24-hour iop telemetry in the nonhuman primate:...

24-Hour IOP Telemetry in the Nonhuman Primate:Implant System Performance and InitialCharacterization of IOP at Multiple Timescales

J. Crawford Downs,1 Claude F. Burgoyne,2 William P. Seigfreid,1,3 Juan F. Reynaud,1,2

Nicholas G. Strouthidis,4 and Verney Sallee5

PURPOSE. IOP is the most common independent risk factor fordevelopment and progression of glaucoma, but very little isknown about IOP dynamics. Continuous IOP telemetry wasused in three nonhuman primates to characterize IOP dynam-ics at multiple time scales for multiple 24-hour periods.

METHODS. An existing implantable telemetric pressure trans-ducer system was adapted to monitoring anterior chamberIOP. The system records 500 IOP, ECG, and body temperaturemeasurements per second and compensates for barometricpressure in real time. The continuous IOP signal was digitallyfiltered for noise and dropout and reported using time-windowaveraging for 19, 18, and 4 24-hour periods in three animals,respectively. Those data were analyzed for a nycthemeral pat-tern within each animal.

RESULTS. Ten-minute time-window averaging for multiple 24-hour periods showed that IOP fluctuated from 7 to 14 mm Hgduring the day, and those changes occurred frequently andquickly. Two-hour time-window averages of IOP for multiple24-hour periods in three animals showed a weak nycthemeraltrend, but IOP was not repeatable from day-to-day withinanimals.

CONCLUSIONS. The measured IOP was successfully measuredcontinuously by using a new, fully implantable IOP telemetrysystem. IOP fluctuates as much as 10 mm Hg from day to dayand hour to hour in unrestrained nonhuman primates, whichindicates that snapshot IOP measurements may be inadequateto capture the true dynamic character of IOP. The distribu-tions, magnitudes, and patterns of IOP are not reproduciblefrom day to day within animals, but IOP tends to be slightlyhigher at night when IOP data are averaged across multiple

24-hour periods within animals. (Invest Ophthalmol Vis Sci.2011;52:7365–7375) DOI:10.1167/iovs.11-7955

IOP and patient age are the most frequently identified inde-pendent risk factors for development and progression of

glaucoma in the major prospective clinical trials.1–5 LoweringIOP is the only clinical treatment that has been shown to retardthe onset and progression of glaucoma, but once damaged,the optic nerve head (ONH) is thought to be more suscep-tible to further glaucomatous progression, even after inter-vention has lowered IOP to epidemiologically determinednormal levels.6,7 In addition to data from prospective trialsin glaucoma and ocular hypertension, every population-based survey conducted to date has demonstrated a strongrelationship between the prevalence of glaucoma and ad-vancing age,8 even though most studies show little or nochange in IOP with age,9 –13 except in black patients.14,15

Furthermore, normal-tension glaucoma is frequently seen inelderly individuals,16,17 but is not seen in children or youngadults other than in a few isolated cases.18

We have previously interpreted these findings to meanthat the aging ONH becomes increasingly vulnerable toglaucomatous injury at similar levels of mean IOP and thatIOP and aging are truly independent risk factors for glau-coma. Although there is evidence that the aged eye is moresusceptible to damage at all levels of mean IOP,19 unmea-sured components of IOP may also be playing a role. High-frequency IOP fluctuations become larger as IOP increases,presumably driven by stretching and stiffening of the ocularcoats, which reduces elastic damping. In addition, we andothers have shown that the ocular coats stiffen with age inmonkeys and humans20 –22 and stiffen after exposure tochronically elevated IOP in a monkey model of glau-coma.23,24 From these data we could conclude that IOPfluctuations will be larger in both elderly and glaucomatouseyes in which the ocular coats have stiffened due to age-related changes and/or IOP-induced remodelling. We there-fore hypothesize that there are unknown age- and disease-related components of IOP (such as greater IOP fluctuation)that independently contribute to the onset and progressionof glaucoma in those at-risk populations.

Very little is known about IOP dynamics, as IOP is typicallymeasured clinically with a snapshot method (Goldmann),which is repeated every 3 to 12 months during clinic hours.IOP fluctuation has been calculated in some studies as theintervisit variation in IOP (monthly) or as diurnal or 24-hourcurves (hourly). In the intervisit IOP studies, the evidence forlong-term IOP fluctuation as a risk factor for visual field pro-gression in glaucoma is inconclusive.7,25–27 Diurnal/nycthem-eral IOP fluctuations with more frequent measurements(hourly) have been considered a risk factor for progression, butagain, results are inconclusive.28–30 The most recent evidence

From the 1Ocular Biomechanics Laboratory and the 2Optic NerveHead Research Laboratory, Devers Eye Institute, Portland, Oregon; the3Electrical and Computer Engineering, University of Missouri, Colum-bia, Missouri; the 4NIHR (National Institute of Health Research) Bio-medical Sciences Centre, Moorfields Eye Hospital and UCL Institute ofOphthalmology, London, United Kingdom; and 5RTOP Consulting,Colorado Springs, Colorado.

Presented in part at the annual meeting of the Association forResearch in Vision and Ophthalmology, Fort Lauderdale, Florida, May2011.

Supported by National Institutes of Health Grant R21-EY016149(JCD) and the Legacy Good Samaritan Foundation, Portland, Oregon

Submitted for publication May 28, 2011; revised June 9, 2011;accepted June 26, 2011.

Disclosure: J.C. Downs, None; C.F. Burgoyne, None; W.P.Seigfreid, None; J.F. Reynaud, None; N.G. Strouthidis, None; V.Sallee, None

Corresponding author: J. Crawford Downs, Associate Scientist,Research Director, Ocular Biomechanics Laboratory, Devers Eye Insti-tute, 1225 NE 2nd Avenue, Portland, OR 97232;[email protected].

Glaucoma

Investigative Ophthalmology & Visual Science, September 2011, Vol. 52, No. 10Copyright 2011 The Association for Research in Vision and Ophthalmology, Inc. 7365

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933248/ on 06/01/2018

suggests that while a nycthemeral IOP rhythm is present inhumans, the IOP curves are not reproducible within individu-als.31,32

The Pascal dynamic contour tonometer is capable of mea-suring IOP continuously for short periods in humans,33,34 butit relies on corneal contact and requires the patient to remainstill and avoid eye movements and blinks, both of which aremajor sources of IOP fluctuation.35 Some studies have shownthat IOP is much more dynamic than can be measured by asnapshot device.35,36 A recent study of a new contact lens–based IOP telemetry system has shown promising results inpatients and shows that IOP is very dynamic.37 The contactlens–based IOP telemetry systems for humans measure therelative change in IOP in arbitrary units, which cannot easilybe translated to absolute IOPs. Several studies have reportedthe results of IOP telemetry in rabbits38–40 and rodents,41 butthe devices used therein (DSI, St. Paul, MN) cannot be used forlong-term studies because of their short service life. Also,although the pressure telemetry systems used in these studies(DSI) are appropriate for rabbit and rodent work, they are notsuitable for use in nonhuman primates because of their veryshort signal transmission range (�18 in.) and the fact that thetransducer cannot be located on or near the monitored eye, adesign that lends itself to hydrostatic pressure errors with headtilt.

To begin investigating IOP dynamics and their relation-ship to glaucomatous onset and progression in the primate,we developed a fully implantable telemetry system thatallows continuous wireless monitoring of IOP measured inthe anterior chamber, ECG, and body temperature, withmeasurements obtained 500 times per second, 24 h/d, 7d/wk, for up to 9 months. The purpose of the present studyis to report the initial performance specifications of thissystem and to characterize the dynamics of IOP over multi-ple time scales for multiple 24-hour periods in three normalnonhuman primates.

MATERIALS AND METHODS

Animals

All animals were treated in accordance with the ARVO Statement forthe Use of Animals in Ophthalmic and Vision Research under a proto-col approved and monitored by our Institutional Animal Care and UseCommittee. Three young adult female rhesus macaques, 4 to 7 yearsold, weighing 5.5 to 7.5 kg and having no ocular abnormality orprevious ocular surgery, were used for the study. After implantation,the animals were kept on a 7 AM-to-7 PM light–dark cycle and fed atapproximately 8 AM and 5 PM daily. Water was provided via continu-ous feed and was always available. Food and water intake were notmeasured.

The Telemetric IOP Monitoring System

The current version of the IOP telemetry implant is based on anexisting, commercially available, battery-powered, implantable pres-sure monitor of proven design42 (T30F-13B; Konigsberg Instruments,Pasadena, CA). To enable IOP telemetry, a standard P4 pressure trans-ducer (Konigsberg Instruments) was integrated into a custom-designedsilicone baseplate that is secured into the orbital wall with bone screwsand connected to the anterior chamber via a 35-mm-long, 23-gaugesilicone aqueous transduction tube (Fig. 1).

The P4-based IOP transducer consists of a thin, deformable, 4.0-mm-diameter titanium membrane, hermetically laser-welded into atitanium housing. When the titanium membrane is placed in contactwith a fluid, it deforms with the fluctuations in pressure of that fluid.Strain gauges on the back of the transducer membrane register thesedeformations, and the resulting electrical responses of the gauges arefed to the transmitter electronics package via silicone-encased, heli-cally wound, braided, stainless steel lead wires. The implant’s trans-mitter electronics package transmits a multiplexed analog signal to anantenna up to 3 m away that is connected to a signal processor (TD-14Basestation; RMISS, Wilmington, DE). The processor digitally samplesthe IOP, ECG, and body temperature measurements 500 times persecond and adjusts for barometric pressure in real time by subtracting

FIGURE 1. (A) A typical T30F totalimplant system (Konigsberg Instru-ments, Pasadena, CA) showing thebattery/transmitter module, radio fre-quency (RF) ring antenna for on/off,transmission antenna, a pressuretransducer, and two ECG electrodesplus ground. (B) The extraorbital sur-face of the custom IOP transducerhousing, which was secured within a1/4-in. hole in the lateral orbital wallwith bone screws, as shown in (C). A23-gauge silicone tube deliveredaqueous from the anterior chamberto a fluid reservoir on the intraorbitalside of the transducer (partially hid-den from view in B); The tube (withappropriate slack to allow for eyemovement) was trimmed, insertedinto the anterior chamber, sutured tothe sclera using the integral scleraltube anchor plate, and covered witha scleral patch graft (not shown).

7366 Downs et al. IOVS, September 2011, Vol. 52, No. 10


the ambient air pressure from the IOP signal (Fig. 2). As the eye and thetransducer housing are closed pressure vessels, this adjustment en-sures that the recorded values represent the true pressures in the eye,over and above the ambient atmospheric pressure. Once decoded andcorrected for ambient air pressure, the IOP data are streamed to adual-redundant industrial data-acquisition system (CA Recorder; DISS,Dexter, MI) that stores the data and allows for visualization and initialprocessing (Fig. 2). The IOP data are then sent to the Ocular Biome-chanics Laboratory server for upload and further in-house custompostprocessing.

Surgical Implantation of the IOP TelemetryImplant System

Each monkey received preprocedural IM analgesics (2.2 mg/kg keto-profen) for 24 hours before surgery, to alleviate pain during theprocedure. On the morning of the surgery, under ketamine/xylazineanesthesia, each animal was shaved and intubated, given IM antibiotics(5 mg/kg Baytril; Bayer Animal Health, Bayer, AG, Leverkusen, Ger-many), and transferred to a sterile surgical suite where isofluraneanesthesia and fentanyl continuous drip analgesia were initiated. Theanimal was ventilated, and blood pressure and oxygen saturation weremonitored throughout the procedure. The abdomen, head, and skinoverlying the SC course of wires (between the abdomen and orbit) wasprepped and draped in a sterile fashion.

Because the transducer can be damaged by intraoperative cautery,the procedure was staged as follows. First, with bipolar cautery used tomaintain hemostasis, the lateral orbital wall was isolated and a 1⁄4-in.-diameter hole was drilled through the bone with a sterile flex-shaftdrill. The animal was then repositioned on its back beneath an oper-ating microscope, allowing the eye and lateral orbital wall to beprepped and draped. A 6-0 Vicryl traction suture was passed throughthe superotemporal peripheral cornea, and the eye was rotated untilthe superotemporal quadrant was centered in the operative field.Lidocaine and epinephrine were injected beneath Tenon’s fascia 8 mmposterior to the limbus. An 8-mm incision was made through theconjunctiva and Tenon’s to bare sclera, followed by dissection of theselayers to the limbus. A small, curved hemostat was passed into thewound posteriorly until the hole in the lateral orbital wall was reachedand the orbital fascia gently protruded. An assistant carefully cutthrough the orbital tissues until a 4-0 nylon suture could be placed intothe working end of the hemostat, after which the hemostat waswithdrawn through the conjunctival incision, leaving the suture tomark the future passage of the 23-gauge aqueous tube of the trans-ducer. All cautery was then removed from the surgical field.

The transceiver module was then inserted through a 5 cm abdom-inal wall incision into a pocket between the peritoneum and abdomi-nal wall muscles created by blunt dissection. The ECG leads were

routed via SC blunt dissection to their appropriate locations on thechest wall and sutured in place. The ocular transducer baseplate(placed in a protective bonnet to avoid contamination and damage)and attached leadwire were threaded SC via blunt dissected tunnelsfrom the abdomen to the back, up the posterior torso and neck, andover the ear to the temple in a serpentine pattern that providessufficient slack to accommodate subsequent animal movement. Multi-ple incisions through the skin were opened and closed to pass thetransducer.

With the animal on its back, the ocular end of the 4-0 nylon suturewas again grasped by the hemostat and used to gently pull the hemo-stat tips through the orbital wall hole. Then, the scleral anchor plate onthe anterior chamber end of the aqueous transduction tube (Fig. 1B)was grasped between the hemostat tips gently pulled back through thehole into the sub-Tenon’s space between the superior and lateralrectus muscles. A loose loop of redundant tube located between thetransducer housing and the scleral anchor plate was pushed back intothe orbit to accommodate subsequent eye movement (Figs. 1C, 2).

The transducer aqueous reservoir was then placed into the orbitalbone window, and the transducer baseplate was secured to the orbitalwall with three size 000 � 3/32-in. stainless steel bone screws (Fig.1C). The aqueous transduction tube was then trimmed and insertedthrough the sclera into the anterior chamber, just superior and parallelto the iris via a 23-gauge needle hole, with care taken to position thetube as close to the iris surface as possible. The aqueous transductiontube was sutured to the sclera with two 8-0 nylon sutures securedthrough the tube’s integrated scleral anchor plate, plus an additionalbutterfly mattress suture. The tube and its fixation plate were thencovered with transgraft human donor sclera. The conjunctiva andTenon’s fascia were closed in a single layer with a running locked 8-0Vicryl suture. A sub-Tenon’s injection of dexamethasone, gentamicin,and cefazolin was given in the inferior fornix. The muscles and fasciaaround the orbital entrance site were closed with deep and superficial3-0 Prolene sutures, and the skin was closed SC with buried 5-0 Vicrylsutures. All wounds were coated with antibiotic ointment. Each animalwas monitored until it regained a swallow reflex, and then it wasextubated and taken back to its cage for recovery.

Each animal received IM antibiotics (5 mg/kg Baytril; Bayer AnimalHealth) once daily and postoperative IM pain medications (0.02 mg/kgbuprenorphine) twice daily for a minimum of 5 days and was exam-ined in its cage each postoperative day for 1 week. Examination at theslit lamp and the first calibration test were performed at postoperativedays 5 to 7, depending on how each animal recovered from anesthesiaand surgery. Slit lamp examinations occurred approximately every 2weeks thereafter (at each IOP transducer calibration test; see below)for the remainder of the study.

FIGURE 2. The IOP telemetry system.

IOVS, September 2011, Vol. 52, No. 10 IOP Telemetry in the Nonhuman Primate 7367


IOP Transducer CalibrationApproximately every 2 weeks, each animal was placed under isoflu-rane anesthesia, and the implanted eye was cannulated with a 27-gaugeneedle placed through the cornea into the anterior chamber. Theneedle was connected to a bottle of sterile isotonic saline solution viaa sterile infusion set, and the connecting tube was fitted with an in-line,digital pressure gauge (XPi; Crystal Engineering, San Luis Obispo, CA).The manometer bottle was lowered to a level such that the in-linepressure gauge read 5 mm Hg and was allowed to stabilize. Thetelemetric IOP reading from the implanted transducer was then re-corded for comparison. The manometer-controlled IOP was raised inincrements of 5 mm Hg up to 45 mm Hg, and the telemetric IOPreading was compared to that from the digital in-line pressure gauge ateach step. The telemetric and gauge IOP values were recorded andcompared, to quantify IOP transducer signal accuracy and drift. Thecalibration file was updated on the signal processor (TD-14 Basestation;RMISS) after each IOP calibration check, such that the transducer IOPread accurately in the system.

IOP Data AcquisitionAll analog IOP data were decoded, adjusted for ambient barometricpressure in real-time, digitally sampled at 500 Hz, and stored on thedata-acquisition system computers (CA Recorder; DISS). The system islimited to recording a maximum of 99 hours of data in each session,but it is recommended that only 24-hour increments be collected tominimize the possibility of write-to-disc errors. Hence, data were re-corded in 24-hour blocks starting at 9 AM, several days per week,generally every other week. This schedule was maintained as closely aspossible, but was periodically interrupted for IOP transducer calibra-tion tests and ocular examinations, where the anesthesia and/or ante-rior chamber cannulation disrupted normal IOP. Data collection wasavoided for at least 24 hours after these events, to allow IOP to recoverto baseline levels. Hence, while the system is capable of recording IOPcontinuously, we captured data in 24-hour blocks approximately 6days per month.

Digital Filtering and Noise EliminationAfter collection and real-time barometric pressure compensation, theIOP signal was extracted from the file and adjusted according to thedata from the IOP calibration tests. The IOP signal was assumed to havedrifted linearly between transducer calibration tests, and the data fromeach acquisition session were continually offset by the drift calculatedfor that period. For example, if the IOP data were acquired at day 7 ofthe 14-day period that ended in a transducer calibration test indicatingthat the IOP transducer signal had drifted �2 mm Hg away fromreference in 14 days, then from 0.86 to 1.0 mm Hg was subtracted fromthe signal for that day (i.e., a 2-mm Hg drift over 14 days equals a0.14-mm Hg drift per day, equating to an 0.86-mm Hg drift at the startof day 7 and a 1.0-mm Hg at the end).

The continuous IOP signal is subject to signal noise and loss whenthe animal moves too far from the antenna or orients its body in a waythat the signal has to pass through most of the animal’s body to reachthe antenna. As a result, the data were digitally filtered to eliminatenoise and signal dropout. IOP data were flagged as suspicious if thetransducer read less than 0 mm Hg or greater than 100 mm Hg. Thesuspect segment, along with five seconds of data both before and afterthe suspect segment, were eliminated from the analysis and resultspresented in this study. This was done to ensure data quality andminimize the possibility of reporting erroneous IOP data. Only 24-hourblocks wherein 80% of the data in all the 2-hour time windowsremained after filtering were used for the analyses. Only one 24-hourperiod (compared with a total of 41 reported herein) did not meet thecriteria.

Data AnalysisAfter the IOP data were adjusted according to the IOP transducercalibration tests and digitally filtered, they were analyzed by using a

time-window–averaging technique. For each 24-hour block, the datawere separated into time windows of 10 minutes and 2 hours, and theaverage IOP for each time window were plotted against time. Theaveraged data points were assigned a continuous color from green tored, with green, yellow, and red indicating that 0%, 50%, and 100% ofthe data were eliminated for noise and/or dropout within that partic-ular time window, respectively. This scheme immediately shows thequality of the data for each time-window average in each 24-hour blockon inspection.

The two-hour time-window averaged data were then statisticallyanalyzed for the presence of a repeatable nycthemeral rhythm. First,each 24-hour block of IOP data were plotted in 2-hour time windows,using a box-and-whisker plots to represent the distribution of thecalibrated and filtered IOP measurements (500 per second) within that24-hour period. These are presented in the Results section for 19 and18 days for two animals, respectively, and 4 days for a third animal inwhich the IOP transducer portion of the implant failed for unknownreasons 5 weeks after implantation. Hence 4 days of reliable data wereall that were acquired for that animal. The data for each 2-hour period(e.g., 9–11 AM, 11 AM–1 PM, and so on) were then averaged across allthe 24-hour blocks for each animal and represented by box-and-whis-ker plots showing the distribution of continuous IOP within each2-hour period across 19, 18, and 4 days for the three animals.

RESULTS

Surgical Outcomes

We implanted the final prototype of the IOP transducer-equipped implanted pressure monitor (T30F; Konigsberg In-struments) in the right eyes of three rhesus macaques. Allsurgical wounds healed well, with minor areas of dehiscence inthe lateral orbital wall skin wounds in two animals that resolvedwith placement of supplemental interrupted sutures where nec-essary. Orbital congestion resolved and chewing of hard foodresumed within the first week of surgery in all three animals. Nogaze restrictions were noted by direct observation or on forcedduction, as performed during each IOP calibration session.

Postoperative anterior segment inflammation was moderate(grade 2� cell and flare, but no fibrin) and resolved completelyin two of the three animals within the first two postoperativeweeks. The third animal (monkey 18012) exhibited continuouslow-grade anterior chamber inflammation (trace anterior seg-ment cell and flare without evidence of vitreous involvement,cataract, or fibrin formation) in the implanted eye for the first3 months of the 7-month period of postoperative observation.Within 3 weeks after surgery, tonometric measures of IOP in allimplanted eyes (Tonopen; Reichert Technologies, Depew, NY)were not markedly different from those in the contralateralnonimplanted eye.

FIGURE 3. Screen capture of approximately 6 seconds of the contin-uous IOP telemetry signal in an awake, unrestrained nonhuman pri-mate, showing the dynamic nature of IOP and demonstrating theability of the system to capture high-frequency IOP fluctuations. Apreliminary analysis of these fluctuations suggested that they resultedfrom blinks and saccades (Seigfreid WP, et al. IOVS 2011;52:ARVOE-Abstract 656), but that has yet to be confirmed.



Total Implant System Performance

Data were collected in two of the three animals for approx-imately 7 months until battery failure. The IOP transducerfailed from unknown causes in the third animal 5 weeksafter implantation. We tested IOP transducer accuracy from5 to 45 mm Hg in the three implanted monkeys by cannu-lating the anterior chamber with a 27-gauge needle con-nected to an adjustable saline reservoir equipped with adigital in-line pressure gauge. We recorded continuous te-lemetric IOP and ECG data during IOP calibration tests whilethe animals were under isoflurane anesthesia, with steadyheart rates between 110 and 130 beats/min, O2 saturation

above 95%, and blood pressures of �110/65 mm Hg. TheIOP transducers in all three animals were accurate to within1 mm Hg, from 5 to 45 mm Hg, with a base noise level of�0.4 mm Hg. The transducers measured transient IOP ele-vations up to 20 mm Hg above baseline in response tocorneal tonometer contact, and were extremely sensitive toany manipulation of the eye or eyelids. High-frequency IOPfluctuations were captured with the system (Fig. 3), whichour preliminary data indicate were due to saccades andblinks. The ocular pulse amplitude (OPA) was readily dis-cernable in the IOP signal when OPA was greater thanapproximately 0.6 mm Hg (Fig. 4).

FIGURE 4. Screen captures of typi-cal telemetric ECG and IOP signalsfor IOPs of 3.5 up to 54 mm Hg inmonkey 18012, showing ocularpulse amplitude (OPA) increasingwith IOP. The IOP is scaled identi-cally for each tracing, and all thesedata were captured within approxi-mately 10 minutes in an anesthetizedanimal during an IOP transducer cal-ibration test. Note the transducernoise of �0.4 mm Hg in the top IOPtrace at 3.5 mm Hg baseline IOP.



We evaluated the effects of head position, manipulation ofthe SC lead wires, hard tapping on the transducer electronicspackage, hard tapping on the orbital ridge, and hard tapping onthe tissue above the transducer itself. None of these perturba-tions induced any change whatsoever in the IOP signal of theanesthetized animals. Forced ductions performed by gentlygrasping the conjunctiva with forceps induced IOP elevationsfrom 8 to 15 mm Hg above baseline in anesthetized animals.Finally, the system records transient IOP elevations as short as20 ms (Fig. 3), which is sufficient to capture IOP fluctuationsfrom blink, saccade, ocular pulse, and other sources (see Sup-plementary Movie S1, http://www.iovs.org/lookup/suppl/doi:10.1167/iovs.11-7955/-/DCSupplemental).

IOP Transducer Signal Noise and Drift

A very important feature of the pressure transducer (T30F;Konigsberg Instruments) design is that its signal noise is lessthan 0.4 mm Hg, and signal drift is rated at less than 3 mmHg/mo monotonic. As such, signal drift errors can be sub-tracted automatically in the software (CA Recorder; DISS) aslong as pressure calibration checks are routinely performed(performed approximately every 2 weeks in this study). Posthoc adjustment for IOP drift between IOP transducer cali-bration tests was accomplished using digital signal compen-sation, as described above. All three transducers exhibitedan initial postimplantation period of high drift (up to 17 mmHg per month) for periods of up to 6 weeks, after which twoof the three implants settled into a monotonic drift patternof 6 mm Hg per month or less. The IOP transducer inmonkey 24251, whose implant failed 5 weeks after implan-tation, never settled into the lower drift phase. IOP signalnoise was �0.4 mm Hg in all three implants for all datacollected (Fig. 4, top IOP trace).

IOP Dynamics

IOP was plotted in 10-minute and 2-hour time windows for19, 18, and 4 24-hour blocks for monkeys 18012, 21356, and24251, respectively. Figure 5 shows a representative plot ofthe 10-minute time-window average plot of IOP for a single24-hour period from animal 18012. Figures 6, 7, and 8, showthe 2-hour time-window average box-and-whisker plots forall 24-hour blocks collected for each animal, respectively.Calibrations were performed routinely, and so these datashould be accurate to within 1 mm Hg of true IOP. The datesof the 24-hour recording periods are shown above each24-hour plot; note that many of the plots are for Monday,Wednesday, and Friday of the same week. When data arecompared within animals across 24-hour periods in Figures

6 to 8, there does not seem to be any particular nycthemeralrhythm or repeatable pattern for any of these animals.

Nycthemeral Rhythm of IOP

All the data for the 2-hour time window plots were com-bined into single box-and-whisker plots for animals 18012,21356, and 24251, as shown in Figures 9, 10, and 11,respectively. When all the data are combined, there is aweak trend of nocturnal elevation in the IOP distributions(in the boxes and whiskers, but not necessarily the mean) inall three animals.

DISCUSSION

In this study, we have described a new, fully implantableIOP telemetry system for nonhuman primates that allowscontinuous monitoring of IOP, ECG, and body temperatureat a rate of 500 measurements per second for up to 7months. Note that the T30F pressure monitor implants thatwe used for this study are designed for short-term use of 6to 9 months, but the equivalent T27G-series implant system(both by (Konigsberg Instruments) equipped with a sepa-rate battery can achieve 2.5 years of continuous monitoring.The transducer sits in the orbital wall adjacent to the eye,and so artifacts due to head movement are minimized—thatis, the transducer moves with the head as the head moves,and there is no hydrostatic pressure head due to differencein height. IOP measurements are taken through a tube in-serted into the anterior chamber, which eliminates the ret-inal damage associated with probes placed in the vitreouschamber. After implantation in three nonhuman primates,the IOP telemetry system exhibited high accuracy (�1 mmHg of error over the physiologic range), low noise (�0.2 mmHg), and the ability to capture IOP fluctuations on the orderof 10s of milliseconds. Two of the three IOP transducerssettled into a monotonic drift phase of �6 mm Hg permonth after an initial 1- to 2-month period of high drift. Thethird IOP transducer failed 5 weeks after implantation andnever entered the low-drift phase.

The continuous IOP signal was digitally filtered for noiseand dropout, and reported by using time-window averagingfor 19, 18, and 4 24-hour periods in three animals. IOPtransducer calibrations were performed routinely, and thesedata should therefore be accurate to within 1 mm Hg of trueIOP (see the Limitations section for a discussion of drift inanimal 24251). Continuous IOP was extremely dynamic anddemonstrated frequent, short-duration IOP fluctuations upto 12 mm Hg (Fig. 3), although the source of these fluctua-

FIGURE 5. (A) Representative plot of the 10-minute time-window average of 24 hours of continuous IOP showing low-frequency IOP fluctuationfrom monkey 18012. The color of the plot points and lines indicates how much data were removed from each 10-minute window after post hocfiltering signal dropout and noise. Green indicates that 100% of the continuous IOP data were used in the 10-minute average IOP plotted in eachpoint, and yellow indicates that 50% were eliminated due to signal dropout or noise. (B) Histogram of IOP for the 24-hour period presented in (A).Note the width of the IOP distribution within this single 24-hour period relative to the median value of 11 mm Hg.



tions has yet to be validated. This observation is consistentwith the relative IOP fluctuation observed in humans fittedwith a contact lens– based IOP telemetry system37 that usescorneal stretch as a surrogate measure for IOP. Ten-minutetime-window averaging for 24-hour periods showed that IOPfluctuated up to �7 mm Hg from baseline during a 24-hourperiod, and those changes occurred frequently and quickly(Fig. 5). Again, this is consistent with human results shownby Mansouri and Shaarawy.37

Two-hour time-window averaging of continuous IOP inmultiple 24-hour periods for three nonhuman primates re-vealed two important observations. First, there did not seemto be any repeatable day-to-day pattern of IOP fluctuation inthese animals (Figs. 6 – 8), which agrees with recent reportsin patients.31,32 The dates of the 24-hour recording periodsare noted above each 2-hour time window plot shown, andmany of the plots are for Monday, Wednesday, and Friday of

the same week. In a significant number of the 24-hourperiods shown in Figures 6 to 8, the whiskers (95% limits)indicate that that IOP was more than 3 mm Hg higher orlower than the mean value for at least 12 minutes during the2-hour block. In some cases, these fluctuations were up to�5 mm Hg from the mean.

Second, 2-hour time window data averaged across multiple24-hour periods in three animals show a nycthemeral trend ofIOP elevation in the early morning hours (Figs. 9–11), althoughthis trend is neither statistically significant nor as strong as thatreported for humans.32,43,44 The lack of a more prominentnycthemeral rhythm in these animals could be caused by theirupright position during sleep. It is also possible that the mon-keys are more active at night than humans and therefore have alesser nocturnal rise in IOP that that of humans, who sleep deeplythroughout the night. Humans sleep supine, which has beenshown to increase IOP,44–46 although one study indicates that the

FIGURE 6. The 2-hour time-windowdistributions of IOP for 19 24-hourperiods in animal 18012. The dates ofdata capture are shown above eachplot. In the box-and-whisker plots,the central bar indicates the meanIOP in each 2-hour segment, the endsof the box show the central 50% ofthe measurements, and the whiskersindicate the 95% limits of the mea-surements in that time window; (E)outliers.



nocturnal IOP elevation in patients is still present in the sittingposition.47 Nocturnal postural differences, nocturnal activity,and/or measurement artifacts induced by waking patients for IOP

measurement could account for some of the difference in themagnitude of nocturnal IOP elevation reported in humans com-pared with the primate data reported herein.

FIGURE 7. The 2-hour time-windowdistributions of IOP for 18 24-hourperiods in animal 21356. The dates ofdata capture are shown above eachplot. In the box-and-whisker plots,the central bar indicates the meanIOP in each 2-hour segment, the endsof the box show the central 50% ofthe measurements, and the whiskersindicate the 95% limits of the mea-surements in that time window; (E)outliers.

FIGURE 8. The 2-hour time-windowdistributions of IOP for 4 24-hour pe-riods in animal 24251. The dates ofdata capture are shown above eachplot. In the box-and-whisker plots,the central bar indicates the meanIOP in each 2-hour segment, the endsof the box show the central 50% ofthe measurements, and the whiskersindicate the 95% limits of the mea-surements in that time window; (E)outliers. Only four 24-hour periods of

continuous IOP for animal 24251 are shown, because the IOP transducer failed prematurely. The data reported for this animal fall within thehigh-drift phase of the IOP transducer. Although we are confident that the IOP patterns shown are representative of the relative 24-hour IOPdistributions, the absolute IOP values should be viewed with caution (see the Limitations section of the Discussion).



Limitations

This study was limited by several important factors. First, weacquired only four 24-hour periods of continuous IOP in animal24251 because the IOP transducer failed prematurely; ourdiscussions with Konigsberg Instruments and with other usersof the company’s pressure telemetry systems indicate that a10% to 15% failure rate is to be expected. The data reported forthis animal fall within the high-drift phase of the IOP trans-ducer, and so we are much less confident that the absolute IOPvalues reported in Figures 8 and 11 are accurate. Although theabsolute IOPs reported for this animal should be viewed withcaution, the IOP distributions and nycthemeral rhythms withineach 24-hour period should be relatively unaffected by thenoted transducer drift. Second, we have not fully validated thesource of the high-frequency IOP fluctuations seen in Figure 3,although preliminary evidence suggests that they are due toblinks and saccades (Seigfreid WP, et al. IOVS 2011;52:ARVOE-Abstract 656). Therefore, the results are focused on averageIOP data for 2-hour time windows, in which these fluctuations

are averaged out. In the future, we will use video monitoring ofblinks and saccades that are time-synced to the IOP signal,along with optokinetic nystagmus (OKN) testing, to confirmthe source of these fluctuations. It will also be necessary toperform in vitro experiments that mimic saccades to con-firm that inertial effects of tube movement do not induce theIOP fluctuations that we recorded. Third, with the excep-tion of the light– dark cycle, we did not monitor environ-mental conditions or food and water intake to isolate thosevariables from the IOP variations noted in Figures 6 to 11. Inthe future, careful monitoring of sound, caretakers enteringthe room, cage changes, food and water intake, and all otherenvironmental factors and events must be logged, to ensurethat we are measuring true resting IOP in the implantedanimals. Finally, we report IOP data for three nonhumanprimates, and thus our ability to apply the conclusionsreached to the primate population at large or to humansubjects is limited.

As with all transducers in the implant system, the bodytemperature signal drifts with time and must be calibratedregularly, which we did not perform to simplify the protocol.Hence, we do not report body temperature data in this article.In addition, the circadian rhythms of body temperature knownto occur in higher primates likely induce temperature-relatedfluctuations in the IOP signal. Konigsberg Instruments rates thebattery and transmitter for the T30F implants used in this studyat less than 0.01 mm Hg/°C of temperature-induced fluctua-

FIGURE 9. The distribution of IOP in 2-hour time windows for 1824-hour periods in animal 18012. In the box-and-whisker plots, thecentral bar indicates the mean IOP in each 2-hour segment, the ends ofthe box show the central 50% of the measurements, and the whiskersindicate the 95% limits of the measurements in that time window; (E)outliers.

FIGURE 10. The distribution of IOP in 2-hour time windows for 1824-hour periods in animal 21356. In the box-and-whisker plots, thecentral bar indicates the mean IOP in each 2-hour segment, the ends ofthe box show the central 50% of the measurements, and the whiskersindicate the 95% limits of the measurements in that time window; (E)outliers.

FIGURE 11. The distribution of IOP in 2-hour time windows for four24-hour periods in animal 24251. In the box-and-whisker plots, thecentral bar indicates the mean IOP in each 2-hour segment, the ends ofthe box show the central 50% of the measurements, and the whiskersindicate the 95% limits of the measurements in that time window; (E)outliers. Note that these data are derived from only four 24-hourperiods of continuous IOP in this animal, because the IOP transducerfailed prematurely. The data reported for this animal fall within thehigh-drift phase of the IOP transducer. Whereas we are confident thatthe relative patterns of IOP distribution shown are representative, theabsolute IOP values should be viewed with caution (see Limitationssection of the Discussion).



tion. The IOP transducer itself is rated for less than a 0.5-mmHg/°C change in IOP, although this temperature-induced fluc-tuation could be significant in light of the 1.8°C circadianvariation in body temperature in female rhesus macaques.48

Unfortunately, there is no way to predict the direction of thetemperature-induced fluctuation in the IOP signal, and it islikely to be different between implants. Hence, the uncertaintyof the temperature-induced fluctuation in the IOP signal mustbe factored into the weak nycthemeral pattern of IOP shown inFigures 9 to 11.

We did not characterize the magnitude and extent of high-frequency IOP fluctuations in this report or focus our results ondata for time windows shorter than 2 hours. The telemetry dataare much more sensitive to high-frequency fluctuations andenvironmental variables in shorter time windows (Figs. 3, 5),and some portion of the high-frequency data may be artifactual(e.g., inertial pressure effects of eye movement on the aqueousin the tube and environmental factors, as discussed above). Inaddition, most human IOP data are collected using snapshotdevices that preclude IOP spikes from blinks and saccades.Hence, to compare our IOP telemetry data in primates to thatfrom human studies, it is necessary to remove the high-fre-quency IOP spikes from the telemetry data and take shortertime-window averages of the resulting baseline IOP. The IOPsignal validation and filtering necessary to accomplish this arenot trivial, however, and so those data will be the subject of afuture report.

SUMMARY

We have successfully measured IOP continuously using anewly developed, fully implantable IOP telemetry system. IOPfluctuates as much as 10 mm Hg day to day and hour to hourin unrestrained nonhuman primates, which indicates that snap-shot IOP measurements may be inadequate to capture the truedynamic character of IOP. The distributions, magnitudes, andpatterns of IOP are not reproducible from day to day withinanimals when IOP is averaged in 2-hour time windows. WhenIOP data are averaged across multiple 24-hour periods withinanimals, a weak nycthemeral rhythm is present in the nonhu-man primate, in which IOP tends to be highest at night. IOPfluctuations of this frequency and magnitude may be an impor-tant yet unknown contributor to IOP-related glaucomatousdamage.

Acknowledgments

The authors thank Shaban Demirel for invaluable assistance in thestatistical analysis of these data and Yi Liang for assistance in thesurgical implantation of the telemetry devices.

References

1. Miglior S, Pfeiffer N, Torri V, Zeyen T, Cunha-Vaz J, Adamsons I.Predictive factors for open-angle glaucoma among patients withocular hypertension in the European Glaucoma Prevention Study.Ophthalmology. 2007;114:3–9.

2. Leske MC, Heijl A, Hussein M, Bengtsson B, Hyman L, Komaroff E.Factors for glaucoma progression and the effect of treatment: theearly manifest glaucoma trial. Arch Ophthalmol. 2003;121:48–56.

3. Gordon MO, Beiser JA, Brandt JD, et al. The Ocular HypertensionTreatment Study: baseline factors that predict the onset of primaryopen-angle glaucoma. Arch Ophthalmol. 2002;120:714–720; dis-cussion 829–730.

4. The Advanced Glaucoma Intervention Study (AGIS): 7. The rela-tionship between control of intraocular pressure and visual fielddeterioration. The AGIS Investigators. Am J Ophthalmol. 2000;130:429–440.

5. The Advanced Glaucoma Intervention Study (AGIS): 12. Baselinerisk factors for sustained loss of visual field and visual acuity inpatients with advanced glaucoma. Am J Ophthalmol. 2002;134:499–512.

6. Musch DC, Gillespie BW, Lichter PR, Niziol LM, Janz NK. Visualfield progression in the Collaborative Initial Glaucoma TreatmentStudy the impact of treatment and other baseline factors. Ophthal-mology. 2009;116:200–207.

7. Caprioli J, Coleman AL. Intraocular pressure fluctuation a riskfactor for visual field progression at low intraocular pressures inthe advanced glaucoma intervention study. Ophthalmology. 2008;115:1123–1129.

8. Rudnicka AR, Mt-Isa S, Owen CG, Cook DG, Ashby D. Variations inprimary open-angle glaucoma prevalence by age, gender, and race:a Bayesian meta-analysis. Invest Ophthalmol Vis Sci. 2006;47:4254–4261.

9. Klein BE, Klein R, Linton KL. Intraocular pressure in an Americancommunity. The Beaver Dam Eye Study. Invest Ophthalmol VisSci. 1992;33:2224–2228.

10. Nomura H, Ando F, Niino N, Shimokata H, Miyake Y. The relation-ship between age and intraocular pressure in a Japanesepopulation: the influence of central corneal thickness. Curr EyeRes. 2002;24:81–85.

11. Nomura H, Shimokata H, Ando F, Miyake Y, Kuzuya F. Age-relatedchanges in intraocular pressure in a large Japanese population: across-sectional and longitudinal study. Ophthalmology. 1999;106:2016–2022.

12. Rochtchina E, Mitchell P, Wang JJ. Relationship between age andintraocular pressure: the Blue Mountains Eye Study. Clin ExpOphthalmol. 2002;30:173–175.

13. Weih LM, Mukesh BN, McCarty CA, Taylor HR. Association ofdemographic, familial, medical, and ocular factors with intraocularpressure. Arch Ophthalmol. 2001;119:875–880.

14. Leske MC, Connell AM, Wu SY, Hyman L, Schachat AP. Distribu-tion of intraocular pressure. The Barbados Eye Study. Arch Oph-thalmol. 1997;115:1051–1057.

15. Wu SY, Leske MC. Associations with intraocular pressure in theBarbados Eye Study. Arch Ophthalmol. 1997;115:1572–1576.

16. Chumbley LC, Brubaker RF. Low-tension glaucoma. Am J Ophthal-mol. 1976;81:761–767.

17. Levene RZ. Low tension glaucoma: a critical review and newmaterial. Surv Ophthalmol. 1980;24:621–664.

18. Geijssen H. Studies on Normal-Pressure Glaucoma. Amsterdam:Kugler Publications; 1991.

19. Burgoyne CF, Downs JC. Premise and prediction-how optic nervehead biomechanics underlies the susceptibility and clinical behav-ior of the aged optic nerve head. J Glaucoma. 2008;17:318–328.

20. Albon J, Purslow PP, Karwatowski WS, Easty DL. Age relatedcompliance of the lamina cribrosa in human eyes. Br J Ophthal-mol. 2000;84:318–323.

21. Girard MJ, Suh JK, Bottlang M, Burgoyne CF, Downs JC. Scleralbiomechanics in the aging monkey eye. Invest Ophthalmol Vis Sci.2009;50:5226–5237.

22. Knox Cartwright NE, Tyrer JR, Marshall J. Age-related differencesin the elasticity of the human cornea. Invest Ophthalmol Vis Sci.June 2011;52:4324–4329.

23. Downs JC, Suh JK, Thomas KA, Bellezza AJ, Hart RT, Burgoyne CF.Viscoelastic material properties of the peripapillary sclera in nor-mal and early-glaucoma monkey eyes. Invest Ophthalmol Vis Sci.2005;46:540–546.

24. Girard MJ, Suh JK, Bottlang M, Burgoyne CF, Downs JC. Biome-chanical changes in the sclera of monkey eyes exposed to chronicIOP elevations. Invest Ophthalmol Vis Sci. 2011;52:5656–5669.

25. Bengtsson B, Leske MC, Hyman L, Heijl A. Fluctuation of intraoc-ular pressure and glaucoma progression in the early manifestglaucoma trial. Ophthalmology. 2007;114:205–209.

26. Musch DC, Gillespie BW, Niziol LM, Lichter PR, Varma R; CIGTSStudy Group. Intraocular pressure control and long-term visualfield loss in the collaborative initial glaucoma treatment study.Ophthalmology. Published online May 19, 2011.

27. Medeiros FA, Weinreb RN, Zangwill LM, et al. Long-term intraoc-ular pressure fluctuations and risk of conversion from ocular hy-pertension to glaucoma. Ophthalmology. 2008;115:934–940.



28. Bengtsson B, Heijl A. Diurnal IOP fluctuation: not an independentrisk factor for glaucomatous visual field loss in high-risk ocularhypertension. Graefes Arch Clin Exp Ophthalmol. 2005;243:513–518.

29. Liu JH, Zhang X, Kripke DF, Weinreb RN. Twenty-four-hour intra-ocular pressure pattern associated with early glaucomatouschanges. Invest Ophthalmol Vis Sci. 2003;44:1586–1590.

30. Asrani S, Zeimer R, Wilensky J, Gieser D, Vitale S, Lindenmuth K.Large diurnal fluctuations in intraocular pressure are an indepen-dent risk factor in patients with glaucoma. J Glaucoma. 2000;9:134–142.

31. Realini T, Weinreb RN, Wisniewski S. Short-term repeatability ofdiurnal intraocular pressure patterns in glaucomatous individuals.Ophthalmology. 2011;118:47–51.

32. Realini T, Weinreb RN, Wisniewski SR. Diurnal intraocular pres-sure patterns are not repeatable in the short term in healthyindividuals. Ophthalmology. 2010;117:1700–1704.

33. Kotecha A, White E, Schlottmann PG, Garway-Heath DF. Intraoc-ular pressure measurement precision with the Goldmann applana-tion, dynamic contour, and ocular response analyzer tonometers.Ophthalmology. 2010;117:730–737.

34. Pourjavan S, Boelle PY, Detry-Morel M, De Potter P. Physiologicaldiurnal variability and characteristics of the ocular pulse amplitude(OPA) with the dynamic contour tonometer (DCT-Pascal). IntOphthalmol. 2007;27:357–360.

35. Coleman DJ, Trokel S. Direct-recorded intraocular pressure varia-tions in a human subject. Arch Ophthalmol. 1969;82:637–640.

36. Kaufmann C, Bachmann LM, Robert YC, Thiel MA. Ocular pulseamplitude in healthy subjects as measured by dynamic contourtonometry. Arch Ophthalmol. 2006;124:1104–1108.

37. Mansouri K, Shaarawy T. Continuous intraocular pressure moni-toring with a wireless ocular telemetry sensor: initial clinical ex-perience in patients with open angle glaucoma. Br J Ophthalmol.2011;95:627–629.

38. McLaren JW, Brubaker RF, FitzSimon JS. Continuous measurementof intraocular pressure in rabbits by telemetry. Invest OphthalmolVis Sci. 1996;37:966–975.

39. McLaren JW, Bachman LA, Brubaker RF. Comparison of effects oftopical ibopamine and epinephrine on the circadian rhythm ofintraocular pressure of the rabbit eye as measured by telemetry. JOcul Pharmacol Ther. 1999;15:107–116.

40. Akaishi T, Ishida N, Shimazaki A, Hara H, Kuwayama Y. Continu-ous monitoring of circadian variations in intraocular pressure bytelemetry system throughout a 12-week treatment with timololmaleate in rabbits. J Ocul Pharmacol Ther. 2005;21:436–444.

41. Li R, Liu JH. Telemetric monitoring of 24 h intraocular pressure inconscious and freely moving C57BL/6J and CBA/CaJ mice. Mol Vis.2008;14:745–749.

42. Markert M, Klumpp A, Trautmann T, Mayer K, Stubhan M, Guth B.The value added by measuring myocardial contractility ‘in vivo’ insafety pharmacological profiling of drug candidates. J PharmacolToxicol Methods. 2007;56:203–211.

43. Liu JH, Kripke DF, Hoffman RE, et al. Nocturnal elevation ofintraocular pressure in young adults. Invest Ophthalmol Vis Sci.1998;39:2707–2712.

44. Liu JH, Kripke DF, Twa MD, et al. Twenty-four-hour pattern ofintraocular pressure in the aging population. Invest OphthalmolVis Sci. 1999;40:2912–2917.

45. Jorge J, Ramoa-Marques R, Lourenco A, et al. IOP variations in thesitting and supine positions. J Glaucoma.

46. Tsukahara S, Sasaki T. Postural change of IOP in normal personsand in patients with primary wide open-angle glaucoma and low-tension glaucoma. Br J Ophthalmol. 1984;68:389–392.

47. Liu JH, Bouligny RP, Kripke DF, Weinreb RN. Nocturnal elevationof intraocular pressure is detectable in the sitting position. InvestOphthalmol Vis Sci. 2003;44:4439–4442.

48. Barger LK, Hoban-Higgins TM, Fuller CA. Gender differences in thecircadian rhythms of rhesus monkeys. Physiol Behav. 2010;101:595–600.



24-hour iop telemetry in the nonhuman primate:...

Documents