When using the MC, if VF progression was present within the central 10°, it was considered to reflect central VF progression. Similarly, if progression occurred in the 10° to 24° area of VF, it was considered to have occurred in the peripheral VF. 

For LA, all individual values at each test location of the total deviation plot within and outside the central 10° arc were averaged (to yield a central MD and peripheral MD, respectively). LA was performed to explore whether the slope of the central and peripheral MD over the follow-up period were statistically significant. A negative slope with a P < 0.05 was considered to be statistically significant, and was taken to reflect progression. 

Analysis

Categorical variables are presented as numbers and percentages, and the χ² test was used to compare clinical characteristics among the three groups (the HMF, MMF, and the LMF groups) classified by baseline 24-hour BP and IOP measurement. Continuous variables are expressed as the mean ± SD, or as the median ± interquartile range, as appropriate, and ANOVA or the Kruskal-Wallis test was used for multiple comparisons. Bonferroni correction was performed in the multiple comparisons. 

The VF progression within central 10° and in the 10° to 24° region determined by two MC and LA was assessed in both the LMF and HMF groups. Various clinical, IOP and BP parameters were compared between central VF progressors and nonprogressors. 

Kaplan-Meier life table analyses were performed to compare time to confirmed central and peripheral VF progression in LMF and HMF groups. 

Hazard ratios (HRs) for associations between potential risk factors and central VF progression based on MC were obtained using Cox's proportional hazards models. Univariate analyses were performed separately for each variable. Variables with a probability value ≤0.20 in univariate analysis were considered significant and were included in a multivariate Cox proportional hazards model. A backward elimination process was used to develop the final multivariate model, and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated. Schoenfeld residuals and the log [−log (survival rate)] test were used to verify that proportional hazards assumptions were not violated. Model fit was assessed using residual analyses (all statistical analyses performed with SAS ver. 9.1; SAS Institute Inc., Cary, NC, and SPSS ver. 15.0; SPSS Inc., Chicago, IL). 

Results

A total of 101 eyes of 101 NTG patients who met the inclusion criteria were analyzed. Among the 101 patients, 48 were men, 53 were women, and all were Asians (Koreans). The average ± SD of age, and PSD at baseline were 54.2 ± 11.9 years, −4.82 ± 5.50 dB, and 5.22 ± 4.03 dB, respectively. The average number of VF examinations was 7.6 ± 2.4 (range, 5–16) and the follow-up period was 74.8 ± 12.1 months (range, 49–120). 

Among 101 eyes, 78% were treated with prostaglandin analog alone, 11% by dorzolamide/timolol fixed combination and prostaglandin analogues, 8% by dorzolamide/timolol fixed combination alone, and 3% timolol alone. Thus, 22% of the eyes were treated with either topical β-blockers alone or topical β-blocker containing fixed combination agents. 

Of the 101 eyes, 33 were classified into the LMF group, 34 into the MMF group, and 34 into the HMF group. Twenty-four hour MOPP fluctuation showed significant differences between the three groups, as expected. Other variables including patients' demographics, IOP parameters, BP parameters, use of systemic hypertensive medications, self-reported history of systemic hypertension, use of topical β-blockers and initial VF severity were not different between the three groups (Table 1). 

Table 1.

 

Demographic Comparisons among the LMF, MMF and HMF Groups

	LMF (n = 33)	MMF (n = 34)	HMF (n = 34)	P
Age, y	52.0 ± 12.3	55.2 ± 13.1	55.0 ± 12.6	0.464
Sex, M/F, n	16/17	15/19	17/17	0.898
SE, D	−0.77 ± 2.1	−0.39 ± 2.4	−0.85 ± 2.0	0.332
CCT, μm	528.2 ± 33.2	539.7 ± 30.5	520.1 ± 32.1	0.112
History of systemic hypertension, n (%)	11 (33%)	8 (24%)	14 (41%)	0.488
Systemic antihypertensive medication, n (%)	9 (27%)	6 (18%)	12 (35%)	0.453
24-Hour mean SBP >140, n (%)	5 (15%)	5 (15%)	7 (21%)	0.552
24-Hour mean DBP >90, n (%)	3 (9%)	2 (6%)	1 (3%)	0.290
VF MD, dB	−4.56 ± 5.12	−4.70 ± 5.50	−5.19 ± 6.73	0.75
VF PSD, dB	5.15 ± 4.31	5.30 ± 4.06	5.42 ± 3.87	0.79
24-Hour mean SBP, mm Hg	123.1 ± 12.9	122.1 ± 13.5	126.8 ± 14.8	0.301
24-Hour mean DBP, mm Hg	75.2 ± 9.28	75.4 ± 9.43	76.0 ± 9.45	0.85
24-Hour mean MOPP, mm Hg	46.3 ± 6.29	46.9 ± 6.45	47.4 ± 6.30	0.47
24-Hour MOPP fluctuation, mm Hg	0.56 ± 4.05	9.3 ± 4.51	13.1 ± 4.48	<0.0001*
Mean baseline IOP, mm Hg	15.5 ± 2.55	15.6 ± 2.31	15.3 ± 2.00	0.85
24-Hour mean IOP, mm Hg	14.4 ± 2.32	14.1 ± 2.45	14.5 ± 2.65	0.79
Diurnal maximum IOP during 24-hour phasing, mm Hg	16.9 ± 3.43	17.3 ± 2.59	16.3 ± 2.80	0.49
Diurnal mean IOP during 24-hour phasing, mm Hg	14.0 ± 2.42	15.3 ± 2.20	14.0 ± 2.19	0.085
Nocturnal minimum IOP during 24-hour phasing, mm Hg	11.1 ± 2.01	13.9 ± 1.65	11.5 ± 2.30	<0.001
				<0.001*
				<0.001**
24-Hour IOP fluctuation, mm Hg	2.55 ± 2.06	1.69 ± 2.19	2.49 ± 1.45	0.073
Mean follow-up IOP, mm Hg	13.6 ± 1.75	13.7 ± 1.77	14.0 ± 1.68	0.55
Follow-up IOP fluctuation, mm Hg	1.94 ± 0.55	1.88 ± 0.63	1.90 ± 0.66	0.79
Use of topical β-blocker, n (%)	9 (27%)	6 (18%)	7 (21%)	0.514

  P < 0.05: * comparison between LMF and MMF and ** between MMF and HMF, by Bonferroni correction.

The prevalence of VF defect within the central 10° was significantly different between baseline and last follow-up in the total and HMF groups (P = 0.023, P = 0.026; Table 2). 

Table 2.

 

Prevalence of VF Defect in Two Locations

	Baseline	Last Follow-up	P
Total (101 eyes)
Center	36 (35.6)	51 (50.5)	0.023*
Pph	82 (81.2)	90 (89.1)	0.083
LMF (33 eyes)
Center	10 (30.3)	13 (39.4)	0.303
Pph	30 (90.9)	32 (97.0)	0.307
MMF (34 eyes)
Center	13 (38.2)	16 (47.1)	0.312
Pph	25 (73.5)	27 (79.4)	0.383
HMF (34 eyes)
Center	13 (38.2)	22 (64.7)	0.026*
pph	26 (76.5)	30 (88.2)	0.170

 Data are expressed as the number of eyes (% of total). Center, within 10°; Pph, 10°–24°.

*   P < 0.05.

The percentages of VF defect were not significantly different between the LMF and HMF groups in both central 10° region and 10° to 24° region at baseline (P = 0.335, P = 0.102). However, a significantly higher percentage of eyes in the HMF group than in the LMF group showed VF defect clusters within the central 10° at last follow-up (P = 0.033, Table 3). 

Table 3.

 

Comparison of VF Defect Prevalence within the Central 10° Region and the 10° to 24° Region between the LMF and HMF Groups

	LMF (33 Eyes)	HMF (34 eyes)	P
Baseline
Center	10 (30.3)	13 (38.2)	0.335
Pph	30 (90.9)	26 (76.5)	0.102
Last follow-up
Center	13 (39.4)	22 (64.7)	0.033*
Pph	32 (97.0)	30 (88.2)	0.187

 Data are expressed as the number of eyes (% of total). Center, within 10°; Pph, 10°–24°.

*   P < 0.05.

Of the 101 eyes, 22 (21.8%) showed VF progression within the central 10° region and 16 (15.8%) within the 10° to 24° region, according to the MC. Fifteen eyes (14.8%) showed VF progression within the central 10° region, and 13 (12.9%) eyes within the 10° to 24° region, by LA. By MC, a significantly higher percentage of eyes in the HMF group progressed within the central 10° than in the LMF group (P = 0.010). The LA of the MD slopes showed a similar result (P = 0.025; Table 4). 

Table 4.

 

Prevalence of VF Progression within Central 10° and 10° to 24° Area by Two Different Criteria

		LMF (33 Eyes)	HMF (34 Eyes)	P
Modified	Center	3 (9.1)	12 (35.4)	0.010*
Anderson	Pph	2 (6.1)	8 (23.5)	0.046*
Linear regression	Center	2 (6.1)	9 (26.5)	0.025*
	Pph	3 (9.1)	5 (14.7)	0.372

 Data are expressed as the number of eyes (% of total). Center, within 10°; Pph, 10°–24°.

*   P < 0.05.

Among the analyzed parameters, use of systemic hypertensive medications, self- reported history of systemic hypertension, 24-hour mean SBP, and MOPP fluctuation differed significantly between central VF progressors and nonprogressors, based on MC. Other investigated variables including age, CCT, IOP parameters, BP parameters, use of topical β-blockers, and initial VF MD and PSD, did not show significant differences between the two groups (Table 5). 

Table 5.

 

Comparison between Central VF Progressors and Nonprogressors Based on Modified Anderson Criteria

	Central Progressors (n = 22)	Nonprogressors (n = 79)	P
Age, y	55.8 ± 12.3	53.8 ± 11.8	0.50
Sex, M/F, n	12/10	36/43	0.63
SE, D	−0.80 ± 1.8	−0.55 ± 1.92	0.59
CCT, μm	492.3 ± 119.8	471.1 ± 165.5	0.52
History of systemic hypertension, n (%)	12 (54.5)	21 (26.6)	0.015*
Systemic antihypertensive medication, n (%)	10 (45.5)	17 (21.5)	0.027*
24-Hour mean SBP >140, n (%)	14 (18.2)	13 (16.5)	0.535
24-Hour mean DBP >90, n (%)	0 (0)	6 (8.2)	0.219
VF MD, dB	−4.82 ± 4.38	−4.77 ± 5.78	0.97
VF PSD, dB	5.18 ± 2.83	5.18 ± 4.31	0.99
24-Hour mean SBP, mm Hg	128.9 ± 9.8	122.7 ± 15.0	0.025*
24-Hour mean DBP, mm Hg	76.6 ± 7.69	75.2 ± 9.74	0.48
24-Hour mean MOPP, mm Hg	48.4 ± 5.06	46.5 ± 6.89	0.16
24-Hour MOPP fluctuation, mm Hg	8.55 ± 4.16	4.06 ± 3.97	0.034
Mean baseline IOP mm Hg	15.5 ± 1.90	15.1 ± 2.33	0.43
24-Hour mean IOP, mm Hg	14.3 ± 1.85	14.3 ± 2.62	0.89
Diurnal maximum IOP during 24-hour phasing, mm Hg	16.9 ± 2.28	16.7 ± 3.12	0.77
Diurnal mean IOP during 24-hour phasing, mm Hg	14.4 ± 1.91	14.3 ± 2.42	0.87
Nocturnal minimum IOP during 24-hour phasing, mm Hg	12.0 ± 2.67	12.0 ± 2.24	0.98
24-Hour IOP fluctuation, mm Hg	2.43 ± 1.67	2.31 ± 1.97	0.82
Mean follow-up IOP mm Hg	13.6 ± 1.52	13.5 ± 1.97	0.80
Follow-up IOP fluctuation, mm Hg	1.95 ± 0.75	1.88 ± 0.60	0.67
Use of topical β-blocker, n (%)	7 (31.8)	15 (19.0)	0.159

*   P < 0.05.

The HMF group showed a greater cumulative probability of central VF progression than the LMF group, according to both the MC and LA, when assessed by Kaplan-Meier analysis (Figs. 2A, 3A). The cumulative probability of nonprogression of VF loss within the central 10° was 82.1% after 80 months in patients in the LMF group and 55.8% in those in the HMF group based on the MC (P = 0.003, Fig. 2A), and 90.9% and 67.2% in the LMF and HMF group, respectively, based on LA (P = 0.015, Fig. 3A). However, in the region of 10° to 24°, the cumulative probability of nonprogression was not different between the LMF and HMF group (P = 0.231, Fig. 3B). 

Figure 2.

 

Kaplan-Meier survival analysis of progressive VF loss as defined by MC (A, within the central 10° region, and B, the 10°–24° region). y-Axis: cumulative probability of VF nonprogression within the central 10°; x-axis: follow-up period (months). Blue line: LMF group. Green line: HMF group.

Figure 3.

 

Kaplan-Meier survival analysis of progressive VF loss as defined by LA (A, within the central 10° region, and B, the 10°–24° region). y-Axis: cumulative probability of VF nonprogression within the central 10°; x-axis: follow-up period (months). Blue line: LMF group. Green line: HMF group.

Twenty-four hour mean SBP, history of systemic hypertension, use of systemic antihypertensive medication, and MOPP fluctuation were found to be significantly predictive of central VF progression by univariate Cox proportional hazard model (Table 6). In multivariate analysis, 24-hour MOPP fluctuation was significantly associated with central VF progression (P = 0.019; Table 7). 

Table 6.

 

Univariate Cox Proportional Hazards Model Data for Prediction of Central VF Loss Progression

Variable	HR	95% CI	P
Age, y	1.025	[0.987–1.025]	0.200
Sex	1.211	[0.510–2.874]	0.665
SE	0.893	[0.694–1.147]	0.375
History of systemic hypertension	1.405	[1.171–1.959]	0.040*
Systemic antihypertensive medication	1.970	[1.827–4.695]	0.126
24-Hour mean SBP >140	0.743	[0.248–2.231]	0.597
24-Hour mean DBP >90	0.838	[0.446–2.678]	0.448
CCT	1.002	[0.999–1.005]	0.256
VF MD	0.993	[0.929–1.061]	0.833
VF PSD	1.028	[0.925–1.144]	0.605
Presence of central VF defect	1.650	[0.661–4.120]	0.283
24-Hour mean SBP	1.027	[0.996–1.059]	0.094
24-Hour mean DBP	1.005	[0.960–1.051]	0.845
24-Hour Mean MOPP	1.040	[0.973–1.111]	0.250
24-Hour MOPP fluctuation	1.061	[1.015–1.109]	0.009*
Mean baseline IOP	1.005	[0.817–1.237]	0.962
24-Hour mean IOP	0.934	[0.772–1.131]	0.485
Diurnal maximum IOP during 24-hour phasing mm Hg	1.00	[0.876–1.142]	0.998
Diurnal mean IOP during 24-hour phasing mm Hg	0.996	[0.829–1.196]	0.963
Nocturnal minimum IOP during 24-hour phasing mm Hg	1.045	[0.847–1.289]	0.682
24-Hour mean IOP fluctuation	0.941	[0.739–0.941]	0.623
Mean follow-up IOP	0.922	[0.711–1.197]	0.543
Follow-up IOP fluctuation	1.040	[0.489–2.215]	0.919
Use of topical β-blocker	0.863	[0.343–2.17]	0.754

*   P < 0.05.

Table 7.

 

Multivariate Cox Proportional Hazards Model with Backward Elimination for Prediction of Central Visual Field Progression

Variable	Hazard Ratio	95% CI	P
Age, y	1.015	[0.977–1.055]	0.446
24-Hour mean SBP	1.014	[0.976–1.053]	0.478
History of systemic hypertension	2.081	[0.875–4.951]	0.098
Systemic antihypertensive medication	0.441	[0.090–2.149]	0.311
24-Hour MOPP fluctuation	1.055	[1.009–1.103]	0.019*

*   P < 0.05.

Discussion

In the previous study, we found that 24-hour MOPP fluctuation was the most consistent prognostic factor for glaucoma progression. ⁴ With the same patient population, we intended to investigate the association between location or area of VF progression and MOPP fluctuation. 

In the present study, we found that the prevalence of VF defects within the central 10° as well as the peripheral 10° to 24° region did not differ between the HMF and LMF groups at baseline. Nevertheless, more eyes in the HMF group were found to have central VF defects when assessed at final follow-up, whereas the two groups did not differ in prevalence in the peripheral 10° to 24° region. This finding was further validated by two types of VF progression measures: MC and LA. By both MC and LA, significantly more eyes in the HMF group showed central VF progression than in the LMF group. 

Using the Kaplan-Meier life table method, we were able to demonstrate that NTG eyes in the HMF group had a greater probability of progressive central VF loss than did eyes in the LMF group. Of interest, in the peripheral 10° to 24° region, the cumulative probability of nonprogression did not differ between the two groups by LA (P = 0.231, log-rank test, Figs. 2, 3). This finding further suggests that functional changes in the central VF may be influenced by disturbances in OPP. 

It is difficult to explain why eyes with unstable 24-hour MOPP (those in the HMF group) showed more progressive VF changes in the central 10° area than did eyes in the LMF group. However, some recent reports may help explain our findings. Unstable 24-hour MOPP, by means of reduced optic nerve head (ONH) blood flow below a critical level during sleep in a vulnerable ONH, may play a role in the progression of central VF defects, as has been noted in ONH ischemic disorders. Hayreh et al. ¹⁷ used hourly averaged BP data to show that a significantly greater decrease in mean DBP was found in eyes with NTG and in those of patients with nonarteritic AION. Therefore, disturbance of retrobulbar hemodynamics in patients with NTG and nonarteritic AION, especially in the short posterior ciliary arteries, might lead to a similar pattern of VF loss and/or progression. 

Although we found that unstable 24-hour MOPP was the most consistent prognostic factor for central VF progression in both univariate and multivariate analyses (P = 0.009, 0.019, respectively), age, high mean SBP over 24 hours, self-reported history of hypertension, and use of systemic antihypertensive medication were significant risk factors in univariate analysis for central VF progression. Moreover, the percentages of eyes with a self-reported history of hypertension, high mean SBP over 24 hours, and use of systemic antihypertensive medication were significantly higher in central VF progressors than nonprogressors. These observations may suggest the influence of a 24-hour systemic BP change and/or subsequent OPP instability on central VF progression. 

Our finding that use of systemic antihypertensive medications was a significant risk factor for central VF progression in univariate analysis appears to be consistent with several population-based reports. Use of antihypertensive medications such as calcium channel antagonists has been shown to be associated with the increased incidence of glaucoma or increased cupping and decreased rim area of the optic disc in patients without glaucoma. ^18,19 A possible explanation of the harmful effect of the use of systemic antihypertensive medication on glaucoma progression may be that blood pressure is decreased, particularly during sleep period, thus reducing OPP, and this decrease may lead to progressive loss of VFs as well as optic nerve fibers. Therefore, our finding suggests that close monitoring and management of systemic BP including in the nocturnal period should be undertaken in conjunction with glaucoma treatment of patients receiving systemic antihypertensive treatment. 

In the mean time, all IOP-related parameters were not found to be related to central VF progression in our analysis. Our result that non-IOP elements are related to progression is in line with the CNTG result. ¹ Our findings that the central VF showed progression in NTG patients, even though IOP was stable, and that progression may be associated with unstable OPP could be of great importance in the management of NTG patients. 

The use of various topical antiglaucoma medications may affect glaucomatous VF progression by having a direct and/or indirect effect on the ONH perfusion by systemic absorption during long-term follow-up. Costagliola et al. ²⁰ have reported that prostaglandin analogues (latanoprost) significantly increases mean 24-hour OPP, whereas β-blockers (timolol) do not improve OPP, despite reductions in BP and heart rate in Caucasian NTG eyes. In the present study, the use of topical β-blockers was not associated with an increased risk of central VF progression in the NTG eyes, despite their reported effects, such as arterial hypotension, bradycardia, and systemic vascular resistance modification. ^{21 –26} Randomized multicenter studies are needed to ascertain the impact of β-blockers on the long-term prognosis of NTG related to its cardiovascular effects. 

Several limitations of our study must be acknowledged, including the relatively small sample size for multivariate analysis. A second limitation of our study is our inability to generalize our findings to all types of primary open-angle glaucoma classified by IOP level or to non-Asian individuals, because our study involved only Koreans. Third, our calculation of MOPP and variations thereof, based on a theoretical formula, may not reflect the real physiological status of ocular perfusion or variation in this parameter. Direct measurement of ocular blood flow (OBF) could result in a different outcome. However, direct measurement of OBF with imaging devices may also have particular inherent limitations. Fourth, measurements of BP and IOP and calculation of MOPP, using data collected in subjects in the sitting position may not provide the best physiological parameters predicting long-term glaucoma progression or disease location. Different results may have been obtained had we measured BP and IOP in a seamless physiological manner. Finally, only baseline 24-hour IOP and BP measurements were included in the analysis. Such measurement may not reflect the general status of BP and IOP parameters. Thus, analysis, including follow-up measurements, will be necessary in forthcoming studies. 

In conclusion, we found that unstable 24-hour MOPP was associated with central VF progression. Systemic hypertension and use of antihypertensive medication were more common in central VF progressors. These findings indicate that NTG eyes with unstable 24-hour MOPP may require more careful monitoring of central VF change. 

Footnotes

 Presented at the annual meeting of the American Academy of Ophthalmology, San Francisco, California, November 2009.

Footnotes

 Disclosure: K.R. Sung, None; J.W. Cho, None; S. Lee, None; S.-C. Yun, None; J. Choi, None; J.H. Na, None; Y. Lee, None; M.S. Kook, None