Differences in Systemic Pulse Waveform Between Individuals With Glaucoma, Glaucoma Suspects, and Healthy Controls

Abstract

Purpose: It has been hypothesized that compromised ocular circulation in glaucoma may be concomitant of systemic changes. The purpose of this study is to test whether systemic blood flow pulse waveform patterns differ between individuals with glaucoma (GL), glaucoma suspects (GLS), and normal healthy controls (HC).

Methods: The study included 35 bilateral GL, 67 bilateral GLS, 29 individuals with unilateral GL who were considered GLS in the other eye, and 44 healthy controls. Systemic pulsatile blood pressure waveforms were recorded using a finger cuff. A continuous 200 Hz plethysmography recording is made to obtain a pulse waveform. Waveform parameters were extracted using custom software from an average of eight pulse cycles. These were compared between GL, GLS, and HC groups on a per-eye basis, using generalized estimating equation models to account for intereye correlations; and plotted against disease severity by visual field linearized mean deviation (MD_lin) and retinal nerve fiber layer thickness (RNFLT).

Results: Averaged blood pressure was significantly lower in the HC group (mean ± standard deviation 91.7 ±11.7 mm Hg) than the GLS (102.4 ± 13.9) or GL (102.8 ± 13.7) groups, with P < 0.0001 (generalized estimating equation regression). Waveform parameters representing vascular resistance were higher in both GLS and GL groups than the HC group; and were correlated with RNFLT and MD_lin (P ≤ 0.05).

Conclusions: The shape of the systemic pulsatile waveform differs in individuals with GL/GLS suspects, compared to HC eyes. Blood pressure changes more rapidly in individuals with GL, which suggests higher arterial stiffness.

It has been long hypothesized that changes in the systemic circulation may play a role in the development and progression of glaucoma.^1–4 However, the relationship between systemic blood pressure and glaucoma is complex and limited studies exist. Systemic hypotension can decrease ocular perfusion pressure, similar to an increase in intraocular pressure (IOP).^5–10 Glaucoma patients exhibit reduced capillary blood flow and systemic vasospastic responses in nailfold capillaroscopy.^11,12 Systemic vasospasm has also been observed in normal-tension glaucoma patients.¹³ Furthermore, two studies have reported greater variability in systemic blood pressure among glaucoma patients compared to normal controls.^14,15 Two additional studies also showed that the disturbance of systemic circulation is one of the possible risk factors for normal-tension glaucoma.^16,17 

Arterial stiffness, a loss of arterial elasticity, is one of the major signs of vascular aging.^18,19 Increased arterial stiffness has been recognized as an independent risk factor for cardiovascular diseases.²⁰ In a small retrospective study, glaucoma patients were found to have higher systemic arterial stiffness compared to age-matched controls in Korean individuals with diabetes.²¹ Conversely this association was not found in a larger, prospective Japanese study involving nondiabetics.²² A Dutch study discovered that glaucoma was associated with high pulse pressure and suggested that individuals with glaucoma may have increased pulse wave velocity, indicating higher arterial stiffness.²³ Arterial vessels have also been reported to be stiffer in other ocular diseases such as pseudoexfoliation glaucoma,²⁴ age-related macular degeneration, retinal arterial narrowing,²⁵ and branch retinal vein occlusion.²⁶ We have recently shown evidence that higher vascular resistance within the eye was associated with more rapid functional progression in glaucoma suspects; it is not yet known whether that represents part of the pathophysiological process within the eye or a manifestation of a systemic risk factor for rapid progression.²⁷ 

The purpose of this study is to test whether systemic blood flow pulse waveform patterns differ between individuals with glaucoma (GL), glaucoma suspects (GLS), and normal healthy controls (HC). To achieve this goal, we recorded the systemic pulsatile blood pressure waveform using a finger cuff, and then used custom software to parametrize the average waveform. In particular, we generated a pulsatility index and a resistivity index, which have been shown to relate to vascular resistance and vascular compliance.²⁸ We then compared pulse waveform–derived parameters between HC individuals and GLS and GL patients. We focused not on blood pressure itself but on derived parameters from the pulse waveform; specifically, the pulsatility index (PI), resistance index (RI), and maximum slope and acceleration of the pulsatile waveform with the aim to establish the relationship between systemic hemodynamics and glaucoma. We expected that the derived parameters of the systemic pulsatile waveform differ in individuals with GL/GLS, compared to HC subjects. 

Methods

Subjects

All study participants were tested at Devers Eye Institute (Portland, OR, USA). All procedures in this study adhered to the tenets of the Declaration of Helsinki, complied with the Health Insurance Portability and Accountability Act of 1996, and were approved and monitored by the Institutional Review Board. All participants provided written informed consent after having the risks and potential benefits of participation explained to them. Data were collected prospectively from the following GL/GLS and HC cohorts to compare parameters derived from the systemic blood pressure waveform. 

P3 cohort: 166 eyes of 172 subjects with open-angle glaucoma or suspected glaucoma, as determined by the subject's clinician. The inclusion criteria included optic disc appearance suspicious for glaucoma, such as a cup to disc asymmetry ≥0.2 between the eyes, a large cup to disc ratio, nerve fiber layer thinning or defect in ophthalmoscopy, a history of disc hemorrhage or rim notching, at least one additional risk factor for glaucoma such as a first-degree family history of primary open-angle glaucoma (POAG), or ocular hypertension defined by IOP ≥ 22 mm Hg; but the final determination of eligibility as a “glaucoma suspect” and/or “glaucomatous eye” was made at the sole discretion of the individual's referring clinician. In addition, if one eye was eligible for P3 study, both eyes were included; on the basis that if one eye has glaucoma then the other is automatically a glaucoma suspect. This cross-sectional data was collected in subjects enrolled in the ongoing longitudinal Portland Progression Project (P3). If data from more than one visit were available, the most recent visit was used. Subjects undergo a set of functional and structural diagnostic tests once every six months, including standard automated perimetry (SAP, using the HFAII perimeter [Carl Zeiss Meditec Inc, Dublin, CA, USA], 24-2 test pattern and SITA Standard test strategy), optical coherence tomography (OCT; using the Spectralis OCT2 [Heidelberg Engineering, Heidelberg, Germany]), and systemic blood pressure measurement using an arm cuff. The systemic pulsatile blood pressure (BP) waveforms were recorded using the Finapres Nova (Finapres Medical Systems BV, Enschede, The Netherlands) finger cuff. A continuous 200 Hz plethysmography recording is made to obtain a pulse waveform (Fig. 1). Subjects were excluded if they had significant visual field loss from causes other than glaucoma. Both eyes were tested, even if only one had extant glaucomatous loss, on the basis that this implies that the fellow eye could be considered as a “glaucoma suspect” that is at increased risk of developing glaucomatous loss in the future. The P3 cohort was then subdivided into 166 eyes without functional loss, defined as a glaucoma hemifield test of either “Within Normal Limits” or “Abnormally High Sensitivity” on that test date; and 110 eyes with existing functional loss, defined as glaucoma hemifield test “Outside Normal Limits” or “Borderline.” Thirty-five individuals had bilateral glaucoma (defined as abnormal visual field), 67 were bilateral glaucoma suspects, and 29 individuals had unilateral functional loss and were considered “glaucoma suspects” in the other eye. 

Figure 1.

 

Generation process of average waveform: (A) Reconstructed brachial blood pressure (BP) waveforms. (B) Peak and trough were identified for each wave in graph A. (C) Represented by the blue dots are the eight consecutive cycles (contained in red box in B) that were most similar to each other, as measured by the cross-correlation between them. (D) The average of the eight cycles is plotted in red.

Healthy controls consisted of 88 eyes from 44 white individuals who volunteered as participants in our study. These control subjects underwent a comprehensive eye examination and were identified as having “normal vision” and “healthy eyes.” Inclusion criteria for these subjects were as follows: age 50 or older; absence of known conditions affecting vision, except for the need for glasses or mild cataracts; vision no worse than 20/30; no ocular surgery (including Lasik) within the month preceding imaging acquisition; absence of diabetes; blood pressure not exceeding 160/100 mm Hg (no uncontrolled hypertension); not currently using Plaquenil (hydroxychloroquine) or Gabapentin (anti-epileptic); not a current regular smoker; ability to concentrate for at least five minutes at a time; IOP < 21 mm Hg; normal glaucoma hemifield test; and OCT data, including retinal nerve fiber layer thickness (RNFLT) and minimum rim width, within normal ranges. 

Noninvasive Continuous Blood Pressure

Finapres Nova (Finapres Medical Systems BV) offers noninvasive continuous blood pressure monitoring, which was used to reconstruct brachial pressure using the generalized waveform filtering and level correction method.²⁹ The participant wears a finger cuff on one finger and an arm cuff on the contralateral arm to measure systemic blood pressure, which was used to correct and calibrate reconstructed brachial BP.²⁹ Care was taken in the selection and application of a proper size finger cuff. The cuffed finger was kept near heart level. Left–right pressure differences were tested and found to be small, indicating they present no major source of error. Recordings were digitized at 200 Hz by the Finometer with a resolution of 0.25 mm Hg and stored internally. 

Data of reconstructed brachial BP waveforms from the finger cuff were exported through a serial interface for further analysis. Reconstructed brachial BP waveforms from the finger cuff were then loaded into Matlab software (MATLAB version: 9.13.0 [R2022b]; The MathWorks Inc., Natick, MA, USA). The software sampled this continuous BP data to find the consecutive eight cycles that were most similar to each other, as measured by the cross-correlation between them. From these eight cycles, the analysis software then generated the average pulsatile waveform for a single representative cardiac cycle. For the primary analysis, the duration of the cycle was normalized to 1 unit (see Fig. 1); secondary analyses were performed without this normalization for pulse rate. A low pass filter was used to reduce noise by removing high-frequency signals. The first derivative (slope) and the second derivative (acceleration) of the waveform were then calculated. We derived the following parameters, as seen in Figure 2: 

• • Average mean BP (BP_ave) of the tissue, defined as the mean BP of the waveform, representing a measure of mean systemic blood flow.
• • PI, a measure of the proportion of blood flow that is maintained in the vessels between heartbeats, calculated from the average BP and the difference of the maximum and minimum BP of the waveform: (BP_max − BP_min)/BP_ave²⁸ (i.e., the magnitude of the pressure pulse normalized by average pressure).
• • RI, defined as the difference between the maximum BP and the minimum BP, divided by the maximum BP (see Fig. 2) (i.e., the magnitude of the pressure pulse normalized by minimum pressure). It is defined as (BP_max– BP_min)/BP_max and depends on both vascular resistance and vascular compliance²⁸).
• • Normalized flow acceleration index (FAI²⁷), an index of momentary power defined as the maximum amount of change in BP over 1/30 of the cycle (see Fig. 2), divided by the average pressure.
• • Normalized maximum slope and acceleration, during the period over which pressure increases before the systolic peak. Defined as MaxSlope_Norm = MaxSlope/BP_Ave and equivalently MaxAccel_Norm = MaxAccel/BP_Ave. MaxSlope is the largest positive slope before the systolic peak of the BP waveform. MaxAccel is the largest positive acceleration (or second derivative) before the systolic peak.
• • Normalized minimum slope and acceleration. Defined as MinSlope_Norm = MinSlope/BP_Ave and equivalently MinAccel_Norm = MinAccel/BP_Ave. MinSlope is the most negative slope before systolic peak of BP waveform. MinAccel is the smallest acceleration or second derivative before systolic peak.

Figure 2.

 

Parameters derived from average waveform: (A) BP_ave is the average blood pressure of the sample points. ∆ BP denotes the change in blood pressure during a 1/30 interval of the total recorded normalized time. (B) The instantaneous slope (first derivative) of the BP waveform in A. Maximum slope before systolic peak and minimum slope after systolic peak are calculated. (C) The instantaneous acceleration waveform (second derivative) of the BP waveform in A. All waveforms in A to C were post-processed by low-pass filter to remove high-frequency waveforms. FAI = Flow Acceleration Index FAI = max ∆ BP ; \(RI = \;\frac{{BPmax - BPmin}}{{BPmax}}\). \(PI = \;\frac{{BPmax - BPmin}}{{BPave}}\).

Analysis

The values of each parameter were formally compared among three groups of eyes: GLS, GL, and HC. These comparisons were performed using generalized estimating equation (GEE) models to account for intereye correlations, including some individuals who contributed one eye to the GL cohort and one eye to the GLS cohort. Each BP parameter was also plotted against linearized mean deviation (MD) from SAP and RNFLT from OCT, collected on the same visit. For perimetry, the linearized metric MD_Lin = 10^MD/10 was used instead of the native decibel-scaled MD, because this has been reported to reduce nonlinearities in structure-function relations.³⁰ Pearson correlations coefficients were reported, and significance was assessed using a GEE model. 

The primary analyses also performed after age-correcting each parameter to their equivalent value for a patient of age 60 years, based on linear regression against age within the HC cohort. The secondary analyses were performed using non-age-corrected BP data. Under each primary and secondary analysis, two individual subanalyses were done for parameter related to time: (1) The first subanalysis was performed expressing time as a proportion of the pulse cycle for that participant. (2) A second subanalysis was also performed expressing time in seconds (i.e., without normalizing for pulse rate). 

Results

Demographics

Demographic information for the P3 cohort, grouped into GL and GLS subgroups of eyes, and the HC cohort, is shown in the Table. HC eyes were significantly younger than the GL and GLS eyes (P < 0.0001, GEE regression). The average RNFLT was significantly thicker in the HC eyes than the P3 GL and GLS eyes (P < 0.0001, GEE regression). Linearized mean deviation is significantly worse in P3 GL eyes than HC eyes and GLS eyes (P < 0.0001, GEE regression). 

Table.

 

Demographic Information Listed as Mean (Standard Deviation) for Each Cohort of Study Eyes

 

Comparisons Between HC, GL, and GLS Subjects

The means of each BP parameter within each group are shown in Supplemental Table S1. The corresponding group distributions and comparison statistics for each parameter are plotted in Figure 3. The average BP_ave in the HC subjects was significantly higher in both the GLS subjects (P < 0.0001 from GEE regression) and the GL subjects (P < 0.0001). The average BP_ave of GLS subjects was not significantly different from that of GL subjects. 

Figure 3.

 

Age-corrected parameterizations of the systemic blood pressure wave compared between HC, GLS, and GL. (A–D) Comparisons between non-normalized waveform parameters. (E–H) Normalized waveform parameters (e.g. maximum slope divided by average blood pressure). Red star: significant difference between GLS versus HC (P < 0.05 GEE model), GL versus HC. GL and GLS were not different for any parameter.

GL and GLS subjects showed similar PI and RI after age correction. Relative to HC subjects, the PI and RI parameters were not greater in the GLS subjects and GL subjects (P > 0.05 for all). GLS subjects were not significantly different from GL subjects for either PI or RI (P > 0.05 for all). 

We also assessed the BP waveform change rate (slope) and acceleration parameters expressing time as proportion or using real time in seconds. We found significantly increased FAI and MaxSlope_Norm in the GL and GLS subjects, indicating that the pulsatile waveform ascends more rapidly toward its systolic peak in these GL and GLS subjects; we also saw more extreme MinSlope_Norm in the GL and GLS subjects, indicating that the pulsatile waveform descends faster in these subjects. In addition to the increase in BP MaxSlope_Norm, the GL and GLS subjects also showed increased MaxAccel_Norm and MinAccel_Norm relative to HC subjects, indicating that the BP pulsatile waveform accelerates and decelerates more rapidly while reaching BP systolic peak of the waveform. FAI in HC subjects was significantly lower than FAI in the P3 GL subjects (P < 0.05, GEE regression). Using real time, MaxSlope_Norm was significantly lower in HC subjects than in P3 GLS subjects but not statistically different than in GL subjects. MaxAccel_Norm in HC subjects was significantly lower than MaxAccel_Norm in P3 GLS subjects only (P < 0.05). MinSlope_Norm magnitude were higher in P3 GL and GLS eyes using both normalized time and real time scale (P < 0.05 for both). MinSlope_Norm was significantly less extreme in HC subjects than in P3 GLS subjects (P < 0.05). Finally, MinAccel_Norm was significantly less extreme in HC subjects than in P3 GLS subjects and GL subjects (P < 0.05 for both). 

Correlations Between Each Parameter Versus MD and RNFLT

All parameters are plotted against MD_lin and RNFLT in Figure 4. Correlation coefficients and P values are listed in Supplemental Table S2. PI significantly increased with worse MD with P < 0.001 and tent to correlate with a thinner RNFLT (P = 0.171). RI tent to correlate with a worse MD as well (P = 0.061). 

Figure 4.

 

Age-corrected parameterizations of the systemic blood pressure waveform plotted against linearized mean deviation and RNFLT. Plotting symbols are red for glaucoma eyes; yellow for glaucoma suspect eyes and green for normal healthy control eyes. Regression lines are shown for relations with significant correlations (P < 0.05, GEE model).

MaxSlope_Norm and MinSlope_Norm all significantly increased with worse MD and thinner RNFLT in all subjects (P < 0.05 for all). MinAccel_Norm significantly increased with worse MD, as well as with a thinner RNFLT but with slightly weaker relationships (P < 0.05). 

Secondary Analyses

The means of each raw BP parameter without age-correction are shown in Supplemental Table S3. The corresponding group distributions and comparison statistics for each age-corrected parameter are plotted in Figure 5. Without age correction, all parameters were higher in GL and GLS subjects than HC subjects (P < 0.002 for all). GLS subjects were not significantly different from GL subjects for any of the parameters (P > 0.05 for all). 

Figure 5.

 

Waveform parameters comparisons between HC, GLS, and GL. (A–D) Comparisons between non-normalized waveform parameters. (E–H) Comparisons between normalized waveform parameters. Red star: significant difference between GLS versus HC (P < 0.05 GEE model), GL versus HC. GL and GLS were not different for any parameter.

All parameters are plotted against MD_lin and RNFLT in Figure 6. Correlation coefficients and P values are listed in Supplemental Table S4. In all subjects, BP_ave increased significantly with declining RNFLT (P = 0.013). However, BP_ave was not significantly correlated with MD_lin (P = 0.510, GEE regression). 

Figure 6.

 

Parameterizations of the systemic blood pressure waveform plotted against linearized mean deviation (left column of panels) and RNFLT (right column of panels). Plotting symbols are red for glaucoma eyes; yellow for glaucoma suspect eyes, and green for normal healthy control eyes. Regression lines are shown for relations with significant correlations (P < 0.05, GEE model).

Both PI and RI significantly increased with worse MD and thinner RNFLT, with P < 0.001. FAI, MaxSlope_Norm and MinSlope_Norm all significantly increased with worse MD and thinner RNFLT in all subjects (P < 0.005 for all). MaxAccel_Norm and MinAccel_Norm all significantly increased with worse MD, as well as with a thinner RNFLT (P < 0.05). 

Discussion

Our parameterizations of the pulsatile waveform analysis revealed a relationship between higher systemic vascular resistance (indicated by higher resistance rate and RI) and worse visual field defects and thinner retinal nerve fiber layer. Moreover, glaucoma and glaucoma suspects exhibited a higher rate of change, acceleration, and deceleration toward the systolic peak within the systemic waveform data compared to healthy controls. We also found significantly higher average systemic blood pressure compared to HC subjects. 

The role of systemic blood pressure has been a subject of great interest in glaucoma research, and several studies have demonstrated that fluctuations in systemic blood pressure can increase the risk of glaucoma. This association is particularly prominent in older individuals and those with normal-tension glaucoma. Systemic hypertension is commonly associated with the development of atherosclerosis, which leads to structural changes in the arterial wall and an increase in arterial stiffness. Initially, an elevation in systemic blood pressure may result in an increase in blood flow. However, as the blood vessel walls become narrower because of thickening of the vessel walls, blood flow eventually decreases. Excessively high systemic blood pressure can also disrupt the regulation of blood flow, thus hypertensive people are at increased risk of developing cardiovascular diseases.³¹ Systemic hypertension may contribute to increased IOP via overproduction or impaired outflow of aqueous humor.^32,33 The association between high systemic blood pressure and IOP remains controversial, with some population studies reporting a positive relationship^34–36 whereas others do not.^37,38 

Although various approaches, such as central blood pressure measurements, arterial waveform analysis derived from pulse tonometry, and dynamic retinal vessel reactivity analysis to flickering light, have been used to study arterial stiffness and glaucoma, the reported relationship between systemic arterial stiffness and glaucoma remains controversial. Our study aligns with a cross-sectional study that reported reduced distensibility coefficient of the common carotid artery and baroreflex sensitivity, as well as higher stiffness in patients with POAG compared to controls using the ultrasound wall tracking system. Furthermore, the Rotterdam Eye Study found that participants with an increased pulse wave velocity and, particularly, a low carotid distensibility coefficient indicative of high arterial stiffness had a higher prevalence of POAG, although the results were of borderline significance. Within the eye, we have recently reported that increased vascular resistance, measured using equivalent parameterizations to the systemic measures in this study (but applied to pulsatile blood flow rather than blood pressure waveforms), was correlated with more rapid functional progression in glaucoma suspects.²⁷ By using a different approach, our study provides significant implications for understanding the role of systemic blood pressure waveform in the pathophysiological processes of glaucoma. 

Because of the cross-sectional nature of our study, we cannot determine whether vessel stiffness is the cause of higher systemic blood pressure or if higher systemic blood pressure leads to vessel stiffening. Hypertension and central vessel stiffness have a complex relationship, and the cause and consequence are not clearly known.³⁹ Several factors contribute to the increase in vascular resistance, including the stiffening and loss of elasticity in blood vessels, with underlying causes including age.^19,40–42 hypertension, and metabolic syndrome, chronic renal disease, chronic inflammation, and diabetes.^43–46 Arterial stiffness typically develops with age as the elastic properties of arteries decline. The elastin of the vessel wall thins, splits, frays and breaks into smaller pieces. The collagen content and ground substance increase, often accompanied with calcium deposits.^42,47 Adiposity and its associated metabolic abnormalities have been shown to have strong relations to worse vascular distensibility function in teenagers.⁴⁴ Chronic low-grade inflammation directly or indirectly (exaggerating the aging process) affects the vessel stiffness.^48,49 Diabetes mellitus is associated with increased risk for CV disease and mortality. The pathophysiological mechanism underlying these associations has not been fully elucidated. However, arterial stiffness may be one important pathway linking diabetes to increased CV risk.⁵⁰ Diabetes may enhance arterial stiffness through pathological changes in the vascular bed, such as reduced nitric oxide bioavailability, increased oxidative stress, chronic low-grade inflammation, increased sympathetic tone, and changes in type or structure of elastin or collagen in the arterial wall.^50,51 Finally these structural changes of the arterial wall are also influenced by molecular and genetic factors.^52,53 

Our custom software enables blood pressure analysis by averaging a series of pulsatile blood pressure waves over multiple cardiac cycles. This allows us to parameterize the shape and size of the pulsatile waveform, similar to parameters used in cardiac literature and ocular blood flow studies such as laser speckle flowgraphy. Additionally, we introduce new parameters that describe the maximum and minimum change rates and accelerations, which have not been extensively investigated in glaucoma research. Our study provides a novel approach to measuring hemodynamic differences in systemic blood pressure and its pulsatile variation in individuals with glaucoma. The development of similar and new parameters, in comparison to ocular blood hemodynamics measured by laser speckle flowgraphy or other methods, will facilitate future investigations into the relationship between local and systemic hemodynamics. These parameters can also be used to study systemic autoregulation and may be more sensitive than just using mean systemic blood pressure. 

A major limitation of our study is its cross-sectional design. As a result, we cannot determine whether individuals with glaucoma and glaucoma suspects had previously experienced an increase in blood pressure and vascular resistance as part of the disease process or if they had consistently high blood pressure and vascular resistance. Distinguishing between these possibilities is crucial because it has mechanistic and diagnostic implications. Therefore a longitudinal study is currently underway to address this issue. 

Our research has other limitations. The sample size in our study is relatively small compared to population studies, preventing us from exploring whether alterations in systemic pulse wave shape are solely linked to glaucoma or are merely associated with the diverse cardiovascular disease profiles within the included groups. We included a small number of subjects exhibiting potential systemic hypertension (approximately 5% with measured systolic BP >140 mm Hg). However, the proportion that had previously experienced hypertension may be higher if medication was used to lower blood pressure. We refrained from adjusting for medication use, including IOP-lowering and diabetic medications, because of the complex nature of medication usage in the elderly population, comprising many different medications and dosages. 

Our primary analyses were performed after age-correcting each parameter to their equivalent value for a patient of age 60 years, based on linear regression against age within the HC cohort, under the null hypothesis that age-corrected BP parameters should not differ between cohorts and hence the same age-correction is appropriate for all cohorts. Normalized data for aging tended to reduce effect sizes. However, this may be obscuring true physiological processes. Blood vessels stiffen with age, but the prevalence of glaucoma also increases substantially with age; if this age-related stiffening is part of the reason for that increase in prevalence, then normalizing by age would be detrimental to uncovering that relation. Secondary analyses were performed repeating the above but using raw parameter values (not normalized to age). We found that the majority of raw parameters (particularly both RI and PI) showed significant and larger differences between GL versus HC and between GLS versus HC; in addition, more parameters showed significant associations with MD and RNFLT. Thus we have emphasized both results in the discussion. 

In conclusion, our study demonstrates that the shape of the systemic pulsatile waveform differs in individuals with glaucoma and glaucoma suspects compared to HC subjects. Individuals with glaucoma exhibit more rapid changes in blood pressure, consistent with higher arterial stiffness. The findings regarding systemic vascular resistance, waveform characteristics, and their associations with visual field defects and retinal nerve fiber layer thickness contribute to our understanding of the pathophysiological processes involved in glaucoma. 

Acknowledgments

The authors thank Cindy Albert for her assistance with data collection and Xiue Jiang for her assistance for cleaning the data. 

Supported by NIH R01-EY0312686, R01-EY020922 (to author SKG); unrestricted research support from The Legacy Good Samaritan Foundation, Portland, OR. The sponsors/funding organizations had no role in the design or conduct of this research. 

Disclosure: H. Yang, None; G. Cull, None; M. Yang, None; L. Wang, None; B. Fortune, None; S.K. Gardiner, None