Wireless and Catheter-Free Bladder Pressure and Volume Sensor

Abstract

Continuous monitoring of bladder activity during normal daily activities would improve clinical diagnostics and understanding of the mechanisms underlying bladder function, or help validate how differing neuromodulation strategies affect the bladder. This work describes a urological monitor of conscious activity (UroMOCA). The UroMOCA included a pressure sensor, urine impedance-sensing electrodes, and wireless battery recharge and data transmission circuitry. Components were assembled on a circuit board and encapsulated with an epoxy/silicone molded package that allowed Pt-Ir electrode feedthrough for urine contact. Packaged UroMOCAs measured 12 × 18 × 6 mm. UroMOCAs continuously transmitted data from all onboard sensors at 10 Hz at 30 cm range, and ran for up to 44 hours between wireless recharges. After in vitro calibration, implantations were performed in 11 animals. Animals carried the device for 28 days, enabling many observations of bladder behavior during natural, conscious behavior. In vivo testing confirmed the UroMOCA did not impact bladder function after a two-week healing period. Pressure data in vivo were highly correlated to a reference catheter used during an anesthetized follow-up. Static volume sensor data were less accurate, but demonstrated reliable detection of bladder volume decreases, and distinguished between voiding and non-voiding bladder events.

Graphical Abstract

I. Introduction

Lower urinary tract neurophysiology research relies on animal models (e.g. feline and porcine) to study how the bladder interacts with the nervous system. Generally, the bladder’s state is determined by the organ pressure and volume. During filling, the bladder volume increases at low pressures as the organ accommodates urine, and during voiding the organ generates a large pressure to expel urine as volume decreases. Remote telemetry systems enable bladder pressure monitoring using wired pressure sensors which cross the bladder wall and are not feasible for eventual human use [1]. Currently there are no sensors available for real-time monitoring of bladder pressure and volume in vivo.

Catheterization is the gold standard method for measuring the bladder during filling, because volume can be controlled by infusing saline (Fig. 1a). Catheter-based measurements require stationary setups, due to the instrumentation connected to the catheters, or to avoid movement-induced measurement artifacts. In research animals, catheterization requires anesthesia which affects neuro-urological pathways, or animal restraint, limiting measurement time and social behaviors surrounding natural bladder filling and emptying. And in humans, stationary catheter measurements do not replicate how people live with incontinence outside the clinic.

Fig. 1.

(A) The Urological Monitor of Conscious Activity (UroMOCA) can replace catheter systems used in animal and human research. (B) UroMOCA provided real time pressure and volume data to characterize bladder function in freely-moving animals.

Sensors have been demonstrated to measure bladder pressure wirelessly in animals [2]–[5] but real-time measurement of bladder volume is also being investigated. Generally, implants measure either bladder volume or bladder pressure, but not both [6]. Bladder pressure is most directly measured from within the bladder, while bladder volume is commonly sensed from outside the bladder. Ultrasonic systems, for example, estimate volume from the external bladder geometry in humans and animals [7]–[9]. Similarly, noninvasive infrared sensors have been shown to estimate bladder volume in humans [10]. Wired electrical impedance sensors have been used to measure bladder volume from outside the body [11]–[13], but have not been demonstrated from inside the bladder. The size and power consumption of external systems limits their use over long time periods, or in freely moving animals, however, improvements in wearable form factors may enable daily use [14].

Implantable systems may enable continuous, real-time monitoring of bladder volume, and have generally focused on instrumenting the organ directly with sensors. Strain-based and magnetic sensors detect the increase in bladder size during filling and require attachment to the outside of the bladder or nearby tissues [15]–[20]. While a combination of internal and external sensors may enable simultaneous measurement of both pressure and volume, in this work we investigated the feasibility of measuring bladder pressure and volume from within the bladder, from a single electronic device combining both sensor modalities. This approach would be much more feasible to enable minimally invasive sensors for clinical use with humans.

The goal of this work was to evaluate a wireless device for sensing both bladder pressure and volume simultaneously in research settings, without catheters (Fig. 1b) – the Urological Monitor of Conscious Activity (UroMOCA). While the underlying technology can be translated to human use, initial feasibility studies were demonstrated in conscious animals. Here we summarize the UroMOCA design and in vivo functionality over 4-week implantations.

II. Wireless Sensor Electronic Design

The UroMOCA construction followed circuitry, experimental, and packaging constraints, which are discussed separately. To enable ambulatory monitoring over a 4-week implantation, a wirelessly rechargeable, battery-powered sensor platform was developed. This enabled fully wireless monitoring of battery power, with wireless recharge occurring in between sessions while animals rested.

A. Urinary Pressure and Volume Sensing Method

The state variables describing the function of the bladder are pressure and volume; measurement of both values describes if the organ is filling (normally low-pressure, slow increasing of volume), contracting (the detrusor muscle contracts to increase pressure) or voiding (the urine volume decreases, with an eventual decrease in bladder pressure as it empties). Therefore, the UroMOCA focused on measuring these pressure and volume simultaneously (Fig. 2A).

Fig. 2.

(A) UroMOCA included a pressure sensor and conductance-measuring electrodes (E_C and E_V). (B) Sensed currents (i_C and i_V) flowed through differential volumes of urine. (C) The ratio of i_C and i_V allowed estimation of volume independent of urine concentration.

The accuracy and sampling requirements for pressure and volume signals are greatly different. Clinically, bladder pressure is sampled at 10 Hz in stationary environments [21]. From prior work in mobile animals [22], we determined that a higher sampling rate avoids aliasing due to motion and abdominal pressure artifacts. Therefore UroMOCA pressure was sampled at 75 Hz, then internally low-pass filtered to a 0.5 Hz bandwidth. The volume sampling rate was much slower to conserve power, and because urine volume changes slowly during filling. Volume measurements were made 10 times per minute.

We previously described prototype functionality of wireless pressure- and volume-sensing circuitry tested in vitro [23], [24] (Fig. 2A). The pressure sensor used on the wireless UroMOCA was the same (STMicro LPS33HWTR). Pressure within the bladder was measured from one face of the device, after encapsulation to protect the sensor from urine. Pressure data were truncated to 16-bit resolution.

Bladder volume was estimated from impedance measurements of urine using two active electrodes with one common cathode (Fig. 2B). Both urine volume and urine concentration in the bladder change (based on diet, and during natural filling/emptying cycles); a measurement of urine impedance alone cannot distinguish between changes in urine volume and concentration.

Two electrodes were used to estimate the urine concentration and urine volume separately. Electrodes were with a concentration-measuring electrode (E_C) 1 mm away from the common cathode (C), while a volume-sensitive electrode (E_V) was placed 18 mm away on the other side of the implant. Due to the great path length difference between electrodes, they demonstrated a differential sensitivity to changing urine concentration and volume (Fig. 2B,C).

B. Sensor Acquisition and Wireless Transmission

To reduce implant size, relatively simple, low-power circuitry was developed to minimize the size of the system battery, which was the largest single component. The UroMOCA used a low-power microcontroller (Texas Instruments MSP430FR2433) to coordinate sensing and transmission. The implant was powered from a single 5.5-mAh battery (Seiko MS621FE). Circuitry was designed to either operate in sleep or active mode; in sleep mode, the microcontroller and pressure sensor were in a low-power state, while in active mode the device transmitted data (Table I).

TABLE I.

UroMOCA Electrical Performance

	Measured/nominal value
UroMOCA sensors	Pressure
	Temperature*
	Urine conductivity (2 channels)
UroMOCA dimensions	18 mm × 12 mm × 6 mm
Pressure sampling rate	75 Hz, 0.5 Hz filtered bandwidth
Pressure accuracy	±1.1 cm H2O
Impedance sampling rate	0.17 Hz
Impedance range	>10 MΩ to 10 Ω
Radio transmission	4 MHz, OOK, 10 Hz packet rate
Transmission distance	30 cm, using 8-cm loop antenna
Wireless battery recharge	2.2 MHz inductive coupling 15-cm range with 16 W power
Current draw (measure)	125 μA
Current draw (sleep)	1.4 μA
Battery lifespan (measure)	44 hours
Battery lifespan (sleep)	163 days

^*measured but not transmitted in presented work

The UroMOCA (Fig. 2) recorded 5 measurements: pressure, temperature, impedance (from 2 active electrodes), and battery voltage. Temperature data were used to compensate pressure sensor response, but were not otherwise transmitted because temperature changes in the bladder are normally minimal [25].

Urine volume estimates were made by exciting E_C and E_V, and measuring the peak-to-peak currents (i_C and i_V) flowing into C. Measures were accumulated for 16 cycles per electrode to produce a 16-bit measure of impedance.

Impedance was measured at 500 Hz [31] with each channel measured independently (Fig. 3B,C). In each measurement, one working electrode (E_C, E_V) was driven by a 500-Hz square wave (φ_V, φ_C) with amplitude V_BAT (system battery voltage). Current flowed through the urine (Z_U), then into the AC-coupled common anode. A resistor divider biased the microcontroller ADC input at V_BAT/2 and was only enabled during electrode measurement to save power (φ_E). For each electrode, 16 cycles of the 500-Hz excitation were applied, and the peak-to-peak amplitude was calculated by sampling the signal at the peak/valley per cycle (φ_S). Amplitudes were accumulated over 16 cycles per electrode to produce a 16-bit measure of impedance using 12-bit ADC conversions.

Fig. 3.

(A) The UroMOCA used a low-power microcontroller communicating with a pressure sensor and stainless-steel electrodes. (B) Urine impedance in low-volume and variable-volume regions was determined from the peak-to-peak current flowing in E_V and E_C when driven at 500 Hz. (C) Currents E_V and E_C were measured separately but immediately after each other. (D) The received voltage at the ADC (proportional to urine impedance, Z_U) was accumulated by ADC samples over 16 waveform cycles to produce digitized values for urine volume and urine concentration.

Passive components (Fig. 3B) limited the urine-sensing test current and balanced any injected charge. Values used were R_1,2 = 2.6 kΩ and C₁ = 10 μF to limit the maximum current during measurement to 1.9 mA. Impedance sensing added about 20 μA current draw from the battery, under worst-case (0 Ω) urine impedance conditions.

C. Wireless Activation and Recharging

The UroMOCA included a 6.5-μH coil resonating at 2.2 MHz for RF battery recharge. Energy was transferred to the implant from a power amplifier with a 16-cm diameter, 16-μH coil. Received energy was rectified using a voltage doubler rectifier implemented with discrete diodes. While simple and inefficient, due to the low coupling coefficient of the inductive link due to implant depths greater than 10 cm, the additional loss in efficiency due to the simple charging circuit was negligible. An NMOS common-source amplifier sensed the rectified voltage and coupled it to the microcontroller, enabling detection of the RF envelope such that pulse-length coded commands could be transmitted to the implant. In the presented experiments, only a command to wirelessly wake up or go to sleep was implemented.

The UroMOCA battery was trickle charged directly from the RF-DC rectifier output. Battery voltage was sensed every second using the internal microcontroller, by setting the ADC reference to the battery voltage and measuring a fixed internal voltage reference. To protect from battery over-charge, the microcontroller shunted excess current by transmitting data continuously and by activating the DC bias pathway for electrode sensing (Fig. 3). To protect over-discharge, the UroMOCA went to sleep while the battery was below 2.0 V.

Between experiments, the UroMOCA remained in a low-power, sleep state. Before recording, the UroMOCA was wirelessly activated with a 50-ms pulse of about 10 W peak power. After activation, the UroMOCA continuously transmitted for 4 hours, or until receiving a sleep command. The sleep signal was transmitted to the UroMOCA by activating the battery recharge field for at least 300 ms. During in vivo studies (described below), animals preferred to rest on the warm recharging coil, effectively recharging implanted sensors between recording sessions.

D. Wireless Data Transmission and Reception

The UroMOCA transmitted data every 100 ms using near-field magnetic communication. Data were Manchester-encoded and modulated via 4-MHz on-off-keying (OOK). The microcontroller directly drove the 10-μH transmitting antenna.

Data were transmitted in individual packets every 100 ms. Each packet consisted of a synchronization frame, an 8-bit header, and a 96-bit payload. The synchronization frame was a 136-μs pulse of unmodulated carrier to prime the receiver for reception of the header and payload bits. The header included a ‘0101’ feature used for reception-side clock synchronization. The payload consisted of 6, 16-bit values relaying pressure and impedance data from all sensors, along with battery voltage and device ID information. Data were transmitted as Manchester-encoded with a bit rate of 58 kbps.

E. Wireless Data Receiver Radio

Data from the UroMOCA were received by a small radio with an external magnetic antenna (Fig. 4). During free-moving data recording, the radio and antenna were attached to a vest which the animal wore (described below). The radio consisted of three circuit boards: a Main Board, BLE Board, and a Teensy 3.6 board (PJRC Electronics, Arduino). The Main Board (Fig. 4A) received transmissions and extracted the transmitted digital data. The BLE Board include a Bluetooth module (uBlox BMD-350) which forwarded received data to a remote laptop during experiments. Finally, the Teensy 3.6 board included an ARM Cortex M4F microprocessor. During data recording, the Teensy 3.6 decoded the transmitted Manchester-encoded data and parsed each sensor channel. All received data were timestamped and stored to a microSD card. The packaged radio measured 44 × 67 × 25 mm and used a rechargeable lithium-polymer battery for wireless data capture during animal studies.

Fig. 4.

(A) Conscious data collection used a 4-MHz OOK receiver. (B) The radio was combined with a Teensy 3.6 and BLE board. (C). The packaged radio stored data to an onboard microSD card and forwarded data over BLE during recording sessions. (D) During conscious data collection animals wore the radio recorder with an antenna near the skin.

III. Implantable Sensor Encapsulation

UroMOCAs were encapsulated after solder assembly, programming, and initial function check. Nonhermetic, polymer-based encapsulation was used to enable production of research devices in typical lab environments [26].

A. Sensor Encapsulation

The UroMOCA used a printed circuit board encased in a medical epoxy package (Fig. 5). Medical silicone rubber was applied to the outside of the device to reduce encrustation when in contact with urine [27]. After solder assembly, three 1 mm × 10 mm strips of Pt-Ir mesh (Alfa Aesar, 150 mesh, 43 μm thick) were soldered to the PCB. Flux residues were removed by soaking boards in a 50 °C bath of distilled water.

Fig. 5.

UroMOCA fabrication: (A) circuit board assembly, (B) epoxy injection molding within silicone mold (C) silicone gel filling of pressure sensor and folding of Pt-Ir electrodes, and (D) final silicone encapsulation of device and pressure sensor gel.

A silicone rubber (OOMOO-25, Smooth-On) mold was fabricated around a 3D-printed model representing the final UroMOCA dimensions. The mold was split along the axis of the UroMOCA PCB allowing insertion of the assembled device for epoxy encapsulation. Protrusions in the mold model were included so that impedance-sensing electrodes were shrouded to avoid contacting the bladder wall.

The Pt-Ir electrode strips were used to suspend the PCB within the mold (Fig. 5A). Next, biocompatible medical epoxy resin (EPO-TEK MED-301) was injected into the mold and cured at 50 °C for 1 hour. After epoxy curing, the mold was peeled away from the device. Next, Pt-Ir strips were folded over the device thickness and adhered using additional MED-301 epoxy (Fig. 5B). The pressure sensor column was filled with silicone gel (Dow Corning Sylgard 527) before curing at 50 °C for 1 hour. A pre-cast silicone membrane (2 × 2 × 0.1 mm, Factor II A-103) was applied to the gel and sealed around the edges using silicone adhesive (Elkem MED-4300) (Fig. 5C). The device was allowed to cure at room temperature overnight before final silicone encapsulation. The outer layer of silicone (Factor II A-103) was applied using a folded lint-free swab (KimTech Science Wipe, Kimberley Clark Professional) selectively by hand, to avoid coating the impedance-sensing electrodes with silicone (Fig. 5D).

Post-packaging UroMOCA dimensions were 12 × 18 × 6 mm (Fig. 6A). This polymeric packaging reduced implant cost, allowed lab assembly, and material layers did not delaminate over 28 days of implantation (Fig. 6B). Pt-Ir electrodes were folded back and trimmed after epoxy molding (Fig. 6C, D).

Fig. 6.

(A) Packaged UroMOCA prior to implant. (B) UroMOCAs showed little deterioration after explant at 28 days. (C) An epoxy-molded encapsulation enabled Pt-Ir electrode feedthrough, shown prior to trimming. (D) UroMOCAs had electrodes attached prior to epoxy molding. Only the sensor side of the device is shown; the opposite side contained the microcontroller and wireless energy recovery circuitry.

B. In vitro Impedance Sensing Calibration

UroMOCAs were bench-tested for impedance sensing performance using discrete resistors and saline in various concentrations. Pressure and impedance sensing were tested with the device submerged in a sealed pressure chamber, coupled to a reference pressure sensor (Fig. 7A). UroMOCA showed optimal sensitivity in the 1 kΩ – 1 MΩ range (Fig. 7B). Measurement of concentration was calibrated using saline from 0 – 4% sodium content; deionized water was used for 0%. Urine concentration is nominally near 0.9% like other bodily fluids [28]. While the primary electrolytes contributing to conductivity in human urine are sodium, chloride, and uric acid [29], we used a simplified in vitro model for urine using saline with variable salt concentration. In vitro testing of early UroMOCA prototypes with collected feline urine showed little difference in response between actual urine and saline [30], [31]; saline was used to calibrate implantable devices for sanitary reasons.

Fig. 7.

Calibration was performed on UroMOCAs prior to implant, to convert digital values measured by the sensors. (A) Pressure calibration of UroMOCAs used a rigid pressure chamber with simultaneous measurement of reference pressure. (B) Impedance calibration used discrete resistors along with varying concentrations of saline in the chamber, from 0–4%, to span the expected range of urine concentration in vivo.

Previously reported in vitro testing used a balloon with saline in concentrations of 0.5 – 4.0% to mimic urine at different concentrations, demonstrating sensor accuracy of about 5 mL [30], [31]. Testing with the UroMOCA submerged in saline showed a measurement response in the most sensitive region of the calibration curve (Fig. 7B). Absolute accuracy was dependent on symmetric placement of the UroMOCA within the balloon. Because the sensor can flip and move within the bladder, reduced accuracy after implant was expected in vivo.

C. In vitro Pressure Sensing Calibration

UroMOCA pressure accuracy was validated on the bench in a water-filled pressure chamber. Reference pressure was recorded using a sensor (Deltran DPT-100) coupled to a data acquisition system (National Instruments CompactDAQ, NI-9218 module). UroMOCA pressure showed excellent agreement with the reference pressure, with an accuracy of about ±1.1 cmH₂O on a range of 0–200 cmH₂O (Fig. 8).

Fig. 8.

(A) UroMOCA in vitro calibration showed linear agreement to a reference sensor. (B) Bland-Altman analysis indicated an accuracy of +/− 1.1 cmH₂O over the range of 0–200 cmH₂O.

IV. In Vivo Demonstration

A. UroMOCA Device Implantations

Implantations were performed at Cleveland Clinic after approval by the Institutional Animal Care and Use Committee in 11 felines. The UroMOCA was implanted into the bladder via an abdominal incision which was closed using sterile monofilament suture (2–0 Prolene, Covidien) (Fig. 9A). One implantation was a sham device to determine the impact of surgery on bladder function, and one was as a terminal non-survival procedure. The remaining nine implantations were active UroMOCAs. Active device experiments lasted 4 weeks before device explant. Implanted UroMOCAs worked for an average of 23 days (range: 12 – 28); 6 out of 9 devices worked for the entire implant duration.

Fig. 9.

(A) UroMOCAs were sutured into the bladder. (B) Fluoroscopy confirmed device location in the bladder. Saline was infused to control bladder volume, with simultaneous pressure and impedance recording. Cross-sectional and longitudinal computerized tomography scans were captured at distinct bladder volumes, with three examples shown: (C) 10 mL, (D) 30 mL, and (E) 50 mL.

Fluoroscopic imaging confirmed device location and bladder volume at maximum capacity (Fig. 9B). Anesthetized check-ups occurred at 14 days and devices were explanted after 4 weeks. During anesthetized procedures, the bladder was filled to known volumes and computerized tomography (CT) imaging was performed to confirm bladder geometry and device location (Fig. 9C–E). During filling, UroMOCA values were recorded to assess pressure and volume sensor accuracy in vivo using catheter-based pressure and known infused volume as references.

Before implant as well as at 14 and 28 days after implant, urinalysis, heavy metal, and toxicology assays were performed on urine samples. No signs of infection or significant changes in metal content of urine were found, suggesting that device packaging remained intact throughout the 4 week study.

All animals returned to normal movement within 24 hours of implant surgery. All animals had a decreased bladder capacity (up to 50%), which resulted in bladder spasticity for 2 weeks after implantation surgery. Spasticity led to frequent voiding attempts with small urine amounts. After 2 weeks all animals resumed normal voiding frequency. All devices remained in the bladder without obstructing for 4 weeks.

B. Anesthetized Bladder Pressure Measurement

Anesthetized cystometry was performed immediately before and after implant, and after 2 and 4 weeks to assess bladder healing response. A catheter infused room-temperature saline into the bladder at a rate of 2 mL/min or through boluses. An external pressure transducer was used to measure reference bladder pressure. Data were transmitted by the UroMOCA throughout cystometry. Wireless pressure data were highly correlated with catheter pressure (R²>0.96). Data demonstrated higher baseline pressure readings at low bladder volumes, possibly due to the force of the collapsed detrusor tissue on the UroMOCA pressure sensor diaphragm (Fig. 10).

Fig. 10.

(A) Simultaneous UroMOCA and reference catheter pressures were highly correlated. Example trace demonstrates bladder phasic contractions during filling. (B) After the filling catheter was removed (marked by arrow), UroMOCA continued to report bladder activity.

C. Anesthetized Bladder Volume Measurement

Controlled in vivo volume testing was only performed in anesthetized procedures (2 per animal) because bladder volume was manipulated by an indwelling catheter. The catheter allowed filling the bladder to known volumes via ramping or bolus fills. Due to the small size of the catheter used (3 French), bolus fills did not immediately flow into the bladder lumen, but entered in a period of several minutes.

Voltages recorded from both UroMOCA electrodes were transmitted, recorded, and converted to estimated volume in post-hoc analysis. Values were converted to estimated volume using a table-based conversion method previously reported [30]. Initial conversion factors for each implanted device were measured prior to implantation using in vitro bench calibration. After implantation, electrode voltages showed unstable baseline offset (discussed below), therefore, estimated volume from anesthetized recordings was only interpreted as a relative change. At the beginning of each bladder fill cycle, the bladder was manually emptied of urine via catheter so that relative volume change from zero was observed (Fig. 11).

Fig. 11.

Two weeks after implant, the bladder was artificially filled while UroMOCA data were recorded. Saline solutions of (A) 0.9%, (B) 1.5%, and (C) 4.0% were infused to simulate urine at different concentrations. UroMOCA values in arbitrary units (digital values as-transmitted) showed increasing range with increasing saline concentration. While the volume sensor demonstrated baseline drift and had limited absolute accuracy, it enabled proportional assessment of changes in bladder volume in vivo.

Volume measurements showed correlation to infused volume, but with more variability than in vitro. Bolus fills with intermittent pauses were used to help distinguish volume changes from baseline drift (Fig. 11, 12). Implanted UroMOCAs showed consistent response to the addition or removal of fluid into the bladder, and reliably distinguished between empty bladders and bladders with more than 20 mL volume. The in vivo accuracy of volume estimation degraded beyond 10 mL, however, changes in volume were still captured (Fig. 11, 12). In some cases, UroMOCAs demonstrated a reversion to baseline in volume measurements (Fig. 12), but were still sensitive in periods of several minutes to changes in volume. The cause of this “high-pass” response was not determined, but still enabled detection of bladder emptying in ambulatory animals.

Fig. 12.

In some animals, bolus infusions of 0.9% saline showed proportional increases in volume readings, but a reversion to baseline after several minutes. Thus, baseline accuracy was not reliable, but bladder emptying events were distinguishable.

D. Untethered Monitoring of Bladder Activity During Conscious Behaviors

Conscious data collection was performed 3 days per week, along with video recording and photography. Data collection used a small radio receiver, which was worn by the animal for up to 2 hours. The UroMOCA transmitted data to the radio receiver; data were continuously logged to an internal memory card and transmitted via Bluetooth to a PC. During recordings, animals moved freely within a small room (approx. 10 × 10 × 10 m). Wireless recharging of implanted UroMOCAs occurred by encouraging the animals to rest on the warm charging coil both supervised or unsupervised.

Conscious data recording captured over 50 hours of catheter-free bladder recordings across all animals. Because animals were untethered with no reference sensor available, data were analyzed qualitatively and correlated to animal behaviors. For two weeks after implantation, animals displayed bladder spasticity, which produced many recordings of low-volume urination attempts with the bladder nearly empty (Fig. 13A). After two weeks of healing, increased bladder capacity showed a more gradual increase in pressure and volume recordings during natural filling (Fig. 13B).

Fig. 13.

(A) Conscious recordings of voiding behavior showed repeated decreases in bladder volume. These low-volume voids were caused by bladder irritation in the week after UroMOCA implant. (B) After recovery of bladder function, UroMOCA pressure and volume sensors allowed catheter-free recording of bladder filling, and differentiation between isovolumetric bladder contractions and activities that led to urine leakage or bladder emptying.

V. Discussion

There has been a great diversity of bladder sensors, all of which have pros and cons. Compared to other recent work (Table II), UroMOCA is unique because it combines both bladder pressure and volume recording without a significant increase in device size or reduction in lifespan or in vivo functionality (like transmission range). While UroMOCA was not implanted in a minimally invasive manner, the underlying sensors can be translated to form factors compatible with minimally invasive insertion in humans [2], [4].

TABLE II.

COMPARISON TO RELATED RECENT INTRALUMINAL SENSORS DEMONSTRATED IN VIVO

	This Work	[32]	[33]	[34]	[2]
Dimensions	18 mm × 12 mm × 6 mm	9 mm diameter 15 mm long	17 mm × 13 mm × 3 mm	8–10 mm diameter† 40 mm long	4.6 mm diameter 30–40 mm long
Sensor modalities	1 pressure sensor 1 temperature sensor* 2 impedance sensors	1 pressure sensor	1 pressure sensor	1 pressure sensor	1 pressure sensor
Sample rate	Pressure and Temperature: 75 Hz Impedance: 0.17 Hz	0.5 Hz	0.5 Hz	< 0.1 Hz†	1–10 Hz
Data Readout	Active transmitter, Wireless recharge	Passive resonant coupling	Active transmitter, Wireless powering	Acoustically powered, pulse transmitter	Load-shift keying, Wireless powering
Wireless communication	4 MHz OOK	114 – 190 MHz	Bluetooth, 2.4 GHz	232 – 250 kHz, pulse frequency encoded	1 MHz back-scatter telemetry
Implant Depth	30 cm	2 cm	7 cm	7 cm	20 cm
Average current draw (measurement mode)	125 μA	-	2 mA	-	-
Battery lifetime	44 hrs	N/A	N/A	N/A	N/A
Untethered recording	YES	NO	NO	NO	NO
Minimally-invasive	NO	NO	NO	NO	YES

^*measured but not transmitted in presented experiments

^†estimated from information presented

While pressure data in conscious animals appeared very similar to recordings made under anesthesia and on the bench, volume sensing was more variable in vivo. While the root cause is not yet known, it may have been caused by bladder wall tissue contact on the impedance-sensing electrodes (preventing accurate measure of urine conductance). This is evidenced by an improvement in volume sensor accuracy observed after two weeks, after bladder inflammation lessened and bladder capacity increased. Despite a loss of static volume sensor accuracy, however, the UroMOCA demonstrated reliable detection of volume decreases associated with voiding events. This critical function allowed the system to distinguish between voiding and non-voiding urinary events, which are an important feature in neurophysiology studies.

The presented experiments were limited in duration, but on enough animals to demonstrate the expected reliability of hand-assembled sensors in a lab environment. Given these limitations, two-thirds of implanted sensors lasted through the entire four-week implant period, and might have functioned much longer. Failed implants demonstrated a rapid loss of battery voltage after recharge, possibly due to battery damage due to moisture ingress. Because UroMOCAs used nonhermetic polymers with a metal feedthrough (for impedance sensing electrodes), this failure mode was likely based on the limited durability of metal-polymer feedthroughs in implanted sensors [26].

While pressure in the bladder is measured with a typical accuracy of 1 cmH₂O to capture the waveform shapes of bladder contractions [31], volume measurement does not need high accuracy. Clinical guidelines, for example, recommend action based on measures of bladder volume of 50–200 mL in humans [35]. Expressed differently, one objective in volume sensing is not absolute accuracy, but determining if the bladder is filling or emptying completely.

In this study, feline bladders had a wide range of capacity from 20–50 mL. Because this is about 10 times lower than human capacity, the relative accuracy of the UroMOCA volume sensing may improve when used in larger animal or human bladders. Controlled calibrations showed the volume sensor achieved an accuracy of about 5–10 mL, as seen in previous wired prototypes [31]. In vivo results showed that pressure was measured very accurately, but volume accuracy was only sufficient for coarse description of bladder fullness (e.g. empty, partly full, very full), and for rapid changes in bladder volume (during emptying).

The variability of in vivo volume data was driven by a loss of baseline (DC) stability of volume readings. Variability persisted both in free-moving and anesthetized animals, and at different levels of urine concentration. One likely explanation is that sensed impedance included other components of abdominal and bladder tissue (and not just urine) [11], [13], or that electrochemical reactions occurring at the Pt-Ir electrode and urine interface [36] changed the contact resistance or voltage potential between electrode and urine. A four-wire impedance approach may therefore be more effective for this sensing modality.

Despite the limited volume sensor accuracy, the UroMOCA allowed for quantitative and qualitative assessment of bladder function over durations long enough to allow for bladder recovery from device implantation. Animals easily adapted to resting on wireless rechargers and wearing recording radios, enabling many, conscious, untethered recordings. The need for conscious recordings techniques is particularly important for neuromodulation and neurophysiology research because anesthetics interfere with nerve function.

VI. Conclusion

Wireless sensors like the UroMOCA may be placed in the bladder to enable monitoring of pressure and volume. This work demonstrated the feasibility of 28-day sensor implantations, suggesting that chronic monitoring is feasible. Pressure sensing from implanted UroMOCAs was highly correlated to reference catheter pressure measured with animals asleep, and described bladder activity with animals awake and moving around. While in vivo volume sensor accuracy was lower than desired, this was worsened by the small capacity of feline bladders, which are ten times smaller than typical human capacity. This suggests relative accuracy gains may be feasible using this technique in larger mammal bladders. Future research with real-time data collection in conscious, freely-moving animals, will aid neurophysiology research especially as it relates to next-generation neuromodulation strategies to restore bladder function.