Congestive heart failure (CHF) is a major cause of mortality, morbidity, and hospitalization worldwide (1). Effective therapy is available (2), and care using disease-management systems improves outcomes (3- 4). Although heart catheterization is the most accurate way to define hemodynamic state, its invasive nature limits its use to patients with severe decompensation (5). The noninvasive detection of hemodynamic abnormalities before clinical deterioration occurs might be helpful to improve care (6- 7). A recent study (unpublished data, COMPASS Investigators, March 2005) demonstrated that CHF management based on continuously monitored intracardiac pressures in patients already receiving the best-available therapy improves outcome.

Clinically available monitoring systems for heart disease use electromagnetic energy to communicate with cardiac sensors (6- 8). Limited energy penetration mandates a wire be used to transmit the signal from the sensor to the body surface. Acoustic energy is not substantially attenuated by body tissue, making wireless communication feasible. This study describes a wireless pulmonary artery (PA) pressure measurement system comprising a miniature PA device implant using right heart catheterization.

The system comprises an implant and an external communication/analysis unit (Remon Medical Technologies/Inc., Waltham, Massachusetts) (Figure 1). The implant is housed in a titanium case measuring 3 × 1.5 × 16 mm. The case contains a pressure sensor, a piezoelectric internal transducer, and a custom-built, low-power control chip. The transducer is omnidirectional because its dimension is much less than the operating wavelength. A self-expanding nitinol anchor adapts to 15- to 25-mm vessel diameters for vessel wall fixation. An external unit transmits and receives signals from the implant via a transducer in bodily contact. This unit includes a pressure sensor for converting absolute into gauge pressure. When interrogated by the external unit, the implant is activated, charged, and subsequently measures pressure and transmit full pressure waveforms for 5 to 10 s.

The implant is designed for the human proximal right PA. This segment is located behind the ascending aorta, providing for minimal acoustic energy attenuation by the air–water interface. A 10-F introducer sheath is advanced to the PA using standard technique, and a pulmonary angiogram is obtained. The optimal implant site is determined by anatomy and vessel size; the sheath is advanced to this site, and the implant is delivered like a self-expandable stent.

Anatomic differences between pigs and humans necessitated device deployment in the angulated distal main PA segment in the animal study. Although this location is suitable for acoustic communication, it has limitations for deploying the nitinol anchor. The anchor was expanded in the curved main PA segment, including the bifurcation in some cases. The animal studies were used to evaluate implant histopathology in addition to sensor and communication performance but were suboptimal for the examination of anchor design.

Eight domestic pigs (55 to 60 kg) underwent evaluation. Animals were given 100 U/kg heparin during catheterization and 100 mg aspirin/75 mg clopidogrel daily for the month after the implant. Invasive hemodynamic assessments were performed immediately after implantation and at 3 and 6 months. A Millar catheter (Millar Instruments Inc., Houston, Texas) was advanced to the PA, and simultaneous pressures obtained for direct comparison. Animals were euthanized by lethal barbiturate injection 17 or 26 weeks (n = 1 and 7, respectively) after implantation. At autopsy, the implant site was carefully examined and the lungs examined for thromboembolism. Sections (20 μm) were obtained from implant and stained with modified Paragon staining. Lung samples were stained with hematoxylin-eosin-safranin.

Patients with class III or IV CHF were approached for study participation at the Apollo Hospital (Gandhinagar, India). Patients were excluded for pacemakers, creatinine >2.5 mg/dl, and expected survival <6 months. The study was approved by the institutional review board of the Apollo Hospital, and patients signed informed consent before the procedure. Demographic parameters were recorded, including age, gender, New York Heart Association functional class, and left ventricular ejection fraction. Simultaneous implant PA and Millar catheter pressure are reported at the implant time.

Diastolic and PA pulse pressure at each implant interrogation were calculated as the mean value obtained during 5 to 10 s of simultaneous waveform recording. In cases with marked respiratory variation, end expiratory values were used. Regression analysis and the Bland-Altman method (9) compared implant and Millar pressures.

Device implantation was successful in all pigs. There were no adverse clinical events at implantation or afterwards, until euthanasia. Fluoroscopy and macroscopic observation verified that the implant was at the desired location in all animals (Figures 2, 3). Anchor struts in contact with the vessel wall were covered by a thin neointimal tissue (Figure 3). Because implants were located close to or within the PA bifurcation, some struts were free in the arterial lumen and were not covered by tissue. This finding was expected based on limitations of pig anatomy (Figure 2). Limited neointimal thickening was observed around the anchor and titanium box, without luminal narrowing, and there was no degradation of the box ((Figure 3)C and Figure 3D). No thrombi occurred in any lung section. Pulmonary artery pressures were easily obtained from all implants. (Table 1) summarizes simultaneous diastolic and pulse PA pressures from the implant and Millar catheter at times 0, 3, and 6 months. Follow-up pressure readings remained as accurate as at implant, and there were no systematic changes over time to suggest sensor drift. Standard deviations of the difference between implant and Millar recordings were 1.45 and 1.16 mm Hg (diastolic and pulse PA pressures, respectively) (Figure 4). This difference is within the error limit of the Millar catheter (2 mm Hg).

Eleven patients with CHF were enrolled in the study. Age ranged from 25 to 58 years (mean 44.6 years), and 10 of the patients were men. Left ventricular ejection fraction range was 20% to 32% (mean 28%). Six patients were identified as New York Heart Association functional class III. Pulmonary angiography was performed, and the implant landing zone was determined ((Figure 2)A and Figure 2C). Implantation was successful in 10 patients (1 patient was excluded after pulmonary angiography revealed a PA diameter >25 mm). Deployment was coaxial in 8 patients, whereas in the other 2, the implant was rotated in the artery because of right upper lobe branch involvement. Part of the implant (in these 2 patients) remained in that branch, but there was no pulmonary flow limitation. Patients were discharged 1 day after the procedure. No device-related complications occurred during the following 6 months. Pressure measurements were successfully obtained from all implants. (Figure 5)A and (Figure 5)B are typical pressure recordings and (Figure 5)C and (Figure 5)D summarize the comparison of the implant and Millar pressure values.

(Figure 6) shows a patient in whom the implanted device was clinically useful. This patient had idiopathic dilated cardiomyopathy and was clinically stable for 2 months, with PA diastolic pressures of 9 to 13 mm Hg. She was subsequently admitted to the hospital complaining of cough and shortness of breath. A clinical diagnosis of viral infection was made but was changed to acute CHF exacerbation when implant interrogation revealed a diastolic PA pressure of 45 mm Hg. She was given a diuretic, the PA diastolic pressure decreased to 15 mm Hg, and she rapidly improved.

Preclinical and human data in this pilot study show that PA pressure can be repeatedly measured by wireless acoustic telemetry using a miniature device in the PA. Device implant is a simple right heart catheterization procedure. The device was functional and accurate for 6 months and sustained no damage from biologic processes.

Wireless communication by electromagnetic energy requires a large antenna and, thus, is not suitable for deep body implants. Long-term bidirectional electromagnetic communication is prohibitive because of large power requirements, typically several milliwatts. Acoustic energy in the kHz frequency range penetrates deep into the body using little energy (tens of nanowatts) and uses small internal transducers (approximately 1 mm). This energy is safe because acoustic pressure is 10 kPa, much lower than typical pressures in diagnostic ultrasound.

Several other implantable systems are under evaluation to monitor patients with heart disease. Intracardiac pressure and pulmonary fluid volume are used to detect early deterioration of CHF (6- 7), and intramyocardial electrocardiography can detect rejection after cardiac transplantation (8). A wire is required for data transfer to a superficially implanted module, similar to a pacemaker, which is interrogated using electromagnetic communication. The advantages of a wireless system are obvious and expand the clinical utility of telemetric systems. The implant is much smaller; implantation is simple and minimally invasive. Also, wireless intrabody communication among multiple miniature implants is feasible using acoustic energy. The unique features of this system enable easy adaptation to other applications.

Our device is designed for use in CHF. Management of these patients still requires frequent hospital admission, with high morbidity and mortality. Telemetric monitoring of simple vital signs improves care (10), and adding telemetered intracardiac pressure provides a quantum improvement (7). The implanted programmable memory provides right ventricular pressure from which diastolic PA pressure can only be estimated. Reynolds et al. (11) validated the estimation algorithm and Magalski et al. (12) demonstrated that the sensor remained stable for more than 1 year. Estimating PA diastolic pressure from right ventricular pressure is not as accurate as direct measurement: a 5-mm Hg difference from control using the estimated value and only a 1.2-mm Hg with direct measurement (our human data).

This was a pilot trial. Safety and accuracy should be determined in additional trials. The current design contains no energy source; therefore, before data transmission, the implant must be charged for 30 s. This device is thus not suitable for home monitoring. Home monitoring is essential to detect impending decompensation by analyzing pressure trends over time (7). Future designs include a miniature battery, which simplifies interrogation and makes the system suitable for home monitoring.

This study demonstrated that wireless communication with a miniature PA pressure sensor is feasible. Repeated, high-quality PA tracings were easily obtained. Acoustic telemetry makes possible multiple novel monitoring and therapeutic interventions based on communication with deeply implanted devices in the heart and elsewhere.