Characterization of intraocular pressure variability in conscious rats

Abstract

Elevation of mean intraocular pressure (IOP) has long been recognized as a leading risk factor for glaucoma. Less is known about the possible contribution of moment-to-moment variations in IOP to disease development and progression due to limitations of tonometry, the prevailing method of IOP measurement. Tonometry provides good estimates of mean IOP but not IOP variance. The aim of this study was to quantitatively characterize IOP variability via round-the-clock IOP telemetry in conscious unrestrained rats. The anterior chamber of one eye was implanted with a microcannula connected to a wireless backpack telemetry system, and IOP data were collected every 4 s for one week. The cannula was then repositioned under the conjunctiva, and control data were collected for an additional week. IOP statistics were computed in 30-min intervals over a 24-h period and averaged across days. All animals exhibited a diurnal variation in mean IOP, while deviations about the mean were independent of time of day. Correlation analysis of the deviations revealed transient and sustained components, which were respectively extracted from IOP records using an event detection algorithm. The amplitude and interval distributions of transient and sustained events were characterized, and their energy content was estimated based on outflow tissue resistance of rat eyes. Transient IOP events occurred ∼231 times per day and were typically ≤5 mmHg in amplitude and 2-8 min in duration, while sustained IOP events occurred ∼16 times per day and were typically ≤5 mmHg in amplitude and 20-60 min in duration. Both persisted but were greatly reduced in control recordings, implying minor contamination of IOP data by motion-induced telemetry noise. Sustained events were also often synchronous across implanted animals, indicating that they were driven by autonomic startle and stress responses or other physiological processes activated by sensory signals in the animal housing environment. Not surprisingly, the total daily fluidic energy applied to resistive outflow pathways was determined primarily by basal IOP level. Nevertheless, transient and sustained fluctuations collectively contributed 6% and diurnal fluctuations contributed 9% to daily IOP energy. It is therefore important to consider the cumulative impact of biomechanical stress that IOP fluctuations apply over time to ocular tissues.

Keywords: Diurnal rhythms; Glaucoma; IOP; Wireless telemetry.

Figure 1.

Event detection algorithm. Schematic pressure waveform containing several peaks, labeled A-F. The algorithm first determines the times and amplitudes of all peaks (filled symbols) and valleys (unfilled symbols) and identifies the subset of peaks (red circles) that satisfies two criteria: minimum peak prominence (minPP) and minimum peak separation (minPS). The former requires peaks to exceed adjacent valleys by a criterion amount (left dashed line), and the latter requires a criterion amount of time between prominent peaks (right dashed line). In the schematic, minPP and minPS are respectively set at 1 mmHg and 1 min so the algorithm would discard peak C because it is not prominent and peak E because it occurs too soon after peak D, which is more prominent. The algorithm then finds the deepest valley between prominent peaks (blue circles), linearly interpolates between the valleys (blue line), and defines events as the area between the interpolated line and pressure waveform (shaded area).

Figure 2.

IOP mean and variance statistics. (A) IOP recorded over one week from a conscious free-moving rat housed under 12-hr light/12-hr dark conditions. (B) Mean IOP change over a 24-hr period averaged across animals in 30-min intervals. Mean changes are relative to the daily resting IOP of each animal. (C) Standard deviation of IOP variability over a 24-hr period averaged across animals in 30-min intervals. Deviations are relative to the mean IOP of each interval. White-black bars indicate the light-dark cycle.

Figure 3.

Temporal correlations in IOP. (A) Data segment showing fast and slow IOP fluctuations in a conscious rat. (B) Control data segment from the same animal showing fluctuations that persist after the cutting the implanted cannula behind the limbus. (C) Auto-correlogram of daytime (10A-2PM) and nighttime (10PM-2AM) IOP data averaged across days for this animal (thick line) and all other animals (thin lines) (N = 12). (D) Left, Auto-correlogram of control cut-cannula data averaged across one week over the same time periods for this and other animals (thick and thin dashed lines, respectively) (N = 4). Right, Average auto-correlogram of cut (dashed line) and uncut (thick solid line) cannula data. The control cut-cannula correlogram was scaled by 0.12 to match the relative variance of the uncut-cannula correlogram. Thin line is a double Gaussian fit of the average auto-correlogram for implanted eyes (σ₁ = 1.2 min, σ₁ = 6.1 min).

Figure 4.

Extraction of transient and sustained IOP events. (A) Representative record of IOP peaks and valleys identified by the event detection algorithm of Fig 1. The algorithm first detected “transient events” using a minimum peak prominence of 1 mmHg and a minimum peak separation of 2 min based on the narrow correlogram peak in Fig 3 (top, red dots/shading), which were then peeled from the IOP record (middle). The algorithm next detected “sustained events” in the record after transient events were peeled off using a minimum peak prominence of 1 mmHg and a minimum peak separation of 20 min based on the broad correlogram peak in Fig 3 (top, blue dots/shading), which were then peeled away as well (bottom). The remainder was considered “baseline” IOP (top, gray shading). (B) 24-hr average transient (top), sustained (middle), and baseline (bottom) waveforms across days and animals. Dashed line marks the minimum baseline level. Shaded areas above and below that line respectively indicate diurnal and basal subcomponents of the baseline waveform. White-black bars indicate the light-dark cycle.

Figure 5.

Statistics of transient IOP events. (A) Probability distribution of transient event amplitudes (upper left) and interevent intervals (lower left) across animals, along with the transient event amplitude (upper right) and interevent interval (lower right) distribution for each animal. (B) Distribution of serial correlation coefficients across animals for transient event amplitude (top) and interevent interval (bottom) records. Filled boxes correspond to the recorded series of transient events and empty boxes correspond to the same series after random shuffling to eliminate potential serial correlation between events. Box edges give upper quartile, median, and lower quartile and whiskers give data range. Asterisks indicate statistical significance (p<0.05).

Figure 6.

Statistics of sustained IOP events. (A) Probability distribution of sustained event amplitudes (upper left) and interevent intervals (lower left) across animals, along with the sustained event amplitude (upper right) and interevent interval (lower right) distribution for each animal. (B) Distribution of serial correlation coefficients across animals for sustained event amplitude (top) and interevent interval (bottom) records. Filled boxes correspond to the recorded series of sustained events and empty boxes correspond to the same series after random shuffling to eliminate potential serial correlation between events. Box edges give upper quartile, median, and lower quartile and whiskers give data range. Asterisks indicate statistical significance (p<0.05).

Figure 7.

Energy content of transient and sustained events. Distributions of IOP energy contained in transient (top) and sustained (bottom) events for individual animals (right) and the population of animals (left). Energy calculations apply to a rat aqueous outflow pathway of standard and fixed resistance. Box edges give upper quartile, median, and lower quartile and whiskers give data range.

Figure 8.

Relative energy in different IOP components. (A) Contribution of transient [T], sustained [S], diurnal [D], and basal [B] components to the total energy in IOP records. Symbols give energy percentages for individual animals. (B) Average daily IOP energy of transient, sustained, diurnal, and basal components (red, blue, gray, and white symbols, respectively) expressed in logarithmic units. (C) Relation of daily transient and sustained energy to diurnal energy. (D) Relation of daily sustained energy to transient energy. Error bars give standard deviations. Dashes are data regression lines.

Figure 9.

Coincident IOP fluctuations in conscious rats. (A) IOP recorded currently over multiple days from two free-moving rats (J105 and J106). White-black bars indicate the light-dark cycle. Asterisks indicate IOP fluctuations that were coincident in both animals. (B) Cross-correlogram of transient (top) and sustained (bottom) events for the pair of IOP records in A (red and blue lines, respectively) and for three random shuffles of event order in one record (gray lines). Dashed line marks the maximum correlation coefficient in 95% of 1000 shuffled cross-correlograms.