Pattern Electroretinograms During the Cold Pressor Test in Normals and Glaucoma Patients

Abstract

Purpose.: To examine the influence of the cold pressor test (CPT) on steady-state pattern electroretinograms (PERG) in healthy subjects and glaucoma patients.

Methods.: Steady-state PERGs to 7.8 Hz pattern reversal stimuli were recorded in 63 subjects. Fifteen healthy control subjects, 14 patients with ocular hypertension (OHT), and 34 patients with open-angle glaucoma (20 normotensive [NTG] and 14 high tension glaucoma [HTG]) were examined. Steady-state PERG amplitudes and latencies were analyzed at baseline, during cold stimulus using a modified CPT, and during the subsequent recovery phase. Blood pressure and heart rate were simultaneously recorded in 10 normals and 11 glaucoma patients.

Results.: During the three test conditions (baseline, ice water, and warm water) glaucoma patients (NTG and HTG) showed significantly reduced PERG amplitudes in comparison with control subjects (P < 0.001) and with OHT patients (P < 0.004). Patients with OHT displayed lower PERG amplitudes than control subjects (nonsignificant, with consideration of Bonferroni). Only NTG patients showed a significant effect of the CPT on PERG amplitude (significant decrease during warm water P = 0.02). Latencies shortened significantly during warm-water period after cold stimulus in control subjects (P = 0.05) and in NTG patients (P = 0.02) with a nonsignificant trend of shortened latencies in the OHT (P = 0.06) and HTG groups (P = 0.3). Systolic and diastolic blood pressure increased during cold water (P < 0.001) and decreased during warming-up conditions (P < 0.001).

Conclusions.: This study shows that a CPT influences the PERG responses particularly in NTG. (ClinicalTrials.gov number, NCT00494923.)

Introduction

The pattern electroretinogram (PERG) is considered to be an early and sensitive objective electrophysiological method for detecting glaucomatous damage of the retinal ganglion cells. ^1–3 The steady-state PERG is a useful tool to study the pathophysiology of glaucoma and its progression. ^1,2,4–6  

Glaucoma is defined as a characteristic optic-neuropathy with a loss of retinal ganglion cells as a result of different well-established risk factors like older age, high IOP, family history of glaucoma, and African ancestry. ⁷ One of the most important and modifiable risk factors is IOP. However, vascular components are considered to be of importance in the pathomechanism of glaucoma, particularly in the absence of elevated IOP, and therefore receive an increasing interest in glaucoma research. ^8–14 Earlier studies have demonstrated a failure of vascular autoregulation in the optic nerve and the retina especially in normotensive glaucoma (NTG) patients. ^10–14 The influence of different conditions, including head-down posture, systemic blood pressure increase, cold pressor test (CPT) on retinal blood flow, and electrophysiological signals indicates the presence of a vascular dysregulation. ^11–18 The CPT is a practical test using a strong and standardized stressor to cause an cardiovascular response. ¹⁹ In a previous study, application of CPT resulted in a significant amplitude reduction of visual-evoked potential measurements (VEPs) of the blue-sensitive pathway in NTG patients, indicating a vascular dysregulation in these patients. ¹²  

The aim of this study was to examine whether the CPT also influences the steady-state PERG and whether this influence is different in healthy subjects and glaucoma patients. 

Methods

Subjects

The study followed the tenets of the Declaration of Helsinki for research involving human subjects and was approved by the local ethics committee. Informed consent was obtained from all participants of the study. Participants of the study were recruited from university staff and from “The Erlangen Glaucoma Registry” (ClinicalTrials.gov number, NCT00494923). ²⁰ The Erlangen Glaucoma Registry is a clinical registry for cross-sectional and longitudinal observation of glaucoma patients and glaucoma suspects that started in 1991. The primary aim of this Registry is the evaluation of diagnostic and prognostic validity of morphometrical, sensory, and hemodynamical diagnostic procedures. Participants are examined annually including standard slit-lamp inspection, Goldmann applanation tonometry, fundoscopy, 30°-white-on-white perimetry, frequency-doubling perimetry technology (FDT), scanning laser flowmetry, VEP, ²¹ and papillometry. ^22,23 The included subjects did not have any eye disease other than glaucoma and did not have any systemic disease with potentially ocular involvement (e.g., diabetes mellitus). They all showed a best corrected visual acuity of 0.5 or better with a myopic or hyperopic refractive error not exceeding 8.0 diopters (D). 

This study included 15 healthy eyes of 15 healthy control subjects (7 females and 8 males, mean age: 55.4 ± 7.67 years), 14 patients with ocular hypertension (OHT, 6 females and 8 males, mean age 53.29 ± 12.54 years), who repeatedly showed IOPs higher than 21 mm Hg without medication, but did not show any glaucomatous functional or morphologic damage, and 34 patients with open-angle glaucoma. The glaucoma group consisted of 14 high-tension glaucoma patients (HTG, 8 females and 6 males, mean age 60.71 ± 7.65 years) and 20 NTG patients (10 females and 10 males, mean age 59.95 ± 13.66 years). The HTG group was characterized by a glaucomatous damage of the optic nerve (defined as an unusually small neuroretinal rim area in relation to the optic disc size and vertical cup-to-disc ratios being larger than horizontal ratios), an IOP higher than 21 mm Hg (30.89 ± 5.59 mmHg) without medication and by an averaged perimetric mean defect (MD) of 6.67 ± 7.6 dB. A white-on-white perimetry was classified as a “nonnormal” visual field when one of the following was present: (1) at least three adjacent test points in the superior or inferior hemifield having a probability of less than 5% and with one test point with a defect of less than 1%, or (2) at least two adjacent test points having a probability of less than 1%. 

The NTG group did not reveal an elevated IOP (20.77 ± 0.87 mm Hg) but showed a glaucomatous damage of the optic nerve and the averaged perimetric mean defect was 6.14 ± 6.31 dB. There was no significant difference in mean MD between the two glaucoma groups (P > 0.1). The eye with the most advanced disease was always chosen for examination. 

Electrophysiology and Stimulus

The steady-state PERG was measured using a Maxwellian view system with a 900-W xenon-arc lamp (OSRAM GmbH, Munich, Germany) as the light source described before in detail by Horn et al. ²⁴ With this stimulus system, retinal luminance is independent of pupil size (as long as the entrance pupil of the system is smaller than the natural pupil) and refraction correction. ^12,21,24,25 All PERGs were measured with natural pupil size, which was always larger than the entrance pupil of the Maxwellian view system (1 mm). We took care that the entrance pupil was centered on the natural pupil. The subject's head was supported by a chin rest and a head band mounted on a three-dimensional micro-adjustment. The pattern was created by inserting a slide of a vertical square wave grating at the retinal plane. The stimulus was focused by adjusting the slide along its optical axis with a remote-controlled step motor (Graupner/SJ GmbH, Kirchheim unter Teck, Germany). The eye's fixation and the position of the electrode were constantly monitored with a television camera (Victor Company of Japan Limited [JVC], Yokohama, Japan). Whenever the entrance pupil did not enter the eye completely the focal point at the pupil plane of the viewing system produced a bright reflex on the edge of the iris, in which case the measurement was interrupted until the head had been readjusted. The circular viewing field was 32° in diameter. The stimulus was a vertical, high contrast (93%), black and white square-wave grating pattern with a spatial frequency of 0.88 cycles per degree. Cross-hairs were placed on the rim of the grating to ensure a stable fixation for the subject. Pattern reversals were created by a mirror connected to a computer controlled motor that switched between two positions, translating the retinal image by half a spatial period of the pattern. The pattern reversal temporal frequency was 7.8 Hz. The mean retinal illuminance was 4263 photopic td. 

The responses were recorded at the cornea with a DTL-microfiber electrode (Statex Produktions & Vertriebs GmbH, Bremen, Germany) placed on the upper side of the lower limbus of the eye. The reference electrode was placed on the ipsilateral temple and the ground electrode was on the forehead. The subjects received local anesthetic eye drops (Conjuncain Edo; Bausch & Lomb GmbH, Berlin, Germany) before placing the DTL electrode and were instructed to suppress blinking. The ERG signals were amplified 10,000 times (EMP88, Electronic Medicine Technique, filter: 0.5–70 Hz; Pölzl, Munich, Germany; no notch filters were used). The responses to 30 sweeps each lasting 512 ms were averaged, digitized (sampling rate: 1000 Hz) and stored. As a result, the averaged responses to four stimulus periods and therefore to eight pattern-reversals were stored. Responses contaminated by eye blinks or eye movements were rejected. The amplitudes and the peak times of the second harmonic component at 15.6 Hz were obtained using fast Fourier analysis of the measured responses. ²⁶ Figure 1 shows examples of steady-state PERG signals in a healthy subject (drawn curve) and in a HTG patient (dotted curve). 

Figure 1

 

Steady-state PERG response during baseline condition of a 53-year-old healthy male participant (drawn curve) and a 53-year-old male HTG patient (dotted curve).

Experimental Procedure

As experimental procedure a modified CPT was used: a 3-minute baseline period, followed by a 3-minute period of cold induction by submerging the right hand of the subject up to the wrists in cold water (between 2°C and 4°C), and finally a 3-minute period of warming up in 30°C to 35°C warm water (Fig. 2). The CPT was modified in order to evoke a strong stimulus for a fast change of blood pressure and retinal blood flow followed by a recovery to baseline conditions by a following warming-up phase. All PERG measurements were performed during the same daytime (between 1 PM and 2 PM) and under constant conditions. Baseline conditions imply a comfortable sitting position in a darkened room with a room temperature of 20°C to 25°C. Four PERGs were recorded per period (baseline, cold water, and warm water). A decrease of PERG amplitude with time is consistent with a slow adaptive change of retinal ganglion cell activity to high-contrast steady state stimuli. ² After approximately 2 minutes, the PERG amplitudes reached a plateau. ²⁷ Therefore, the first PERG recording under baseline condition was discarded and served as an adaptation stimulus. Similarly, the first measurements during the cold induction and the warming-up phases were excluded from analysis. As a result, three PERG recordings were averaged for every period to obtain one averaged PERG. Blood pressure and heart rate were simultaneously recorded with an automatic sphygmomanometer once per experimental period in a subpopulation of 21 subjects (10 normals and 11 glaucoma patients). 

Figure 2

 

Test set-up. Pattern electroretinogram recording was performed during three periods of conditions. Four recordings occurred at baseline and four recordings each while submerging one hand into ice water (2°C–4°C) for 3 minutes and subsequently dipping the same hand into warm water (30°C–35°C) for the rest of time. Because of habituation, recordings .1, .5, and .9 were always excluded and not calculated for the results. Measurements during the three different experimental conditions were averaged for statistical comparison of PERG at baseline (.2, .3, .4), during ice water (.6, .7, .8), and warm water (.10, .11, .12).

To prove short-time intra-individual reliability of the PERG-recordings, nine PERGs were recorded repeatedly during in total 9 minutes under baseline conditions in five healthy control subjects. 

Statistical Analysis

Pattern electroretinogram data (given as means and SDs) were subjected to nonparametric tests because a normal distribution could not be confirmed for some parameters. Comparisons between the groups were performed using Mann-Whitney U test. Comparisons between individual PERG data within the groups were performed using the Wilcoxon and Friedman test. Statistical significance, after Bonferroni correction for multiple testing, was defined as P less than or equal to 0.05. 

Cardiovascular parameters (blood pressure and heart rate) were normally distributed and thus were subjected to Student's t-tests. 

Results

Reproducibility of PERG Recording

The intra-individual reproducibility of the PERG was studied in five healthy subjects. The recording scheme was identical as sketched in Figure 2 with the difference that cold induction and recovery were not performed. Three averaged amplitude and latency values were analyzed using the Friedman test to compare the repeated PERG recordings in the same subjects. The mean amplitude differences were −0.02 μV (±0.26 μV) between the first and second measurement groups and 0.05 μV (±0.48 μV) between second and third measurements groups. Mean latency differences were 0.14 ms (±1.23) and 0.01 ms (±0.67) between first and second and between second and third groups, respectively. These averaged values did not show any statistically significant fluctuation (amplitude: P = 0.819; latency: P = 1.0). Based on these results any statistically significant variation of PERG recorded under experimental conditions was assumed to be generated by the CPT procedure. 

Effects of the Cold Pressor Test on Systemic Cardiovascular Parameters

The effects of the CPT procedure on cardiovascular data are summarized in Figure 3. There were no different reactions between control subjects and glaucoma patients. Combined data of all measured subjects showed an increase of systolic and diastolic blood pressure during cold induction (P < 0.001). A significant decrease of systolic and diastolic blood pressure (P < 0.001) was observed under the subsequent warming-up conditions. The same applies to mean arterial blood pressure. There is an increase from 99.12 ± 11.23 mm Hg to 109.32 ± 11.62 mm Hg (average increase of 10.29% ± 6.06%) during the ice-water period. Afterwards, there was a decrease to 98.58 ± 10.66 mm Hg during the warm-water period (average decrease of 9.82% ± 10.63%). No statistically significant influence was found on heart rate. 

Figure 3

 

Cardiovascular parameters (mean ± confidence limits 95%). Pairwise comparisons show significant increase of systolic, diastolic, and mean arterial blood pressure during ice-water and significant decrease during following warm-water period. No significant influence on heart rate was found (*P < 0.001; Wilcoxon test after Bonferroni correction).

Effects of the Cold Pressor Test on the Pattern Electroretinogram

Comparison Between the Groups.

The PERG amplitude data are summarized in Figure 4. Pattern electroretinogram amplitudes of glaucoma patients were statistically significantly smaller (P < 0.001, Mann-Whitney U test) than those of the healthy subjects during all test conditions (baseline, ice or warm water; Fig. 4). High-tension glaucoma and NTG patients showed similar amplitudes (P = 0.42). Amplitudes of OHT patients were between those of controls and glaucoma patients. However, the amplitude reduction in OHT subjects in comparison to normals was not statistically significant (during baseline P = 0.354, during ice water P = 0.16, during warm water P = 0.310). The amplitude differences in OHT subjects and glaucoma patients were statistically significant (P < 0.004, Mann-Whitney U test). There was no statistically significant latency difference between the groups. 

Figure 4

 

Amplitudes at 16 Hz of the steady state PERG (mean ± confidence limits 95%) of the healthy, OHT, NTG, and HTG groups.

Comparison Within the Groups.

Pairwise comparison of individual amplitudes and latencies, taking the different magnitude into account, within the groups under CPT is presented in Tables 1 and 2 (including paired statistic). 

Table 1

 

PERG Amplitude Change During CPT

Amplitude 16 Hz	Positive	Negative	Equal	P
Normal
Baseline vs. ice	6	8	1	n.s.
Ice vs. warm	6	9	0	n.s.
Warm vs. baseline	5	10	0	n.s.
OHT
Baseline vs. ice	5	9	0	n.s.
Ice vs. warm	6	8	0	n.s.
Warm vs. baseline	3	11	0	n.s.
NTG
Baseline vs. ice	13	7	0	n.s.
Ice vs. warm	5	15	0	0.02
Warm vs. baseline	8	12	0	n.s.
HTG
Baseline vs. ice	9	5	0	n.s.
Ice vs. warm	4	10	0	n.s.
Warm vs. baseline	5	9	0	n.s.

 Pairwise individual comparison of amplitudes at 16 Hz taking the magnitude of difference into consideration. In NTG group there is a significant decrease of amplitude after ice-water period during warm-water condition (P = 0.02, Wilcoxon test with correction according to Bonferroni). Positive, increased amplitude; negative, decreased amplitude; equal, stable amplitude; n.s., nonsignificant.

Table 2

 

PERG Latency Change During CPT

Latency 16 Hz	Positive	Negative	Equal	P
Normal
Baseline vs. ice	4	11	0	n.s.
Ice vs. warm	6	9	0	n.s.
Warm vs. baseline	3	12	0	0.05
OHT
Baseline vs. ice	4	10	0	n.s.
Ice vs. warm	6	8	0	n.s.
Warm vs. baseline	2	12	0	n.s.
NTG
Baseline vs. ice	8	12	0	n.s.
Ice vs. warm	5	15	0	n.s.
Warm vs. baseline	3	17	0	0.02
HTG
Baseline vs. ice	3	10	1	n.s.
Ice vs. warm	6	8	0	n.s.
Warm vs. baseline	4	10	0	n.s.

 Pairwise individual comparison of latency at 16 Hz taking the magnitude of difference into consideration. In the healthy and NTG group there is a significant shortening of latency after ice-water period during warm-water condition (normal group: P = 0.05, NTG group: P = 0.02, Wilcoxon test with correction according to Bonferroni). The other groups show tendency of shortening in latency during test conditions. Positive, prolonged latency; negative, shortened latency; equal, stable latency; n.s., nonsignificant.

Healthy subjects showed no detectable statistically significant amplitude variation in the course of the recordings. Neither the OHT group nor the HTG group displayed a statistically significant change in the amplitude. Pattern electroretinogram amplitude was not significantly influenced by the effect of CPT in these groups. In the NTG group, there was no statistically significant effect on amplitude during the ice-water period, but a significant decrease of the amplitude of 0.19 μV (±0.20 μV; observe that the changes are substantially larger than the changes obtained in the test–retest comparisons given above) during the following warming-up conditions (P = 0.02). 

Pattern electroretinorgram's latencies of healthy subjects were significantly reduced during the warming-up condition (P = 0.05; 0.99 ± 1.02 ms mean differences ± SD). Also, NTG subjects showed a statistically significant latency reduction during the warming-up period (P = 0.02; 1.21 ± 1.43 ms). These latency changes are substantially larger than those found in the reproducibility study (see above). There was a nonsignificant trend of reduced latencies in the OHT (P = 0.06) and the HTG group (P = 0.3) during the warming-up condition. 

Discussion

Glaucoma is a neurodegenerative disease that mainly affects the retinal ganglion cells. The function of these cells can be examined objectively using the steady-state PERG. ^1,2 Disturbed retinal blood flow is an inherent part of the multifactorial pathogenetic concept of glaucoma disease. ⁹ Vascular dysregulation or some vasospastic components might be involved in glaucoma development and are possibly of interest for therapeutic intervention. ^{8,11,18,28,29} This is supported by a previous study that demonstrated changes in the VEP of the blue-sensitive pathway under CPT in NTG patients. ¹² The aim of this study was to investigate the influence of CPT on ganglion cell activity measured with PERG. We hypothesized that there may be differences in CPT-induced PERG changes between normals and glaucoma patients. 

The CPT is a strong and standardized stressor that evokes a cardiovascular response. The CPT was first introduced by Hines and Brown in 1932 ¹⁹ and was applied in cardiovascular research as well as in research on stress and pain. The CPT causes arterial vasoconstriction and an increase in blood pressure by activating the sympathetic autonomic nervous system. ³⁰ In the eye, Nagaoka et al. ¹⁷ detected a transient increase in retinal blood flow and blood velocity induced by an acute increase in mean arterial blood pressure caused by CPT. In the present study, the CPT procedure was modified by adding a following warming-up period to evoke a strong change of blood pressure and retinal blood flow followed by a recovery to baseline conditions during the warming-up phase. 

In agreement with previous data, ^2,5,31–35 we found that PERG amplitudes are reduced in glaucoma patients. Previously, an alteration of the PERG was found in OHT. ^{1,5,6,31,32,36,37} In the present study, a reduced PERG amplitude was also found in OHT patients, but this reduction was not statistically significant. In a long-term observational study, Bach et al. ¹ interpreted the reduced PERG signal as an “early indicator” for OHT eyes to convert to glaucoma. The reduced signal may be caused by dysfunctional ganglion cells. Ventura ² and North ³¹ claim that there is preliminary ganglion cell dysfunction prior to apoptosis possibly related to remodeling processes with shrinking of the dendritic branches and a functional reduction in neuronal sensitivity. ^38–42 Possibly, OHT patients show reduced amplitude due to a reduced mass response of retinal ganglion cells without detectable perimetric defects. As mentioned by North et al. ³¹ electrophysiology might be used to quantify retinal ganglion cell dysfunction that occurs before structural cell death. Our results show a continuous decrease of PERG amplitude depending on the degree of retinal ganglion cell dysfunction. 

In the present study, it is shown that CPT results in reduced PERG amplitude in the NTG group (Table 1). This supports the hypothesis that vascular dysregulation plays an important role in the pathogenesis of NTG. ¹⁰ Maybe there is no, or only a reduced, initial transient increase of retinal blood flow in NTG as a result of a vasospastic reaction to an increase of mean arterial blood pressure. Alternatively, the reaction of the arterial blood pressure to the CPT may be altered in glaucoma patients. ¹¹ The present data do not allow a conclusion on which of the two possibilities is true. For that an analysis of individual IOP, blood pressure, retinal blood flow, and PERG data is needed. Several studies showed disturbed neurovascular coupling in patients with NTG. For instance, Flammer and Orgül, ¹⁰ found an increased prevalence of vasospasms in patients with glaucoma without increased IOP. Kóthy et al. ¹⁴ found an immediate decrease in retinal and optic nerve head perfusion after cutaneous cold provocation in a part of vasospastic subjects. In this study, we found a statistically significant decrease of amplitude during warming-up condition after cold provocation in NTG subjects, which was not found in healthy-, OHT- and HTG-subjects. Possibly, a disturbed response to CPT leads to larger amplitude changes after provocation tests. 

Pattern electroretinogram latency differences between normals and glaucoma are less frequently reported. ⁴³ Korth et al. ^44,45 found an increase of latency with age. We excluded this influence by matching the groups by age. Porciatti et al. ³⁴ suggested that dendritic dysfunction or delay of axonal transport may be responsible for a response delay, because electrical signals take more time to be generated. In this study, no statistically significant latency difference during baseline conditions was found between normals and glaucoma subjects. However, during CPT, we found a trend toward shorter latencies in all groups, with statistical significance only in the OHT and NTG groups. This may be explained by a faster generation of the PERG signal because of a transient increase of retinal blood flow as a reaction to temperature stimuli. Nagaoka et al. ¹⁷ showed that CPT induces an increase of mean arterial blood pressure and a transient increase of the retinal blood flow. Porciatti and Ventura ³⁴ proposes that the increase latency was caused by a slower generation of the electrical signal in activated neurons. Conversely, a CPT-induced latency decrease suggests a faster electrical activity signal possibly caused by an increased retinal blood flow. 

In conclusion, our data confirm that PERG amplitude is significantly smaller in glaucoma subjects during normal test conditions and remains reduced during a simultaneous CPT. In the NTG subgroup, the amplitude changed in the warming-up phase of our modified CPT protocol, possibly as a result of a disturbed vasomotoric reaction. In OHT subjects, the PERG-amplitude is slightly reduced to a level between that of normals and glaucoma subjects. In all subjects, CPT possibly leads to a decreased latency. In this regard, further electrophysiological studies combined with perfusion influencing tests could help to learn more about the vascular pathogenic component of glaucoma. 

Acknowledgments

The present work was performed in fulfillment of the requirements for obtaining the degree Dr. med. 

Disclosure: A. La Mancusa, None; F.K. Horn, None; J. Kremers, None; C. Huchzermeyer, None; M. Rudolph, None; A. Jünemann, None